Technical Field of Invention

This invention relates to a linear motor and a compressor incorporating such a motor. It also relates to a pump and a cooling system suitable for refrigeration or air-conditioning, using the compressor and pump Background of Invention

It is known that a conventional reciprocating compressor is driven by a rotary motor, in which a crank or cam mechanism is used to convert motor's rotational movements into reciprocating movements of one or more pistons. Such an compressor has the following drawbacks. Firstly, the arrangement is not efficient because its electric input has to drive a chain of mechanical parts, including unavoidably a motor rotor, a crankshaft, a piston-rod and a piston-head. Keeping such a chain of pans in operation per se consumes a lot of power before any useful work can be done to the working medium. Secondly, due to the conversion fro rotary to reciprocating movements it is unavoidable that die piston-head is subject to sideways forces, which cause friction and wear and produce unwanted heat. To cope with this problem, it is necessary to incorporate further arrangements for supplying lubricant oil to the moving pans and keeping the mechanism cooled, which add more complications to the structure and further burden to the motor Furthermore, due to the above reasons the parts used in such a compressor have to be made of high strength materials by precise machining processes, therefore high costs. Finally, the compressor, once built, has to work within a narrow range of rated working conditions with little flexibility to cope with changes of operation conditions or load. It is also known that an axial flow pump having a propeller driven by a motor can produce an axial flow in a pipeline. The pipeline has to be angled so that the propeller can be fitted into the pipeline with the motor shaft extending through die wali of the pipeline at the angled position, otherwise a more complicated gear arrangement has to be used. Such arrangements make it difficult to arrange a pipeline to suit a pump, while on the other hand, a reliable sealing of the rotary shaft which passes through the wall of the pipeline is difficult to achieve but easy to wear out.

It is also known that conventional vapour-compression type cooling systems are responsible for serious damages to the environment. Generally speaking, the damages are in die following three main aspects. Firstly, these systems consume a huge quantity of power because they operate 24 hours a day for all the year round (in case of refrigerators or freezers) or for the whole season (in case of air-conditioners). This huge consumption contributes indirectly to the accumulation of greenliouse gases produced during power generation. Secondly . these systems use CFCs or HCFCs as refrigerant which cause direct damages to tlie ozone layer. Thirdly, a conventional system uses metal materials such as copper or aluminium for making thermally conductive components, i.e. condenser and evaporator, which consume natural resources. Summary of Invention It is, therefore, a first general object of the invention to provide a linear motor and a compressor which overcomes the above problems and disadvantages.

According to one aspect of the invention, there is provided a reciprocating motor and/or a compressor using such a motor, comprising: two opposing magnetic driving members with coaxial poles; a reciprocating member, serving as a piston in the compressor, with magnetic poles disposed ovably between and coaxially

widi said driving members; and means for energising said driving members and/or reciprocating member to generate an alternating electromagnetic field therebetween; wherein said poles of each said driving member include an outer one of a first polarity and an inner one of a second polarity located wid in said outer one, and said reciprocating member has complementary poles arranged dierebetween to form a push-and-pull driving force pair for said reciprocating member to reciprocate.

It is preferable that each driving member is an electromagnet having conical pole faces, and the reciprocating member has pole faces formed by permanent magnet, so as to provide concentric driving forces to suspend the reciprocating member magnetically.

The piston can have a permanent magnet with a radially arranged magnetic field and a shunt mechanism to regulate its magnetic flux to die driving members. Furdiermore, sliding pole pieces can be arranged to improve tlie piston's flexibility and agility for a better energy efficiency. It is preferable to provide a lubricant circulating circuit in the compressor to keep the piston's sliding pans lubricated. Fluid dynamic and/or magnetic means can be used to cause die piston to rotate during its reciprocating movements, so as to improve its lubrication and reduce its wear. Also, buffer mechanisms can provide cushion effects to protect the reciprocating member at die end positions of its movements, including fluid cushion, spring cushion and/or magnetic cushion. Furthermore, die buffer mechanisms can also be used to balance tlie reciprocating member ^' s movements in opposite directions to improve its energy efficiency. The driving members can have magnetic coupling means to selectively adjust die magnetic flux of each driving member according to die position of the reciprocating member. The mechanism can be powered by either AC or DC current. It further comprises sensor means which adjust die AC current or reverse die DC current in response to the movements of said reciprocating member to each end position. At least one driving member can be arranged in a manner diat its position is adjustable in response to die output pressure to provide automatic compensation to load changes.

The fully "floated" piston reduces friction to ensure a long service life. In case of high pressure application, a plurality of die reciprocating compressors can be connected in series to build up output pressure, so a wide range of application requirements can be met by combining standard modules.

Furthermore, it is a second general object of die invention to provide an axial-flow pump/marine propeller which overcomes die problems of die conventional products.

According to anodier aspect of the invention, there is provided an axial flow pump/marine propeller comprising: a hollow body, a stator carried by die body and defining a cylindrical inner space, a hollow rotor member fitted in die inner space with means for propelling fluid dieredirough, and electromagnetic means for generating a rotational magnetic field to drive die rotor to rotate; wherein die rotor is supported by a suspension bearing for providing rotational and dirust bearing when the rotor rotates. The bearing can include means sensitive to die rotor's rotational and/or axial movements to retain it at a balanced position. Preferably, die bearing includes at least one spiral vane formed on die outer surface of the rotor for producing a peripheral flow of a fluid in die gap between die stator and die rotor when die rotor rotates.

Also, diere is provided a suspension bearing mechanism for levitating a rotary member during its rotation. Advantageously, die mechanism includes magnetic axial registration and suspension arrangements formed by annular magnets. Also, it can include flow dividing means for keeping die rotary member self- balanced during its rotation. Furthermore, diere is provided an axial flow pump/marine propeller having fluid

driving means formed by one or more flexible spiral blades which can be compressed axially in response to die changes of power input or working load. The axial flow pump can have two coaxial driving members widi driving means in opposite spiral directions so that when tiiey have opposite rotational movements, they cancel each odier's swirling effects. Tlie central driving member can be supported by a pivot bearing. Also, a conical impeller can be fitted to promote die axial flow. The pump can be used as an electric generator or flow-meter.

For the sack of easy understanding, die axial flow pump/marine propeller of the invention is to be described hereinbelow as a pump unless specifically stated odierwise.

Furthermore, it is a diird general object of die invention to provide a cooling system, and components for it. which are environment-friendly, i.e. easy to manufacture, use environmentally less hazardous materials, and operate widi high energy efficiency.

According to one aspect of the invention, there is provided a cooling system comprising: a mechanism for transferring heat from a cold-generating member to a heat-rejecting member, and a circuit dieπnally engaged with said heat-rejecting member; wherein a coolant widi at least one evaporable component is used to circulate in said circuit which has means for promoting evaporation of said evaporable component. Preferably. die coolant comprises at least one endotiiemiic salt for cooling said heat-rejecting member by endothermic dissolution. The system can have a brine circuit for dissipating cold energy. Preferably, said brine is also used as a cold storage material when mere is low cost electricity. Environmentally benign refrigerant, such as carbon dioxide, ammonia or propane, can be used in die system to avoid environment damages.

Also, there is provided a compression assembly comprising: a plurality of compressors serially connected for a refrigerant to be compressed progressively dierein, and a coolant passage for circulating a coolant fluid ^* , wherein said gas passage and coolant passage are diermally coupled to each odier so that die heat generated in the compressed refrigerant can be absorbed and carried away by said coolant.

Further, there is provided a circuit for cooling a heat source by circulating a coolant of at least two components, said circuit comprising: a heat absorbing portion adapted in diermal coupling with said heat source, and a heat dissipating portion in fluid communication widi said heat absorbing portion: wherein said heat absorbing portion is arranged to have an upper part widi a coolant outlet, an lower part with a coolant inlet, and an intermediate zone engaging said heat-generating member, said intermediate zone has baffling means for stabilising a temperature gradient in die coolant flow from said lower end to said upper end so that at least one component of die coolant solution can evaporate and be circulated to said heat dissipating portion via said outlet. Preferably a thermoelectric member is fitted in the intermediate zone to provide an elevated temperature to promote said evaporation. A relatively low pressure can be formed in said heat dissipating portion and/or a vapour compressor can be used to promote said evaporation.

Furthermore, there is provided a defrosting arrangement comprising a thermoelectric member widi one thermal pole coupled to a cold-generating member and die other thermal pole coupled to a heat exchange member, and a unit for control electric supply to said diemioelectric member to reverse its heat transfer direction, diereby selectively changing die operation of said arrangement from a frost accumulating mode to a defrosting mode.

Finally, there is provided a method of operating a cooling system having a brine circuit, comprising steps of: (a) setting a control unit for selecting one of two operation modes in response to die availability of

low-cost electricity, (b) operating die system in a first mode when said low-cost electricity is not available, in which die bnne is cooled to a first temperature range and circulated, and (c) operating die system in a second mode when said low-cost electricity is available, in which die bπne is cooled to a second temperature for it to be frozen so as to store latent cold energy therein It is to be understood diat the terms "low cost " niatenals or electricity are used here m the sense of both low conunercial and ecological costs, I e causing less environment damages and being easily recyclable or renewable

Brief Description of Drawings

The above described objects, aspects, features, advantages, and structural and functional details of the invention will become apparent in the following descnption of the embodiments with reference to the accompanying drawings, in which

Fig 1 is a cross-sectional view taken along the central axis of a compressor according to a first embodiment of the invention.

Fig 2 is a cross-sectional view taken along die plane B-B shown in Fig 1. Fig is a cross-sectional view of a compressor of a second embodiment, Fig 4 is a cross-sectional view of a compressor assembly according to a third embodiment.

Fig 5 is a cross-sectional view taken along the plane C-C shown in Fig 4, Figs 6A and 6B are cross-sectional views of a compressor according to a fourth embodiment. Figs 7A and 7B are sectional views of die piston 560 shown in Figs 6A and 6B, Figs 8A and 8B are sectional views of the electromagnet 5^0 and its cushion arrangements 540 and 550, as shown in Figs 6 A and 6B

Figs 9A and 9B are cross-sectional views of a piston according to a fifdi embodiment.

Figs 10A and 10B are sectional views of a magnet disc 610 and a magnet nng 640' taken along die planes C-C and D-D, respectively, as shown in Fig 9A,

Fig 1 1 is a sectional view illustrating a first embodiment of die punip/maπne propeller according to die invention

Fig 12 is a perspective view of die rotor 40 in Fig 11 ,

Fig 13 is a sectional view of a second embodiment of the pump/maπne propeller. Fig 14 is a sectional view of a third embodiment of die pump/manne propeller. Figs 15 A and 15B are sectional views of a fourth embodiment of die pump/manne propeller, Fig 16 is a perspective view of die rotor assembly used in die fourth embodiment.

Fig 17 is an enlarged sectional view showing details of a localised portion of Figs 15A and 15B. Fig 18 is a sectional view showing a fifdi embodiment. Figs 19A to 1 D show different parts of the fifdi embodiment. Fig 20 is a sectional view showing a sixdi embodiment, Fig 21 is a block diagram illustrating basic concepts and die operational and functional relationship between die components of a cooling system according to the invention.

Figs. 22A to 22C are sectional views of a freezer according to die invention,

Figs. 23A to 23C are different views showing details and different anangements of die heat storage tank 400 shown in Figs 21 and 22A,

Figs. 24 A to 24C are sectional views of different embodiments of die coolant column 310; Figs. 25A and 25B are sectional views of die defrosting arrangement; and Fig. 26 is a flow-chart illustrating die control method. Detailed Description of Preferred Embodiments Firstly, details of die embodiments according to die first general object are described. In die following description, the illustrated embodiments are described as compressors for die sake of easy understanding. The same concept can be used for a liquid pump or a vacuum pump. Tlie term compressor should be interpreted as covering all these applications, unless stated otherwise specifically. Also, die same concept can be used to build a linear motor by simply removing valves and seal members. Figs. 1 and 2 show a compressor 10 having a housing 20, two electromagnets 30 fitted to two ends o die housing, and a free piston 50 in die housing, dividing its iiuier space into chambers I and II. The housing 20 is a cylinder which can have external ribs and/or fins to improve its mechanical strength and heat dissipation, especially when it is used as a gas compressor. It can be made of any noil-magnetic material, such as plastics, fibre-reinforced resin, ceramics, aluminium, copper, brass, stainless steel etc. In case of the application as a gas compressor in a cooling system, metals are preferable for good heat dissipation. A low friction and wear resistance coating (not shown) can be formed on its inner surface by conventional mediods.

