Energy Storage by pumped storage based on liquids having different densities

Cross-Reference to Related Applications

[0001] This patent application claims the benefit of US provisional patent application US

Serial Number 62/049,341 filed September 11, 2014; and US Serial Number 62/123,278 filed November 12, 2014; and US Serial Number 62/115,091 filed February 11, 2015; and US Serial Number 62/138,357 filed March 25, 2015. All of these are incorporated by reference herein in their entirety.

Field of the Invention

[0002] Embodiments of the invention pertain to the storage of energy.

Background of the Invention

[0003] There is a general need for energy storage technologies, such as to help balance

supply and demand in the electric power grid. This need becomes more prominent with the increasing use of renewable energy sources such as wind and solar, because the pattern of energy generation from those sources tends to be irregular due to causes such as the weather, and not necessarily well matched to demand.

[0004] Currently, the energy storage technology that has by far the largest base of experience at significant size scales is pumped storage hydro (Figure 1). Pumped storage hydro uses two reservoirs of water at different elevations. To store energy, water is pumped from the lower reservoir up to the upper reservoir, and to recover energy the water is allowed to flow back to the lower reservoir through a turbine or similar energy extractor. Most commonly, the working liquid is freshwater. Currently the generating capacity of pumped storage hydro amounts to several percent of the total generating capacity of the entire electric power grid. However, new sites for pumped storage hydro are often non-optimal for reasons of topography, location or environmental sensitivity. It would be desirable to increase the applicability of pumped storage technology by, among other factors, providing more possible sites that can be used with this technology. Summary of the Invention

[0005] In an embodiment of the invention, there may be provided an energy storage system, comprising: an upper storage region; a lower storage region, located at a lower elevation than the upper storage region; a connecting conduit connecting the upper storage region and the lower storage region; a more-dense liquid contained in at least some of the upper storage region, the lower storage region and the connecting conduit; a pump in communication with the connecting conduit, to pump the more-dense liquid from the lower storage region toward the upper storage region; and an energy extractor in communication with the connecting conduit, to extract energy from motion of the more-dense liquid from the upper storage region toward the lower storage region, wherein the upper storage region, the lower storage region and the connecting conduit are surrounded by or are floating in a less-dense liquid having a density that is less than a density of the more-dense liquid.

[0006] In an embodiment of the invention, there may be provided an energy storage system, comprising: an upper storage region, located on land; a lower storage region, located at a lower elevation than said upper storage region, said lower storage region being located on land or underground or at a surface of an ocean or beneath a surface of said ocean, said lower storage region being at an elevation lower than said upper storage region; a connecting conduit that connects said upper storage region and said lower storage region; a pump in communication with said connecting conduit, disposed to pump a working liquid from said lower storage region toward said upper storage region; and an energy extractor in

communication with said connecting conduit, to extract energy from motion of said working liquid from said upper storage region toward said lower storage region, wherein said working liquid has a density that is greater than a density of seawater.

Brief Description of the Illustrations

[0007] Embodiments of the invention are further described in the following illustrations.

[0008] Figure 1 shows a conventional on-land pumped hydro installation.

[0009] Figure 2A shows a geometry of a single column of liquid for calculation of pressure.

Figure 2B shows a U-tube geometry for calculation of differential pressure due to two columns of liquid of different densities. Figure 2C shows the same as Figure 2B, but with one side of the U-tube not shown because an open body of water is its equivalent. [0010] Figure 3 A shows an embodiment of the invention in which the upper storage region, the lower storage region and the connecting conduit all rest on the seafloor. In Figure 3 A, the system is shown in a configuration in which energy is stored.

[0011] Figure 3B shows the same system as Figure 3 A, but in a configuration in which

energy has been recovered.

[0012] Figure 4 is a diagram that shows values of stored energy as a function of storage volume and elevation difference, for particular values of working fluid density.

[0013] Figures 5A-5E are a set of diagrams characterizing the topography of the seafloor at certain places along the coast of the United States.

[0014] Figure 6 is an illustration of various regions of a typical meeting of sea and land, and also shows a possible location of an embodiment of the invention.

[0015] Figure 7 is a phase diagram of sodium chloride in water.

[0016] Figure 8 is a diagram illustrating the density of the liquid that remains from seawater as evaporation of water from seawater progresses.

[0017] Figure 9 shows an embodiment of the invention with two pillow tanks on a sloping seafloor.

[0018] Figure 10 shows an embodiment of the invention with storage regions that each

contain a manifold and sub-regions.

[0019] Figure 11 A shows a bag that is bounded by posts that are anchored to the seafloor.

[0020] Figure 1 IB shows a bag on a sloping seafloor, wherein the tendency of the bag to roll is reacted at its lower edge by supports that are anchored to the seafloor.

[0021] Figure 12 shows an embodiment of the invention that has underwater impoundments for both the upper storage region and the lower storage region.

[0022] Figure 13 shows an embodiment of the invention is which the upper storage region is on land and the lower storage region is any of several kinds of storage regions underwater.

[0023] Figure 14 shows an embodiment of the invention that is located entirely on land.

[0024] Figure 15A shows an embodiment of the invention having bags for both the upper storage region and the lower storage region, with the system in a configuration in which energy is stored.

[0025] Figure 15B shows the system of Figure 15B, in a configuration in which energy has been recovered. [0026] Figure 16A shows an embodiment of the invention having a bag for the upper storage region and an impoundment region for the lower storage region, with the system in a configuration in which energy is stored.

[0027] Figure 16B shows the system of Figure 16A, in a configuration in which energy has been recovered.

[0028] Figure 17A shows an embodiment of the invention having a barge for the upper

storage region and an impoundment region for the lower storage region, with the system in a configuration in which energy is stored.

[0029] Figure 17B shows the system of Figure 17A, in a configuration in which energy has been recovered.

[0030] Figure 18 shows a strain relief device that can allow change in the effective length of the connecting conduit.

[0031] Figure 19 shows an upper bag having flotation chambers and tethers.

[0032] Figure 20 shows an upper bag having a non-uniform cross-section and a slightly sloping bottom.

[0033] Figure 21 A shows the upper bags in a position at the surface of the ocean.

[0034] Figure 21B shows the upper bags redeployed to a position at a depth below the

surface of the ocean such as for protection from a possible storm.

[0035] Figure 22 shows an upper bag with separate flotation devices and shock absorbers between the flotation devices and the bag.

[0036] Figure 23 shows an upper bag maintained in a position slightly below the ocean

surface, with both flotation devices and anchoring cables.

[0037] Figure 24 shows a spray evaporation system for concentrating liquid.

Detailed Description of the Invention

[0038] For an understanding of embodiments of the invention, it is helpful to discuss certain underlying information. Referring now to Figure 2A, there is shown a column of liquid that is bounded at its top surface by the atmosphere, and this situation results in a pressure at the bottom of the column. In conventional pumped storage hydro, there is created a pressure due to the height of a water column due to an elevation difference, and that pressure is p*g*H , where p is the density of the liquid, g is the gravitational acceleration, and H is the height of the column of liquid. The pressure difference usable for power generation is the pressure difference between the pressure of that water column and the pressure of the surrounding air, i.e., essentially the quantity just calculated, because the pressure of the surrounding air is negligible in this situation.

[0039] Referring now to Figure 2B, there is shown the geometry of a U-tube, which is a more general way in principle of illustrating pressure relationships. A U-tube representation can be used to illustrate pressure difference that may occur in the presence of two fluids of unequal densities. The two fluids can be referred to as a more-dense liquid and a less-dense liquid. For sake of illustration here, it can be assumed that at the bottom of the U-tube there is a barrier between the two fluid regions. It can be assumed that the fluid in the left-hand leg of the U-tube has a density of p ^ and the fluid in the right-hand leg has a density of The heights of the two legs are assumed to be equal to each other, having a value of H. In this situation, the pressure difference measured by a differential pressure gage at the indicated location is

ΔΡ = p _j *g*H - p2*g*H = (p _j -p2)*g*H .

[0040] This is the pressure difference that would be available for generation of electric power if power generating apparatus were placed at the location of the barrier and the differential pressure gage.

[0041] It can be noted that the illustrated shape of a U-tube having a pair of identical tubular legs is a convenience for sake of discussion, as is commonly done in textbooks. The conclusion about pressure difference is independent of the shape, cross-sectional dimensions, etc. of the legs of the U-tube. In Figure 2C, the lower-density leg of the U-tube is replaced by a body of water of the same density and the same top level or elevation as what had been illustrated in Figure 2B for the leg of the U-tube. The liquid of the naturally occurring body of water, which is shown as the light shading, is the lower-density liquid. Figure 2C is completely equivalent to Figure 2B, and pertains to some embodiments of the invention.

[0042] For applications and embodiments of the invention that are located offshore in the ocean, the less-dense liquid may be considered to be ordinary seawater of the ocean. In the temperature range commonly encountered outdoors, at or near atmospheric pressure, the density of typical seawater is about 1.025 g/cc. If the system were located in freshwater, the density of the less-dense liquid would be the density of pure water, which is very close to 1.0 g/cc.

[0043] As a thought example, if the more-dense liquid had a specific gravity of 2.0 while the less-dense liquid was water having a specific gravity of 1.0, then the pressure difference and the energy stored would be the same as for conventional pumped storage hydro using water having a specific gravity of 1.0 pumped against air, for equivalent elevation difference and volume of liquid stored. Aqueous solutions do exist having a specific gravity of 2.0 or even greater. However, for economic and other practical reasons, a solution might be chosen having a specific gravity between 1 and 2. Examples of such solutions are discussed elsewhere herein.

[0044] For an application in the ocean, if a liquid column of more-dense liquid is compared to a liquid column of ordinary seawater, as in a hypothetical U-tube arrangement, for the same liquid column heights, the pressure at the bottom of the more-dense liquid column will be larger than the pressure at the bottom of the seawater column. This provides the opportunity to perform a form of pumped storage of energy by displacing the more-dense liquid upward in relation to the less-dense liquid. The less-dense liquid may be ordinary seawater.

[0045] Referring now to Figure 3A and 3B, there is shown an embodiment of the invention in which there may be provided an upper storage region 100 and a lower storage region 200 and a connecting conduit 300, such that all of these components rest on and are supported by a seafloor having topography suitable to provide an elevation difference between the upper storage region and the lower storage region. The seafloor is shown as having a small slope. (In Figures 3A and 3B, the vertical scale is exaggerated for convenience of illustration, so the illustrated slopes are not realistic.) Such a topography could be found by appropriate site selection as is discussed elsewhere herein. As illustrated in Figures 3A and 3B, the less- dense liquid may be the ocean. Figure 3A shows an embodiment of the invention in a condition where the upper storage region is almost full and the lower storage is almost empty, and energy is stored. Figure 3B shows the opposite condition, in which energy has been regenerated and discharged.

[0046] The upper storage region 100 may be defined by a substantially impermeable

boundary, such as bag 101, or otherwise may have a variable volume of liquid contained therein. Upper storage region 100 may be defined by a boundary that is suitable to contain an identifiable mass of liquid within the boundary, such as bag 101. The boundary may be substantially impermeable to the passage of liquid through it. The boundary may be a flexible or deformable boundary. The upper storage region 100 such as bag 101 may be immersed in or surrounded by a body of the less-dense liquid. The upper storage region 100 is illustrated in Figures 3 A, 3B as resting on the bottom of the ocean or body of less-dense liquid. [0047] The lower storage region 200 may be at a lower elevation than the upper storage region 100. Similar to the upper storage region 100, the lower storage region 200 may be defined by a substantially impermeable boundary, such as bag 201, or otherwise may have a variable volume of liquid contained therein. The lower storage region 200 such as bag 201 may be immersed or surrounded by a body of liquid such as the ocean. The lower storage region 200 is illustrated in Figures 3 A, 3B as resting on the bottom of the ocean or body of less-dense liquid.

[0048] The bags 101 and 201 may be flexible while also being substantially impermeable to the passage of liquid through the membrane. In general, the bags could be made of fabric and polymer. It is further possible that the bag could have bands built into it in appropriate directions to strengthen it against possible tensile stress or for any other reason. As discussed elsewhere herein, it is possible that even if the upper storage region 100 such as bag 101 or the lower storage region such as bag 201 has a minor amount of holes or leakage, the system could still continue to function appropriately to be used as an embodiment of the invention. However, an amount of leakage that would result in significant loss or significant dilution of the more-dense liquid would be detrimental.

[0049] It can be observed in Figure 3 A that in the energy storage condition, the lower storage region 200 is small and the more-dense liquid is mostly in the upper storage region 100 and the upper storage region 100 is large. It can be observed in Figure 3B that in the discharged condition, the more-dense liquid is mostly in the lower storage region 200, and the lower storage region 200 is large and the upper storage region 100 is small.

[0050] Connecting the upper storage region and the lower storage region may be a

connecting conduit 300. Using the connecting conduit 300, the more-dense liquid may be able to move back and forth between the upper storage region 100 such as bag 101 and the lower storage region 200 such as bag 201 in order to either store or give back energy depending on the direction of flow of the more-dense liquid.

[0051] At some point in connecting conduit 300 there may be provided a valve 350. There may also be a pump 400 and an energy extractor 500. The pump may perform work against a pressure difference that results from the difference in density between the more-dense liquid and the less-dense liquid. The energy extractor 500 may recover at least some of the energy represented by the work that has been performed by the pump 400. The pump 400 and the energy extractor 500 may be the same device, which may be capable of operating in both directions or modes of operation. For example, the pump 400 and the energy extractor 500 may both be turbomachinery. Alternatively, they could both be a positive-displacement type device. The pump 400 and the energy extractor 500 may be collectively referred to as energy conversion device 600. The pump 400 and energy extractor 500 may be similar to those of conventional pumped storage hydro. The materials of construction may be appropriately corrosion-resistant materials, in view of the fact that the environments and liquids that embodiments of the present invention may be exposed to may be more corrosive than the freshwater that is usually handled in conventional pumped storage hydro installations.

[0052] The pump 400 and the energy extractor 500 may be located at or near the lower

storage region 200. The pump 400 and the energy extractor 500 may be designed and installed in such a way that they are able to be brought back up to the surface of the ocean for maintenance and repair. For example, there may be demountable joints connecting the pump 400 and the energy extractor 500 to other components of the system. Such joints may allow components such as pump 400 and the energy extractor 500 to be brought up to the surface of the body of water for purposes such as maintenance, repair or replacement. It is also possible for the pump and the energy extractor to be located at some other appropriate place along the connecting conduit 300 as discussed elsewhere herein.

[0053] There may be an electrical cable 700 connecting the pumped storage system to the electric power grid or to a power source and a load. During energy charging or pump-up, the cable could bring electrical power from a source of electrical energy such as the electric power grid on land, or alternatively from an at-sea energy source such as wind turbines. This energy could be used to store energy in the energy storage system such as by pumping more- dense fluid into the upper storage region. During energy recovery, the cable could deliver energy from the energy storage system to the electric power grid. The cable may carry either Alternating Current or Direct Current. Alternating Current would be convenient for interfacing with the electric power grid. Direct Current would have an advantage in that it would not electromagnetically couple with seawater surrounding the cable, and therefore would avoid certain losses. However, the use of Direct Current would involve certain additional power conversion apparatus. A coaxial cable would also be possible. Such electric power transmission and corresponding equipment are known.

Attributes of a completely-seafloor-supported system

[0054] For simplicity of explanation, Figures 3A and 3B show horizontal seafloor underlying upper bag 101 and lower bag 201, and sloping seafloor is shown underneath connecting conduit 300 between upper baglOl and lower bag 201. It is possible that such topography may be found occurring naturally, but another option for installing a bag-type storage region in a flat horizontal orientation on a generally sloping seafloor would be to modify the local seafloor so as to create a flat region where the bag is located. This could be done, for example, by dredging or digging. In such a situation, material could be removed from the seafloor corresponding to a portion of the bag footprint so as to create a relatively flat horizontal surface at the dredged location, and the dug (dredged) material could be deposited nearby to build up adjacent spaces so as to also be flat and horizontal and also provide support for another portion of the same bag footprint. This technique for construction on a slope is sometimes referred to as "cut and fill." The option of placing a bag directly on a sloped surface is also discussed elsewhere herein.

A completely-seafloor-supported system may provide substantially continuous structural support for those components (upper bag 101, lower bag 201 and connecting conduit 300) substantially everywhere where those components exist. It can be expected that structurally speaking, this would be a reliable and safe situation with relatively few possible structural problems or failure modes. Also, in such an embodiment, it can be expected that ocean currents and localized motion of the water at those components would be relatively small because of the components would be located so close to the seafloor, as compared to some other possible situation of being surrounded by ocean currents and motion that might exist in the open water of the ocean at various depths. In particular, if the uppermost component, i.e. the upper bag 101, rests on the seafloor and is located at least a modest distance below the surface of the ocean (such as at least 10 meters below the surface of the ocean) for all fill levels of the upper bag 101, it is likely that most of the fluid disturbances that exist at the ocean surface, such as waves, motion of water due to storms, other effects of weather, etc., will be significantly attenuated at that depth, and thus the upper bag 101 will be shielded from most of the effects of weather and ocean surface phenomena. For example, with respect to ocean currents, it can be expected that there would be a boundary layer near the seafloor such that the velocity of the ocean current adjacent to the seafloor is relatively smaller than it is at various other depths in the open ocean, and hence the velocity of the ocean current adjacent to the system components is relatively smaller than it is at various other depths in the open ocean. Of course, components that are located at even greater depths can be expected to be even more shielded. For example, if the upper storage volume is located approximately where the continental slope meets the continental shelf, the depth at that location would probably be a hundred meters or so, and that location would be well shielded from the influence of waves, storms and other near-surface phenomena. Of course, the lower storage region would be located at a much greater depth and would be even more shielded from the influence of waves, storms and other near-surface phenomena.

[0056] Basically, a completely-seafloor-supported system can achieve good structural

reliability in return for relying on the naturally occurring topography to provide the required slope, and this to some extent requires that the connecting conduit 300 be at least moderately long.