The two electromagnets 30 have basically die same structure, i.e. each has an annular core with an outer pole 34 and a coaxial inner pole 36, which is connected to die outer pole by a connection portion 31. A toroidal coil 32 is fitted in the space between die poles 34 and 36. The pole 34 has an outer surface which is formed by a conical portion 33. and a cylindrical portion engaging the inner surface of die housing 20. Two seals 35 are fitted to this cylindrical portion to ensure an airtight engagement between die electromagnet 30 and the housing 20, which are secured by fixing members 21. The inner pole 36 has a central hole formed by a small dirough-hole portion 37. a cylindrical middle portion 38 and a conical portion 39 extending towards die free end of the pole 36. An one-way valve 61 or 65 is fitted to one end of the through hole 37 of each electromagnet 30 to cover die hole 37 forming an one-way fluid communication between die exterior of the housing 20 and each of die chambers I and II. The leading edges of the poles 34 and 36 are separated by a non¬ magnetic member 321, which can be a part of die bobbin for die coil 32. The core of die electromagnets 30 can be made of any "soft" magnetic material, such as iron, steel or ferrite. and it can be made by moulding, casing or die-pressing, using particles of magnetic material held together by a bonding agent, e.g. a resin. The piston 50 is axisynunetrical and also symmetrical about the central plane perpendicular to its axis, i.e. its left-hand half is a minor image of its right-hand half. The piston 50 lias a supporting frame formed by a non-magnetic material, including an outer cylinder 51 , an inner tube 53 and a central disc 55 connecting the outer cylinder 51 to die inner tube 53. The material of the disc 55 is thicker than that of die cylinder 51 or tube 53. and a number of ribs 57 and 59 are formed to further strengthen die connection between the cylinder 51 and die disc 55 and between die tube 53 and die disc 55. The central disc 55 is peφendicular to die central axis of the cylinder 51 and the tube 59. which are coaxial. At each end of die outer surface of die cylinder 51 diere is a seal 52 made of low friction and wear resistance material for forming an airtight engagement with the inner surface of the housing 20. To further reduce the frictional resistance between them, the seal 52 can have by a number of spiral ridges. At die central point of die disc 55 there is a through hole 47 which forms the only

fluid communication between the two chambers 1 and II. via an one-way valve 63. The piston 50 carries, on each side thereof, a permanent magnet 40 formed in a structure which is generally complementary to die configuration of the two poles of a conespouding electromagnet 30. That is to say that each permanent magnet 40 has a central pole 46 with a conical outer surface portion 49 conespouding to the conical surface portion 39 of die electromagnet's inner pole 36. and an annular outer pole 44 with a conical inner surface portion 43 corresponding to die conical outer surface portion 33 of the outer pole 34. The central and outer poles 46 and 44 are magnetically connected to each other by a magnetic layer 41 formed on the non-magnetic disc 55. so that the two magnets 40 are magnetically insulated from each odier. As shown in the left-hand side of Fig 1. the peniiaiient magnet 40 is attracted towards die electromagnet 30 at left end of die compressor 10 so that their corresponding poles are brought close to each odier. The leading end of die tube 53 has a seal 54 which forms an airtight contact with the middle portion 38 of the inner pole 36 of the left-hand electromagnet 30.

The supporting frame of the piston 50 can be formed by moulding, casting or die-pressing. Preferably it is die same material as that of the housing 20 so that they have the same theπnal expansion during operation. The magnets 40 can be formed by a separate casting or pressing step after the supporting frame has been made. It is preferable to use rare-earth material for forming the magnets 40, so as to have high magnetic strength and low weight. The magnets 40 can be formed as a thin layer on die frame or even as a film of magnetic coating, as long as they provide a strong enough magnetic field and the required general configuration. An extra coating of a magnetic conductive material can be formed on the surface of the central tube 53 and die cylinder 1 before the magnet is formed on top of it. to make their axial ends magnetically conductive. Since both the housing 20 and the supporting frame of the piston 50 are non-magnetic, the two pairs of the permanent magnets 40 and the electromagnets 30 are magnetically separated. The polarity of the poles of die electromagnets 30 and die permanent magnets 40 are arranged in a manner diat when die two electromagnets are energised at the same time, diey provide a combined push-and-pull force onto die piston 50, forcing it to move towards one end of die compressor 10 (the left-hand end as shown in Fig 1 ). When the cunent is changed, the push-and-pull force will also reverse to force die piston to move to the odier end. One way ot arranging the magnetic polarity is to keep die inner pole 36 of die two electromagnets 30 at die same polarity while the two inner poles 46 of die magnets 40 are in different polarity. For example, die left-hand one is S and the right-hand one is N as shown in Fig 1.

The conical portions 33 and 39 of die electromagnets 30 and die conical portions 43 and 43 of the magnets 40 have the effects of reducing die air gaps between them, while at die same time allowing a relatively long stroke length for the piston movement. As shown in Fig 1, die leugdis of die maximum air gaps are /, and Λ respectively, while the stroke lengdi is / ₍ , which is significantly longer. When the piston is at an end position, one pair of magnets would have a minimum gap, as the left-hand pair shown in Fig. 1. At diis status die pair would have die maximum repelling force when the cunent dirough die coil 32 is reversed. This would compensate die weak attraction of die odier pair which have die maximum gap. As die gap increases at one side of die piston die odier side decreases so die combined push-and-pull force by die two sides would be stable. The anangement has die effects of reducing magnetic leakage to die exterior of die housing for the flux is evenly distributed around die central axis. This will in turn form an evenly distributed concentric driving force between each magnet pair. By such a driving force, the piston is magnetically suspended and axial ly aligned.

dierefore die friction between die piston 50 and die housing 20 is minimized. Since die piston 50 is die only movable component in die compressor, the above "floating" effects of die annular magnetic field would prolong the service life of die compressor significantly.

The fluid circuit through die compressor 10 is formed by three one-way valves 61 , 63 and 65 which connect die inlet (right-hand end as shown in Fig. 1) to chamber II. then to chamber I and finally to the outlet. The valves 61 , 63 and 65 can be any conventional type, such as flap valve, disc valve or ball valve.

As shown in Fig. 1. die seal 54 at left-hand free end of the piston's central tube 53 also serves as a valve to stop fluid flow from the chamber 1 to the outlet valve 65. therefore the remaining space of the chamber I becomes a "dead" space because the fluid in the space is trapped and caiuiot flow out. This is designed to work as a cushion to prevent die piston 50 from hitting the electromagnet 30 at the left-hand end of die compressor. The seal 54 at die odier end of die piston works in the same way in an opposite stroke of the compressor. The size of this buffer space depends on whether the fluid is gas or liquid, and it can be easily adjusted by changing the axial position of the seals 54 along the tube 53. In case of a gas compressor, the space needs to be relative larger to allow the gas trapped in the space to be compressed so the seals 54 are positioned at leading edge of each end of die tube 53. and die space would work as a gas spring, while in case of a liquid pump, the space can be smaller and the seal 54 can be positioned backwards or even omitted because die narrow gap between each end of die tube 53 and the inner wall of the corresponding hole 38 would be small enough to restrict flow rate of the liquid so as to effectively buffer die piston's movement. When die two strokes of the piston 40 are subject to different resistance, e.g. due to higher output pressure which produces higher resistance to the output stroke, the seals 54 at the two opposite ends of die tube 53 can be located at different positions relative to the respective leading edge have different buffering effects which balance die piston's movement.

It is clear now diat aldiough the piston 50 itself is made of light material and carries highly fragile permanent magnets, it is nevertheless fully protected and highly durable due to die evenly distributed driving forces and the fluid buffer arrangement. For control puφoses, a sensor 70 is fitted in die dirough-hole 38 of each electromagnet 30. for detecting die position of die piston 50. The sensor 70 can be simply a series of electric contacts which are connected when in contact widi die seal 54 which is made conductive. The sensor can also be capacitive or magnetic type, which produces sensing signals widiout physical contact. Tlie ouφut signals from die sensor 70 are used to control die cunent to die coil 32 of die electromagnet 30 in response to a change of load. For example, when die electromagnet 30 is powered by an AC input, die signals from the sensor 70 are compared widi d e phase of die input cunent to see whed er die cunent is too large or too small. In case the input cunent is too large, die magnetic driving force to die piston would also be large so die piston 50 would move faster and arrive at die end position earlier, while when input is too small, die magnetic force would be small and move die piston slowly, so it arrives at die end position later, or even not arrive at die position of die sensor at all. This would be checked by die sensor, and die input cunent can be adjusted accordingly. When DC power is used, die signals from die sensors 70 are used to switch the DC input, causing die piston to reciprocate.

Fig. 3 shows a compressor 100 has a cylindrical housing 120, two electromagnets 130 fixed via supporting plates 121 to each end of die housing and a piston 150 between die two electromagnets.. Inlet valves 161 are also fitted on the plates 121. Each electromagnet 130 has an annular core providing two poles

134 and 1 6 and a coil 132 fitted between them. A central comcal cavity 139 is foπned in the central pole 136. The piston 150 has a central permanent magnet 140 having two comcal tips, a cylindrical permanent magnet 145 and an annular member 153 of non-magnetic material for securing the tow magnets 140 and 145 togedier. The magnet 145 is sunounded by a cylinder 151 which is non-magnetic, preferably die same material as the housing 120. The cylinder 151 has seals 156. similar to that of the first embodiment. The tips of the permanent magnet 140 are shaped to fit into the central cavity 139 of the electromagnets 130 when the piston is attracted to one of the electromagnets. The pole direction of die two electromagnets 130 are ananged in die same manner, that is to say, when an electrical cunent is supplied to each of the coils 132, the central poles 136 are both North. In this case the piston 150 is attracted to left as shown in Fig. 3 and repelled from right. By changing the current direction in die coils 132, die piston moves back. Tlie inlet valves 161 are ananged at two ends so diat when die piston 140 is moved from right to left, die valves 160 at right end are opened to let the fluid into the chamber II while the outlet valve 165 at the left end is opened to let the fluid to be forced out of the chamber I . When the piston arrives at the end position as shown in Fig. 3, die leading edge of the outer cylinder 151 blocks the outlet valve 165 to form a buffer space, similar as the first embodiment. In tliis embodiment die second permanent magnet 145 is not essential for its operation. When die magnetic field provided by the magnet 140 is strong enough, die magnet 145 can be replaced by a magnetically conductive member. This embodiment provides a well balanced operation in the sense that the piston's left and right strokes have exactly die same fluid compression function, and are subject to the same moving resistance. When two identical compressors 100 are ananged end to end along the same axis, and their pistons are made to move always in opposite directions, e.g. by coiuiecting the electric supply in opposite direction, die pair would provide a highly balanced anangement in which die impacts by the two pistons always cancel each odier so die pair as a whole cause virtually no vibration. Obviously, the anangement would be highly beneficial to the applications where low noise and vibration are required.

Fig. 4 shows a compressor assembly according to a third preferred embodiment. Fig. 5 on sheet 2 ot the drawings is the cross-sectional view taken along the plane C-C in Fig. 4. showing details of the piston 250. In Fig. 4. the assembly is a multistage anangement formed by two serially connected compressors, 200 and 300, which are of die similar structure. The difference between them is mat die first stage compressor 200 is one size larger dian die next stage compressor 300, so a gaseous working medium can be progressively compressed. If die working medium is liquid, die chain should be made of pumps of die same size for the liquid is not compressible, but die operating principles are die same. Obviously, further stages of compressors can be connected for a higher ouφut pressure, and adapter members can be used between compressors. In die following description, only die details of die compressor 200 are explained. Tlie compressor 200 has a tubular housing 210 widi a lining member 220, two electromagnets 230, a free-piston 250 and a movable support 280. The electromagnets and die piston work in die similar manner as die previous embodiment. In Figs. 4 and 5. die lining member 220 is secured inside die housing 210 by thermal fitting, i.e. by fitting a cool lining 220 into a heated and expanded casing 210 so diat when it cools down die housing grips die lining firmly. The one-way fluid communication through die compressor 200 is formed by die inlet valves 261 , piston valves 263 and an outlet valve 265. As shown in Fig. 5, die piston valves are formed by a number of angled ώrough-holes 262 formed on the piston disc, directing fluid flow in outward tangential directions. A

flap member indicated by the dot line encourages such outward flow, causing die piston to rotate, as shown by die anow sign R.

Between die piston 250 and each of the two electromagnets 230 and 230 ^' , a buffer mechanism is formed by seals 271 and 273. When die piston 250 moves upwards to the end position as shown in Fig. 4, the seal 271 , on the inner surface of the piston cylinder, acts with the outer cylindrical surface of the top electromagnet 230 to form an annular gas cushion between the leading edge of the piston cylinder and the inlet valves 261 on the supporting plate of the top electromagnet 230. At the same time, the inner seal 273 engages the central cavity of die electromagnet 230 to form a central gas cushion in the top cavity 237. These two cushions protect the piston from hitting the upper electromagnet 230. When the piston moves downwards to die lower end, the other inner seal 273 would block the outlet valve 265, t πiiing a cushion between the piston disc and die top end of die lower electromagnet 230'. Lubricant 285 collected around die electromagnet 230 ^' would also have buffer effects.

The movable support 280 canies the lower electromagnet 230' and the outlet valve 265. The support 280 has a base 281 , a sleeve 282 and a biasing spring 284. which urges the base away from the piston. The leading edge of the sleeve 282 forms a sliding contact with die lower edge of die lining 220, and defines a chamber 212. in which is located die spring 284. The sleeve 282 also serves as a stopper, when it contacts with the top edge of the chamber 212, to define the upmost position of the support 280. The range of movement for the support 280 is shown by the double-arrow sign 286. In operation, when the output pressure is low, e.g. during die start-up of the compressor, the spring 284 urges the support 280 downwards against the leading end of die next stage compressor 300. and at diis position the piston has the longest stroke, so die compressor has the largest throughput. As the ouφut pressure is gradually built up, the total force on die bottom surface of the base 281 would eventually become larger dian die biasing force of the spring 284, so the support 280 would be forced to move upwards. This movement reduces die stroke lengdi dierefore reduces the output by each stroke. Furthermore, due to the reduced stroke lengdi, die average air gap in die magnetic circuit is also reduced, with die effect of an increased driving force, producing a higher ouφut pressure. When diere is a big enough drop of die output pressure, e.g. due to an increased release of die compressed medium at die system outlet, die support 280 would immediately resume its original position under the biasing force of die spring 284, dien die compressor is ready to work at its maximum output rate. That is to say, the compressor changes automatically from a low-pressure high-ouφut operation to a high-pressure low-output operation, or vice versa. This automatic adjusmient becomes more beneficial when a number of compressors are comiected serially in a multistage anangement, in which each of iem can adjust its own rate to match with the others in die chain so diat die load is eveidy distributed over die whole chain.

A lubricant 285 used in die compressor 200 keeps die outer surface of die piston 250 and die inner surface of die lining 220 lubricated, and it would end up in die collecting area around die lower electromagnet 230'. A lubricant circulating circuit is formed by die grooves 283 formed on die inner surface of die sleeve member 282, the chamber 212, the grooves 211 formed between the housing 210 and die lining 220, and die ώrough-holes 221 which return die lubricant back to die piston. The grooves 21 1 and 283 are made small enough so diat die lubricant is sucked into die grooves mainly by capillary effects. When die piston 250 moves downwards, the increased pressure on die liquid surface would help die liquid into die chamber 212 then to

enter the capillary grooves 21 1 , while on the other hand die holes 221 would be exposed to die low pressure side of die piston, allowing die liquid to come out of die grooves. When die piston moves upwards in a retuπi stroke, die valves 263 would open, causing die piston to rotate as mentioned above, dierefore to spread die lubricant evenly around the whole inner surface of die lining 220, forming a film of lubricant between the outer surface of the piston and the ii er surface of the lining 220. This film also improves the sealing around the piston. The holes 221 are positioned so that they are covered when the piston is at its upper position, not causing gas leakage between the two sides of die piston.

Figs. 6A to 8B show a compressor 500 according to a fourth embodiment. Figs. 6A and 6B are cross- sectional views showing the positional changes of die different parts of die compressor during its operation. Fig. 7 A shows more details of the piston 560. and Fig. 7B is the cross-sectional view taken along the plane B-B in Fig. 7A. Similarly, Figs. 8A and 8B show the details of the electromagnet 530 and buffer anangenients 540 and 550.

In Fig. 6A, the compressor 500 has a housing 510 with a lining 51 1 similar to that shown in Figs. 4 and 5. a free piston 560 with an active magnetic shunt mechanism 570, two electromagnets 530 and 530' . one supported by a plate 520 and the other by a movable support 580. Each electromagnet has outer and iiuier buffer mechanisms 540 and 550, or 540' and 550' for cushioning the piston 560. The end plate 520 and the movable support 580 also provide die fluid inlets 521 and 523 and outlets 582, respectively. The lining member 51 1 , which is made of non-magnetic steel, and die movable support 580 are similar to and operate in die same way as die members 220 and 280 shown in Fig. 4 to provide lubricant circulation and automatic stroke lengdi adjustment. When the compressor 500 is used to work widi a liquid, e.g. it can be used as a hydraulic pump for high pressure and leak-free applications, die piston can be lubricated by die working medium therefore the lubricant circulation arrangement would not be needed. In this case further outlet holes can be formed close to die outer periphery of die support 580 to match die holes 521 on die top end plate 520.

When analysed from die viewpoint of die fluid compression, die piston's down strokes suck in fluid through the inlet holes 521 and 523 on die top plate 520, and at the same time drive die fluid already in the compressor out of die outlet holes 582 in die bottom support 580. In contrast, die upward strokes merely force the fluid from upper side of die piston to its lower side widiout producing any ouφut. That is to say, the two strokes are unbalanced widi most of die actual work done by die downward strokes and with its upward strokes simply as return movements. Such unbalanced piston movements reduce die compressor's energy efficiency. In order to tackle diis problem, anangemeiits are made in die compressor 500 to balance die two types of strokes by converting die kinetic energy of die upward piston movements into energy reserve in different buffer mechanisms, which is men released during the piston's downward strokes. These anangemeiits include the inner and outer buffers 540. 550, 540' and 550' associated widi die two electromagnets 530 and 530' and die springs in the shunt mechanism 570. In Fig. 7A, die piston 560 has a main permanent magnet 561 in die shape of an annular disc. The polarity of die magnet 561 is in radial direction, i.e. with its outer periphery as die soudi S, and die inner periphery as the north N. The thickness of die disc 561 increases from its outer periphery inwards to ensure diat die inner and outer pole faces are of the similar sizes and die cross sectional area perpendicular to die magnet flux maintains unchanged along die flux direction to avoid local flux saturation. A cylindrical magnetic

-π- member 562 is secured to the outer periphery of the disc 561. serving as the pole piece for the outer pole and a generally tubular pole piece 563 is secured to die iiuier pole, so as to form generally radially directed magnetic flux between them, as shown by die dash lines in Fig. 7A.

The disc 561 is sandwiched between two protection member 564 or 564 ^' to protect diis relatively brittle member from being damaged by mechanical shocks and also to hold the disc and its two pole pieces together. The two members 564 and 564' can be formed by injection moulding, preferably by using gasified plastics or foamed rubber, after the disc and die pole pieces having been assembled so as to foπn an integrated structure of good mechanical strengdi and light weight. The members 564 and 564' also define smooth surfaces which closely match with the pole faces of die two electromagnets, to reduce any "dead space" between die piston faces and the driving electromagnets. The member 564' also has a groove fitted with a valve member 566. As also shown in Fig. 7B, a number of angled dirough-holes 565 are formed in the disc 561 and the protection members 564 and 564', which are covered by the valve member 566 for causing the piston 560 to rotate in direction R. similar to the previous embodiment.

The inner pole piece 563 has a central supporting part 567 which carries a magnetic shunt mechanism 570. The shunt 570 is formed by two permanent magnets 571 and 571 ' comiected together by a non-magnetic bar 572. The bar 572 is movably supported by the part 567 which has seals to prevent any leakage dirough the sliding engagement between them. Two springs 573 and 574 are each used between the part 567 and one of die magnets 571 and 571 ' to provide biasing forces. Low friction bushing is used for die springs to allow die shunt to rotate relative to the piston. Tlie springs 573 and 574 are made of magnetic steel so diey also provide magnetic comiection, via die inner pole piece 563, between die magnet 571 or 571 ' and die disc magnet 561. The polarity of die shunt magnets are ananged in a way that both of diem are magnetically attracted by the iiuier pole of disc magnet 561. The leading face of the shunt magnet 571 or 571 ' is covered by a soft layer, such as rubber, to protect it.

The piston's magnetic circuit has two branches, each includes an air gap between a leading edge of die pole piece 562 and the pole face of die magnet 571 or 571 ' for acting widi a conesponding electromagnet, and the main magnet 561 serves as die common route shared by the two circuit branches. In Fig. 7 A. the magnetic shunt 570 is at a neutral position relative to die disc magnet 561, where die two circuit branches have equal flux distribution indicated by die dash lines. However, diis balanced flux distribution is unstable because a very slight axial movement of die shunt 570 relative to die disc 561 would increase die flux in one branch at die expense of the odier. In odier words, from die viewpoint of die flux distribution die shunt 570 serves as a magnetic switch which decides by its axial position that which circuit branch would have a bigger share of the total flux. This anangement is also used as a buffer mechanism by having a natural status predetermined for the magnetic switch by die two springs 573 and 574 which are ananged to form a push-and-pull pair, i.e. die upper spring 573 is a compression one which tends to expand and push die magnet 571 upwards and away from the disc magnet 561 , while the lower spring 574 is an extension one which tends to contract and pull the magnet 571 ' towards the disc magnet 561. When not subjected to external forces the shunt 570 would end up at a position where die magnet 571 ' engages with the lower end of die imier pole piece 563, pushing die odier magnet 571 , via die bar 572, to a position far away from the top end of die inner pole piece 563. At diis position die flux of die lower branch would be much larger than that of die upper branch. In operation the

piston has to overcome die biasing forces of die spπngs 573 and 574 in its retuni strokes and dien release the energy during its output strokes Furthermore, the axial movement of the shunt 570 switches die main magnetic flux from one end of the piston to the other so as to act with electromagnets 530 or 530' in a more efficient way. as to be descnbed later Fig 8 A is a partial sectional view with die right half of die electromagnet 530 and the outer buffer 540 shown in section and their left half in front view It also shows details of the top end plate 520 and the inner buffer 550 Details of the conespouding members 530' 540' and 550' at the bottom end ot the compressor are basically the same, unless to be descnbed odierwise

More specifically, the end plate 520 is secured onto the housing 510 and it carnes die electromagnet 530 and its butter niecha snis A series of outer inlet holes 521 and a series of inner inlet holes 523 are formed on the end plate 520, conesponding to the outer penphery and the central hole of the a iular electromagnet 530. and being covered by flap member 522 or 524 to form one-way valves Details of die flap members 522 and 524 are not shown in Fig 8B but they are similar to the flap member 566 shown in Fig 7B The new features ot the electromagnet 530 are the angled slots 536 and 537 toπned on its outer and inner comcal pole faces, and die slots 534 and 535 formed on the opposite end all in axial direction The slots are toπned to facilitate die fluid flow from the holes 521 and 523 into die compressor chamber and at the same time helping to uiol the electromagnet 530 and to reduce eddy cunents in the core material As also shown in Fig 6B that the slots 534 and 536 on the surface of the outer pole 5 1 are not connected to one other neither the slots 535 and 537 on the surface of the inner pole 533 On each of the surface a narrow band free ot an> slots remains between the two sets of slots, which co-operates widi a seal member 547 or 557 of die two buffer 540 and 550 to provide air cushion effects

Associated with the electromagnet 530 are the outer buffer 540 surrounding it and an inner buffer 550 in its central hole The buffer 540 includes a cushion nng 541 biased downwards by an expansion spnng 549 The nng 541 has a number of retaining fingers 542 extending in axial direction, each having a hooked tip 543 fitted into one of the slots 534 to define die range of the buffer's movement by the length of die slots 534 ax shown in Figs 6A and 6B The ring 541 with its finger members 542 is made ot a non-magnetic and rigid matenal, such as aluminium or plastics, to provide good mechaiucal streugdi Tlie πng 541 carnes a flux coupler 544 made of a magnetic matenal, which is in die shape of a nng carrying a number of fingers 545 each widi a hooked tip 546 The fingers 545 are made of flexible and resilient matenal so diat die fingers can bend easily when attracted by magnetic force The tips 546 are rigid and they also have the function of limiting the buffer's upward movement by engaging the leading edge of the outer pole 531 The fingers 545 are configured to fit closely to the comcal pole face of die outer pole 531 when die cushion nng is in die contracted position as shown in Fig 8A Shallow groves can be formed on die comcal surface so that die fingers 545 can fit into diem to provide a smoodi pole face to match widi die complementing face of die piston 560, as mentioned above Tlie sealing nng 547 is clamped between die nngs 541 and 544, to form an air cushion when it engages die annular band on die outer pole 531

Similar to die outer buffer 540. die inner buffer 550 has a cushion nng 551 of a non-magnetic and πgid material, biased by a spnng 559 The nng 551 has a number of retaining fingers 552 each having a hooked tip 553 fitted into one of die slots 535 to define its movement A flux coupler 554 has a number of

-ι.v resilient fingers 555 each widi a hooked tip 556 to match with a conesponding tip 546 of die outer buffer 540. A sealing ring 557 is fitted between die rings 551 and 554, for forming an air cushion in the central hole of die electromagnet 530. The imier buffer 550 also includes a permanent magnet 558 widi a polarity repulsive to that of die shunt magnet 571 for providing magnetic cushion to die piston 560. The operations of die buffers 540 and 550 are to be explained below.

Now the operations of the compressor 500 are to be described with reference to Figs. 6A and 6B, in which different positions of die piston 560 are shown together with the changes of die cushion and driving mechanisms.

Firstly, the operation of die buffer mechanisms is explained in details. Fig. 6A shows diat the piston 560 is oved to its up end position where it is well cushioned by the two buffers 540 and 550, which provide air cushion by die effects of die seals 547 and 557, spring cushion by the compressed springs 549 and 559 and magnetic cushion by the central magnet 558 which expels die shunt magnet 571. In Fig. 6B, the piston 560 is in the middle of its downward stroke, as indicated by the arrow A. At the this particular position, in addition to the magnetic driving force by the electromagnets, the piston's outer cylinder pole piece is pushed downwards by die outer cushion ring 541 and the shunt 570 is forced down by die inner buffer 550. The piston is moving to die bottom end, where the main cushion effect will be provided by the resistance of the compressed fluid, which is being forced out through die outlet holes 582, togedier widi die effects of the buffers 540' and 550' at the end of the movement. In addition to diese, the lubricating liquid surrounding the lower outer periphery of die electromagnet 530' also helps to buffer die piston's final movement, dierefore air cushion effects are not provided by die buffers 540' and 550'. i.e. dieir rings 547' and 557' do not have sealing effects.

Since die working load of die piston's upward strokes is smaller dian diat of die downward strokes, as mentioned before, the upper cushion springs 549 and 559 are selected to be stronger dian the springs 549' and 559'. That is to say, die piston's upward strokes are powered basically by the push-and-pull force of the electromagnets alone and the piston's upward kinetic energy is convened into die elastic energy of the springs 549 and 559 for helping die piston's down stroke. Further compensation is made by die permanent magnets 558 and 558', which are ananged with die permanent 558 opposing die piston's upward movement while die magnet 558' attracting its downward movement, forming anodier push-and-pull pair. Finally die springs 573 and 574 of die shunt mechanism 570 also urges the piston downwards, as mentioned above. Due to diese compensating anangemeiits. a significant portion of the push-and-pull force by die two electromagnets for piston's upward strokes is converted to make the piston's downward strokes more powerful, dierefore effectively balanced die piston's movements.

In Fig. 6A, die piston 560 is driven to die top where die upper branch of die piston's magnetic circuit is included into die magnetic circuit of die electromagnet 530. At the same time, die magnetic circuit of die bottom electromagnet 530' is "closed" by die magnetic fingers 545' and 555' of the buffers 540' and 550' because die fingers are raised by springs and magnetised by die electromagnet to attract one anodier until diey bend and dieir tips contact one anoώer. This magnetic coupling increases die electromagnet's self-induction, dierefore reduce die electric cunent passing through its coil at this particular moment. Since at diis moment die distance between die electromagnet 530' and die piston is large, die electromagnet has no important influence to the piston's movement. In odier words, electric power consumption by die electromagnet 530' is saved when it

is lest effective. In Fig. 6B, here die electric cunents to die electromagnets are reversed, the piston 560 is driven downwards so die lower shunt magnet 571 ' contacts the buffer 550' first to break die magnetic coupling between the fingers 545' and 555'. Once they are separated, the fingers 545' are attracted to die lower edge of piston's outer pole piece to form a more effective magnetic comiection between the outer pole of the electromagnet 530' and the outer pole of die piston, and die fingers 555 ^' are attracted to die shunt magnet 571 ' . Such contacts foπn direct comiection between die lower branch of piston's magnetic circuit and the electromagnet 530' , causing stronger attraction between them. On the other hand, once the piston has moved away from the top electromagnet 530. it makes enough room for the fingers 545 and 555, attracted to one anodier by dieir magnetic polarity, to bend until diey touch one another. When this happens, the magnetic flux from die electromagnet 530 is effectively "turned off from the piston 560.