[0057] In a completely-seafloor-supported system, the upper bag 101 and the lower bag 201 could be designed and manufactured having many features in common with each other. One possibility is that both the upper storage region 100 and the lower storage region 200 could be relatively simple bags of the style sometimes referred to as "pillow tanks." A "pillow tank" is a bag that has two relatively large dimensions and a third dimension that, when the tank is full, is still somewhat smaller than the other two dimensions. In conventional usage, the surface formed by the two large dimensions usually rests on a generally horizontal support surface. Such bags are available in a wide range of sizes for the bulk storage of liquids on land.

[0058] If the pillow tank rests on a surface that is substantially horizontal, the pillow tank would tend to assume a shape that is approximately uniform thickness for most of the pillow tank, except at the edges.

[0059] In regard to the connecting conduit 300, it can first of all be appreciated that the

connecting conduit 300 is subject to an internal pressure that varies with position along the conduit. This is discussed elsewhere herein. Specifically, the pressure difference across the wall of the connecting conduit 300 increases with increasing depth below the upper storage region 100, from the upper storage region 100 as far as the pump/turbine 400, 500 or valve 350 (which would typically be located at or near the lower storage region). Accordingly, it is possible that the deeper portion of the connecting conduit 300 may be designed more robustly than the shallower portion of the connecting conduit 300. For example, the wall thickness of the connecting conduit 300 could be greater at the region that is subject to a larger internal pressure.

[0060] If the connecting conduit 300 is manufactured of a flexible material in somewhat the same way as the upper bag 101 or lower bag 201 is manufactured, it can be expected that the connecting conduit 300 may assume a cross-sectional shape that is responsive to and variable as a function of the local pressure inside the connecting conduit 300. If the local pressure inside the connecting conduit 300 is sufficiently large, which would probably be the case for much of the length of the connecting conduit 300, it can be expected that the connecting conduit 300 will assume a cross-sectional shape that is approximately circular. If the local pressure inside the connecting conduit 300 is somewhat smaller, as might occur in a somewhat upper-elevation portion of the connecting conduit 300, it can be expected that the connecting conduit may assume a cross-sectional shape that may be partially D-shaped, unless the connecting conduit 300 is designed to maintain a desired cross-sectional shape, such as with stiff eners.

[0061] Embodiments of the invention are further described but are in no way limited by the following Examples, which present certain numerical calculations and other specific information.

Example 1: Hydrostatics

[0062] Displacing a more-dense liquid against seawater will require work and will store energy.

[0063] A baseline case can be considered to be conventional pumped storage hydro on land using freshwater. For this situation, the specific gravity of the pumped water is 1.0 and the water is displaced against air. For purposes of the present calculation, air has essentially zero density. The water is displaced through a specified elevation difference. It can be thought that this action would store one unit of energy per unit volume of displaced water per unit of elevation difference.

[0064] In comparison to the baseline case, if a more-dense liquid of specific gravity 2.0 were displaced through the same specified elevation difference against water having a specific gravity of 1.0, the density difference or pressure difference against which work is performed is 2.0 - 1.0, or 1.0. So, that operation would also store one unit of energy per unit volume of displaced water per unit of elevation difference. This is exactly the same result as for the baseline case of conventional pumped storage hydro with water compared to atmospheric air. Similarly, in regard to a system that operates in and is surrounded by the ocean, the surrounding liquid is seawater having a specific gravity of 1.025. If a working fluid had a specific gravity of 2.025 and was displaced relative to seawater having a specific gravity of 1.025, then the energy storage per unit volume of displaced working fluid would be the same as for conventional pumped storage hydro using water.

[0065] Aqueous solutions having a specific gravity of 2.025 or even greater are physically achievable as described elsewhere herein. However, in view of the fact that large quantities of the more-dense liquid may be required, the choice of solution to be used for the working fluid could also be influenced by economics and other factors. [0066] If a working fluid having a specific gravity greater than 2.025 were used in an ocean- based application (with seawater having a specific gravity of 1.025), then the requirement for (elevation difference * volume), to match a corresponding conventional pumped storage hydro installation, would actually be less than for the corresponding conventional pumped storage hydro installation.

[0067] If a working fluid having a specific gravity less than 2.025 is used, in an ocean

application, it will not store as much energy for the same volume for the same elevation difference, as would a conventional pumped storage hydro system. In this situation, the requirement for (elevation difference * volume), to match a corresponding conventional pumped storage hydro installation, would be greater than for the corresponding conventional pumped storage hydro installation.

[0068] Based on some of the possible compositions of solutions discussed elsewhere herein, it may be appropriate to assume that the working fluid has a specific gravity that is somewhere between 1 and 2.

[0069] Because of this calculation, it becomes relevant that the depth of many potential sites in the ocean, for deployment of embodiments of the invention, is larger than an elevation difference that would typically or easily be available for a conventional pumped hydro site on land. Also, because of the vast size of the oceans, the volume of stored liquid at sea is not constrained by the type of siting and environmental considerations as are encountered with conventional pumped storage hydro at sites on land. The constraints on stored volume in embodiments of the present invention may be more of an economic and engineering nature relating to construction of the storage regions such as bags or other related facilities. So, the disadvantage that may come from the density of the more-dense liquid not being as large as might be desired, can be made up for with a larger depth than would be typical of conventional pumped storage hydro, or with a larger volume of liquid, or with some combination of both factors.

[0070] For introductory discussion using round numbers, some possible working fluids are: effluent from desalination plant, specific gravity approximately 1.1; nearly salt-point brine, specific gravity approximately 1.2; nearly-saturated magnesium chloride or maximum- density bittern, specific gravity approximately 1.3; and nearly-saturated potassium carbonate solution, specific gravity approximately 1.5 .

[0071] On the same scale on which water lifted against atmospheric pressure is 1.0, the

amount of energy stored by pumped storage of salt-point brine relative to seawater is about

1.2185-1.025, or 0.1935, which may be rounded downward to 0.18 to allow for a slight non- saturation as discussed elsewhere herein. In this case, the energy storage would be about 1/5.5 (about one-fifth) as much as for conventional pumped storage hydro of comparable elevation and volume. So, to store the same amount of energy as conventional pumped storage hydro by using displacement of nearly-salt-point brine relative to seawater, it would be necessary that the product of elevation difference and storage volume be approximately a factor of 5.5 times as large as the corresponding values for conventional pumped storage hydro. In regard to the use of magnesium sulfate or bittern at or near its maximum density, the amount of energy stored by pumped storage of such fluid relative to seawater is about 1.33-1.025, or 0.305, i.e., about 1/3.3 (about one-third) as much as for conventional pumped storage hydro of comparable elevation and volume.

[0072] There is opportunity for the depth in the ocean to be larger than the largest elevation difference that is available or convenient for conventional on-land pumped storage hydro. It is, for example, possible for the elevation difference of systems located at sea to be several times the elevation difference of conventional pumped storage hydro. So, for embodiments of the invention, the volume of more-dense liquid needing to be stored might be comparable to or not much greater than the volume of reservoirs used for conventional pumped storage hydro on land.

[0073] This comparison may be summarized in an Energy Storage Parameter, which may be defined as (rhofreshwater - 0) / (rhoworkingfluid - rhoseawater) /, which is (rhofreshwater) / (rhoworkingfluid - rhoseawater). This parameter compares the required volume*elevation difference for embodiments of the invention, relative to the volume*elevation difference for conventional pumped storage hydro using freshwater in order to store the same amount of energy.

[0074] For desalination plant effluent having a specific gravity of 1.1, this factor is 13.3. For nearly-salt-point brine having a specific gravity of 1.2, this factor is 5.7. For a solution such as a magnesium sulfate solution or bittern having a specific gravity of 1.3, this factor is 3.6. For a solution such as a potassium carbonate solution having a specific gravity of 1.5, this factor is 2.1. As already discussed, for a hypothetical solution having a specific gravity of 2.025, this factor is 1.0. This energy storage parameter is summarized in Table 1.

Table 1

SG working fluid Typical solute Energy Storage Parameter

1.1 Desalination effluent 13.3

1.2 sodium chloride 5.7

1.3 magnesium sulfate 3.6 1.4 calcium chloride 2.7

1.5 potassum carbonate 2.1

1.6 ferric sulfate 1.7

1.7 ferric sulfate 1.5

2.025 heavy metal halide 1.0

Example 2. Further Quantitative guideline about stored energy

[0075] One quantitative guideline relevant to the design of an energy storage system is to express how much energy is stored as a function of the relevant parameters. Relevant parameters include elevation difference between the two reservoirs, difference in specific gravity between the working fluid and the surrounding fluid, and the volume or mass of the two reservoirs that hold the working fluid.

The stored energy is calculated as:

Stored Energy = (rhoworkingfluid - rhoseawater) * g * H * Volume

where

rhoworkingfluid is the density of the working fluid

rhoseawater is the density of the seawater

g is the acceleration of gravity

H is the difference in elevation between the upper storage region and the lower storage region

Volume is the stored volume that shifts between lower and upper storage regions

[0076] It is assumed for this formula, for simplicity, that the elevation of the working fluid in the upper storage region and the lower storage region are constant. If there is some variation of level of the working fluid in one or both of the storage regions during a charging cycle or a discharging cycle, that variation can be accounted for in known ways.

[0077] A particularly simple numerical example is available for the situation of a working fluid whose specific gravity is 1.39 when surrounded by seawater. In this situation, if working in appropriate units of measurement, there is a proportionality factor that is exactly 1. The proportionality factor expresses a relation between a storage parameter that may be expressed in m ^A 4, and stored energy expressed in Watt-hours. The storage parameter m^ represents volume of working fluid stored in a storage region, which may be expressed in m^, multiplied by the difference in elevation between the upper storage region and the lower storage region, which may be expressed in m.

[0078] For the situation in which the described proportionality factor is 1 , a storage

parameter value of 1 m^ corresponds to a stored energy of 1 W-hr. So, for example, a situation of 1000 m^ in combination with an elevation difference of 1000 m would give a storage parameter of 1.E6 m^ and would store 1.E6 W-hr (1 MW-hour) of energy. The same storage parameter of 1.E6 would also be obtained for a situation of 10,000 m^ in combination with an elevation difference of 100 m, and this situation also would store 1 MW-hr of energy. It may be appreciated that for values of specific gravity of the working fluid other than 1.39, the stored energy in Watt-hr would still be proportional to the storage parameter in m^, but the proportionality constant would be different from 1.

[0079] This calculation is simply a calculation of energy storage and does not include any allowance for inefficiency.

Example 3: Storage Regions and Volumes

[0080] It is also useful to discuss numerical examples and properties of a system associated with sample numerical values of volumes. Also, because of the size of the required storage volume, it is useful to know what commercial experience exists.

[0081] A category of flexible bags for which there is commercial experience is flexible

storage tanks sometimes referred to as "pillow bags" because of their shape. Such bags are generally used on land on a flat substrate to store any of various liquids including aqueous liquids, petrochemicals, and other types of liquids. A typical proportion of such a bag when full is 1 unit by 1 unit by 0.1 unit. Footprint shapes of the bags also can, if desired, be rectangular rather than square. Such bags are available from Granite Environmental, Sebastian, FL; Aero Tec Labs, Ramsey NJ; Husky Portable Containment, Bartlesville, OK; Labaronne Citaf (France); and other vendors. The upper end of the range of sizes of pillow bags that are commercially available is in the range of 1000 m^ (for example, 28 m x 25 m x 1.6 m) (265,000 gallons). Sometimes the upper limit sometimes is determined by the size of a manufacturing facility rather than any limit inherent in the product itself. In one catalog, a

1000 m^ pillow bag is listed as having an empty (dry) weight of 2 metric tonnes. Again, it can be noted that the design condition for such bags is to contain liquid in conditions when the outside of the bag is exposed to the atmosphere. [0082] There is also some experience with other large flexible bags that are sometimes referred to as flexible barges, mainly for the purpose of transporting freshwater across distances in the ocean, generally from freshwater sources such as rivers to places that are in need of fresh water. Such bags have been made or demonstrated by Spragg Associates, Manhattan Beach CA (sometimes referred to as Spragg Bags) (US patents 5,413,065 and 5,488,921); by Aqueous International (Alaska); by Nordic Water Supply ASA, Oslo, Norway, which has transported water commercially from Turkey to Cyprus using such bags; and by Aquarius Water Trading and Transportation (England, Greece, Australia). Typically such bags are cylindrical and have a length that is about an order of magnitude greater than the diameter. Typically the ends are hemispherical. Typically a number of such bags are connected together in series to form a train, like links of a sausage. The train of bags is then pulled through the ocean to its destination by a tugboat. For purposes such as water supply, the bags contain fresh water, which is slightly buoyant with respect to seawater. Thus, such bags naturally float at or near the surface of the ocean, although typically a large portion of the volume of the bag would be submerged (like an iceberg). The buoyancy situation of such bags is different from that of the embodiments of the present invention, in which the contained fluid is more dense than the surrounding fluid (seawater) and so is not inherently buoyant with respect to the surrounding liquid (seawater). Nevertheless, the experience of such bags or flexible barges being used in the open ocean is still relevant.

[0083] One numerical example of a volume is a bag whose volume is the equivalent of a shape that is a hemisphere-capped cylinder 25 feet diameter, 250 feet overall length including hemispherical caps at both ends. The volume of such a bag is 3359 m^ (119,000 ft^, which is 887,000 gallons, corresponding to a weight of 4.45E6 kilograms of a liquid whose specific gravity is 1.325, which corresponds to an unbuoyed weight in the ocean of 1.008E6 kg). (Such bags are typically filled only to slightly less than full capacity, but what is calculated and reported here is the full geometric volume.) In this example, the less-dense liquid is assumed to be ordinary seawater having a specific gravity of 1.025, and the more-dense liquid is assumed to have a specific gravity of about 1.325, which can refer to bittern at nearly its maximum density as discussed elsewhere herein. This gives a difference in specific gravity of approximately 0.3, or a density difference of 0.3 g/cc.

[0084] It can be assumed for purposes of calculation and example that there is an elevation difference between the two such bags for pumped energy storage of 1000 meters. The pressure difference due to the elevation and to the density difference is 2.94 MPa (29 atmospheres, 427 psid). The stored energy is 9.9 GigaJoules, which is 2.74 MegaWatt-hour. This quantity of energy represents 1 MW for a time period of 2.74 hours. Such a discharge time period would be representative of an energy storage system that is intended to even out variations during the course of a day.

[0085] As a representation of inefficiencies that are typical of pumped storage hydro, it can be assumed that it might require about 110% of that amount of energy to pump the more- dense liquid to the upper storage region, and around 90% of that amount of energy might be recoverable in terms of being converted back into electricity. Such efficiency ratios are in the range typically discussed in the literature of Pumped Storage Hydro.

[0086] Another calculation can be performed using a larger storage volume than was just used. A large ocean-going tanker such as an oil tanker may typically have a storage capacity of 1 million barrels (of oil). In the type of size description that is used to describe shipping using large ocean-going ships, a 1 million barrel tanker is intermediate in size between the limiting size of a ship that is able to travel through the Panama Canal as originally constructed (referred to as Panamax) and the limiting ship size of a ship that is able to travel through the Suez Canal (referred to as Suezmax). So, a tanker having a cargo capacity of 1 million barrels would be able to go through the Suez Canal, but it would be too large to go through the Panama Canal as the Panama Canal was originally constructed, prior to the expansion project that is currently in progress. Of course, a corresponding barge could also be envisioned, having similar cargo capacity as a tanker ship, but needing to be towed. For purposes of a sample calculation, a volume of 1 million barrels is 42 million gallons or

159,000 m^. For a liquid whose specific gravity is 1.325, the mass is 210 E6 kilograms (464 E6 lbm), which corresponds to an unbuoyed weight in the ocean of 47 E6 kg.)

[0087] Another representation of this same volume of liquid, which may be a useful

representation of a lower storage region as discussed elsewhere herein, the lower storage region can be assumed to be a shallow cylindrical region having a height of 10 meters and an appropriate diameter. In this case, the region would have a diameter of 142 meters.

[0088] For purposes of this example, an elevation difference of 2000 meters may be assumed for calculation purposes, in view of the larger overall size of the system. In this case, the differential pressure of the liquid column is 5.88 MPa (58 atmospheres, 853 psid). The stored energy is 935 GigaJoules, which is 260 MegaWatt-hour. This represents 100 MW for a period of 2.6 hours. Again, it can be assumed that it might require about 110% of that amount of energy to create the stored energy, and around 90% of that amount of energy might be recoverable in terms of being converted back into electricity, as is typically discussed in the literature of Pumped Storage Hydro.

[0089] For further comparison, in some of the largest commercial pumped storage hydro installations, the amount of water that is capable of being transferred between the two reservoirs is in the range of 10,000 acre-feet which is 12 million m ^A 3. Such flowable volumes are typical, for example of two conventional pumped hydro installations that are currently in operation, namely, the Castaic Pumped Storage Plant in California and the Northfield Mountain Pumped Storage Project in Massachusetts. The power rating of the Castaic Pumped Storage Plant is 1.25 GW, and the power rating of the Northfield Mountain Pumped Storage Project is 1.1 GW. These volumes identified here are the volume of water that is capable of flowing back and forth between the two reservoirs at design conditions. This is different from the entire volume of a particular reservoir (which is generally larger than the volume of water that is designed to flow back and forth during energy storage or energy recovery).

[0090] It is possible to operate a storage system in which the upper storage region 100

comprises multiple individual bags and the lower storage region 200 comprises multiple individual bags or storage regions. Each individual bag can be a convenient size for fabrication. Additionally, the use of individual bags or storage regions provides for the ability to isolate bags from the rest of the system or from each other in order to deal with a possible leak in an individual bag.

[0091] Representative storage volumes are given in Table 2.