It is clear from the above description diat by incoφorating the magnetic coupling mechanisms, each of the electromagnets is automatically "turned off when it no longer has effective contribution in driving the piston, therefore the consumption of electric energy at such ineffective moments is significantly reduced and the compressor's overall energy efficiency is improved. Fuitheniiore. the anangement also improves magnetic comiection between the piston and the electromagnet by reducing the average size of the air gaps between them, i.e. increasing the magnetic flux and die effectiveness of the magnetic comiection. This is shown by the facts diat when in operation die piston has a stroke lengdi S shown at die left side of Fig. 6A, while the maximum air gap between the shunt magnet 571 ' and die magnetic fingers 555' is merely #, shown in Fig. 6A, which is much smaller compared with S. In addition, each of the buffer mechanisms 540. 550, 540' and 550' has a movable range of d. which is slightly less dun a half of die stroke lengdi 5. That is to say, for most of time in operation die piston 560, including die shunt 570, is in direct contact widi at least one of die buffers which serves as pole extension to act between die piston and electromagnets.

Figs. 9A to 10B show a piston 600 according to a fifth embodiment. This piston can be used in die compressor of the previous embodiment to replace die piston 560. and it is suitable when the compressor's dimensions are large. Figs. 9 A and 9B are cross-sectional views showing die positional changes of its different parts during its operation. Fig. 10A is the cross-sectional view taken along the plane C-C in Fig. 9A, showing details of a magnet disc 610. and Fig. 10B is die cross-sectional view taken along the plane D-D in Fig. 9A. showing details of a magnet ring 640' .

In Figs. 9A and 9B, die piston 600 has an annular permanent magnet disc 610 sandwiched between protection layers 620 and 620'. An outer pole piece 630 engages die outer periphery of die disc 610 and an inner pole piece 660 engages its inner periphery. Tlie inner pole piece 660 canies an active magnetic shunt 650 which works in a similar way as diat of die previous embodiment. More specifically, as also shown in Fig. 10A, the disc 610 has a number of inner channels 612 and outer channels 613 formed in axial direction on its inner and outer periphery. The protective layers 620 and 620' on two sides of die disc are made by moulding plastic or rubber material to cover bodi sides and diey are connected to each odier via diese channels to foπn an integrated body widi die disc "caged" inside. This structure protects die brittle disc and also defines angled valve holes 61 1 covered by a valve member 621.

Now back to Fig. 9A, die outer pole piece 630 includes a cylinder member 631 made of a soft magnetic material, which is ananged to have sliding engagement widi die outer periphery of die disc 610 fitted

with two sliding seals 614. Two permanent magnet rings 640 are each fitted to an end of die cylinder 631 , to provide a secondary magnet in each of the two circuit branches of the piston's magnetic circuit. A seal 634 is clamped between each ring 640 or 640' and the cylinder 631. A pair of springs 633 and 633' are each fitted between a ring 640 and a seal 614, to keep die disc 610 biased relative to die rings 640. The imier pole piece 660 includes a magnetic cylinder 662 engaging die iiuier periphery of the disc 610, two magnetic end caps 661 each fitted to an end of the cylinder 662, and a non-magnetic filling stuff 663. such as gasified plastics, filed in the cylinder 662. The shunt 650 includes two permanent magnet caps 651 and 651 ' comiected to each other by a non-magnetic tube 652, which is supported on the inner pole piece 660 by die filling stuff 663 and sealed by seals 664. A pair of springs 653 and 653' keep die shunt 650 biased relative to the pole piece 660. The magnet cap 651 or 651 ' is shaped to have a leading surface providing pole face matching widi that of an electromagnet as shown in Figs. 6A and 6B, and a back pole face matching diat of the cap 661 or 661 '. The magnet caps 651 and 651 ' foπn the secondary magnets at the odier end of the piston's two circuit branches. The polarities of die main and secondary magnets are ananged, as shown in Figs. 9A and 9B. to make each secondary magnet being attracted by the main magnet 610. When in operation, the disc 610 would slide along the cylinder 631 towards one end, under the effects of die two electromagnets. Fig. 9A shows the disc at its lower end, where the lower protection member 620' engages both the magnet ring 640' and die magnet cap 651 ' to define a snioodi piston face matching that of die pole faces of die lower electromagnet (not shown). At this position, die inner pole of the disc 610 is magnetically comiected to die cap 651 ' via the cylinder 662 and the cap 661 ' of the iiuier pole piece 660. so is the outer pole to the magnet ring 640' via the cylinder 631 and the spring 633'. From the viewpoint of the lower electromagnet, the piston 600 behaves as one powerful permanent magnet.

As shown in Fig. 9B, once die electric cunents in the two electromagnets are reversed, die piston assembly is forced to move upwards, as indicated by an anow A to the left. At the same time, since die disc 610 is no longer attracted by the lower electromagnet, it is forced by the springs 633 and 633' to move upwards along die cylinder 631. Similarly, die imier pole piece 660 is also forced upwards by die springs 653 and 653' until die upper edge of die cap 661 ' engages die disc 610 and carries die same to move with it. Since die inner pole piece 660 is very light and it is not subjected to any significant fluid resistance to its movement, it can move much faster than die disc 610, so as to effectively switch die disk's main magnetic flux from its lower branch to its upper branch, as described before in relation to die previous embodiment. Again, diese springs can be arranged to provide more downward biasing forces for balancing the piston's movements. Also die cushioning and magnetic coupling for die piston 600 can be die same. In operation, die disc 610 with its protection layers 620 and 621 ' can slide a distance 57 along die outer pole piece 630, in additional to the piston's stroke lengdi 5, as shown in Fig. 6A. That is to say, die piston's effective stroke lengdi is increased wid out increasing the pole pieces' movements, so die piston has better agility in response to die driving forces by die electromagnets. Tlie reduced movements of die outer pole piece 630, which contributes to a very large portion of die piston's total weight due to die steel cylinder 631, ensure that the overall energy efficiency is improved. To keep the piston parts lubricated, a number of small holes are made Uirough the cylinder 631, so that lubricant can enter the gap between the outer periphery of the disc 610 and die inner surface of die cylinder 631. Furthermore, a fabric layer 615 is sandwiched between each side face of die disc 610 and a conespouding

protection layer 620 or 620' to provide capillary passages for the lubricant to enter the gap between the imier periphery of the disc 610 and the imier pole piece 660.

Finally, an additional feature is that the leading pole face of the magnet rings 640 and 640' has a number of channels 641 and 641 ' as shown in Fig. 10B. which match the chaimels 536 fanned on die outer pole face 531 in Fig. 6B. During the operation, these chaimels have the effects of causing the piston to rotate so that they can align with one another, in a way similar to the toothed patterns in the magnetic core of a stepping motor. By arranging the chaimels on the two rings 640 and 640' at staggered annular positions, they will ensure, togedier with the effects of the angled valve holes 61 1 , continued rotation of the piston 600 during its reciprocating movements, so as to improve its lubrication and reduce its wear. The chaimels 641 are filled by a non-niagiietic material to define a smooth piston face. Industrial Applicability

The reciprocating mechanisms disclosed above can be used in many different applications. Due to the tubular structure of the embodiments, they are especially advantageous in meeting the needs of an existing system design or hardware pipeline because such a compressor or pu p can be easily fitted into an existing system at any position that a driving force is needed, without any need to re-design die system or change the existing hardware. Standard models can be comiected serially or in parallel to meet a wide varieties of requirements to output pressures and/or flow rates. The tubular structure of the compressor makes it easy for heat dissipation, therefore ideal for applications in refrigeration and air-conditioning.

The above description is made by way of examples, which also indicates diat different features described above, such as the polarity and/or general shape of the poles of the magnets or the flow direction of die valve anangemeiits. can be modified or changed by combining die described features in different ways. For example, the cross-sectional shape of the anangement is described as cylindrical, but it can be replaced by odier generally symmetrical polygonal shapes, especially when die number of sides of a polygon is large. Furtheπnore. the pistons widi permanent magnets have die advantage of reduced weight, therefore high efficiency. Obviously, electromagnets can be used which have advantages that die magnetism will not be affected by temperature changes, desirable for high temperature applications. On the odier hand, when the piston canies an electromagnet, die two driving magnets fitted to the housing can use permanent magnets so as to reduce die total weight of the motor /compressor as a whole.

Now details of the invention according to the second general object are described. Figs. 1 1 and 12 show a pump 10 having a cylindrical stator 20 and two end caps 30, each fitted to one end of the stator 20, for defining a cylindrical imier space in which is fitted a rotor 40. Each cap 30 has a central opening which is coaxial widi the cylindrical stator 20 and the rotor 40 so diat a fluid passage is formed along die rotational axis of die pump 10. The stator 20 has an outer housing 22. a magnetic core 24 with windings 26 and bores 28 indicated by dash lines. An air gap 72 is formed between die core 24 and die rotor 40 to generate, when electric cuπents are supplied to die windings 26. a rotational magnetic field as in a conventional electric motor. Tlie caps 30 are of die same general structure, i.e. having an annular flat wall pan 32. an annular slope wall part 36 which defines a central opening, and a sleeve part 34. Each cap is secured at its outer edge to die stator 20 by a number of fast members 82. while at die inner edge near die central opening it is secured to a tube 90 by fast members 84. On the inner surface of the slope wall part 36, diere are a

number of chaimels 38 extending radially from the iiuier edge of the central opening of the cap 30 to and passing through die sleeve 34. to enter a chamber 74 formed between the cap and the stator. The two chambers 74 are in fluid communication via the bores 28. and with die gap 72 via die channels 38. Sealing members 61. 62 and 64 are used to ensure sealing between the cap 30 and the stator 20, and between the cap and tubes 90, so diat die comiection between die pump and pipeline is leak-free.

In Fig. 12, the rotor 40 has a tube 42 carrying a magnetic core 44 on its outer surface secured between two supporting members 46. The core 44 can be constructed as for a squirrel cage motor, to rotate under the influence of the rotational magnetic field generated by the stator 20. The tube 42 has on its imier surface an propelling member 48 in the form of a screw blade, so that when die rotor 40 rotates die blade 48 drives die fluid in the central passage to flow along its axial direction, as shown by the anow A. On the outer surface of the rotor 40. there are a number of spiral vanes 49 fanned by cutting the surface of the magnetic core 44. or by adding a sleeve with the vanes formed on its surface, and this member can be made of an elastic material. The rotor 40 is supported by two mechanical bearings 50. each having an imier bearing member 52 fitted to each end of the rotor 40, an outer bearing member 54 fitted to the cap 30 between the sleeve part 34 and the slope wall part 36. and a number of rollers 56. The purposes of die bearings 50 are to provide both thrust bearing and rotary bearing to die rotor 40 to keep it coaxial with the stator and to ensure its smooth rotation. At each end of the rotor 40, further seals 63 are fitted in die seal grooves 43 (shown in Fig. 12) to provide a fluid-tight contact between the rotor 40 and the cap 30. This arrangement ensures that the air gap 72 is sealed from the main-stream axial flow indicated by die arrow A, so that no dirt carried by the flow can enter the gap 72. In case a conosive fluid is traiisfened in die pipeline, effective sealing anangemeiits also protect die odier components of the pump from being conoded since only die caps 30, die seals 63, die tube 42 and die blade 48 are in direct contact with the corrosive fluid.

A lubricating-cooling anangement is fanned between the rotor 40 and the i ier surfaces of the pump body by filling the gap 72 widi a lubricating-cooling medium. Since die gap 72 is in fluid communication with chambers 74, die same medium can circulate via the channels 38 and die dirough bores 28 in die magnetic core 24 so as to work as lubricating fluid to die outer surface of die rotor 40 and die bearings 50 and as cooling fluid to die windings 26 and the magnetic core 24. Tlie spiral vanes 49 on die outer surface of the rotor 40 have a spiral direction opposite to diat of die inner blade 48, to drive die medium in die gap 72 to flow in die opposite direction through die channels 38 into die chamber 74 at die left end of die pump. Conventional transformer oil can be used as die medium for this puφose. It should be noted that the peripheral cushion of die medium also serves as a hydraulic bearing to help keeping die rotor coaxial with die stator core 24 and to provide a counter- thrust to die rotor, which reduces die burden of die mechanical bearings 50. To enhance the hydraulic bearing effects, spring biased one-way valves (not shown) can be fitted, e.g. in die clumber 74 to block die oil flow into it so diat only when die pressure of the oil in die gap 72 is built up by die rotation of the rotor 40 to a value higher than diis biasing force, i.e. when the rotor 40 rotates beyond a certain speed level, die circulation of the oil staπs. A small oil supply tank (not shown) can be fitted to die pump body for supplying oil into diis lubricating-cooling system to compensate any leakage into die main-stream flow via the seals 63, for diis purposes, a stable oil pressure need to be maintained in die tank, e.g. by a spring loaded piston. Such

anangement is known in die an and do not need further description. When the stator windings 26 are comiected to an electric power supply, a rotational magnetic field is formed between die stator and die rotor which drives the rotor 40 to rotate. As shown in Fig. 1 1. when the rotational direction of die rotor is diat shown by die anow R, the flow direction is that of the arrow A. The blade 48 forces die fluid in the central passage to flow in its axial direction while at die same time die rotor itself is subject to a counter dirust in the opposite direction. The flow direction is reversible by controlling the electric power supply, hence the rotational direction of the magnetic field. On the odier hand, as the rotor 40 rotates, the spiral vanes 49 on the outer surface of the rotor 40 would produce a peripheral flow of the lubricating-cooling oil in the direction as shown by the anow B. which is opposite to diat of die axial flow inside the rotor because their spiral direction is opposite to that of the blade 48. therefore it always provides a counter thrust which helps to balance the rotor's radial and axial position and reduce the thrust force on the bearing. It should be noted that when the rotor's rotating speed is increased, i.e. when the rotor is subject to an increased thrust force, die vanes 49 would produce an increased peripheral flow with an increased balancing effect. That is to say the arrangement is sensitive to the changes of the rotor movements and it provides self-balancing effects which compensate such changes automatically. The back flow of the lubricating oil also keeps the oil pressure higher at the upstream side (left-hand side in Fig. 1 1 ) to prevent dirt from entering the gap 72 , and also keeping the seals 63 well lubricated. In this anangement, as long as the seals 63. especially those at the upstrea side which are more subject to the thrust force of the axial flow, are intact, the pump would be able to operate normally because all the other components are less eligible wear. Fig. 13 shows a second embodiment of a pump 100 having a general structure similar to that of die first embodiment. A stator 120 and a rotor 140 are ananged to foπn a brushless DC motor, widi an array of permanent magnets 144, preferably rare earth magnets, fitted on die rotor. This makes it suitable for applications requiring easy control of both rotational speed and direction, such as in case of a marine propeller. The rotor 140 is fitted in die pump body as a "free" rotor, in the sense diat during its operation, the rotor is fully suspended and levitated by magnetic and hydraulic forces widiout direct physical contact with any support member. At each end of die rotor 140, diere is a space of a distance "d ". as shown in Fig. 13, which allows die rotor to have a limited axial movement. This fully suspended or "floated" state of die rotor is maintained by die following suspension bearing mechanisms.

Firstly, a magnetic axial registration mechanism is formed by having a number of non-magnetic or high magnetic resistance rings 127 on the inner surface of die stator's magnetic core 124, and the same number of similar rings 147 on die outer surface of the rotor, and die two sets of rings match each other when the stator and die rotor are in axial alignment. These rings separate die magnetic coupling between die stator and the rotor into separated zones so that a maximum magnetic coupling can be achieved only when die zones have a perfect axial registration, which forms a neutral position, as shown in Fig. 13. Any axial movement of die rotor will cause misalignment of diese magnetic zones dierefore increase die total magnetic resistance in die magnetic circuit formed by die stator and the rotor hence reduce die magnetic flux. The mechanism is highly sensitive to any axial movement of the rotor and an axial force would be generated between die stator and die rotor which tends to return die rotor to its neutral position. When the magnets 144 are of strong rare earth

material. diey produce strong registration force against any misalignment of the rotor even when the currents to die windings 126 are switched off.

Secondly, each magnetic suspension bearing 150 has a comcal ring shaped electromagnet 154 fitted in die cap 130 and a matching conical ring shaped permanent magnet 152 fitted to an end of die rotor 140. The electromagnet 154 has a magnetic core 157 with a generally U-shaped cross-section, a toroidal coil 155 fitted in the chaiuiel of the core 157 and a non-magnetic member 156 covering die coil in the core. When there is an electric current in the coil 155. a magnetic flux forms between the two anus of the U-shaped core across the member 156, to form an evenly distributed annular field if viewed in the axial direction. The polarity can be changed by changing die cunent direction in die coil. The annular permanent magnet 152 has a similar U- shaped cross-section with a non-magnetic member 153 in it to separate die two poles S and N of the magnet. Obviously, the two magnets 152 and 154 will produce between them a repulsive or attractive force according to the current direction in die coil 155. These two magnets are ananged to have their opposing surfaces in generally conical shape (similar to that of the bearing member 52 shown in Fig. 12) which are complementary to each other so that a centring farce formed between them would tend to keep them in a coaxial relationship. The non-magnetic members 153 and 156 are preferably made of low friction materials, such copper, bearing steel or Teflon, so that diey can have between them low friction sliding contact as a further buffer anangement to the rotor ^' s axial movement.

The operation of this suspension mechanism is as follows. Before the stator 120 is switched on. each of the electromagnet 154 is supplied with a current of the same value, to generate from two ends of the rotor two opposite repelling forces which are of die same streiigdi. These opposite forces would "clamp" die rotor at its neutral and balanced position and at die same time keep it radially suspended in the pump body. When an operating current is supplied to die stator, which drives die suspended rotor to rotate, a control signal which is proportional to the stator's average operating cunent is supplied to each of the electromagnets 154 with the effects of increasing the repelling force of the one at the upstream end of the rotor (left-hand end in Fig. 1 ) while reducing the repelling force at die other end. The combined effect by this control signal is a net force directing towards the downstream end of die rotor, which is sensitive to die rotor's movement and tends to balance the dirust force on die rotor produced by die axial flow of die liquid. In diis way. die rotor's axial position is automatically maintained. When die rotation is reversed by changing die direction of the stator's operating cunent, so is the direction of die control signal supplied to die electromagnets 154, widi the effects of having a reversed balancing force automatically. In order to make further accurate adjustment of this sensitive suspension bearing, sensors 166 are fitted to each end of die pump body which constantly monitor die axial position of the rotor. The ouφut signals from die sensors are sent to a control unit (not shown) to further adjust die cunent supplied to the electromagnets 154, so diat when a significant displacement of die rotor is detected, die bearings 150 can be further adjusted, e.g. having one generating an repelling force and the odier an attracting force to help to retuni die rotor to its neutral position. The sensors 166 can be any conventional type, such as capacitive sensors. Hall's effect sensors, optical or infrared sensors or ultrasonic sensors.

Thirdly, the spiral vanes 149 formed on the outer surface of the rotor 140 work in a way similar to diat of die first embodiment, i.e. producing a peripheral fluid cushion which operates as a hydraulic bearing. This

hydraulic bearing is formed by sucking water into die gap 172 from the downstream end of the rotor, as shown by anows In at the right-hand end of die rotor 140 in Fig. 13, driving it to flow backwards and returning it at the upstream end of the rotor, as shown by the anows Out. By arranging a throttle ring 160 at each end of die rotor 140, this hydraulic bearing also works to damp any axial oscillation of the rotor. For example, when the rotor experiences a sudden increase of die thrusting force, e.g. due to a sudden increase of the output resistance or die increase of the input electric power, it would be forced to move backwards. This brings the throttle ring 160 at the upstream end closer to the member 154. therefore partially block the outlet for the back flow circulation. The reduced outlet will at the same time lead to an increase of the water pressure in the space between die magnets 152 and 154 which is a part of the gap 172, that in turn resists the rotor's further axial movement. This damping effect works when there is a sudden movement of the rotor so it protects the free rotor from hitting the pu p body. The gradual change of the rotor ^' s axial position is mainly counter-balanced by the forces provided by the magnetic registration and the suspension bearings 150 as explained above.

In Fig. 13, it is also shown that the tube 142 has an increased thickness compared with that of the first embodiment. This is made to form a buoyant structure by using light materials, such as plastics or resin, or by using etal material with chaimels or cavities filled with light material to reduce its total weight. The intended effects are to make the rotor 140 as a whole to have a gravity close to the liquid to be pumped, so that when the pump is filled with the liquid, the rotor "floats" in the liquid. This will help to suspend the rotor in the liquid and reduce its radial and axial oscillation relative to the stator when it is suspended. The portion 143 at each end of the tubular member 142 forms a narrowed entrance for liquid to enter the rotor, which has the effect of increasing the flow speed at each end of the rotor. This is intended to reduce the opportunity for the solid particles carried by the flow, such as sand or rust, to enter die gap 172 between die rotor and the stator. When the solid particles are carried by die flow to die inlet end of die rotor 140 (the left-hand end in Fig. 3), they are driven to the centre of the flow by the back flow coming out of the gap 1 2. as shown by the anows Out. This annular outlet of the back flow joins the mainstream and furdier increase die flow speed at the narrow entrance defined by the portion 143, which helps to carry the solid particle, if any, into the hollow rotor. Once the panicles are inside the rotor, because diey are heavier d n die liquid, they would be urged by the centrifugal force of the rotor against die imier surface of the fluid passage and be moved forwards by the axial flow. As diey are moved to die outlet end, die shape of die portion 143. which accelerates die flow and "shot" the particles out along die line shown by the anow S, before they have a chance to enter die gap 172. Fig. 14 shows a third embodiment of the pump, which is actually die second embodiment adapted to be used in upright position. Therefore, the reference numerals in this figure are same as those in Fig. 13 for those parts which are not changed. When die pump 100 is used in diis upright position, die need for preventing die solid panicles from entering the gap 172 is not so important and the nanow entrance formed by die portion 143 is not necessary. In this case, die rotor 140 is under a much greater counter force for it has to support die weight of the whole column of liquid above it. In order to compensate die weight of die liquid on die rotor 140, an additional magnetic bearing 190 is fitted to die lower end of die pump, which includes an electromagnet formed by an annular magnetic core 194 and a coil 195 carried by die cap 130, and a permanent magnet 192 carried by die rotor 140. The magnet 192 is arranged not in perfect axial registration relative to the

electromagnet 194, so that when die electromagnet is energized die magnet 192 hence the rotor 140 is subject to an upward magnetic force which counter-balance die weight of die liquid on die rotor. Obviously more bearings can be fitted to one or bodi ends of die rotor to produce enough balancing force. The arrangement of Fig. 14 can be used as a generator when water flows down to drive the rotor so the operation of the magnets 144 and the coil 126 would generate electricity. It can also be used as a flow-meter when made small and light.

Figs. 15A to 17 show a pump 200 having a housing 220 and two cap members 230, each canies a bearing electromagnet 250. similar to those shown in Fig. 13. Inside the housing 220. there are two stators 210 and 210' , separated by an aiuiular separator 221 which carries a central bearing member 240. to be fully described below with reference to Fig. 17. Widiin the cylindrical imier space defined by the two stators there is a rotor assembly formed by two hollow rotors 260 and 260' each matching a conesponding stator. an aiuiular connector 270 for coiuiecting the two rotors to each other, and a driving mechanism 280 fitted inside die hollow structure of the rotor assembly. Electric connections (not shown) are made to the two stators so that they produce same rotary electromagnetic fields to drive the two rotors to rotate togedier. A perspective view of the rotor assembly is shown in Fig. 16 with the deuils of the comiection between the coi ector 270 and the two rotors shown in Fig. 17. The general structure and bearing arrangements for each rotor 260 or 260' are similar to that shown in Fig. 13.

A driving mechanism 280 includes two securing rings 281 and 281 ' fastened to the rotor assembly, each at one end; a lining tube 282 clamped between the two rings: a number of screw blades 283 of flexible and elastic material fitted inside the tube 282: and tow spiral wire springs 284 and 286 fitted around the outer surface of the tube 282. The blades 283 have an outer diameter slightly larger than the i ier diameter of the tube 282 so diat when die blades are fitted into die tube 282 they tend to expand, forming a tight fit. Each blade 283 has a number of studs 288 projecting radially outwards from its outer edge and being evenly located along the blade ^' s length, which studs are fitted into corresponding slots 287 formed in die wall of the tube 282 to keep die blades physically engaged widi the tube 282. The two springs 284 and 286 outside the tube 282 are arranged in opposite spiral senses, widi die spring 284 in die same spiral manner as die blades 283, and each spring has one end secured to an end of die tube with die other end joined to a connecting ring 285, which is also slidably fitted around the tube 282. Such an anangement allows smooth sliding movements of the springs and die blades. Tlie studs are engaged by die springs so die blades are biased by die springs. Under die conditions as shown in Fig. 15, die spring 284 tends to expand along its axial direction while die spring 286 tends to contract, dius diey produce a joint biasing force on the blades to urge them towards die downstream end of die pump (i.e. the right-hand side in Fig. 5A) and keep diem diere in their fully expanded status. The shape and orientation of die slots 287 are different from one anodier and the anay of slots are ananged in a way diat ώeir lengdis are progressively increased along die flow direction shown by die arrow A. That is to say diat die downstream end of each blade is relatively "free" because its stud is fitted in a long slot which gives it more room to move, compared widi die stud at die upstream end, which is virtually fixed. The puφoses for this arrangement is to make die blades 283 compressible during die pump's operation by allowing die studs 288 to slid widiin and along the slots 287, so as to keep die blades always engaged widi die tube 282. When die blades 283 are compressed, diey tend to expand radially but this radial expansion is restricted by die tube 282.

The result is for die blades to twist while being squeezed between it two ends. The shape and length of each slot is made to accommodate dus twisting factor to ensure blades' smooth movement during their compression.

In operation, when the rotating rotor assembly produces a forward driving force, the fluid would produce a backward thrust to compress the blades. In a low load operation, this counter force is balanced by the elasticity of the blades and die two springs 284 and 286. When the counter dirust is increased beyond a limit, for example when there is a big increase of die output resistance or a big increase of input driving power, the counter thrust on the blades would overcome the biasing forces and cause the blades to be compressed. Once this happens it increases die biasing forces by the springs so a new balance would be established at a new position, where since die blades are compressed to a smaller pitch (see P2 in Fig. 5B in contrast with PI in Fig. 15A), the pump operation would be stabilised again for a smaller flow rate under a higher output pressure, without affecting the rotor's rotating speed. The pump with such an adjustable driving mechanism can automatically and instantly response to changes of output resistance or input power, or both, from an operation of low pressure and high flow rate to one of high pressure and low flow rate, or vise versa, without compromising its energy efficiency. This makes die pump operable over a much wider range of working conditions and capable to provide smooth operation during its start-up or slow-down. This adjustability is particularly desirable when the pump is used as a marine propeller for, e.g. a speedboat, where quick acceleration/deceleration and smooth transformation under changing conditions are essential for good performance.

In Fig. 17. the driving mechanism 280 is removed to show a cross-sectional and partially exposed view of the details of die central bearing member 240 and the annular coiuiector 270. The combination of die central bearing member 240 and die annular connector member 270, togedier widi die magnets 264 and 264" , provide both magnetic and hydraulic suspension bearing effects. For die sake of easy understanding, diese two aspects are to be explained separately.