Table 2

Large commercial pillow tank 1000 m^

Spragg bag 25 ft dia 250 ft long hemispherical caps 3400 m^

1 M bbl tanker (larger than Panamax but smaller than Suezmax) 159,000 m^

Castaic or Northfield Plants flowable volume 12,000,000 m^

[0092] Referring now to Figure 4, there is shown a graph relating storage volume, elevation difference and stored energy. This graph is for a working fluid having a density of 1.39 g/cc. What is shown in this graph is the value of stored energy calculated according to the equation Energy = (rhol - rho2) * g * H * Volume, without taking into account any inefficiencies of either charging or discharging. [0093] It can further be noted that a scaling relationship can be used to compare pillow tanks with cylindrical storage vessels. In view of the fact that the cost of constructing a bag-like storage region can be expected to vary with the amount of fabric used, and the amount of fabric varies according to the surface area of the storage region, it is appropriate to quantify the volume -to-surface ratio of a storage region. For a pillow tank, typical proportions of such tanks are approximately 1 unit by 1 unit by 0.11 unit, so if the length of a side is s, then the volume is s * s * 0.1 *s, or 0.1 * s^. The surface area of such a pillow tank can be approximated as 2*s^. So, the volume-to-surface ratio is about 0.05 *s. For a cylinder having diameter D and length L, the volume is 0.25*pi**D*D*L, and the surface area is pi*D*L. So, the volume to surface area is 0.25 *D. For purposes of storing a quantity of liquid in a storage region, a larger value of the volume-to-surface is more economical. It is possible to make various assumptions about comparable values of s (the side length of a pillow tank) and D (the diameter of a cylinder); they could be compared as being roughly equal to each other. In any event, it can be seen that a cylindrical bag has an advantage over a pillow tank by a factor of at least several times, in terms of volume-to-surface ratio.

Example 4. Power

[0094] It is also useful to have a formula describing the instantaneous power, as a function of flow parameters. In general, power is volumetric flowrate times pressure difference. If normalized on the basis of unit cross-sectional area of the connecting conduit, the power per unit of cross-sectional flow area is velocity times pressure difference. As discussed elsewhere herein, there are various practical considerations combine to constrain a realistic range of fluid velocities in the connecting conduit. Thus, power can be described in terms of the volumetric flowrate of the working fluid in the connecting conduit.

[0095] For sizing the connecting conduit based on the desired power, the power is given by

Power = volumetric flowrate * pressure difference

= velocity * cross-sectional area * (rhol-rho2)*g*H

[0096] As illustrative values, for a working fluid having a specific gravity of 1.3, for an

elevation difference of 1000 meters, the power per unit area is 2.7 MW per m^ of cross- sectional flow area. For an elevation difference of 100 meters, the power per unit area is 0.27

MW per m^. This formula is not adjusted for possible inefficiency. Example 5. Quantitative guideline about stored energy loss in connecting conduit

[0097] For energy storage systems, the term round-trip efficiency refers to how much energy is recovered compared to how much energy is put in to the storage system. So, the term round trip efficiency is essentially refers to a summation of all of the sources of energy loss in the storage system. Generally it is desirable that the round-trip efficiency of an energy storage system be as large as possible, such as around 80%. This requires total energy losses of around 20% or not much more than 20% of the energy that is put into the system.

[0098] In a pumped storage hydro system, typical sources of energy loss include inefficiency of the pump; losses due to pressure drop due to flow in the connecting conduit during charging; losses due to pressure drop due to flow in the connecting conduit during discharging; and inefficiency of the turbine.

[0099] A pump or turbine efficiency parameter compares the energy of the fluid delivered by the pump or received by the turbine, in relation to the electrical or mechanical power delivered to the pump or extracted by the turbine. Efficiency factors in the range of greater than 0.9 are achievable. The efficiency of turbomachinery is partly a function of the geometry of the blades and other internal components of the pump or turbine. Another feature that is known to be helpful in regard to improving the efficiency of a pump/turbine is the use of variable speed drives.

[00100] In order to stay within the desired range of round-trip efficiency, it may be

appropriate to assume an energy loss budget that allocates an inefficiency of approximately 7%) to the pump, an inefficiency of around 7% to the turbine, an energy loss due to pressure drop due to flow in the conduit during discharging of no more than approximately 5% of the stored energy, and an energy loss due to pressure drop due to flow in the conduit during charging of no more than approximately 1% of the stored energy. It should be understood that these are simply round numbers for sake of example.

[00101] In somewhat greater detail, the energy loss due to pressure drop due to flow in the connecting conduit 300 in either direction of flow has an approximately quadratic dependence on the flowrate of the liquid in the connecting conduit 300. This means that if one of the phases of operation, such as charging, could be performed more slowly than assumed or can be performed relatively slowly, it would be advantageous. For example, it is possible that the charging of the system could be done more slowly than the discharging of the system. In this case, losses due to flow in the conduit during charging would be smaller than losses due to flow in the conduit during discharging. This strategy could significantly reduce the energy loss due to flow in the connecting conduit 300 during charging. Such loss could be expected to scale with the square of the velocity of the working fluid in the connecting conduit 300. So, charging at a rate that is half of the discharge rate would reduce the energy losses due to pressure drop during charging by a factor of four.

[00102] As a numerical target, it may be desirable for an energy loss due to flow in the

connecting conduit during discharge to be no more than about 5% of the stored energy. It is believed that this characteristic may be typical for applications such as load-leveling on a daily cycle, i.e., the period of heavy demand may be in the range of several hours to less than half a day, while the period of charging may be a majority of the day, which would allow the charging rate to be smaller than the discharging rate.

[00103] As a reference point for making comparisons, it is useful to compare any design

against a form of pumped storage that could be considered the simplest and most ideal design. Such a simple ideal design is a land-based pumped storage (water surrounded by atmosphere) with a connecting conduit that is exactly vertical. Such a design is the most optimistic design possible in regard to minimizing energy loss. It can be appreciated that compared to this baseline, embodiments of the invention may be less favorable in certain parameters, but a way of compensating for this less-favorable situation is by appropriately sizing the dimensional parameters of the system, such as the cross-sectional area of the connecting conduit 300.

Fluids properties parameter

[00104] A first scaling exercise may be performed by considering only fluid properties, i.e., densities or specific gravities. A Fluid Properties Parameter, SGfluid/ (SGfluid - SGseawater), can be considered to be a dimensionless quantity that is descriptive of the physical properties of the combination of the working fluid and seawater.

[00105] As a baseline situation, for a simple on-land pumped hydro installation using ordinary fresh water surrounded by the atmosphere, the fluid density that appears in the calculation of loss due to pressure drop is the density that corresponds to a specific gravity of 1.0. Also, for this situation, the driving density difference is 1.0 (the specific gravity of fresh water compared to a negligible density or specific gravity for the surrounding atmosphere). A comparison can be made in the form of a ratio of these two quantities. This gives a Fluid Properties Parameter that is 1.0/ (1.0-0), or 1. [00106] In contrast, for an embodiment of the invention, using the same scaling, the fluid specific gravity that is used in the calculation of loss due to pressure drop might be 1.3, as an example value. Also, for this situation, the driving density difference would be 0.275 (i.e., 1.3 for the specific gravity of the working fiuid, compared to 1.025 for the specific gravity of surrounding seawater). It can be seen that for this at-sea system with this working fiuid, the same Fluid Properties Parameter is 1.3/0.275, which is 4.73. A range of examples is given in Table 3:

Table 3

SG working fiuid Typical solute Fluid Properties Parameter

1.1 desalination effluent 14.7

1.2 sodium chloride 6.86

1.3 magnesium sulfate 4.73

1.4 calcium chloride 3.73

1.5 potassum carbonate 3.15

1.6 ferric sulfate 2.78

1.7 ferric sulfate 2.52

[00107] It can be noted that this parameter is not the same as the Energy Storage Parameter discussed elsewhere herein.

[00108] It can be appreciated that a smaller Fluid Properties Parameter is more favorable for minimizing losses due to pressure drop. It can be seen by inspection algebraically that the Fluid Properties Parameter is always greater than unity. This is true for all values of specific gravity of the working fluid, in a situation of embodiments of the invention. This value is greater than the previously discussed value of 1.0 for a conventional on-land pumped hydro system. If the working fluid has a relatively large density then the Fluid Properties Parameter is not as unfavorable, but the Fluid Properties Parameter is always greater than unity. This fact can be mitigated in other ways through sizing of appropriate dimensional parameters of the system, such as the cross-sectional area of the connecting conduit 300.

Length Parameter

[00109] Another factor in quantifying energy loss is the length of the connecting conduit 300.

Comparison in this regard can be performed taking into account the varying lengths of connecting conduit 300 associated either with a vertical connecting conduit (which can be considered to be a baseline for comparison) or with a connecting conduit having some defined slope or angle. If the connecting conduit is not exactly vertical, the connecting conduit length will be longer than the vertical elevation difference between the two storage regions. The ratio of these two may be quantified by a parameter referred to herein as the Length Parameter, which is equal to the length of the connecting conduit 300 divided by the elevation difference between the two storage regions.

[00110] In most on-land pumped hydro installations, the connecting conduit is non-vertical, as determined by the available locations of the reservoirs, and this sometimes results in moderately long tunnels or conduits. Table 4 shows data about some conventional on-land pumped hydro installations:

Table 4

Project Tunnel length Head Lratio

Castaic Pumped Storage Plant (California) 11.6 km 320 m 36

Glendoe Hydroelectric Power Plant (Scotland) 8.6 km 600 m 14.3

Helms Pumped Storage Plant (California) 3889 m 495 m 7.9

Taum Sauk (Missouri) 2100 m 240 m 9.0

Yards Creek Pumped Storage (New Jersey) 1800 ft 700 ft 2.6

Swan Lake North (proposed) (Oregon) 2100 m 1300 ft 5.3

Tantangara-Blowering (proposed) (Australia) 53 km 875 m 60

Black Canyon Pumped Storage (proposed) (Wyoming) 6600 ft 1000 ft 6.6

[00111] If an embodiment containing a free-floating upper storage region is used at sea, it might indeed have a vertical connecting conduit, but an embodiment that is built as a completely seafloor-supported system may have a connecting conduit that is sloped an angle that is not very steep. For a seafloor-supported installation that rests on a sloping seafloor, the Length Parameter or pipelength/elevation ratio is determined by available slopes of the seafloor. A Length Parameter of 5 corresponds to a slope angle of about 1 1.54 degrees. A Length Parameter of 10 corresponds to a slope angle of about 5.74 degrees. A Length Parameter of 20 corresponds to a slope angle of about 2.87 degrees. The availability of sloping ocean floor is discussed elsewhere herein. The Length Parameter is the reciprocal of the sine of an appropriately defined angle of the sloping connecting conduit, which is the cosecant of that angle. The angle is defined such that a connecting conduit that is horizontal has a slope angle of zero. This Length Parameter is applicable both to the case of a straight vertical connecting conduit (in which case it has a limiting value of one) and to the case of a sloping connecting conduit.

Friction factor

[00112] Flow in pipes has been characterized by the classic Moody diagram, which provides a value of the friction factor. The friction factor appears in the pressure drop formula

Deltap = (K + f*L/D)*0.5*rho*V ²

where:

f is the Fanning friction factor (dimensionless)

K represents losses due to geometric features such as entrances, exits and elbows

L is the length of the pipe

D is the diameter of the pipe

rho is the density of the flowing fluid

V is the velocity of the flowing fluid

[00113] According to the Moody diagram, at large Reynolds number (which is characteristic of conditions of interest here), and for a surface roughness that would be typical of the connecting conduit 300, the value of friction factor is approximately 0.02. The friction factor is nearly constant within the range of conditions of interest for present purposes.

Combined Derivation for Frictional Loss Fraction

[00114] Based on considerations such as the desired overall efficiency, a guideline can be derived that pertains to fractional loss of energy due to flow friction in the connecting conduit. This derivation includes dependence on the slope of the connecting conduit but does not require explicit knowledge of the actual elevation difference between the two storage regions. (This calculation assumes that the connecting conduit is a single conduit, not multiple conduits in parallel.)

[00115] The pressure drop of fluid flowing in the conduit is given by

Deltap flow loss = f * (Lconduit /D) * 0.5 * (rho) *V ² Driving pressure difference = (rholiquid-rhoseawater) * g * H [00116] As a representation of the Frictional Loss Fraction, the ratio of these two quantities is given by:

Fractional loss = ( f * (Lconduit /D) * 0.5 * (rho) *V2 ) / (rholiquid-rhoseawater) * g * H Variables in this equation can be rearranged and grouped as follows:

Fractional loss = (rho/( rholiquid-rhoseawater)) *(Lconduit/H) * f * (0.5 *ν ² /(D*g))

[00117] The first group is the Fluid Properties Parameter that has already been discussed. The second group is the Length Parameter that has already been discussed. The third parameter is the Fanning friction factor, whose value is typically nearly constant at a value of 0.02 for conditions of interest here. This leaves a fourth dimensionless group, which is (0.5 *V ² / (D*g)). This dimensionless parameter may be referred to as a velocity head dimensionless parameter.

[00118] As discussed herein, for reasons of overall system efficiency, it may be desirable that the frictional loss fraction be limited to about 0.05. The Fluid Properties Parameter may be taken to be 4.73 as representing a working fluid that is a magnesium sulfate solution. The Length Parameter may be taken as 10 for a completely-seafloor-supported system having a typical slope of approximately 6 degrees. As discussed elsewhere herein, it is believed that such a slope is available at a reasonable number of sites along the continental slope of

continents, as well as at other locations that would not be considered part of the continental slope. The Fanning friction factor may be assumed to have a value of 0.02 as discussed. For a fractional loss to have a value of 0.05 as discussed, the product of 4.73 * 10 * 0.02 *

velocity head dimensionless parameter must equal 0.05. This means that the velocity head dimensionless parameter has to be 0.052. Using the value of g as 9.8 m/sec^, and assuming a connecting conduit diameter of 1 meter, and solving for Velocity, this implies that the

velocity in the connecting conduit 300 has to be about 1 m/s.

[00119] If the Length Parameter were 1 (representing a purely vertical connecting conduit)

instead of 10, then the allowable velocity would be about 3.1 m/s. Possible variables that could affect these calculated velocities are the density of the working fluid, and the diameter of the connecting conduit. It can be noted, for scaling purposes, that if the diameter of the connecting conduit were quadrupled, for example 4 meters rather than the 1 meter just assumed, all of these velocities would be doubled, respectively, to 2 m/s and 6 m/s. Also, if one were to use a working fluid that is more dense than the assumed specific gravity of 1.3 that is typical of a magnesium sulfate solution, then the allowable velocity would also increase. It can be appreciated that choosing a working fluid with as large a density as possible not only increases the stored energy for a given volume of bag (as described by the Energy Storage Parameter), but also has an effect on the sizing of the connecting conduit (as described by the Fluid Properties Parameter). In both respects, a larger density of the working fluid is beneficial.

[00120] If the system is a completely-seafloor-supported system, it requires that the Length

Parameter be somewhat large (depending on the seafloor slope), but the value of that Length Parameter could still be in the same range that already exists in some of the longer-tunnel- length conventional pumped hydro power installations that have already been built on land. In embodiments of the invention, providing the connecting conduit 300 might be easier than for a conventional on-land pumped hydro system, because the system of embodiments of the invention would only involve laying connecting conduit (possibly even connecting conduit that is flexible), as opposed to boring a tunnel through rock as is often the case for conventional pumped hydro storage installations on land. However, compared to on-land conventional pumped storage hydro, the at-sea hydro system does suffer from the increased value of the Fluid Properties Parameter. Given these known values or constraints and the desire to limit the Frictional Loss Fraction to a desired value, the velocity head dimensionless parameter has to be chosen appropriately to achieve an appropriate frictional loss fraction, such as by providing for operation using a fluid velocity that is sufficiently small.

Parameters involving time, and sizing of bag and connecting conduit

[00121] Another quantity that is useful in characterizing and comparing marine pumped

storage systems, especially those with non-vertical connecting conduits, is the quantity of how long it takes for working fluid to flow the length of the connecting conduit 300. If conduit is 10,000 m long, liquid flowing at 1 m/s takes 10000 seconds or 2.8 hours to flow the length of the connecting conduit.

[00122] It may also be useful to compare the size, in some sense, of the connecting conduit

300 to the size, in the same sense, of the upper and lower bags 101, 201. From an overall system perspective, it may be desirable that the bags 101, 201 be a more major part of the system than the connecting conduit, in terms of volume of working fluid and other factors that are descriptive of the relative size of the bags 101, 201 and the connecting conduits 300.

In this sense, it may be desirable for the discharge time of the system to be greater or even substantially greater than the transit time for liquid to transit the connecting conduit 300. For example, a 3 hour transit time for liquid to transit the connecting conduit 300, as just

calculated, could be used with a duty cycle that has a discharge time of 9 hours during a day, with a corresponding charging time of 15 hours. In this context, discharge time refers to fully emptying the upper storage region.

[00123] Another descriptive ratio is the ratio of the volume of the connecting conduit to the

storage volume of the storage region. It can be appreciated that this ratio is equal to the

transit time of a particle of working fluid to transit the length of the connecting conduit, divided by the ratio of the operating time to drain or fill the storage region. The ratio of operating time to transit time describes how many times the volume of the connecting conduit passes a given point during the discharge cycle.

[00124] Some simple relations exist among the transit time for fluid to transit the length of the

conduit.

Transit time = Lconduit / velocity

Time duration of discharge (or charge) = volume of storage region / volumetric flowrate

Transit time = Lconduit / [ (volume of storageregion /dischargetime)/conduitcrosssecarea ]

Volconduit / Volstorageregion = conduittransittime / dischargetime volfrac= conduitxsec*Lconduit / volstorageregion

The volume of the storage region can further be described as

Volstorageregion = Energy / (deltarho*g*H)

[00125] Taken together, these considerations suggest that it might be desirable to limit the

length of the connecting conduit 300 to not much more than approximately 10 km.

[00126] It is of interest to know how the volume of the connecting conduit 300 compares to

the volume of the storage region because the volume of the connecting conduit 300 has

associated with it some cost of preparing the working fluid that occupies the connecting

conduit 300, but that fluid is unavailable for energy storage because only the volume of the storage region itself is available for energy storage. So, a preferable value of Volconduit /

Volstorageregion is less than one and preferably well under one.

[00127] The above relationships are useful relationships for some purposes, but none of them

takes into account the need for the connecting conduit 300 to be sized appropriately so that the lossfraction is such as to produce a system that has a useful efficiency. In general, the value of Volconduit / Volstorageregion can easily be reduced by reducing the diameter of the connecting conduit 300, but if that goal is pursued too far, that is unfavorable for reasons related to the pressure drop and the Lossfraction.

[00128] So, it is possible to start with the equations

Volconduit = crosssecarea * Lconduit

Volstorageregion = Energy / (deltarho * g * H) Volratio = Volconduit/V olstorageregion = (crosssecarea * Lconduit) / [Energy / (deltarho * g * H)] Volratio = (crosssecarea * Lconduit)* (deltarho * g * H) / Energy

[00129] Next, it is possible to take the equation

Lossfraction = (FPP)) *(Lfactor) * f * (V ² /(2*D*g))

and make the substitutions Power = V * crosssecarea* deltaPressure or

Power = V * crosssecarea* deltarho * g * H

Solving this equation for V which is velocity

V = Power / (Crosssecarea * deltarho * g * H)

and substituting into the equation for Lossfraction,

Lossfraction = (FPP) *(Lfactor) * f * ( V ² /(2*D*g))

Substituting this expression for V into the equation for lossfraction

Lossfraction = [FPP *Lfactor * f] * ( 1 /(2*D*g)) * [ Power / (Crosssecarea * deltarho * g * H)] ²

Lossfraction = [FPP *Lfactor * f] * ( 1 /(2*D*g)) * Power ^A 2 / [(Crosssecarea ^A 2 * (deltarho * g * H] ² ]

Solving for crosssecarea ²

Crosssecarea ^A 2 = [FPP *Lfactor * f /(2*D*g)] * Power ² / [(Lossfraction * (deltarho * g * H) ² ]

Crosssecarea ^A 2 = [FPP *Lfactor * f /(2*D*g)] * Power ² / [(Lossfraction * (deltarho * g * H) ² ]

Crosssecarea ^A 2 = [Power ² /(deltarho * g * H ) ² ]* [FPP *Lfactor * f / (Lossfraction*2*D*g)

Crosssecarea = [Power /(deltarho * g * H )]* [FPP *Lfactor * fj 0.5 / (Lossfraction*2*D*g)0-5

Substituting this into the equation for Volratio, Volratio = (crosssecarea * Lconduit)* (deltarho * g * H) / [Energy]

Lconduit* (deltarho * g * H) *crosssecarea

Volratio =

Energy

Lconduit* (deltarho * g * H) * Power [FPP *Lfactor *

Volratio ^:

Energy (deltarho * g * H ) (Lossfraction*2*D*g) 0-5

Power * Lconduit [FPP *Lfactor * fj 0.5

Volratio

Energy (Lossfraction*2*D*g) 0-5

we obtain the equations

Power * [FPP *Lfactor * fj 0.5 *Lconduit

Volratio

Energy [2*Lossfraction] 0-5 *g0.5 DO-5

Using the definition Lconduit = Lfactor * H,

Power [FPP *Lfactor * *Lfactor * H

Volratio ^:

Energy [2*Lossfraction] °- ⁵ *g ⁰ - ⁵ D ⁰ - ⁵

Power * [FPP * fJ ⁰ - ^{5 ♦} Lfector ¹ - ⁵ * H

Volratio ^:

Energy [2*Lossfraction] °- ⁵ *g ⁰ - ⁵ D ⁰ - ⁵

Power *Lfactor ¹ - ⁵ * H * [FPP

Volratio ^:

Energy _D 0.5 * [2*Lossfraction]°- ⁵ *g ⁰ - ⁵

*Lfactor ⁰ - ⁵ * H * [FPP * fj °- ⁵

Volratio

dischargetime * * [2*Lossfraction] 0-5 *g0.5 It can be seen that the Volratio is a function of a number of design parameters simultaneously, seemingly without further opportunity for simplification of the equation. However, what can be observed is that one of the strongest dependencies of Volratio is that Volratio is proportional to the 1.5 power of Lfactor. Lfactor (which is the Length Parameter discussed elsewhere herein) is closely related to the slope angle of the seafloor on which the system is installed. Thus, a steeper seafloor slope is beneficial.

[00131] The fact that the Volratio increases with at least the first power of Lfactor is not

surprising. The fact that Volratio is related to a power of Lfactor greater than one may be due to the fact that as the length of the connecting conduit 300 increases, some further increase in the diameter or cross-sectional area of the connecting conduit 300 is also necessary in order to keep the lossfracton at a desired value. That diametral increase would further contribute to the volume of the connecting conduit (Volfrac) beyond just the contribution due to the increased length of the conduit.

[00132] Also, as would be expected, increasing the volume of the storage region improves

(i.e., decreases) the value of Volratio.

[00133] The value of Volratio is proportional to H. This could be expected, because the

volume of the connecting conduit 300 would increase directly with the length of the connecting conduit 300, while the loss fraction (which entered into the derivation) would be independent of such a change.

[00134] Yet another comparison of design variations could be made on the basis of surface area of the respective components (upper bag 100, lower bag 200 and connecting conduit 300). Surface area is relevant to the amount of fabric needed for construction, which is relevant to cost.

[00135] It can be realized that both the bags and the connecting conduit 300 may require

significant amounts of fabric or related material. However, the service conditions in regard to internal pressure are a little bit different, when comparing the bags 100, 200 on the one hand, and the connecting conduit 300, on the other hand.

[00136] The bags 100, 200 only have to withstand an internal pressure that is essentially

determined by the vertical dimension of the bag (i.e., the elevation difference between the top and the bottom of the bag, as installed), multiplied by the density difference of the fluids (working fluid density minus seawater density).

[00137] In contrast, the connecting conduit 300 has to withstand internal pressure related to the entire elevation difference between the two bags 100, 200. This internal pressure inside connecting conduit 300 is not constant throughout the connecting conduit 300, but rather increases with increasing depth below the upper bag 100. Fortunately the connecting conduit 300 has cross-sectional dimensions that are not too large, as it relates to calculation of hoop stress in the wall of the connecting conduit. [00138] It can be expected that the trend of how large the connecting conduit 300 is compared to the bags in terms of surface area of the respective components will be generally consistent with the trend indicated by Volratio, i.e., how large volumetrically the connecting conduit is compared to the bags.

Example 6: Bathy metric information

[00139] In general, for siting a system of embodiments of the invention, what is desirable is a combination of deep water, and closeness to shore, and (for a completely-seafloor-supported system) the slope of the seafloor.

[00140] Depth of the water is desirable because the storage of energy increases in direct proportion with increasing depth, while it is believed that the capital cost of the system increases much more slowly with increasing depth.

[00141] Distance from shore is relevant because being close to shore reduces the difficulty of transmitting electricity between the on-shore electric power grid and the energy storage system. In particular, it is known that when Alternating Current cables are submerged in seawater (which is somewhat electrically conductive), the Alternating Current can induce eddy currents in the seawater, resulting in some loss of the electrical power that is being transmitted. The use of coaxial cables could also be considered. Such loss of power as a result of eddy currents does not occur with Direct Current, which can be transmitted in the form of High Voltage Direct Current. Of course, there is a tradeoff in that the use of HVDC introduces the need for power conversion equipment such as rectifiers and inverters in order to produce HVDC and to convert it back into Alternating Current. The possible need for HVDC, rather than AC, depends on the distance over which the power needs to be transmitted, particularly the distance for which the cable is submerged in seawater. It is believed that as an estimate, for Alternating Current, the losses would be tolerable for an underwater cable length less than about 100 km. In the case of a completely-seafloor- supported system, the electrical components might be located as far away from shore as the lower bag 200, or at least would typically be located farther from shore than the upper bag 100.

[00142] The difficulty of power transmission to and from shore may be more or less relevant depending on a particular installation. If a system of an embodiment of the invention is deployed near or is used in connection with offshore wind turbines, such as floating offshore wind turbines that are located some distance offshore, then the problem of transmission of power to shore is already inherent in the use of the wind turbines themselves. [00143] Finally, for a system that is a completely-seafloor-supported system as described herein, it is advantageous to have as steep a seafloor slope as possible. As already discussed herein, steepness of slope helps to reduce flow losses by reducing the length of the connecting conduit 300 as described by the Length Parameter (Lfactor).

[00144] Furthermore, having a seafloor slope that both is steep and also continues steeply for a substantial elevation change is desirable for reasons that are consistent with discussions herein.

[00145] For conventional pumped storage hydro installations on land, a typical elevation

difference between the two reservoirs or bodies of water is about 300 to 400 meters. Some pumped-storage installations have smaller elevation differences, and a few pumped storage installations with the largest elevation differences reach an elevation difference of 1000 meters or slightly more. As discussed, because in embodiments of the invention the difference in specific gravities between the working fluid and seawater is likely to be less than one, it may be desirable to have an elevation difference that is greater than the elevation difference of pumped storage installations on land. As general information, the Atlantic Ocean has an average depth of about 3 km and a maximum depth of 8 km. The Pacific Ocean has an average depth of about 4 km and a maximum depth of 10 km. In the Gulf of Mexico, the maximum depth is about 3 to 4 km.

[00146] From the study of oceans it is known that in general, near the coasts of continents, the seafloor topography includes (in sequence going away from land) the following features: the continental shelf, the continental slope, the continental rise, and the abyssal plain.

[00147] First, and closest to shore, there is the continental shelf. The continental shelf usually extends to a depth of at most only several hundred meters, and tends to have a slope that is fairly shallow. The extent of the continental shelf, in terms of distance from shore, varies quite widely among various continents and places.

[00148] Much further out at sea, the abyssal plane also tends to have slope angles that are very small or even substantially flat. In general, both the continental shelf and the abyssal plane tend to have seafloor slopes that are less steep than what might be desired for a completely- seafloor-supported pumped storage system.

[00149] However, the situation is different at the continental slope, which is available off the shore of most continents. At the continental slope of the ocean, slope angles in the desired range have a somewhat widespread availability.

[00150] Finally, the continental rise is located between the continental slope and the abyssal plain, and is sort of a transition region between the continental slope and the abyssal plain. The continental rise tends to have a slope that is less steep than the slope of the continental slope, and is not considered further herein.

[00151] Some literature suggests that, as a very general observation, with respect to plate

tectonics of the earth, for shorelines that are near boundaries of subduction zones of tectonic plates, the width of the continental shelf is small, i.e., the continental slope is close to the shoreline, and such places are more likely to have a seafloor slope that is relatively steep. This would suggest, for example that the West Coast of the United States is a relatively good candidate for suitable underwater sites, as compared for example to the East Coast of the United States. Similarly, it is believed that islands, especially islands that are of volcanic origin, are likely to have a seafloor slope that is relatively steep and/or close to land. For example, within the United States, Puerto Rico and Hawaii could be such candidates for having suitable seafloor topography. In general, islands also tend to have inherently high energy costs due to lack of fossil fuel natural resources within their territory.

[00152] A journal article ("What is the slope of the U.S. continental slope?" by Lincoln F.

Pratson and William F. Haxby; Geology; January 1996; v. 24; no. 1; p. 3-6) compiles a large amount of data about slope of the continental slope for five regions along the coasts of the United States. The regions are New Jersey - Maryland; Florida; Louisiana; California; and Oregon. The cited paper finds distinctive differences between the five regions, and the article attributes the differences among the five regions variously to geology and to history.

Specifically, the cited paper suggests that along the entire western U.S. margin, seismicity is an important trigger for failure and slope readjustment, and this accounts for slopes in those locations being shallower than would otherwise be expected. (This observation is somewhat different from general suggestions that appear in some other literature regarding plate boundaries and slopes.) The cited paper also suggests that in regard specifically to the coastline in the region of Louisiana, irregular movement related to the salt that underlies the Louisiana region contributes to relatively shallow slope near Louisiana. The slope in Florida region is attributed to geological composition that is unique to the Florida region. The cited paper finds in general there is somewhat greater slope on the East Coast of the United States compared to the West Coast. Graphical data from the cited paper is reproduced in Figures 5A-5E.

[00153] A general conclusion from the cited paper is that in any region of the continental slope of the United States, there is a range of slopes at different local places, and, in any of the five regions surveyed, it is possible to find at least some places having slopes greater than, for example, 6 degrees. As discussed elsewhere herein, 6 degrees gives a length-to-elevation ratio or Length Parameter or Lfactor of about 10, which is considered useful for purposes of a completely-seafloor-supported embodiment of the invention. Reading from graphs in the cited paper, in the New Jersey - Maryland area, a slope of at least 6 degrees is available at about half of the continental slope. In Florida, approximately the same is true. In Louisiana, a slope of at least 6 degrees is available at only about 12% of the continental slope. In California, a slope of at least 6 degrees is available at about 30% of the continental slope. The same is also true for Oregon. It can be appreciated that an availability of a suitable slope at even a fraction of the continental slope still provides a large number of sites that could be suitable for a system of the completely-seafloor-supported variety. Further selection or ranking of sites could additionally take into account other site features such as closeness to shore, water depth and overall change of elevation from the top of the continental slope to the bottom of the continental slope. In Figures 5A-5E, the second graph in each group provides information about available depth.

[00154] In general, based on the cited paper, it can be expected that in many regions of the seafloor off the coast of the United States and elsewhere, there are at least some places, and potentially fairly many places, at the continental slope that have seafloor slopes that are potentially useful for a completely-seafloor-supported system of embodiments of the invention. The possible positioning of an embodiment of the invention on the continental slope is illustrated in Figure 6. It can be noted that in Figure 6, the vertical dimension is exaggerated, so that the slopes as they appear in the figure are not realistic.

[00155] Additionally, apart from the continental slope (which was the focus of the cited

paper), there are various specific places in the ocean that would not be considered part of the continental slope, where, as a function of local topography, the local slope of the seafloor is appropriately steep. For example, islands may have appropriate seafloor topography while not being considered part of the continental slope. Also, there may be local steeply-sloped features that are not part of the continental slope, such as underwater canyons, seamounts, and sinkholes.

[00156] Examples of localized underwater features having significant depth, significant slope, closeness to shore, or combinations of these features, include the following locations. Many of these topographical features are underwater canyons, some of which are located at the edge of the continental shelf. With respect to the mid-Atlantic coast of the United States, and close to the New York City area in particular, there is the Hudson Canyon, which is an extension of the Hudson River underwater. Hudson Canyon has water depths in the range of

1 km to 2 km. Off the coast of California there are at least the following underwater canyons and features: San Pedro Basin (900 meters deep, while being close to shore); Mugu Canyon (Point Mugu, 730 meters deep); Guide Seamount near San Francisco (adjacent to water 1600 meters depth); Los Angeles Canyon; Scripps Canyon (164 meters deep); Hueneme Canyon (near Ventura, CA, 300 m deep); Lompoc Canyon; Redondo Canyon (400 meters deep), Dume Canyon (300 meters); Newport Canyon; Santa Monica Canyon; and La Jolla Canyon (300 meters). Still other examples are places (Tacoma, for example) in Puget Sound, Washington state, and Fishers Island near Connecticut.

[00157] It is even possible that within a desired generally sloping topography, some local flat spots could perhaps be found, at which the bags could be placed, so that it would not be necessary to provide either bag design for use on a slope or site preparation to create a flat space. Alternatively, such local flat spots could be created through dredging or similar means. Alternatively, the bags could be placed on resting spots that are sloped, with the bags having appropriate design features as described herein.

[00158] It can also be noted how close the continental slope is to shore (i.e., how short the continental shelf is) in various places. On the East Coast of the United States, some places have a fairly wide continental shelf. On the East Coast, one of the closest approaches of the continental slope to land is at Cape Hatteras, NC. On the West Coast of the United States, there are places in southern California where the continental slope is close to land. In island locations such as Hawaii and Puerto Rico, there are also places with deep water close to shore.

Example 7: Composition of Fluid

[00159] Another aspect of interest for embodiments of the invention is the composition of the working fluid.

[00160] In embodiments of the invention, the working liquid may be an aqueous solution that is more dense than seawater. It is possible that brine directly as discharged from a desalination process could be used as the working liquid in embodiments of the present invention. Brine as discharged from a desalination plant is a form of concentrated seawater that is somewhat easily available and is sometimes considered a waste product. Typically brine that is discharged from a desalination plant is discharged at a salinity of close to 13% by weight, which is roughly half of the saturation concentration of NaCl in water. The corresponding specific gravity of that brine solution is about 1.1 , which is not as large as might be desired. [00161] It is further possible that brine as discharged from a desalination plant could be used as the starting point for further concentration to larger densities beyond the composition typically present in desalination plant effluent. It is possible that the working liquid could be a form of concentrated seawater. Seawater can be concentrated by evaporation or other processes or by a combination of processes to form a liquid having a specific gravity larger than that of ordinary seawater. As discussed herein, concentrated seawater can reach a specific gravity of 1.2 (for salt-point brine) or even as high as 1.3 (for bittern).

[00162] Producing the working liquid directly from seawater may have several advantages.

The working liquid may be created on-site from the ocean, which is a resource that surrounds or may be located close to the apparatus and is essentially infinite in size. (The terms ocean and seawater are intended to also include inland bodies of water that have significant salt content.) Seawater would be readily available for production of additional working liquid as needed, and use of seawater would avoid the need for provision of specialized chemicals such as from on-shore facilities. Production of the working liquid directly from seawater could be done either for initial filling of the system or for make-up of any loss or dilution of any of the working liquid. Associated with the idea of producing the working liquid directly from seawater would be an expectation that the production of the working liquid directly from seawater would be relatively inexpensive. The starting material, seawater, would be essentially free and in essentially infinite supply, although processing steps would be required. Another consideration is that, if there were ever a leak or any type of discharge of the working liquid into the ocean, such leak or discharge should not be considered to be a toxic pollutant, because the substances that would be entering the ocean would be substances that originally came from the ocean, and after mixing and equilibration of the leaked or discharged liquid, the composition of the ocean would be essentially indistinguishable from the ordinary ocean. There would not be the presence of any substance that is any more toxic than any substance that occurs naturally in ordinary seawater. It is true that marine life is affected by brine concentration, but eventual dilution of brine by seawater would leave essentially seawater. If the seawater were concentrated to less than the salt point (which is a specific gravity of 1.2185), then its composition would very closely resemble that of ordinary seawater, while just being somewhat more concentrated, and with only calcium carbonate and calcium sulfate, both of which are minor constituents of seawater, having been removed. If the seawater were concentrated until it is bittern, then its composition would be fairly different from that of seawater, because the largest-proportion solute component of seawater (sodium chloride) would have been removed, but still the bittern would only contain substances that are constituents of naturally occurring seawater.