The aiuiular coiuiector 270 has a base ring 271 nude of a magnetic material, such as iron, which connects die two rotors 260 and 260' to each odier by snap engagements 274, so as to form an integrated rotor assembly. The base ring 271 also serves as a magnetic bridge for coiuiecting die two magnets 264 and 264' to foπn a complete nugnetic ring with conesponding poles matching diat of the two annular poles of the core member 241 of die electromagnet 240, separated by a non-magnetic member 243. In operation, when the coil 242 is energised, die electromagnet 240 and the permanent magnets 264 and 264' form a suspension pair with dieir conesponding poles opposing one odier, causing mutual-repelling, so as to "lock" die rotor assembly to its neutral position as described above.

When a liquid is filled into die pump structure, it enters die gap between die tube 282 and die inner surface of die rotor assembly via die slots 287. The base ring 271 has several rows of small holes 273, also shown in Fig. 6, which are covered by a filter member 272 which soaks up die liquid once it enters die gap. When die rotor assembly sαrts to rotate, liquid soaked up by die member 272 will be spun off by centrifugal forces dirough die holes 273 to form a tangential flow indicated by die anows S in Fig. 6. This tangential flow of the spin-off liquid is equally divided by a dividing edge 244 on die inner surface of die dividing member 243 to form two separated flows, each of a width "d", and to be directed by the guiding fins 245 or 245' towards

-2.V die conespond spiral vanes 269 or 269' to foπn two opposite bearing flows represented by the arrows B. The remaining parts of each bearing flow passage are similar to diat of Fig. 3. Since diese two bearing flows produce equal-sized but oppositely directed bearing forces, diey make further contributions in keeping die rotor assembly close to its neutral position. The widdi d to each side of the dividing edge 244 is about the same as die widdi of the axial registration members 217 and 267. In operation, the rotating rotor assembly is always kept within a small axial range of "d" to each side of its neutral position although there is no physical support except the bearing anangements. Whenever the rotor assembly is forced to move away from its neutral position, e.g. due to the counter thrust caused by the axial flow, the base ring 271 would also move relative to die fixed dividing edge 244 so dut the holes to die one side of die dividing edge 244 would be increased widi a conesponding decrease to the odier side. This change will cause an imbalance between die two bearing flows, which is enhanced due to die effects of die dirottle ring 268 or 268' at the downstream end of each bearing flow B. Because of this shift of balance, any further movement of the rotor assembly form its neutral position would meet increased resistance until the rotor stabilised at a new balanced position. Furtheπnore, when the rotor assembly is moved away from it neutral position for a disunce close to die size of "d", die misalignment between die rings 217 and 267 would be at its maximum, causing a maximum registration force to return the rotor assembly to its neutral position, further restricting die rotor assembly's deviation from its neutral position. Finally, it should be mentioned diat additional suspension control can be conducted according to the sensing signals provided by die sensors (166 in Fig. 3) fitted in each cap members, as described before.

The assembly of the pump 200 can be made in the following procedure. First of all. die electromagnet 240 and die separator 221 are made as one member by moulding die separator 221 around the electromagnet and forming die member 243 at die same time. The two stators 210 and 210' are dien fitted to each side of die moulded member, and die diree of diem are inserted into die housing 220 to form die tubular stator structure.

After this is done, the coiuiector 270 is located in the imier space of die stator structure and the two rotors 260 and 260' are siup-fitted to die coiuiector from two ends to toπn the rotor assembly. Once die rotor assembly is fixed, die driving mechanism 280 can be assembled by first fitting the end ring 281 ' to one end of the rotor assembly, then sliding die tube 282 togedier widi die springs and die blades into die rotor assembly from the odier end. Tlie tube 282 engages die ring 281 ' by the projection 289' at its leading end and also engages with die channels formed on die imier surface of die rotor 260 by die ribs indicated by the dash lines 289 at die odier end, so dut the tube 282 is carried to rotate by die rotor assembly. Then die end ring 281 is fasted to the other end. Finally die two cap members 230 are fitted to die housing to produce a finished pump.

Fig. 18 shows a fifdi embodiment of a pump 300. Fig. 19A is an enlarged view of a bearing mechanism 380 shown in Fig. 18; Fig. 19B is a cross-sectional view of die bearing mechanism 380 taken along the line B-B in Fig. 19A; Fig. 19C is a cross-sectional view of an impeller 350 shown in Fig. 18: and Fig. 19D is a cross-sectional view of the impeller 350 taken along the line D-D in Fig. 19C. The pump 300 is to be kept upright. It has a housing 320, a top cap member 330 and a base member 340. Inside die housing 320, diere is a stator 310 which has axial registration rings 317, as described with reference to Fig. 13, and a rotor assembly formed by a hollow rotor 360, an impeller member 350 connected to die bottom end of die rotor 360

and a pivot bearing mechanism 380 fitted inside die rotor close to its upper end. A supporting shaft 370. which is not in section, is also fitted inside the rotor extending between die base 340 and die bearing 380.

More particularly, die housing 320 has a cylindrical portion defining an upper motor chamber for the stator 310 and the rotor 360, and an extended lower portion 326 defining, together with die base member 340. an impeller chamber for the impeller 350. An aiuiular part 324 of the housing 320 separates die two chambers and foπns a support for die stator 310. Liquid inlet holes 328 are fanned around the lower part of the portion 326. which holes are covered by a filter 391. The base member 340 has a central portion 342 of a generally conical shape protruding into the impeller chamber to define with the housing portion 326 an annular and generally conical iiuier space for the similarly shaped impeller 350. The base 340 also has adjustable supports 349 for keeping the punip upright. Liquid inlet holes 344 are fanned in die base member and covered by a second filter 392. The puφose of using the filters 391 and 392 is to prevent any solid particles from entering the interior of the pump which may cause damages to hydraulic bearing vanes and surfaces. They are not incorporated when the pump is used for pumping clean liquids or highly viscous or pasty stuff. At the centre of die base ponion 342 there is a cylindrical protrusion which has a central hole 346 of a polyhedral cross-section for receiving the low end of die shaft 370, to be described later. A throttle ring 348 is fitted around the cylindrical protrusion far hydraulic bearing effects, also to be described later. The cap 330 is similar to diat of the previous embodiments, and it has a top portion with a screw thread 332 far coiuiecting to a pipeline or hose.

The rotor 360, similar to diat shown in Fig. 14, has magnets 364 and axial registration rings 367 matching that of the sutor 310, i ier screw blades 366 for driving the liquid flow, outer spiral vanes 369 for hydraulic bearing and a throttle ring 368 fitted at its top end. Inside the hollow space of the rotor 360 is fitted the supporting shaft 370 which has screw blades 372 in a spiral direction opposite to that of the blades 366. The lower end of the shaft 370 has a polyhedral coiuiector which is snap-fitted in die polyhedral hole 346 of the base member 340 to ensure diat the shaft is not rotatable. The top end of the shaft is fitted with a bearing base 376 of a very hard material, allowing the tip of the pivot bearing 380 to foπn a gimbal mount. In Figs. 19A and 19B, the bearing mechanism 380 includes a ring member 381 widi a number of engaging teeth 384 for snap-engaging widi die end portion 362 of die rotor 360. The teedi 384 ensure d t the ring 381 is fixed to die rotor and die mechanism 380 as a whole would rotate with the rotor 360 during punip operation. A central part 383 is comiected to die ring 381 by a number of coiuiecting members 382 which also serve as propelling blades when diey rotate widi die rotor. The central part 383 has a hexaliedral hole far acconunodatiiig a pin 385 of the same cross section, so dut die pin rotates with die rotor. Tlie pin 385 is secured by a nut 387 at its upper end and biased downwards by a spring 388 at its lower end. The tip 386 at die lower end of die pin 385 is made of a very hard material, such as ceramic, glass or super-hard metal, for forming a single point bearing contact with die bearing base 376 of a similarly hard material. The use of die spring 388 ensures d t die tip 386 is always biased against the base 376, so dut although the rotor assembly as a whole may have a small degree of axial movement, of a distance "d" as shown in Fig. 18, it is prevented from any direct impact widi die hard bearing base due to sudden changes of ouφut load or electric driving input. It should be noted that when in operation, die whole rotor assembly formed by die members 380, 360 and 350 is home, except the effects of die hydraulic bearings and die axial registration anangement, solely by

die tip 386 on die base 376 which provides a ginibal mount of high stability and low wear and rotating resistance, and no magnetic bearing is used in diis embodiment.

Figs. 19C and 19D show the details of die impeller 350 which is to be secured to die lower end of the rotor 360 by an engaging ring 358, so as to rotate widi die rotor. The engaging ring itself can be made as a separate member which is dien secured to die impeller by a similar engaging anangement. The engagement can be of die similar construction as dut of die ring 381 shown in Fig. 19A. The impeller 350 has a hollow comcal body 352 with outer blades 354 and imier vanes 356. The blades and vanes are in the same spiral direction so that when the impeller rotates with die rotor, die blades 354 produce a generally upward liquid flow from the holes 328 towards the hollow rotor while at die same time the vanes 356 produce an upward bearing flow which keeps die impeller "floated" on die outer surface of the base portion 342. The bearing flow is eventually forced to pass over the throttle ring 348 and enters the hollow rotor via the aiuiular gap between the inner surface of the cylindrical portion 357 and die outer surface of the central protrusion of the base member 340. That is to say, when die rotor assembly rotates at a stable speed, the hydraulic bearing flow fanned by the vanes 369 on the outer surface of the rotor 360 and that by the vanes 356 on the imier surface of the body 352 would provide joint bearing effects which keep the whole assembly "floated" and also lubricated, therefore significantly reduce the bearing load on the pivot tip 386.

In operation, once the rotor 360 starts to rotate under die influence of the magnetic driving force produced by die stator 310, it would carry die impeller 350 and die bearing mechanism 380 to rotate widi it. The liquid sucked into die impeller chamber via the holes 328 and 344 would be forced to flow upwards by die blades 354 and die vanes 356 to enter die hollow rotor, where it would be forced to flow upwards by die effects of die blades 366 which are rotating and die blades 372 on die supporting shaft 370, which are not rotating. Since die blades 372 are in a spiral direction opposite to dut of the blades 366, they work together to reduce the swirling factor of the liquid flow passing between diem and to increase the upward driving force, dierefore producing a significantly increased output pressure. This upward flow is further promoted by the blades 359 of die impeller 350 and die blades 382 of die bearing 380, bodi sets rotate widi die rotor.

When diere is a sudden change of load or electric input, die rotor assembly would tend to move axially. This tendency is compensated by the elastic bearing force by die pivot bearing 380 and die two throttle rings 348 and 368. For example, assuming diere is a sudden drop of ouφut pressure due to the fact the hose connected to die top cap 330 is burst under die ouφut pressure, die whole rotor assembly would unavoidably move upwards, which leads to the sitiution that die upper dirottle ring 368 would block die bearing flow inlet to the hydraulic bearing on die outer surface of die rotor 360, and at die same time the gap defined by die lower dirottle ring 348 is increased so much diat die vanes 356 on die imier surface of die impeller 350 would not be able to produce bearing effects. That is to say bodi hydraulic bearing anangemeiits stop to provide upward bearing forces so die whole rotor assembly would move to a new balanced position. When die rotor assembly moves downwards, e.g. due to increased ouφut pressure, the arrangement works in die opposite way to balance die system automatically. Since die pump 300 has two sets of liquid inlet holes 328 and 344, it can he conveniently used as a mixing pump widi die holes 328 for the main liquid component while die holes 344 far adding a second component which would be fully mixed widi die main flow in die hollow rotor. This is also

useful in die case where die main flow is a highly viscous or diick mixture and a lubricant and/or diluent liquid can be introduced via the holes 344 to keep die system lubricated to reduce flow resistance. All die structural components of diis pump, except die electric or magnetic parts and the hard gimbal bearing tip and base, can be made by moulding plastics or resin, so it is easy to achieve high precision and low manufacturing costs. Fig. 20 shows a cross-sectional view of a sixth embodiment of a pump 400 having a housing 420 accommodating two stators 410 and 410' separated by a separator 421 , and two rotors 460 and 460', which are kept between two cap members 430 and 430", each canying a gimbal bearing 480 or 480'. Generally speaking, the stator/rotor combiiution 410 and 460 work in a way similar to that of Fig. 14. with hydraulic and magnetic bearings and an magnetic axial registration anangement farmed between them. The operation of the stator/rotor combiiution 410' and 460' is similar. New features of die rotor 460' include the central propelling member 470' which is integral with die screw blades 466' and the hollow rotor body. The member 470' has a polyhedral central hole. The hole engages at its upper end the coiuiector part 474 of a central propelling shaft 470. which is similar to the central shaft 370 shown in Fig 8. The low end of the hollow member 470' engages the bearing base of the gimbal bearing 480' supported by the cap 430 ^" , while die upper end of the propelling shaft 470 carries the bearing base of die odier gimbal bearing 480 fitted in the cap 430. That is to say. the two bearings 480 and 480' "clamp" between them die combiiution of the propelling shaft 470 and the whole rotor 460'. which are made to rotate together. The structure of die bearings 480 and 480' is similar to that of the bearing 380 shown in Figs 8A and 8B, except that the coiuiecting members 382 are arranged along the axial direction because they do not work as propelling blades. In operation, the rotor 460 rotates in direction of the arrow "R". which produces an axial flow in direction "A" . On the odier hand, die rotor 460' and die shaft 470 rotate in die opposite direction shown by die anow "r" at die bottom, and produce an axial flow in direction "a". This arrangement ensures that the two sets of screw blades always cancel each odier's swirling effects to the liquid flow and at die same time enhance the common axial direction driving force. That is to say. a significantly increased output pressure is achieved widiout sacrificing die flow rate. The rotor 460' can be manufactured by first making a central propeller 470 ^' widi integrated screw blades 466'. It can be made as a separate pan widi spiral chaimels fanned on its iiuier surface, matching the outer edge of die blades 466'. Then die propeller 470' is fitted into the rotor 460' by screwing the blades in die channels. Adhesive is used to bind die blades 466' to die rotor body. Since these pans can be easily nude of plastics or resius, the manufacturing costs are low. The pump 400 is suitable for forming a high pressure fluid jet of a stable flow rate. In case of high pressure applications, a number of diem can be comiected by simply screw one to another. Because of its very compact structure, it can be conveniently fitted to die outlet of a pipeline and die whole pipeline can be operated under a much lower internal pressure for transferring die fluid medium, dierefore die costs and difficulties of using and maintaining high pressure pipes and associated connectors are avoided without compromising die operational requirements. Industrial Applicability

The simple and highly symmetrical structures disclosed in die present application make diem suitable for mass-production and high quality control, therefore low cost per item. The caps and die housing should be made of non-magnetic materials, such as aluminium, stainless steel, copper, plastics or fibre reinforced resin.