[00163] In order to produce bittern at or near approximately its maximum density, it is

necessary that quite a large portion of the liquid water of natural seawater be removed, and it is also necessary that a majority of the dissolved solutes (especially NaCl) be precipitated out and removed. It is, of course, possible that all of the required removal of water could be performed by evaporation. One such process is spray evaporation described elsewhere herein. Another evaporative process would be evaporation from shallow pools that are exposed to sunlight or warm air. If evaporation were used for the entirety of the

concentration process starting with ordinary seawater and ending at highly concentrated bittern, the amount of water evaporation required would be a large multiple of the quantity of the working liquid or bittern that is needed, and that needed amount of liquid or bittern may itself be a large quantity. Another possibility would be the use of a desalination process such as reverse osmosis in order to achieve some of the initial concentration by a means other than evaporation. The desalination process might achieve the initial increase in salinity from the salinity of ordinary seawater to the salinity of desalination effluent (such as to a specific gravity of about 1.1). After that, to achieve salinity greater than the salinity of desalination effluent, an evaporation process or other process may be used to further concentrate the desalination effluent.

[00164] It is also possible that the working liquid may be an aqueous solution that may contain various components of seawater and may further contain an amount of some other added constituent, such as any of those mentioned herein or others. It is also possible that the working liquid may be an aqueous solution of two or more of the substances mentioned herein. It may be that combinations of more than one solute may be more advantageous (achieve higher density) than a single-solute solution of either of the solutes alone. When combining more than one solute in a solution, it is possible that some combinations of solutes might undergo a reaction with each other such as a double replacement reaction, resulting in reaction products that are different from the starting components. It may be desirable to avoid combining solutes that will undergo a double replacement reaction with each other, or it may be desirable to avoid combining solutes that undergo a double replacement reaction that produces a reaction product that precipitates out of solution.

[00165] Another consideration may be to avoid providing a solution that is perfectly fully saturated. Providing a solution that is slightly less than fully saturated might lessen the possibility of precipitation of solute in undesired places or at undesired times if there should be any unanticipated evaporation or a change of temperature that might cause precipitation. For example, if it is desired to have a working liquid that resembles salt-point brine, it might be possible to concentrate seawater up to the salt point and then add back in a small amount of pure water or ordinary seawater or similar additive, in order to dilute the concentration to just slightly below saturation, or the concentrating process could be stopped shortly before saturation is reached. Similarly, if the working liquid is or resembles bittern, it would be possible to produce bittern and then add back in a small amount of pure water or ordinary seawater or similar additive, in order to dilute the concentration to just slightly below saturation.

[00166] In view of the large amount of liquid that would likely be needed for large-scale energy storage, the actual composition of the working liquid would likely be chosen based on a combination of its physical properties and economic considerations.

[00167] It is known from oceanographic studies that the composition of seawater is quite consistent throughout the oceans of the world. There are slight variations in overall salinity such as due to local differences in evaporation or due to freshwater inflow or rainfall, but the relative proportion of the various solutes among themselves is quite consistent throughout the oceans of the world. By far the largest solute constituent of seawater is NaCl.

[00168] As an approximate indication of the properties of seawater, and for conditions at which NaCl remains dissolved in water, the solution properties of NaCl in water provide some starting information about the properties of seawater. Figure 7 is a phase diagram of NaCl in water, at atmospheric pressure. It can be seen that at temperatures higher than 0 C, the saturation concentration of NaCl in water is almost independent of temperature within the temperature range of interest. This saturation condition occurs at a weight fraction of about 26% (mass of NaCl / (mass of NaCl + mass of water)), and the density of that aqueous solution is about 1.2185 g/cc. At conditions beyond saturation, for which some undissolved solid is present, the solid can be in the form of either undissolved NaCl, or undissolved solid hydrate of NaCl, depending on the details of the conditions. The saturated aqueous solution of NaCl has a viscosity that is about four times the viscosity of pure water. This is for solutions of pure NaCl in water.

[00169] Although NaCl is the major dissolved solute in seawater, actual seawater is a much more complicated solution containing many solutes and constituents. So, it is worthwhile to consider the conditions that occur starting with natural seawater and progressing as water is gradually evaporated from seawater. Information about sea water and the progressive evaporation of seawater is available in the following references: The Composition of Sea Water and its Concentrates, by Gino Baseggio, Morton Salt

Company, Fourth International Symposium on Salt - Northern Ohio Geological Society (1974);

"Salt for New Zealand" by Jensen Reid, published by Dominion Salt Limited, Pegasus Press, Christchurch, New Zealand; and

Evaporites: Sediments, Resources and Hydrocarbons by John K Warren Springer- Verlag Berlin Heidelberg New York; Library of Congress Control Number 2005932089; 2006

[00170] It can be noted that these references describe the progressive evaporation of seawater by reporting the specific gravity of the liquid that remains. If any precipitation has happened up to a particular point, the result is reported on the basis that the precipitated solids have been removed or are ignored, and the specific gravity that is reported is the specific gravity of the remaining liquid. Figure 8 is an illustration taken from Warren reference.

[00171] According to these references, there are several observable points during the process of water being gradually removed from seawater by evaporation. In some of the references, it is reported that the first substance to precipitate out, as water is gradually evaporated, is CaCC^. Not all of the references report this CaCC^ precipitation as a step in the evaporation process, and apparently the amount of CaCC^ available to be precipitated out is small and possibly variable among seawaters from various sources, presumably due to variations in the amount and type of organisms living in the seawater. After that, at a specific gravity of 1.0897, there is precipitation of the CaSC^ that is present in seawater, and this point is sometimes referred to as the gypsum onset. Baseggio identifies this precipitate as the hydrate of calcium sulfate, CaS04*2H20. This precipitation occurs when the volume of remaining liquid is about one-fifth of the original volume of seawater. After that, at a specific gravity of 1.2185, there comes the point at which NaCl is saturated and any further removal of water would result in the beginning of precipitation of NaCl. This point is referred to as the salt point or halite onset, and the liquid in that condition is referred to as salt-point brine. At the point when salt-point brine begins to exist, the amount of liquid remaining is about one-tenth of the amount of liquid of the original seawater. If evaporation is continued beyond that point, the NaCl precipitates out and the solutes that remain in solution after the beginning of precipitation of the NaCl are a mixture of calcium compounds, magnesium compounds and other substances. Liquid that results from continued evaporation beyond the salt-point is referred to as bittern. Upon continued evaporation of water, this bittern can reach a density that is still larger density than the density of salt-point brine. Warren reports that the bittern reaches a maximum specific gravity of about 1.33 . This condition occurs after an amount of evaporation such that the amount of liquid remaining is about one-seventieth of the amount of liquid of the original seawater. He describes the viscosity of the liquid in this condition as resembling that of olive oil. Warren further reports results for even further evaporation beyond that point, up until a condition where the remaining liquid is about one-hundredth of the amount of liquid of the original seawater. However, this last stage of evaporation seems to result in specific gravity of the remaining liquid decreasing slightly from the just- mentioned maximum value of 1.33 .

[00172] Apart from the just-described compositions that are derived directly from naturally occurring seawater, it is also known that various other substances can form fairly dense aqueous solutions. Many of these are inorganic salts that are chemically rather simple and are somewhat readily available. An aqueous solution of magnesium sulfate can achieve a specific gravity of about 1.3. This is of similar magnitude as the density of the bittern, and in fact magnesium sulfate is a major component in the bittern.

[00173] Another possibility is that the solution could be or could contain a simple inorganic substance that could be produced from seawater, such as a salt that contains anions and cations that are found in some form in seawater. For example, potassium carbonate is not a major constituent of seawater, but potassium compounds and potassium ions are found in seawater, and carbonate compounds and carbonate ions are also found in seawater. With appropriate chemical processing, which could be carried out at sea, a substance such as potassium carbonate could be produced from seawater. Similarly, the specific gravity of an aqueous solution of calcium chloride can be at least as large as 1.4 . Even though calcium chloride is not naturally present in seawater in significant concentration, it has chemical similarity to sodium chloride, which forms a major part of the solutes in seawater, and other calcium compounds are naturally present in seawater.

[00174] Within the category of simple inorganic salts is the category of metal halides. A

sodium bromide (NaBr) solution can reach a specific gravity of about 1.41 . The specific gravity of a saturated lithium chloride (LiCl) aqueous solution is about 1.47. The specific gravity of a saturated solution of potassium iodide (KI) is about 1.72 . Calcium bromide (CaBr2) can form a brine of specific gravity 1.81 . Zinc bromide (ZnBr2) can form a brine of specific gravity 2.3. Most of these just-mentioned chlorides, iodides and bromides are commercially used as brines in drilling in the oil and gas industry. It is believed that in general, halide salts are somewhat more corrosive than other solutes. [00175] The working liquid (more-dense liquid) could be any of the solutions discussed herein, and combinations thereof.

[00176] Still other dense aqueous solutions are possible with some other slightly more unusual salt solutes. A category of salts that sometimes have large solution densities is formates (salts of formic acid) or acetates (salts of acetic acid). Saturated solutions of a density of 1.6 to 2.26 g/ml are obtained when cesium- rubidium formate solution is mixed with a saturated potassium formate solution. An 83.1% solution of cesium formate in water solution has a specific gravity of 2.3775. A cesium formate/cesium acetate blend can achieve a brine density to 2.42 g/cm3. An 85.6 gram solution of cesium tungstate in 100 grams of water has a specific gravity of 3.0133. A cesium-, rubidium-tungstate solution has a density of 2.95 g/ml at 20° C. An aqueous solution of sodium polytungstate can achieve an aqueous solution density as great as 3.1 . A solution of lithium metatungstate (Lig (¾Wi 2 04o)*3H20) is reported in US Patent No. 5,328,035 as having a maximum specific gravity of 3.5. Some of these just-mentioned high-density aqueous solutions have been used in mineral separation by flotation.

[00177] It is also possible that the more-dense liquid may be an aqueous solution that may contain various components of seawater and may further contain an amount of some other added constituent, such as any of those mentioned above or others. It is also possible that the more-dense liquid may be an aqueous solution of two or more of the substances mentioned herein. It may be that combinations of more than one solute may achieve higher density than a single-solute solution of either of the solutes alone. When combining more than one solute in a solution, it may be desirable to avoid combining solutes that will undergo a double replacement reaction, or at least avoid combining solutes that will undergo a double replacement reaction that produces a reaction product that precipitates out as a solid.

[00178] Properties that may be desirable for a working fluid include having a fairly high

density in aqueous solution, being not excessively corrosive, being not toxic or otherwise harmful, being readily available and inexpensive, etc. It is possible that several of these factors in combination could determine a choice of working fluid.

[00179] Table 5 presents a list of some possible solutes and their solution densities.

Table 5

Specific gravities of aqueous solutions of candidate substances Note that most of these data points are simply

the largest specific gravity that is listed in the reference,

with no indication that the conditions are fully saturated solutions

Density listed is at 20C if more than one temperature is available

Substances are loosely arranged so that substances toward the top of this list are believed to be more realistic candidates in terms of some combination of

density, availability, price, and not being too corrosive

Specific gravity reported for solution

Chemical Chem Phys 66

Substance Formula Handbook substances Other source

Ferric sulfate Fe2(S04)3 1.80 1.70

Ferric chloride FeC13 1.55 1.42

Potassiumcarbonate K2C03 1.54 1.54

Magnesium sulfate MgS04 1.30 1.30

Calcium chloride CaC12 1.40 1.39

Magnesium

chloride MgC12 1.27 1.28

Sodium chloride NaCl 1.20

Calcium nitrate Ca(N03)2 1.75 tempdep

Sodium thiosulfate Na2S203 1.38 1.38

Aluminum Sulfate A12(S04)3 1.31 1.37

Ammoniumsulfate (NH4)2S04 1.28 1.23

Cesium chloride 1.88

Lithium chloride LiCl 1.18

Potassium

bicarbonate KHC03 1.17

Potassium bromide KBr 1.38 1.37

Potassium chloride KC1 1.16 1.16

Potassium

hydroxide KOH 1.54 1.50

Sodium hydroxide NaOH 1.53 1.43

Potassium iodide KI 1.40

Potassium sulfite K2S03 1.24

Sodium bromide NaBr 1.41 1.41

Sodium carbonate Na2C03 1.33 1.16

Sodium sulfate Na2S04 1.23 varies widely w temperature Sodium nitrate NaN03 1.37

Strontium chloride SrC12 1.41

Barium chloride BaC12 1.28

Ferrous sulfate FeS04 1.21

Zinc bromide ZnBr2 2.00 2.31

Zinc chloride ZnC12 1.96

Sodium bromide 1.52

Calcium bromide 1.82 Cesium formate 2.31

Refs.: Solution densities for 66 substances, which references

1. Wolf, A. V., Aqueous Solutions and Body Fluids, Hoeber, 1966.

2. Sohnel, O., and Novotny, P., Densities of Aqueous Solutions of Inorganic Substances, Elsevier, Amsterdam, 1985.

[00180] The solutions listed in Table 5 are all single-component solutions. Further, it is

contemplated that there might be multi-component solutions, such as solutions of simple salts such as the above substances, that achieve higher solution density than is achieved by any of the individual components in that solution.

[00181] It is further noted that there is one simple way to make a multi-component solution for which there is no worry about a double replacement reaction occurring. That is to have one ion of the salts be the same for the various components of the solution, i.e., a common ion. For example, if all of the solutes are sulfates, no replacement reaction will take place. For example, two sulfates that have fairly large solution densities are magnesium sulfate and ferric sulfate. A possible working fluid would be a two-component solution containing ferric sulfate and magnesium sulfate. Or, it would be possible to use a solution in which all constituents are magnesium salts or all potassium salts. Replacement reactions would not take place in those situations, either.

[00182] In embodiments of the invention, there may be provided a solution that is slightly less than fully saturated, so as to lessen the possibility of precipitation of solute in undesired places or at undesired times if there should be any unanticipated evaporation or a change of temperature or some other factor that might cause precipitation. For example, if it is desired to have a solution whose density is close to the maximum density of bittern, it might be possible to concentrate seawater to form bittern having a density slightly larger than the desired density and then add back in a small amount of pure water or ordinary seawater or similar additive, in order to dilute the concentration to just slightly below saturation. In any of these situations, the composition of the more-dense liquid could be described as having a density that is more than 90% but less than 100% what would be the density of a saturated solution of the same solutes in aqueous solution. The composition of the more-dense liquid could be described as having a specific gravity that is greater than 1.18, possibly with all solute constituents being substances that occur in seawater. The composition of the more- dense liquid could be described as having a specific gravity that is greater than 1.30, possibly with all solute constituents being substances that occur in seawater. The composition of the more-dense liquid could be described as having a specific gravity that is greater than 1.50, possibly with all solute constituents being substances that can be produced from seawater as a starting point.

[00183] In view of the large amount of liquid that would likely be needed for large-scale

energy storage, the actual composition of the more-dense liquid would likely be chosen based on a combination of the physical properties of the liquid, economic considerations, and environmental considerations.

Example 8: Turbines

[00184] Pumps and turbines used in embodiments of the invention may be designed and sized according to the available head (pressure difference) and other operating characteristics. This can draw on experience in existing hydropower applications. Possible designs include Francis turbines, Pelton turbines and Axial turbines, each of which has particular parameter ranges to which it is best suited. It may be understood that for purposes of discussion of operating parameter ranges, a parameter such as "head" is conventionally discussed or described in terms of the height of a column of ordinary water that is delivered to or from the pump or energy extractor. This is done for convenience because in many instances the liquid being pumped is either water or an aqueous liquid whose properties (especially density) are close to those of water. However, more fundamentally, the parameter "head" refers to pressure of the liquid delivered to or from the pump 400 or energy extractor 500, essentially the pressure drop across the pump 400 or energy extractor 500 when it is operating. For purposes of embodiments of the present invention, the relevant pressure would be the pressure produced by the column of the more-dense liquid that is in excess of the pressure of the less-dense liquid (which may be seawater) at a the same location or depth. So, for embodiments of the present invention, "head" does not generally refer to actual physical depth of the pump 400, energy extractor 500 or lower storage region 200 below the surface of the body of water. Rather, approximately, the pressure expressed as head (having units of depth) may refer to the height of an equivalent water column that would produce the same pressure difference. For example, if the more-dense liquid has a specific gravity of approximately 1.325 and the less-dense liquid has a specific gravity of 1.025, giving a difference of specific gravities of 0.30, as is used in certain calculations elsewhere herein, then the actual "head" as a representation of pressure would be about 30% of the actual physical elevation difference between the upper storage region 100 and the lower storage region 200. Additional features of embodiments of the invention

[00185] Returning now to description of embodiments of the invention, there may be

described further features of this embodiment, and also there may be described additional embodiments.

[00186] It can be noted that for embodiments of the present invention, the conduit 300 may be thicker- walled or stronger towards its lower end. This is because the pressure that is contained inside the conduit 300 (relative to the corresponding external pressure of the ocean or body of water) increases with depth below the upper storage region 100.

[00187] A design factor possibly affecting the design of the bag 101, 201 is the possible

location of the bag 101, 201 on a seafloor that has some slope to it. This contrasts with conventional applications of pillow tanks on land, which are generally installed on surfaces that are substantially flat horizontal surfaces. A conventional pillow tank on a flat supporting surface can be expected to fill uniformly and to stay in place. In a conventional pillow tank that rests on a flat level surface, the upper surface of the bag can be expected to be horizontal and parallel to the ground on which the pillow tank is supported. However, if a simple pillow tank is mounted on a sloping surface, it can be expected that as the pillow tank fills in a manner such as to produce a triangular cross-section of the pillow tank, with the size of the triangle increasing as the fill progresses. It can be expected that the upper surface of the pillow tank will be approximately horizontal but it will not be parallel to the bottom surface of the pillow tank or the seafloor. It can also be expected that this could result in a tendency for the bag to roll downhill.

[00188] Referring now to Figure 9, for a bag 101, 201 that is intended for use on a sloping seafloor, it is possible to provide internal tethers 153 inside the bag 101, 201. The bag 101, 201 when at least partially full may have a generally flat top surface and a generally flat bottom surface, and the internal tethers 153 may connect the generally top surface with the generally bottom surface. The internal tethers 153 may be such as to be able to carry tensile load along their length between the top surface and the bottom surface. The internal tethers 153 may be flexible and need not be capable of carrying compressive load along their length. Capability of carrying shear load may also be provided, depending on whether the internal tethers 153 are strips of small cross-sectional area, or whether the internal tethers 153 are sheets or other designs. Internal tethers 153 that are sheets or similar designs may take a situation in which the pillow tank would tend to assume a shape that is a triangular cross- section, and make that into a situation in which the tank maintains a somewhat constant thickness and stays in place rather than trying to roll downhill. If there is any significant elevation difference between different portions of the bag, compared to the nominal thickness of the bag in the vertical direction when full, this situation may create a pressure difference that the bag has to be designed for. This is illustrated in Figure 9.

[00189] It can also be expected that in a sloping pillow tank, the pressure will be nonuniform and may vary over a greater range than would be the case for a flat-mounted pillow tank. In a conventional flat-mounted pillow tank, the pressure is essentially the same at any position along the flat surface on which the tank rests and is determined by the height of the tank when full. For a sloping pillow tank, some of the pillow tank (located at a lower elevation) will be exposed to a larger-than average pressure, while other parts of the pillow tank will not be exposed to so much pressure. The bag 101, 201 (such as its wall thickness) would have to be appropriately designed for such pressure.

[00190] Of course, even though many conventional pillow tanks are made having a footprint that is square or approximately square, for embodiments of the invention it is not necessary that this be the case. In embodiments of the invention, the footprint of the bag 101, 201 could, for example, be rectangular. For example, the proportions could be such that the bag is wider in the transverse direction and not as long in the direction along the slope. Such proportioning may be useful for providing storage volume while not creating as large a pressure in the lower-elevation regions of the bag 101, 201 as might be the case for a bag 101, 201 that is of square footprint.

[00191] If the lower storage region 200 is a pillow tank that is supported on a sloping surface, the lower storage region 200 could have a built-in sump tube. As such, the connecting conduit 300 could connect to the lower bag at a location that is a relatively higher-elevation portion of the lower bag, because such a location may be a convenient place to connect, given the likely arrangement of various components. There could be a sump tube built into the bag such that the connecting conduit 300 can draw liquid from the lowest-elevation part of the bag even if the connecting conduit connects to the upper part of the lower bag 201. This is illustrated in the lower bag 201 in Figure 9.

[00192] Referring now to Figure 10, it is illustrated that upper storage volume 100 and lower storage volume 200 could have a shape that could be described as resembling a glove.

[00193] In such a design for the upper storage region 100, the upper storage region 100 may comprise a manifold 136 and a plurality of sub-regions 138 in communication with the manifold 136. The sub-regions 138 may, for example, be generally cylindrical when full. A cylindrical shape may be chosen, for example, based on the fact that a cylinder stores more volume for a given amount of surface area (i.e., fabric) than does a conventional pillow tank that has a typical aspect ratio of 1 : 1 : 0.1. As illustrated, if such an upper storage region 100 is installed on a sloping seafloor, the manifold 136 could be located at a lower elevation than the sub-regions 138, and the connecting conduit 300 could connect to the lower portion of the manifold 136. The sub-regions 138 are illustrated as extending upward away from the manifold 136 along the sloping seafloor. If desired (not illustrated), there could be provided isolation valves where the individual sub-regions join the manifold.

[00194] In such a design for the lower storage region 200, if the lower storage region 200 is installed on a sloping seafloor, the manifold 236 could be located at a lower elevation than the sub-regions 238. The sub-regions 238 are illustrated as extending upward away from the manifold 236 along the sloping seafloor.

[00195] In this situation, it could be considered that connecting conduit 300 extends to the manifold 236, or alternatively, it could be considered that there could be provided a sump tube 266 that runs parallel to the tubes that form the sub-regions 238 such as approximately the same length as the sub-regions 238, allowing connecting conduit 300 to be connected to that sump tube 266. It can be noted that it may be desirable for such sump tube 266 to be collapse-resistant to aid in emptying the lower storage region 200 if the pump 350 is located at a higher elevation than the lower storage region 200. The sump tube 266 could be designed with appropriate internal stiffness to maintain its shape.

[00196] Figure 10 shows the upper storage region 100 and the lower storage region 200 in a condition such that the system is mostly charged but not completely charged. It can be appreciated that if the upper storage region 100 that comprises a manifold 136 and a plurality of sub-regions 138 that are generally cylindrical is installed on a slope, as the upper storage region 100 fills, the sub-regions 138 will fill starting at the lower elevations and progressing upward. In Figure 10, portions of a sub-region 138 that are filled are shown with curved accent lines on the surface of the sub-region 138, indicating that due to internal pressure of the more-dense liquid, that portion of the sub-region 138 assumes a shape that is generally cylindrical. It can be expected that at higher elevations of the sub-regions 138, portions of the sub-region 138 that are not filled with liquid will assume a generally flat (uninflated) configuration. This is indicated by accent lines that are flat. In Figure 10, the sub-regions 138 of the upper storage region 100 are illustrated as being approximately three-quarters full.

[00197] Correspondingly, the lower storage region 200 is illustrated such that its sub-regions

238 are approximately one-quarter full. Again, portions of a sub-region 238 that are filled with the working liquid (more-dense liquid) are shown with curved accent lines on the surface of the sub-region 238, indicating that due to internal pressure of the more-dense liquid, that portion of the sub-region 238 assumes a shape that is generally cylindrical. At higher elevations, it can be expected that portions of the sub-region 238 that are not filled with liquid will assume a generally flat (uninflated) configuration. This is indicated by accent lines that are flat. In Figure 10, the sub-regions 238 of the lower storage region 200 are illustrated as being approximately one-quarter full. As discussed, the sump tube 266 is illustrated as being of a curved shape along its entire length, which is different from the situation of the sub-regions 238. Such curved configuration could be accomplished by appropriate structural design such as by appropriate stiffness or ribs. Such design could ensure that possible suction involved in emptying the lower storage region 200 would not collapse the sump tube 266.

[00198] It would be possible to provide external attachments to the sub-regions 138 or 238 so as to provide some attachment between adjacent sub-regions, in order to aid in maintaining the relative or absolute positions of the sub-regions. For example, such attachment could comprise straps or cables or a net that loosely surrounds the sub-regions and exerts some constraint on the relative or absolute positions of the sub-regions 138 or 238.

[00199] It is further possible that a water hammer damper or absorber may be provided in the circuit of the working fluid, typically near the valve 350. The water hammer damper may comprise an expandable volume that is in fluid communication with connecting conduit 300 between valve 350 and a portion of connecting conduit 300. Because flow in connecting conduit 300 may be bidirectional, a water hammer damper connection may be provided on both sides of valve 350. Such a water hammer damper may make it possible for the overall system to be able to respond to faster transients in the electric power grid. Alternatively or in addition, it is possible that the valve 350 in the system may be designed or operated so that it closes gradually, especially when it is almost fully closed, thereby potentially reducing water hammer effects.

[00200] In a further embodiment of the invention, some or all of the submerged components, such as the upper bag 101, the lower bag 201 and the connecting conduit 300, may be covered with a layer of mud, sand, or other particulate material. Such material may be material that occurs naturally on the nearby seafloor. However, such a material still may be sufficiently deformable or fluent to allow the upper bag 101 and the lower bag 201 to easily expand or contract as they fill or are emptied. Such covering material may serve as a form of protection against possible damage to the bags 101 , 201 or to the connecting conduit 300, such as by impact from any possible cause. Such covering material also may protect those components against encrustation or other damage by marine life.

[00201] An aspect of this embodiment of the invention is the question of exactly where the pump/turbine 400, 500 is located. In conventional on-land pumped storage hydro

installations, the pump/turbine is typically located at one of the lowest points in the fluid flow system. Conventional on-land pumped hydro storage often involves construction of tunnels, whose location and route can be chosen as needed. It is typical that the pump/turbine is located at a slightly lower elevation than the intake from the lower storage region, or that the route of the tunnel slopes slightly downward as one goes away from the lower storage region. A reason for this is to prevent cavitation of the water as the water is being pumped during charging of the energy storage system. Cavitation can occur if the local pressure in a flowing liquid is sub-atmospheric and more specifically is lower than the vapor pressure of the flowing liquid. The local pressure that determines the possible occurrence of cavitation is the local hydrostatic pressure as further influenced by local flow dynamics within the

pump/turbine. If the local pressure is sufficiently low, this can allow the formation of bubbles of the vapor of the flowing liquid. Subsequent collapse of those bubbles can create an erosive action that is damaging to blades and other components of a pump/turbine. It may be appreciated that the possible formation of such bubbles is determined by the local pressure measured in absolute units of pressure measurement.

[00202] In embodiments of the invention, it is of course desirable to prevent cavitation at all locations where liquid flows. This could be accomplished by locating the pump/turbine very close to the lower reservoir or lower storage region, or even slightly lower than the lower reservoir or storage region, similar to what is typically done in conventional pumped storage hydro on land. However, because of the local absolute hydrostatic pressure, it may not be necessary to use that exact location for the pump/turbine. In embodiments of the invention, especially in completely-seafloor-supported embodiments of the invention, it may be possible to locate the pump/turbine at some intermediate location along the length of the connecting conduit. Advantages possibly available from such a design choice are twofold. First, the pump/turbine could be located at a depth that is less extreme than the lower reservoir itself. This might make maintenance and repair of the pump/turbine slightly easier (although it still might be necessary to use remotely operated devices). Second, assuming that the system is located on a sloping seafloor, this would permit the pump/turbine to be located somewhat closer to the shoreline. This could reduce the length of power transmission between land and the pump/turbine, and specifically could reduce the required length of underwater power cable.

[00203] The possibility of cavitation can be considered for several modes of operation. One mode is when the pump is running, another mode is when the turbine is running, and a third mode is when the valve 350 is closed with neither the pump nor the turbine is running. It is helpful to consider first the relatively simple situation of the closed valve. It can be asked what is the absolute pressure just below the closed valve 350. For this, a hypothetical U-tube can be constructed. One leg of the U-tube would contain the height of the column of the more-dense liquid from the lower bag to the closed valve 350. The other leg of the U-tube would contain the entire vertical distance in the ocean from the ocean surface to the lower bag 201. A mathematical constraint would be that there is no pressure difference across the boundary of the lower bag 201. This is realistic given that the lower bag 201 would typically have some looseness and flexibility in its shape even when holding the more-dense liquid. The pressure just below the valve 350, in absolute pressure terms, would be atmospheric pressure plus the pressure of the height of the ocean column from the ocean surface to the lower bag 201, minus the pressure of the column of the more-dense liquid from the bag 201 to the valve 350. If the pressure of the column of the more-dense liquid from the lower bag 201 to the valve 350 essentially equaled the pressure of the height of the ocean column from the ocean surface to the lower bag 201, then the absolute pressure just below the valve 350 would be essentially atmospheric pressure, which would not result in cavitation. If desired, some safety margin could be built in to the system by locating the valve 350 and

pump/turbine 400, 500 at a slightly lower elevation than suggested by this criterion. Similar criteria could be developed for operation of the pump and operation of the turbine, and would involve an additional calculation of pressure drops due to flow. These criteria could be expected to differ from the criterion just derived to a slight extent, but not in any major way. The criterion just derived could, for example, allow the placement of the valve 350 and pump/turbine 400, 500 at some fraction of the water depth at which the lower bag 201 is placed, and could correspondingly shorten the length of the electrical cable by some number of kilometers in a typical completely-seafloor-supported installation.

[00204] On a topic of structural support of a bag, Fig. 11A illustrates a bag 101 or 201, which may be resting on the seafloor. The bag 101, 201 may have horizontal dimensions that are larger than its vertical dimension when full. For example, the vertical dimension when full may be less than 10 meters tall, and the bag 101, 201 might have horizontal dimensions of tens of meters or more. There may be provided some structural supports in the form of posts that are embedded in the ocean floor like fence posts. These supports or posts 182 may provide some reaction force resisting the spreading of the bag 101, 201 radially outward due to the excess density of the liquid within the bag 101, 201.

[00205] The posts 182 may allow the bag 101, 201 to bulge radially outward between the

posts 182. This may impose less stress in the bag (hoop stress) than would occur if the bag 101, 201 were relatively tight between posts. As yet another possibility, it would be possible to build local supporting walls as a perimeter between the posts 182, and the walls themselves would experience bending. In any such configuration where a bag 101, 201 is restrained by isolated posts 182, the local hoop stress in the bag 101, 201 due to radially outward pressure forces would be pressure * local radius / bag thickness. If the bag 101, 201 were stretched somewhat tightly between posts (not as illustrated), then the hoop stress in the bag would be correspondingly larger. On the other hand, if the situation were as shown, the local radius of curvature R of the bulges in the bag 101, 201 would be fairly smaller, and the local hoop stress in the bag 101, 201 would be smaller.

[00206] The posts 182 could have at their lower ends, i.e., the ends that are embedded in the solid material of the ocean bottom, some attachment features. For example, the attachment feature could be a helical thread, which would allow the post 182 to be advanced into the ocean bottom by rotating the post 182 like a screw. Similarly, the thread could be a corkscrew type of thread i.e., a helix not having a solid central region. Such posts could act like a fence, or a containment wall, or a dam.

[00207] Referring now to Figure 1 IB, it is illustrated that for a bag that rests on a sloping

seafloor, posts 182 could form an anchorage that can react some load to counteract the tendency of the bag to roll downhill. The posts could act somewhat like a dam.

Fixed Impoundments

[00208] Referring now to Figure 12, there is shown yet another embodiment of the invention.

In this embodiment, both the upper storage region and the lower storage region are confinement regions that are underwater and that rest on the floor of the body of water, at different elevations. The confinement regions are shown as being substantially rigid structures rather than bags. Such structures could, for example, resemble dams. In Figure 12, the upper storage region is almost full and the lower storage region is almost empty, and energy is stored. The opposite condition would exist when energy has been regenerated and discharged. [00209] This embodiment of Figure 12 is suitable for particular geometries of the floor of the body of water, and in this sense it resembles pumped storage hydro installations on land. It assumes the availability of two places on the floor of the body of water, that are each suitable for use as a confinement region, with the two confinement regions being separated from each other by a suitably large elevation difference, while being separated from each other by a horizontal distance that is suitably small. This is a more specialized seafloor topography than is required by some of the other embodiments of the invention. A seafloor geometry that may be appropriate is a geometry in which there is a sharp change of depth. An example of such would be a canyon on the seafloor. Geographic examples of submarine canyons are discussed elsewhere herein. A more-dense liquid may be stored in both the upper

confinement region and the lower confinement region, and may be moved back and forth between those two confinement regions in much the same way as in conventional pumped storage hydro on land. As illustrated, the upper confinement region is a portion of the seafloor, which may be generally somewhat flat, that is surrounded by a fence that forms a closed perimeter and is suitable to restrain a quantity of the more-dense liquid in its interior. The lower confinement region is illustrated as being similar to the upper confinement region. However, another possibility is that the lower confinement region could be a naturally occurring submarine canyon, essentially a valley, that has been supplemented with a dam-like structure in an appropriate place, and the upper confinement region could be on a relatively flat portion of the ocean floor near the top of the canyon. Still another possibility would be a naturally-occurring basin. For any such combination of upper and lower confinement regions, it is possible that the more-dense liquid may be in contact with the less-dense liquid and may remain separated by virtue of density stratification, as illustrated.

[00210] Alternatively, it is possible that there may be a separation boundary between the

more-dense liquid and the less-dense liquid. Such a separation boundary need not be a perfect, impermeable separation boundary; it may be sufficient if the separation boundary is merely effective enough for separation that it reduces the amount of mixing that would otherwise occur between the more-dense liquid and the less-dense liquid. For example, the separation boundary could be thought of as a tarpaulin that defines a boundary between the more-dense liquid and the less-dense liquid in most places even if not every place of potential contact between the more-dense liquid and the less-dense liquid. Such a separation boundary may be free to rise and fall as the interface between the more-dense liquid and the less-dense liquid rises and falls as the more-dense liquid moves back and forth between the two confinement regions. Of course, it would also be possible for one of the confinement regions to be replaced by a bag.

[00211] It is further possible that a lower storage region 200 could exist in the form of a

confinement region 210 could be provided by a naturally occurring depression having the general shape of a bowl or a crater. Also, it is possible that such a confinement region 210 could be provided partially by a naturally occurring depression and partially by an

underwater structure that is man-made. This could be in the nature of a naturally occurring valley on the ocean floor that is at least partially dammed at an appropriate place by a man- made structure on the ocean floor. It is still further possible that such a confinement region 210 could be provided by underwater structures that are entirely man-made. Any of these could be augmented by dredging.

Storage on Land

[00212] In yet another embodiment of the invention, referring now to Figure 13, it would be possible to provide a lower storage region that is a bag 201 as already discussed, or an underwater impoundment as just discussed, in combination with an upper storage region 100 that is located on land. In this embodiment, the upper storage region 100 could resemble the upper reservoir of a conventional pumped hydro storage installation. The working fluid could be a liquid that is more dense than ordinary water or ordinary seawater. The upper reservoir, which might be located on land near the shoreline, could be a type of reservoir that is sometimes referred to as a "turkey nest," in which essentially the top of a mountain is converted into a lake.