Similarly. the tubular member of the rotor togedier widi the propelling blade(s) should also be made of non¬ magnetic material. The screw blade(s) can be made as a separate member first dieu welded to a tube or casting/moulding die tube around it. Alternatively, die whole structure can be formed by casting around a cylinder die with a spiral channel for forming die blade(s), dien the final product can be unscrewed from die die. A continuous casting process can also be used to from a long tube with an inteπul blade, then it can be cut to required length for making die tubular rotor member. The number of die blades, its height h. pitch p and propelling angle α, as shown in Fig. 11, can be changed or adjusted to suit die needs of a particular application.

The internal blade structure of the rotor provides a clog-free propelling structure which is able to propel through it anything that can by all means enter the pipeline at the first place, therefore it is suitable to a wide variety of applications, especially when used for handling liquids canying a high proportion of solid contents, e.g. applications as marine propeller, sludge pump, sewerage pump or pump for impelling or injecting pasty or viscous mixtures. The relatively long screw or spiral blade is able to distribute stress evenly over its whole length so it can undergo large load without causing over-stress. Because the pump can be wholly sealed. it is suitable for submerged applications and/or leak-free applications. Furthermore, it is easy to fit such a pump into an existing pipeline at any desirable position since die pump can be configured in roughly the same dimension as a piece of tube, and a number of pumps can be coi ected in series into the same pipeline to increase the total driving force, or to have one or more of them kept idle as back-up units. When an idle pu p is fitted to a pipeline, its rotor will be able to rotate freely dierefore not causing significant resistance to the flow through it. The hydraulic bearing vanes can also be formed on the iiuier surface of die stator, instead of or in addition to dut formed on the outer surface of die rotor. However it is easier to form diem on the outer surface of the rotor from die point of view of machine operation. In the above description the teπn "spiral vane" should be inteφreted to include both die form of projecting ribs or grooves, which are easier to manufacture and can produce die satisfactory effects, especially when made of elastic materials. It is worth mentioning diat in the accompanying drawings the size of the bearing vanes is exaggerated for the sake of easy recognition. In practice diey are made very small to ensure a close fit between die moving parts. In a relatively small pump die magnetic suspension bearing 150 can be formed by two permanent magnets ananged to repel each other, instead of using an electromagnet 154. This will reduce power consumption and weight.

Now, details of the embodiments according to the third general object are described below.

In Fig. 21, a cooling system 10 lus four functional units, represented respectively by dash-line blocks 100, 200. 300 and 400, each for a different working medium, thermally coupled to form a cascade heat transfer chain. Tlie system 10 also includes a control unit 500 which is electrically comiected to these blocks.

The block 100 is a primary cooling mechanism in die form of a vapour-compression circuit. It has a compressor assembly 110, a condenser 120, bodi physically adapted widiin a coolant column 310, a dryer 130. an expansion valve 140, an evaporator 150 positioned in a cold storage tank 210, and an accumulator 160. Except die compressor assembly 110 and die condenser 120, the remaining parts of die circuit 100 can use commercially available components. The block 200 represents a brine circulating circuit located widiin a thermally insulated and closed space. The circuit includes a cold storage tank 210 holding an antifreeze liquid which submerges the evaporator 150 of the above circuit 100. A flow passage is formed by an outlet

coiuiection 220. a circulating pump 230 and a retuni comiection 240 which is comiected back to die cold storage tank 210. The block 300 is a secondary cooling mechanism in the foπn of a coolant circulating circuit which includes a coolant column 310. a compressor 320 comiected between a coolant outlet of die column 310 and an inlet end of a coolant passage 330 ananged in die block 400 which is a heat storage tank. The outlet end of die passage 330 is connected via tubing 340 and a control valve 350 to a coolant inlet of the column 310. A control uiϋt 500 is comiected via, respectively, a control line 501 to the compressor assembly 1 10, a control line 502 to the antifreeze circulating pump 230. a control line 503 to the compressor 320, a control line 504 to the control valve 350 and signal lines 505, 506, 507 and 508 to sensors 51 1 , 512. 513 and 514, each fitted in the heat storage tank 400, closed space 200, cold storage tank 210 and coolant column 310. Generally speaking, heat is traiisfened in die cascade system 10 from its right-hand side. i.e. die block

200, to its left-hand side. i.e. die block 400. where it is dissipated to the ambient air. The basic concept of the invention is to improve the heat exchange of the primary cooling mechanism 100 at both die heat-receiving side. i.e. die evaporator 150 where the heat is absorbed, and the heat-rejecting side. i.e. the assembly 110 and die condenser 120 where the heat is rejected. When in operation, the evaporator 150 transfers cold energy (i.e. negative heat energy) directly to die antifreeze liquid in the cold storage tank 210, and the liquid in turn dissipates, under a forced convection driven by the pump 230, cold energy to contents in die space 200. e.g. when it is a freezer, or to an air flow when it is an air-conditioner. Since a basically water-based liquid is used, which has better theπiul conductivity and higher specific heat than air, the arrangement works better d n using die evaporator to cool air directly by conduction and lutural convection. On die other hand, the heat generated by die compressor assembly 110 is directly traiisfened to die coolant in the column 310, which again provides a more efficient heat transfer than in the case where the heat is traiisfened directly to die ambient air. Tlie heat exchange efficiency is further improved because the heat rejected by the circuit 100 is used to circulate in die circuit 300 the coolant solution having evaporable components and/or endothermic salt components, so that the coolant circuit 300 can take heat away from the compressor assembly 1 10 in die form of latent heat of vaporisation which is more efficient than do it in the foπii of sensible heat, and also cause endothermic dissolution of the salt to cool the refrigerant liquid in die condenser 120.

Since diese beneficial effects are produced by using die heat which is conventionally rejected to die ambient as waste, die overall energy consumption is reduced. Any excessive heat is to be temporarily stored in the heat storage tank 400 which serves as a heat buffer to ensure that no part of the system will overheat during its normal operation. By having this heat buffer, die system as a whole can dissipate heat continuously although the compressor assembly operates intermittently. Because of improved efficiency for both the cold and hot sides, physical size of die circuit 100 can be reduced significantly for it no longer needs large heai exchange surfaces and diis leads to a much reduced length of the circulating route actually travelled by refrigerant during a compression cycle. It means die overall flow resistance is reduced, hence a further improvement of efficiency and reduction of cost. Finally, since die size of die circuit is mimmised, so is die amount of die refrigerant needed in the circuit, making it easier to meet safety requirements when an environmentally acceptable but toxic and/or flammable refrigerant, such as ammonia or propane, is used in die system. The above explanation regarding the basic concept of die invention is made widi reference to a primary cooling mechanism of vapour-compression type. However, the same principles would also apply if a different

cooling mechanism is used, e.g. an absoφtion type, a thermoelectric one (by Peltier effect) or a magnetic one (by tliermomagiietic effect). The compression type is preferable because it is by far die most energy efficient and also die most commonly used in existing facilities.

In die following description, details of die block 200 are explained with reference to Figs. 22A to 22C. the heat storage tank 400 with reference to Figs. 23 A to 23C, the coolant column 310 widi reference to Figs 24A to 24C, a defrosting arrangement 600 with reference to Figs. 25A and 25B. and a control mediod is illustrated in the flow-chart of Fig. 26.

Fig. 22A is a cross-sectional view taking along the plane A-A shown in Fig. 22B, while Fig. 22B is a cross-sectioiul view taking along the plane B-B in Fig. 22A, and Fig. 22C is a cross-sectioiul view uking along the plane C-C shown in Fig. 22B. In Fig. 22A. a freezer 200 lus an insulated casing 201 , an insulated door 202 and a number of shelf members 203 for supporting goods. For die puφose of easy illustration, a coolant column 310 is shown to be attached to the back of die casing 201 , which in practice can be built into the insulating wall of the casing 201 , as shown in Fig. 22B. A thermal storage tank (heat tank) 400 is on top of the freezer 200, but again it can be located at other places. For example, in a large system it is practical to tit the unk outdoor for maximum efficiency. The column 310 is comiected via a compressor 320 to a coolant passage 330 in the tank 400, which passage is comiected in turn, via die tubing 340 and a solvent collector (collector) 341. to die bottom end of the column 310. The collector 341 is positioned to maintain a proper liquid level in die column 310, as to be described later. The compressor 320 is ananged to be thermally coupled with the heat tank 400, so that the heat generated during its operation is absorbed by the heat storage material in the tank 400. The control unit 500 is fitted to the top front face of the freezer 200.

Widiin die freezer 200, die cold storage tank (cold tank) 210 is fitted to die ceiling of the imier space, which tank 210 has a bottom wall 211 made of a d em lly conductive material and the evaporator 150 is fitted on die bottom wall 21 1 so d t diey form an integrated cold-generating member. A brine 212 is filled in the cold tank 210. which keep die evaporator 150 submerged. An commercially available antifreeze solution can be used as the brine 212. The concentration of the solution is controlled to ensure it has a freezing point a few degrees below die temperature to be maintained in die freezer. This nukes it possible to use die liquid as cold storage material when the temperature in die freezer is deliberately brought to its freezing point. It is worth mentioning mat once ice crystals are formed, the remaining liquid will have a higher concentration of antifreeze compound so a lower freezing point. Tlie liquid does not have a single freezing point as pure water, instead it will freeze over a temperature range. Evenmally the iiquid will freeze into a slush which stores latent cold energy, to be released later when die slush melts. A headroom is kept in the cold tank 210 to cope with the expansion of die liquid 212 when it is frozen. The temperature of die liquid is monitored by die sensor 513 in die tank 210. while the air temperature inside die freezer 200 is monitored by the sensor 512. for control puφoses to be explained later. Underneadi die cold unk 210 is fitted a deep-freezing/defrosting system 600. In Figs. 22B and 22C, on each of die four side-walls, including die inner surface of die door 202, is formed a circulating padi 220 which connects the cold tank 210 to a collecting clumber 221 formed on the bottom wall of die casing 201. Flexible tubes 225, one of them being shown in Fig. 22B, are used to comiect die channel 220 formed on die inner surface of die door so diat die brine circulation is not affected by door movements. A small circulating pump 230 is connected to die collecting chamber 221 for returning die liquid

-.10- back to the tank 210 via a return pipe 240. The pump 230 and pipe 240 are embedded in die insulating material of die casing 201. Two vertical corner channels 633 are shown in Fig. 22B, which provide air circulation passages. The casing 201 and die door 202 are made by moulding plastic materials, preferably gasified plastics, which provides good strengdi and thermal insulation. A number of internal ridges 204 are formed to enhance their mechanical strength, which also provide good atuchments for the foamed insulating material. The pump 230 can be of any conventional type, but it is preferable to us a small axial flow and block-free pump as described above regarding the second general object of the invention.

As shown in Fig. 22C. the channel 220 is defined by channel members 223 formed on a flat panel, which are also nude by moulding a plastic material, and covered by a flexible sheet 222 which is preferably a lamiiuted sheet with at least one metal foil layer. A central support member 224 is foπned in the middle of the channel to enhance the attachment of the flexible sheet to the cluiuiel base. The members 223 and 224 provide main support to the weight of die liquid in the cluiuiel 220, which reduces die stress subjected to by the flexible sheet 222. In this way. a layer of antifreeze liquid 212 is formed by the seφentine chaimels 220 which cover virtually the entire inner surface of the freezer 200. and in a large system, such channels can also be formed on die shelf members 203, to further increase die overall size of the heat exchange surface. By using the flexible sheet 222, which is very diin and thermally conductive, a good heat exchange surface is formed between die liquid 212 in the cluiuiel 220 and the interior of the freezer 200. On die other hand, the flexibility nukes the sheet 222 well adapted to cope with the expansion of the liquid when it is frozen in its cold storage ode, as mentioned above. Since the lamiiuted sheet 222 has plastic cover layers, it is stable against any potential conosive effects of the antifreeze compound to its metal foil layer(s). Such sheets are used, e.g. for food and beverage packaging, and they can be easily attached to die supporting members 223 and 224 by adhesive or diernul welding, with or without further fastening means. It is clear from the above description that except the cold tank 210, the evaporator 150 and the defrosting system 600, the whole freezer body 201 and the door 202 can be made of plastics by moulding, therefore having a better thermal insulation and also lower costs of material and manufacturing, k also makes die whole casing easily recyclable after its service life.

Figs. 23A and 23B are top views showing two embodiments of the heat tank 400 of Fig. 22A. The differences between diese embodiments are dut Fig. 23 A shows a seφentine vapour passage 330 while Fig. 23B shows a coil passage. Fig. 23C is a cross-sectional view taking along die plane C-C shown in Fig. 23A. As shown in Figs 23 A to 23C, die heat tank 400 has a generally flat casing formed by two casing members 401 which can be identical in structure. Each member 401 has a number of exteπul ribs 402 and internal section walls 403 which separate the interior of die unk 400 into a number of chambers as shown in Fig 3A or a central chamber 401 ' and an outer coil chamber as shown in Fig. 23B. In use die central chamber 401 ' provides a hot spot convenient for diawing frozen product taken out of die freezer. A coolant passage 330 is formed by two flexible and dieπiully conductive sheets 332, which can be of die similar type as the sheet 222 shown in Fig. 22C. The two sheets 332 are clamped between die two casing members 401 and separated by a supporting member 331 in die form of a perforated pipe or a coil of a spiral wire. The function of the supporting member 331 is to separate die two sheets 332 so as to prevent die passage 330 become blocked when diere is a low pressure in die passage, as to be explained later. A number of intenul ridges 404 are formed on die inner surface of die casing 401 , to increase its strengdi and provide support to die sheets 332 when diey are inflated

by an internal pressure, as shown by die dash lines 332' in Fig. 23C. A phase change heat storage material 410 is filled in each chamber for receiving heat via die sheets 332 and storing it bodi as sensible heat and latent heat of fusion when its temperature is raised to its melting (fusion) point.