[00213] In an embodiment of the invention, with continued reference to Figure 13, there may be provided an upper storage region 100. The upper storage region 100 may be defined by a boundary that suitable to contain an identifiable mass of liquid within the boundary, such as a dam, walls, impoundment or other construction. The upper storage region 100 could be produced by constructing a dam, by excavation, or by a combination of excavation and construction. In an impoundment or construction, there may be a liner or coating to reduce or prevent seepage of the working liquid into the adjacent ground or soil. Alternatively, the place of construction may be chosen such that the ground itself has a sufficiently low permeability for seepage. Alternatively, the upper storage region could be a bag, whose boundary may be substantially impermeable to the passage of liquid therethrough. The bag may have a flexible or deformable boundary, or the boundary may be in the form of a container that is less than fully flexible. [00214] There may also be a lower storage region 200, which may be at a lower elevation than the upper storage region 100. The lower storage region may be located in or on or adjacent to the ocean. The lower storage region 200 may rest on the bottom of the ocean. It is also possible that the lower storage region 200 could be in the ocean at approximately sea level. Several possibilities are shown in Figure 13. As shown on the right of Figure 13, the lower storage region 200 could be a bag 201 as discussed herein. Alternatively, the lower storage region may be defined a man-made wall or an impoundment, which may enclose a region. The lower storage region may be defined in part by solid boundaries such as the ocean floor and in part by a man-made boundary resembling a dam underwater. This option is shown having an upper boundary that could be thought of as a covering or separator 222 like a tarpaulin, to discourage mixing between the working liquid and the ocean, even if the covering or separator might not form a perfectly impermeable boundary. Such a covering or separator 222 could be made of a material having a density such that it floats on the more- dense liquid but sinks with respect to seawater. As shown farthest left, the lower storage region could be a simple depression in the seafloor, either naturally occurring or man-made. This option is illustrated without a covering, although of course a covering could be used. These options are similar to options discussed in other embodiments, or could be used in other embodiments.

[00215] It is further possible that in connection with a lower storage region 200 such as

confinement region, the discharge and/or intake of the working liquid could be provided with a sparger 245 or diffuser component that spreads out the flow discharge into a number of smaller and more localized jets that might not all be in the same direction. Such a device could reduce the amount of motion and local disturbance that might be created in the pool of the working liquid. Use of a sparger 245 or diffuser component could help to maintain desired stratification and separation of the working region with respect to the rest of the body of water. Such a sparger 245 geometry may discharge liquid through a plurality of small paths or jets rather than in a single large jet. The diffuser or sparger 245 may contain a plurality of holes. The holes in the sparger 245 or diffuser could be oriented so that the jets or discharge paths are at least partially horizontal, or are closer to horizontal than they are to vertical. Any such design features may reduce the creation of or may localize the direction of jets turbulence and other fluid phenomena that might be detrimental to the goal of maintaining the stratified layers. In terms of elevation, the sparger 245 or diffuser could be located toward the bottom of the confinement region 210 of the working liquid. [00216] Withdrawing the dense liquid from the lower confinement region could be done through the same sparger 245 or diffuser used for discharge. Benefits of a sparger 245 or diffuser geometry can be experienced during intake of working liquid from the pool of working liquid, similar to the benefits of using a sparger 245 or diffuser during discharge of the working liquid into the pool of the working liquid. Of course, other designs of sparger or diffuser, or a simple tube, could also be used.

Bittern on land

[00217] For conventional pumped storage hydro installations on land, a typical elevation difference between the two bodies of water is about 300 to 400 meters. Some pumped- storage installations have smaller elevation differences, and a few pumped storage installations with the largest elevation differences reach elevation difference of 1000 meters or slightly more.

[00218] In an embodiment of the invention, referring now to Figure 14, it is possible that the energy storage system could be a pumped storage hydro system that has an upper storage region that is a defined reservoir and a lower storage system that is a defined reservoir, and neither of the two storage regions is or is located in or on the ocean, and yet the working liquid could be a liquid as described herein that has a density greater than the density of seawater. The two reservoirs or storage regions could both be surface bodies of water on land. As yet another alternative, the lower reservoir could be an excavated cavern on land such as an abandoned mine. Another possibility would be a cavern in a salt bed, in which case the described working liquid could have the feature of being saturated and unable to dissolve any further salt from the salt bed, because of the already high concentration of solutes in the working liquid.

[00219] In embodiments of the invention, if the working fluid has for example a specific gravity of 1.3 rather than 1.025 or 1.0 as would be the case for seawater or freshwater, respectively, as the working fluid, then the amount of energy stored for the same geometric conditions would be about 30% larger than would be stored in the same facility by pumped storage of seawater or freshwater. This would be suitable for use in a closed pumped storage facility, i.e., a facility that is mostly separate from other bodies of water such as rivers.

Depending on the composition of the working fluid as discussed herein, the working fluid could be chosen to have a specific gravity that is greater than 1.1 , or greater than 1.2, or greater than 1.3, or greater than 1.4, or greater than 1.5 . [00220] In some embodiments, the working liquid in the confinement region is shown as being in physical contact with the ocean, without a physical boundary separating the two fluids. Because the density of the working liquid (such as concentrated seawater or a specially prepared solution) may be significantly larger than the density of the surrounding less-dense liquid (such as ordinary seawater), there will be a tendency for the working liquid to settle into and occupy as low a position as possible, and it is possible that the two fluids may tend to remain stably separated in such a situation, i.e., stratified and not mixing with each other. It is believed, although it is not wished to be limited to this explanation, that as the overall size scale of the amount of the working liquid increases, whatever tendency toward mixing does exist will become progressively less important in the overall scale of the system. It is still further possible that the working liquid and the seawater could be separated by a membrane or flexible cover. The membrane material may have a density that is intermediate between the densities of the working liquid and the seawater, thereby causing it to float at the interface between the working liquid and the seawater. The membrane could completely or only partially separate the working liquid from the seawater, and could be free to move as the position of the interface changes due to introduction or withdrawal of the working fluid. However, none of this is essential. Similarly, it is possible (not illustrated) that between the working liquid and the seawater there could be a layer of an immiscible fluid (such as an oil) whose density is intermediate between the densities of the working liquid and the seawater, thereby causing the immiscible fluid to float at the interface between the working liquid and the seawater. The ability to maintain a stratified situation over a long period of time may be influenced by the amount of local motion, turbulence etc. that naturally occurs in the ocean.

[00221] Yet another possibility is that the upper storage region could be a floating marine vessel of substantially rigid construction, such as a ship or barge.

Upper storage region Floating

Most of the embodiments disclosed so far have had their storage regions resting on the seafloor or on land. However, it is also possible that the upper storage region 100 could be in the open ocean such as in the form of a barge or a ship or a floating bag.

[00222] Referring now to Figures 15A-17B, there are described embodiments in which the upper storage region is floating. In Figures 15A-15B, the upper storage region 100 is a bag and the lower storage region 200 is also a bag. In Figures 16A-16B, the upper storage region 100 is a bag and the lower storage region 200 is an impoundment region. In Figures 17A- 17B, the upper storage region 100 is a barge or ship, and the lower storage region 200 is an impoundment region. In each of these cases, there is shown a configuration in which energy is stored, and a configuration in which energy is discharged.

[00223] It can be observed in Figures 15 A, 16A and 17A that in the energy storage condition, the lower storage region 200 is small and the more-dense liquid is mostly in the upper storage region 100 and the upper storage region 100 is large and is sitting low in the body of water. It can be observed in Figures 15B, 16B and 17B that in the discharged condition, the more- dense liquid is mostly in the lower storage region 200, and the lower storage region 200 is large and the upper storage region 100 is small and floats higher in the body of water.

[00224] The lower storage region 200 may rest on the bottom of the ocean or body of less- dense liquid, or may have some other defined location and may be as described elsewhere herein.

[00225] The upper storage region 100 may be defined by a flexible or deformable,

impermeable boundary that can contain an identifiable mass of liquid within the boundary, such as bag 101. The upper storage region 100 may be floating in the body of less-dense liquid, or may be partially floating and partially submerged in the less-dense liquid.

[00226] It can be appreciated that the upper storage region 100 such as bag 101 may contain a more-dense liquid that is to be stored at a high elevation in relation to a less-dense liquid, and therefore, left to itself, the more-dense liquid would have a natural tendency to sink in or with respect to the less-dense liquid. The upper storage region 100 such as bag 101 may have to bear or support the unbuoyed weight of a substantial mass of the more-dense liquid above the less-dense liquid. Accordingly, the upper storage region 100 such as bag 101 may be provided with flotation devices 120 that are appropriate to maintain the overall combination of the more-dense liquid, the bag 101 and the flotation devices 120 in a condition that will float with respect to the body of the less-dense liquid such as the ocean. The flotation devices

120 may be distributed among a plurality of places in the bag 101. The flotation devices 120 may be built into or integral with bag 101 , or may be attached to bag 101. The flotation devices 120 may contain low-density porous material, which may be somewhat rigid. The flotation devices 120 may contain pockets of air or other gas in either a rigid pocket boundary or a non-rigid pocket boundary. (It can be realized, however, that air pockets that are surrounded by a non-rigid boundary could reduce their volume as they submerge some distance below the surface of the body of water, which might be undesirable for providing consistent flotation.) To the extent that the materials of construction themselves of the bag

101 influence the flotation properties of bag 101, those construction materials also may be chosen to have a density that is less than the density of the less-dense liquid such as seawater. For example, polypropylene is less dense than some other candidate polymers. It is possible that a disproportionate portion of the flotation devices 120 could be at the top of bag 101, but the lower portions of the bag 101 do not have to be devoid of flotation devices 120. The distribution of locations of flotation devices may determine what surface is the top surface of the bag 101 when the bag 101 is floating.

[00227] The bags 101 and 201 may be flexible while also being substantially impermeable to the passage of liquid through the membrane. In general, the bag could be made of fabric and polymer. It is further possible that the bag could have bands built into it in appropriate directions to strengthen it against possible tensile stress.

[00228] It is further possible that any of the bags described anywhere herein could be enclosed in a cage for protection against possible damage such as due to debris or accidents, or for possible structural connection such as for anchoring or to constrain the position of the bag in any direction. This is true whether the bag floats exactly at the ocean surface or some distance under the ocean surface.

[00229] The lower storage region 200 has been illustrated as being a physical bag 201, similar to the upper bag 101 of the upper storage region 100. However, that is not the only possible option or embodiment. In general, any of the lower storage region options illustrated in Figure 13 could be used in combination with an upper storage region that is floating or surrounded by the ocean.

[00230] It can be noted that for embodiments of the present invention in which the upper

storage region 100 is floating, the conduit 300 may be vertical or nearly vertical. It is possible that the conduit 300 may be thicker or stronger towards its lower end. This is because the pressure that is contained inside the conduit 300 (relative to the corresponding external pressure of the ocean or body of water that is exerted on the corresponding outside surface of the connecting conduit 300) increases with depth below the surface of the body of water. It is further possible that the connecting conduit 300 may have flotation built into it so that the weight of connecting conduit 300 does not pull down the upper storage region 100 as much as would otherwise be the case. It is also possible that the weight of the connecting conduit 300 (assuming that the connecting conduit has a net downward unbuoyed force in the seawater environment) may hang from the upper storage region. The connecting conduit 300, especially the upper portion, may be structurally designed to bear that load.

[00231] It can be appreciated that for an upper storage region that floats at or near the surface of the ocean, as the upper storage region 100 empties or fills, the vertical elevation of the connection point between the conduit 300 and the upper storage region 100 may rise or fall. For example, if the upper storage region 100 is a bag 101, the vertical dimension of the bag 101 may change with filling/emptying of the bag 101. If the upper storage region 100 is a barge 105 or ocean-going vessel, the barge 105 or vessel itself will tend to rise and fall in the water depending on the amount of more-dense liquid stored in the barge 105 or vessel at a particular time. Other possible influences include tides; waves; storms; drifting of one region horizontally relative to the other region; change in elevation of the conduit connection to upper storage region 100 due to change in filling; change in elevation of the conduit connection to the lower storage region 200 due to change in filling. Any of these motions could change the effective length or required length of the conduit 300. This could make it useful to provide some sort of taking-up of slack or letting-out extra length of conduit 300 as may be needed.

[00232] Referring now to Figure 18, in an embodiment of the invention, there may be

provided a form of strain relief for the connecting conduit 300 that goes between the upper storage region 100 and the lower storage region 200. Such a strain relief may allow for relative motion between the upper storage region 100 and the lower storage region 200, such as may be due to any of various causes as just discussed. Such a strain relief device may have a top and a bottom and a nominal distance between the top and the bottom of the strain relief device, and the connecting conduit 300 may have a length between the top and the bottom of the strain relief device that is longer than the nominal distance between the top and the bottom of the strain relief device. The nominal distance between the top and the bottom of the strain relief device may be determined by a spring element, but the spring element may be capable of stretching as needed to achieve a range of lengths. Between the top and the bottom of such a strain relief device, the connecting conduit 300 may have a nominal shape such as helical or zig-zag, that is capable of undergoing an overall increase or decrease of length between its end at the top of the strain relief device and its end at the bottom of the strain relief device, and the conduit 300 may be flexible at least in the region between the top and the bottom of the strain relief device.

[00233] It can be appreciated that in order to store a quantity of the more-dense liquid at or near the surface of the body of water, it may be necessary to provide flotation that is built into or is attached to or is in some way force-transmittingly connected to the upper bag.

[00234] If the upper bag 101 were simply one large open volume, it would be possible that flotation could be provided by trapping some air inside the upper bag 101 and never completely fill the upper bag 101 with liquid. For example, if the more-dense liquid is 30% more dense than the surrounding seawater, the air pocket might amount to approximately or slightly more than 30% of the total volume of the bag 101. However, in a large bag 101, the location of a simple air pocket inside the bag might be unpredictable.

[00235] As an alternative, the bag 101 could have one or more flotation elements 120 attached to the fabric of the bag 101 or the outside of the bag 101 in distributed locations. Such a flotation element 120 may be simply an air pocket or may be a region containing a low- density material such as a porous solid. Such low-density material may be rigid, although it does not have to be rigid. There could be a plurality of flotation elements 120 such as air pockets distributed within the fabric of the bag 101 or attached to bag 101.

[00236] In this situation, the mass of the more-dense liquid would essentially be suspended from the flotation provided by that flotation 120. The wall of the upper bag 101 would have to carry weight of the more-dense liquid that is in excess of the buoyancy provided by the body of water. It is believed that for an operating value of wall tensile (hoop) stress that might be 10,000 psi (69 MPa) typical of polymers such as polypropylene, polyethylene and nylon, for a bag diameter in the range of 25 ft diameter containing more-dense liquid whose unbuoyed density is the equivalent of a liquid having a specific gravity of 0.3, and assuming that the bag hangs from its flotation without a large force magnification due to the angle at which the bag wall hangs, the wall thickness needed in order to not exceed such a stress would be about 1 to 2 mm. Of course, if desired, it would also be possible to provide bands or a network of reinforcing material, which might be built into the bag 101.

[00237] In an embodiment of the invention, there may be provided a flotation element 120 that forms itself, when inflated, into a shape other than a circular-cross-section. Figure 19 shows, in cross-section, a bag 101 that has flotation elements 120. The flotation element may have a perimeter and may also have one or more internal tethers 130 that connect a point on the perimeter with another point on the perimeter. More specifically, there may be several tethers 130 connecting an upper part of the flotation chamber 120 with the lower part of the flotation chamber 120. Tethers 130 may be flexible straps or webs that connect one portion of a flotation chamber 120 or bag 101 with another portion of a flotation chamber 120 or bag

101. As illustrated, a tether 130 is capable of carrying tensile load, but may be incapable of carrying a compressive load because of its flexibility. These tethers 130 may help to constrain the shape of the flotation chamber 120 when the flotation chamber 120 is filled. As illustrated, the flotation chamber 120 is constrained to have a shape that is more horizontal than vertical, generally somewhat flat, but with small bulges permitted between the tethers

130. If the bag 101 is elongated in one direction, such as being generally cylindrical, it is not necessary that the tethers 130 or the tethered shape be uniform or constant from one end to the other; rather, the tethers 130 and the tethered shape could vary from one end to the other.

[00238] Referring now to Figure 20, in a bag 101 that might be generally cylindrical with its axis generally horizontal, and which may be long in comparison to its vertical dimension, and if the bag contains flotation elements 120 above its mid-elevation, there might be a chance that the more-dense liquid might not empty out of the bag 101 in a uniform manner. This could make it difficult to completely drain the more-dense liquid from the upper bag 101.

[00239] Accordingly, it may be desirable to design the bag 101 or system of bags 101 so as to encourage the bag(s) 101 to empty completely due to gravity, with minimal likelihood of a pocket of the more-dense liquid being isolated somewhere and blocked off by a region of the bag 101 that has floated and lifted to a higher elevation than some other part of the bag 101. So, the upper bag 101 could be designed so that when it is positioned in the body of water in its nominal position and being at least partially filled with the more-dense liquid, the bottom surface of the bag 101 could be sloped toward a point where a drain such as a conduit 300 connects with the bag 101. The slope could be for example at least 2%, which is a typical slope requirement for drainage due to gravity. For example, as illustrated in Figure 20, the upper bag 101, or the liquid-containing region thereof, could be frustoconical. The connection or drain point would be at a relatively lower point in the frustoconical cross- section. At the more distal (away from the drain or connection point) end of the bag 101, due to the smaller cross-sectional dimension there, the bottom of the bag 101 could be at a higher elevation than the more proximal end of the bag 101. It would also be possible to provide a non-uniform distribution of flotation 120 along the length of the upper bag 101, and thereby create slope of the bottom of the bag. For example, there could be more flotation provided toward the distal end of the upper bag 101. The portions of the bag 101 having more flotation would ride higher in the water than the portion(s) with less flotation, causing a slope of the bottom surface of the bag 101. This could be done even if the interior of the bag 101 were of generally constant cross-section.