Tlie heat storage capacity of the tank 400 should be large enough to cope with die need of hot weadier. To increase the storage capacity and the heat exchange efficiency of the tank 400. it is preferable to have different heat storage materials 410 in different chambers, so d t the chamber at the vapour inlet end of die passage 330 has a material of a higher phase change temperature (fusion point) dun diose to the downstream end of die passage 330. In such an anangement, a temperature gradient is formed along the passage 330 and each chamber may absorb heat from die vapour flow over a temperature range so dut heat is stored evenly along die whole length of the passage 330. For control puφoses. this temperature is momtored by the sensor 511 embedded in the heat storage materiel. The phase change temperature at the inlet end can be about 75°C which is below die vapour output temperature from die column 310, while at die outlet end it can be about 30°C which should be above a practical high ambient temperature, to ensure that the material will not absorb heat from ambient air during a hot summer day. These values can be easily adjusted by selecting different heat storage materials to meet die needs of different climatic conditions. For this puφose, many materials can be used in die tank 400, which can be hyrated inorganic salts and dieir eutectic mixtures or reciprocal salt pairs. The suiuble examples include: calcium chloride hexaliydrate (of a fusion point of 29°C); sodium sulphate decaliydrate (of a fusion point of 32°C); calcium bromide tetraliydrate (of a fusion point of 33.8°C); calcium bromide hexaliydrate (of a fusion point of 34.3°C); zinc nitrate hexaliydrate (of a fusion point of 35°C): sodium carbonate decaliydrate (of a fusion point of 35°C); disodium hydrogen phosphate dodecahydrate (of a fusion point of 35.5°C); sodium diiosulplute pentaliydrate (of a fusion point of 50°C); sodium acetate trihydrate (of a fusion point of 58°C): and barium hydroxide octaliydrate (of a fusion point of 75°C).

When die tank 400 is cool, i.e. all die heat storage materials are in frozen status, a vacuum space 420 as in Fig. 23C is formed above die surface of die storage material 410 for coping widi die expansion of die passage 330. as shown by die dash line 332'. Fluid communication (not shown) is formed between die spaces above and below die passage 330 to allow die material 410 to flow in each chamber when it melts. The interior of die passage 330 is also in low pressure under diis condition so diat the tank is ready to accept vapour output from die column 310. When vapour enters die passage 330, it condenses on die imier surface of die sheets 332 and give up heat to the heat storage material 410, mainly in die form of heat of fusion, and eventually dissipate it to ambient air by natural convection via the casing 401 and its inner ridges 403 and 404 and outer ribs 402 which are thermally conductive. The casing can be made of metal materials, but to reduce cost, it also can be made of plastic materials which are moulded to form die casing member 401 , dien covered by a diermally conductive coating, or covered on each surface by a metal layer to increase its thermal conductivity. The checked pattern of die inner ridges and outer ribs ensure that the casing has enough mechanical strengdi to undergo die internal pressure changes and diey also serve as fins for heat dissipation. The casing wall is made relatively thin to further improve thermal conductivity. The plastic material also has the advantage diat it is stable against conosive effects of die heat storage material 410. Again, die main considerations are to provide high efficiency together with low cost and recyclable structure.

Fig. 24A is a sectioiul view of one embodiment of die coolant column 310. widi die compressor assembly 310 in it fully exposed but not in section, and Fig. 24C is a partial view of another embodiment of die column. Fig. 24B is a cross-sectioiul view taking along die plane B-B shown in Fig. 24C. In Fig. 24A. die coolant column has a tubular housing 310. The upper end of the housing 310 is a vapour outlet comiected to the compressor 320 and its lower end is a coolant inlet connected to the collector 341 via the control valve 350. Inside the housing 310 is ananged the compressor assembly 1 10 with its upper end comiected to a suction line 108 covered by a thermal insulation layer 109, which leads to the accumulator 160, and its lower end comiected to the condenser 120 which is formed by a spiral coil of metal tubing leading to the dryer 130, as shown in Fig. 21. A further compressor can be used between the accumulator 160 and the pipeline 108 to pre-compress the refrigerant thus to increase its pressure and temperature before it enters the housing 310.

The compressor assembly 1 10 has a series of compressors 1 10A, HOB and HOC and chambers far compressed refrigerant, 1 1 1A, 1 1 I B and 1 1 1C, forming a multistage chain through which the refrigerant is compressed progressively. The compressors 1 10A to 1 10C are preferably a free-piston type as described above regarding the first general object of the invention. The compressor 1 10A is connected to the chamber 1 1 1 A. which in turn is connected to die next suge compressor 1 10B, and so on. Their outer surface can be fanned by a single cylindrical member nude of diermally conductive material, to achieve good heat exchange. Internal tins can be fanned inside each of the chambers 1 1 1 A to 1 1 1C, as shown in Fig. 4B. to further enhance heat exchange efficiency, and also mechanical strength so that the side wall of the cylindrical member can be made relatively diin. The number of compressors can be changed to meet the needs of different refrigerants or application requirements. That is to say, die whole anangement is highly flexible for different needs.

From the upper end of the assembly 1 10 downwards are fitted in sequence a cup member 1 15, a collar member 1 14. an inner heat exchange member 1 17 sunounding the chamber 1 1 1 A. and a spiral fin 1 16 extending from die compressor HOB down to die last chamber 1 1 1C, which in turn is comiected to die condenser 120. A sensor 514 is fitted close to the top end of the fin 1 16. As also shown in Fig. 24B, the four theπiioelectric members 1 12 are circu ferentially fitted between the outer aiuiular surface of the imier heat exchange member 1 17. which has fins extending inwards to form tlieπnal contact with the chamber 1 1 1 A, and the inner aiuiular surface of an outer heat exchange member 113 which has radially outward extending fins. Their functions are to be explained later. Widiin the tubular housing 310. it is fitted close to its upper end a vapour separator 31 1 which is a porous or perforated board allowing vapours to pass through but not any liquid droplets carried up by a vapour flow. Below the heat exchange member 1 13. there is a flow guide member 312 which has a tubular portion 316. a collar portion 313 at the upper end of the tubular portion 316. and a number of parallelly ananged spiral flanges 315 formed on die outer surface of die tubular portion 316. At die bottom end of die housing 310 is fitted a nozzle 352 which has an internal one-way valve 351 allowing a coolant outflow from the nozzle but not a flow back into it. A particle separator 353 is fitted to cover die nozzle 352, which separator functions as a filter allowing liquid from die nozzle 352 to pass through but not allowing solid particles to settle into die nozzle.

The coolant is preferably formed by a cockuil of aqueous ammonia widi one or more endodiermic salts dissolvable in it. Suitable salts include ammonium nitrate, potassium thiocyanate, ammonium chloride, poussium nitrate, urea etc. or a mixture of any of diem. The concentration of die aqueous ammonia is made

below saturation within the normal range of ambient temperature, i.e. in the concentration range of 20-35 wt. %, its exact value is to be determined according to local climate. A salt mixture of ammonium nitrate and urea is preferable for dieir low commercial cost (bodi are used as common chemical fertiliser). Carbon dioxide can also be included in this cocktail to increase its evaporable contents. From the operational point of view, die internal space of the column housing 310 is divided into die following functioiul areas. Firstly, the top portion forms a vapour chamber 301 in which ammonia and water vapours rise upwards to the vapour outlet. Below die vapour chamber 301 is a boiling zone 302 which is formed by die thermoelectric members 112 and its outer heat exchange member 113. The space below the boiling zone 302 is divided by the flow guide member 312 into an outer precipitation zone 303 which is die aiuiular space between the housing 310 and the outer surface of the tubular portion 316 of the flow guide member 312, and an inner evaporation zone 304 which is the aiuiular space between the iiuier surface of the tubular portion 316 and die outer surface of the compressor assembly 110. and the zone extends upwards to the cap 1 15. Below the flow guide member 312 is a salt chamber 305 which is comiected at its bottom to the coolant inlet nozzle 352. The operation of die circuit 300 is now described with reference to Fig. 23A to 23C. 24A and 24B.

When die compressor assembly 110 suns working, it sucks in gaseous refrigerant from the suction line 108 and compresses it progressively through the chain of the compressors 110A to HOC and the chambers 1 1 1 A to I I IC. During this process the temperature of the assembly 1 10 and condenser 120 begins to rise because of the pressure increase. At the same time, the valve 350 is opened to allow the aqueous ammonia in die collector 341 to enter die salt chamber 305 dirough die nozzle 352 and die salt separator 353 and to form a upward counterflow to reach a liquid level in the lower part of the boiling zone 302. The Iiquid level is stabilised within the zone 302 due to die relative vertical position of the collector 341 which supplies the solvent.

During diis process a quantity of salt particles in the chamber 305 is dissolved by aqueous anuiionia. producing endotheπiiic effects in the salt chamber 305 to about 0°C. therefore die part of the condenser coil 120 widiin the chamber 305 is cooled and the refrigerant in it is properly condensed to liquid aldiough the total lengdi of die coil is very short. The saturated salt solution f rmed in the salt chamber 305 flows upwards into die inner evaporation zone 304 defined by die inner tubular surface of die flow guide member 312. Once this upward flow comes into contact with the upstream pan of the condenser 120 and the compressor assembly 1 10, which are at a higher temperature as explained above, the temperature of diis saturated solution begins to rise: This has the effects d t on die one hand it reduces die solubility of die ammonia in die solution, which causes ammonia evaporation, as indicated by die small circles shown in the inner evaporation zone 304, while on the other hand, it increases die solubility of die salts so die solution becomes less saturated and die salts remain dissolved although its relative concentration becomes higher after die evaporation of die ammonia. Due to die existence of die spiral heat exchange fin 116, die upward flow of die coolant solution in the imier evaporation zone 304 has a very good diermal contact with die compressor assembly 110, dierefore keeps it cooled. The heat transfened from the assembly 110 to die coolant flow in die zone 304 causes a temperature increase along die flow padi. hence more evaporation of the ammonia, forming a strong bubbling flow. Finally diis bubbling flow passes die gaps between die imier heat exchange member 117 and die clumber 111 A, and is turned by die cap 115 downwards into die boiling zone 302. The temperature of diis bubbling flow is constantly monitored

by die sensor 514, so dut the control unit 500 in Fig. 21 can adjust die flow rate via die control valve 350. and also die current through the thermoelectric member 1 12 to keep the temperature of the assembly 1 10 stable, so as to compensate the changes of its load.

The vapours in the bubbling flow rise immediately into the vapour chamber 301. as shown by the small circles therein, while the liquid part of the bubbling flow coming out of the cap member 1 15 enters the boiling zone 302. The outer heat exchange member 1 13 in the boiling zone 302 is theπi lly engaged with the hot side of the thermoelectric members 1 12 to receive heat traiisfened by dieir Peltier's effect from the wall of die chamber 1 1 1 A via the imier member 1 17, as shown in Fig. 24B. This arrangement is made to maintain the outer member 1 13 thermally elevated to die boiling temperature of the coolant solution, which is also the highest temperature in the whole column 310. so a significant amount of the water in the boiling zone is evaporated. Because of the high temperature in the zone 302, the solubility of the salt components in the solution is further increased to fonn a solution of very high salt concentration after the water evaporation. This strong solution is heavier than the solution below it in the outer precipitation zone 303, which causes it to sink through the tliroughholes 314 foπned in the collar portion 313 of die flow guide member 312. After the strong solution enters the zone 303, it is guided by the flanges 315 to flow downwards and at the same ti e losing its heat to the solution in the imier zone 304 through the diin wall 316. As this solution is cooled progressively during the downward flow, the solubility of its salt contents decrease so the salts begin to form crystal particles, as represented by the small cross signs in the zone 303. These particles would first settle onto slope surfaces o the flanges 315 then slip down into the salt chamber 305. to be re-dissolved by the incoming aqueous ammonia flow for a new cycle of die salt circulation.

It is clear from the above description d t under die control of the unit 500 based on the value sensed by die sensor 514. a dynamic circulation is carried on between the i ier evaporation zone 304, the boiling zone 302, die outer precipiution zone 303 and die salt chamber 305. A temperature gradient along the axial direction of die column 310 is dierefore subilised between the salt chamber 305, which is about 0°C. and the boiling zone 302 which is about 100°C. due to die baffling effects of the spiral fin 1 16 in the zone 304 and also the spiral flanges 315 in die zone 303. which prevent turbulent mixing of the liquid in each of the zones. The flow guide member 312 is nude by moulding a plastic material to form the flanges 315. which are relative thick compared widi die tubular wall 316, which can be nude of dieπnally conductive materials, e.g. by moulding die flanges 315 around a metal tube 316. so that die heat exchange in die radical direction between the imier and outer zones 304 and 303 is promoted against the heat exchange in die axial direction within each zone. The column housing 310 is made of plastics and has a heat reflective inner surface, so it provides good thermal insulation which helps to maintain die temperature gradient in its axial direction. The system efficiency is dierefore significantly improved due to two mutually supporting factors which reduce energy losses during its operation. Firstly, die whole compressing chain is cooled by die coolant flow which reduces die compressor loss and also takes a significant amount of heat away from die refrigerant before it enters the condenser 120. Secondly, die cooling effects in the salt chamber lower the condensation pressure of die refrigerant in die condenser, hence reduce condenser