[00240] As yet another possibility, it is possible to design flotation chamber 120 so that it is subdivided into distinct sub-chambers that may be positioned respectively along the lengthwise direction of upper bag 101, and it is possible that means could be provided to independently vary and control the air content, i.e., flotation, of individual sub-chambers.

[00241] The flotation chamber 120 may be provided with sub-regions as desired. This could be done for redundancy, for reasons of controlling the position or shape of the upper bag, etc.

Regions of the flotation chamber 120 could be passively filled with air, or could be actively filled by pumping air into them. Referring now to Figures 21A-21B, it is possible to reposition bags 101 as desired such as for protection in the event of a major storm. If the flotation chamber 120 is deflated so that the bag 101 substantially sinks, then the bag 101 would be significantly below the surface of the ocean and would experience much less of the effect of a storm, compared to what it would experience if floating at the surface of the ocean. For example, if a major storm were anticipated, some or all of the air could be released from the flotation chamber 120 so as to permit the upper bag to ride lower in the water or even be submerged, in order to experience less of the conditions of the storm. For example, during the storm, the bag could be positioned at a desired distance below the surface of the body of water.

[00242] In order to position the bag submerged in the water such as for protection from a

storm, it may be desirable to provide discrete internal chambers and by deflating the internal chambers in a desired sequence. In order to keep the overall apparatus near the surface of the ocean even when some internal chambers are deflated enough for the bag 101 to sink below the surface of the ocean, there could be provided an alternate flotation device that does stay at the surface of the ocean even if the bag 101 is allowed to sink for storm protection. That alternate flotation device could be connected to the rest of the apparatus by a strain relief that might absorb motion of the storm without transmitting it to the portions of the apparatus that have been submerged for storm protection.

[00243] Another design possibility for the upper bag is that a bag and its flotation do not have to be integral with each other and do not have to be made from fabric that is directly connected to each other. The illustrated bag could be the upper storage region in the embodiments of the invention. This is illustrated in Figure 22. Instead of the bag and its flotation being integral with each other, the bag and its flotation could be discrete

components and the bag could be connected to its flotation by cords, straps, suspenders or tensile connectors, so that the flotation elements would be at the surface of the ocean but the bag itself could be located entirely underwater a modest distance below the surface of the ocean. At a modest distance below the surface of the ocean, the water can be expected to be quieter and there would not be waves crashing over the bag. This could provide a less severe environment for the bag, and this could extend the useful lifetime of the bag. It is further possible that the tensile connectors could have built into them some form of spring or shock absorber so that if the flotation bounces up and down due to waves storms and other disturbances at the surface of the ocean, the bag itself could ride at a more steady position a modest distance below the surface of the ocean. If the flotation elements at the ocean surface do get damaged by waves or storms before the bag itself reaches the end of its useful life, then the flotation elements could be replaced without replacing the entire bag. It would also be possible to replace, if necessary, the cords or straps or suspenders or tensile connectors or springs or shock absorbers, without replacing the bag itself.

[00244] Also illustrated in Figure 23 is the possibility that that part of the load path for

suspending load from flotation could be a suspending, generally vertical web. Such a web also could continue on through the bag from top to bottom of the bag, i.e., the top somewhat horizontal part of the bag could be sewn to the web, and the web could be sewn to the bottom somewhat horizontal part of the bag. In this way, the web could be a continuous fabric carrying vertical load continuously all the way from the bottom of the bag (which is the part of the bag that first feels the downward weight of the dense liquid), and the web continues on upward through the bag and to the place where the top of the bag is sewed to connect with the web, and the web continues on further upward to where there might be connection features to connect to suspender cables that go on upward to floats. The portion of the web that is inside the bag could have holes or openings through it, so as to help the bag to empty out its liquid or fill up with liquid, to avoid creating trapped pockets. Or, the web could be interrupted in places inside the bag for the same purpose. In the illustrations, the web is illustrated as extending upward from the bag. However, it would also be possible that the web could instead be upside down from what is shown and could extend downward from the bottom of the bag. This could be used for connecting to cables that extend up from ocean floor. Such a web could also extend out sideways for use in positioning against sideways drifting of the bag.

[00245] Referring now to Figure 23, it would be possible to have some flotation attached to the bag such that all of the flotation stays submerged all the time and never gets all the way to the ocean surface, but keeps the bag buoyant in all circumstances ranging from full to empty. In this situation, the bag may be anchored by tension cables to the bottom of the ocean, such as for the bag-empty (more buoyant) condition. The tension cables would have appropriate length to maintain the position of the upper bag at the desired elevation, which may be a short distance below the ocean surface. This could also be done without any flotation that extends all the way to the surface of the ocean, i.e., the Flotation that is at the ocean surface could be omitted. It would also be possible to use catenary cables if desired for control of horizontal position of the bag.

[00246] It is also possible that a floating bag could comprise tethers that transfer load from one part of the bag to another part of the bag. Such tethers could be fabric straps, which could be sewn into or onto the bag, or could be other construction. Connection points for the tethers could include roughly the same location as where the suspenders connect to the bag, although this is not essential. Other connection points for the other end of the tether could be on the underside of the bag. The tethers could be inside the bag. Alternatively, the bag itself could simply be reinforced near places where the suspenders connect to the bag. The reinforcement could be in the form of rings or bands that go around the bag in whatever pattern is desired.

[00247] If the upper storage region is a free-floating bag, the anchor to seafloor could be a cable that is in the form of a catenary, i.e., it rests on the seafloor for some distance before being actually anchored to the seafloor. Such an arrangement allows for some change of position of the floating anchored object by changing the tension on the cable or the angle of its approach to the seafloor, thereby changing the point at which it loses contact with the seafloor.

[00248] It is also possible that the weight of lower storage region such as bag could be used to anchor the upper storage region. When the upper storage region is buoyant because of not containing much of the more-dense liquid, the lower storage region will be heavy due to containing a large quantity of the more-dense liquid, and will be able to contribute to anchoring. This could be done in connection with other, fixed anchoring.

[00249] Referring now to Figure 24, there is illustrated an example of a facility that could be used to concentrate an aqueous solution, which could be seawater or previously-processed seawater, to a higher concentration of dissolved solutes, through evaporation of water. In such a facility, there may be a spray nozzle, and a pump to pump an aqueous solution through the spray nozzle, to create a mist or spray. The spray nozzle may be at a high enough elevation so that the mist or spray may spend some time in the air as it gradually falls.

During this time of falling, the mist or spray may be exposed to ambient air so that some, although not all, of the water content of the mist or spray evaporates. The mist or spray may also be exposed to sunlight to provide some of the energy to cause evaporation, although exposure to sunlight is not essential. The mist or spray may be allowed to fall to a screen or filter. During falling and the evaporation that occurs during falling, it is possible that some dissolved solids may precipitate out of solution because of the reduction in the amount of liquid water remaining in individual droplets in the mist or spray. Thus, the droplets of mist or spray that reach the filter or screen may contain particles of solid. Such particles of solid may be separated out by the screen or filter, while remaining liquid may seep through the screen or filter, and may form drops that further fall to a collection sump. Thus, the sump may contain liquid, but as the process continues, the remaining liquid in the sump may become more dense (and smaller in quantity due to evaporation of some of the water).

Liquid from the collection sump may be pumped back through the spray nozzle for multiple trips through the spray system to be further concentrated. Alternatively, other liquid such as seawater or other concentrated liquid could be provided to the spray nozzle, or a mixture of liquid from the collection sump and other liquid could be provided to the spray nozzle. A system such as this may be advantageous, because at least some of the energy representing the heat of evaporation of water may obtained from the atmosphere or sunlight or both, which would lessen or eliminate the need for a heat source to cause the evaporation. This could be desirable because of the large specific heat of evaporation of water, and because of the large quantities of water that might need to be evaporated. Of course, it would be possible to use any desired combination of supplied heat, evaporation to the atmosphere and solar heat input. The precipitated solid that collects on the screen or filter could be taken away for use for some other purpose, or could be put back into the ocean, especially if the initial liquid was seawater. It is also possible that liquid that has been concentrated by spray evaporation could be further concentrated by some other means. It is possible that any such concentration either could be done once to charge the energy storage system with more-dense liquid at the beginning of operation, or could be done on an ongoing basis to make up for leakage or dilution, or could be done on both an initial basis and an ongoing basis. Such a concentrator facility could be located adjacent to the energy storage system, either temporarily or permanently. Such a concentrator facility could be located on a ship. An embodiment of the invention may comprise a desalination plant for such purpose. The concentrator could be use any process or combination of processes such as reverse osmosis, multi-stage flash distillation, spray evaporation, drying using the sun's heat, and drying using supplied heat to cause evaporation, such as immersed heater, radiant heat.

Further Comments

Boundaries such as the boundaries of upper or lower storage regions may be substantially impermeable. A substantially impermeable boundary generally prevents flow of liquid therethrough or prevents mixing of the substances that are located on opposite sides of the boundary. However, it is still possible for a boundary to function adequately if there is a small amount of leakage present. It can be noted that while a completely impermeable boundary is preferred, it is nevertheless possible to operate the described system in the presence of some leaks or holes, if the passage in either direction of amounts of fluid through such leaks or holes is minor. Furthermore such a boundary could be a material such as cloth that has pores that might not be completely sealed but might still effectively discourage mixing between the working liquid and seawater.

[00251] Similarly, it is possible that the upper storage region 100 or the lower storage region

200 or both could be an open system depending on density stratification to maintain itself separate from the ocean in general. It is possible that a minor amount of mixing could be permitted between the working liquid and the surrounding ocean. However, a major amount of leakage or mixing, resulting in significant loss or significant dilution of the working liquid, would be detrimental.

[00252] A storage region could be or could include an impoundment on the seafloor, such as a fence enclosing a circular region, or a dam that impounds a sloping valley on the seafloor, or simply a naturally occurring or man-made depression on the seafloor that is suitable to contain a quantity of the more-dense liquid. In such a situation, the more-dense liquid may contact the seafloor. It is possible that such a containment region could be covered with a sheet of a material that is at least somewhat impervious to the passage of liquid therethrough, in order to maintain at least some separation between the more-dense liquid and the surrounding liquid such as seawater,

[00253] It is possible that the upper storage region and the lower storage region could be any desired combination of containment designs described herein. For example, the upper storage region could be an impoundment and the lower storage region could be a bag, or vice versa.

[00254] It can be noted that in embodiments of the invention, there is never storage of the entire volume of the working liquid at an elevated pressure relative to its surrounding less- dense liquid. There is no pressure vessel that needs to contain the entire volume of working fluid against a large pressure difference (either tensile stress or compressive stress in the wall of the pressure vessel). The only components that have to withstand full pressure difference are the lower portion of the connecting conduit and the pump and energy extractor and a valve, all of which only contain a fraction of the overall mass of the working liquid at any given time. However, these components are relatively small, and the volume of the working liquid contained inside these components at any instant of time is relatively small, which corresponds to requiring only a small amount of highly-reinforced pressure-containing material of construction for these components. If there is a dam or impoundment, portions of that structure would also have to be suitably designed for appropriate loads. [00255] In the overall system of embodiments of the invention that involve a floating (not completely-seafloor-supported) storage region, there is a need for supporting of the unbuoyed (differential) weight of the volume of the more-dense liquid. In embodiments of the invention that are completely-seafloor-supported, this weight is borne by the seafloor much as in conventional pumped storage hydro.

[00256] In general, embodiments of the invention are believed to be advantageous (although it is not wished to be limited to this explanation) because there is never a need to contain the entire volume of liquid times the full pressure difference that is used in energy storage. This is in contrast to conventional Compressed Air Energy Storage (CAES), in which, when the system is fully charged, all of the energy is present as high pressure exerted upon the walls of the storage vessel(s). A related system using rigid vessels such as spheres located deep underwater, which are pumped empty of internal water in order to store energy (US patent 8698338), so that the walls are in compression as opposed to tension as in the case of conventional CAES. For such a system also, when it is fully charged, all of the energy is present in the form of pressure exerted on the walls of the storage vessel(s) at one time.

[00257] It is further believed, although it is not wished to be limited to this explanation, that in general, fabric and polymer are more economic materials of construction than the materials needed for construction of rigid pressure vessels (in either tension or compression). In embodiments of the invention in which the upper storage volume floats at or near the surface of the ocean, there is a need for the walls of the upper storage region and related components to support the unbuoyed weight of the working fluid.

[00258] In a completely-seafloor-supported system, the upper bag would only see a maximum outward pressure difference, across the wall of the bag, at the lowest elevation point on the bag, that is given by the density difference between the working liquid and the surrounding liquid, times the acceleration of gravity, times the elevation difference between the lowest point on the bag and the highest elevation on the bag that contains any liquid. This elevation difference can include the effects of slope of the upper storage region if the upper storage region is resting on a sloping seafloor. Similarly, for the lower bag, in the situation that the valve in the connecting conduit 300 is closed, the lower bag would only see a maximum outward pressure difference, across the wall of the bag, at the lowest elevation point on the bag, that is given by the density difference between the working liquid and the surrounding liquid, times the acceleration of gravity, times the elevation difference between the lowest point on the bag and the highest elevation on the bag that contains any liquid. Again, this elevation difference can include the effects of slope of the lower storage region if the lower storage region is resting on a sloping seafloor.

[00259] In many conventional pumped storage hydro installations, the pumped liquid is

freshwater. Seawater is a more corrosive medium and environment than is found in conventional pumped storage hydro. However, there is experience handling seawater or various compositions of brine in applications such as desalination plants, and drilling for oil and gas, and seawater pumped storage hydro. For such purposes, appropriate materials of construction are known and can be chosen, such as corrosion-resistant metals and polymers. Furthermore, corrosion protection measures, such as cathodic protection and a sacrificial anode, may be undertaken. Corrosion-resistant coatings can be used. It is possible that some relatively dense aqueous solutions, such as some forms of bittern, actually may be less corrosive than naturally occurring seawater. It may be desirable, in the formulation of the working fluid, to avoid or minimize the content of chloride ions.

[00260] Also, it can be noted that if the working liquid is substantially concentrated seawater

(having the same solutes as seawater concentrated into less water), then the eventual outcome of any loss or leakage of the working liquid into the ocean would not really be a contaminant. Loss or leakage of concentrated brine would cause an increase in the salinity of a local region of the ocean for a limited period of time, which could have a localized and temporary impact. When the lost or leaked liquid is eventually mixed with substantial quantities of seawater, it will be indistinguishable from other seawater. There would be no added substances that are not already present in ordinary seawater. Alternatively, if the composition of the working liquid does include additional substances beyond those typically found in seawater, it is still possible for those additional substances to be chosen so as to be relatively innocuous if they are released into the ocean or escape into the ocean.

[00261] Yet another possibility is that the working liquid could be a suspension that is a liquid containing very small solid particles suspended in it, which might achieve a relatively high overall density. As yet another possibility, dense non-aqueous liquids are also known and could be used. Assuming operation in the ocean, it also would be possible to pump freshwater, or other liquid that is less dense than seawater, down to a lower storage region in order to store energy, rather than pumping more-dense liquid up to an upper storage region to store energy. Then, energy would be recovered by allowing the freshwater to come back up to the surface of the ocean while passing through an energy extractor such as a turbine.

However, the relatively small density difference available with freshwater is not favorable. [00262] Any combination of type of storage region is possible for the upper storage region and the lower storage region: bag(s) 101 at the upper elevation in combination with either bags 201 at the bottom; or confinement region 210 on bottom; vessel 105 on top in combination with either bags 201 on the bottom or confinement region 210 on the bottom.

[00263] It is possible that components such as the upper storage region 100 such as bag 101 or barge 105 may have some kind of anchor to at least somewhat fix their positions especially in horizontal directions, such as against drifting due to any cause. Multiple such anchor cables may be attached to the floating components and to the ocean floor from various different directions to better define and control the horizontal position of the upper storage region 100. Such anchor may comprise a cable attached to the bottom of the body of water. Such cable could be a relatively tight relatively straight cable, or it could be in the form of a catenary, in which the cable hangs in somewhat of a curve and touches the floor of the body of water somewhat tangentially to the floor of the body of water. In such a catenary arrangement, if a horizontal force is applied to the curved cable, the cable may partially straighten and may allow some horizontal motion while changing the location of the point of tangency of the cable to the bottom of the body of water. Yet another possibility is that there already exist drilling rigs for offshore commercial oil and gas drilling. Many of these are already attached to the ocean floor and have heights in the range of 1000 to 2000 feet. It is possible that at least some of the components of an embodiment of the invention could be attached to a rig that is attached to the ocean floor.

[00264] It can be noted that embodiments of the invention are substantially free of the siting constraints that are so important and influential in the selection of sites for conventional pumped storage hydro installations on land. Embodiments of the invention similarly are free of the environmental impact on land such as are associated with possible flooding of dry land due to the construction of impoundments. The ocean, in appropriate places, is significantly deeper than the elevation difference that is attained with most of the existing or realistically available pumped hydro storage sites on land. Also, the vast size of the ocean provides many opportunities for siting.

[00265] Features disclosed herein may be combined in any combination.

[00266] A system of an embodiment of the invention could be used on a large scale for bulk storage of energy, such as at the scale of the electric power grid for extended periods of time such as hours. On a smaller scale, a system of an embodiment of the invention could be used to help compensate for shorter-term fluctuations in the conditions of the electric power grid to maintain grid voltage or grid frequency. Energy storage can also be useful as protection against interruptions in supply and for so-called "black start" capability. It is possible that if the period of operation is short, it might be shorter than the transit time needed for a parcel of liquid to transit the length of the connecting conduit. Nevertheless, the system would still work as described herein.

[00267] All references cited herein are incorporated herein by reference.

[00268] Although embodiments and examples have been disclosed, it is desired that the scope of the invention be limited only by the appended claims.

[00269] We claim: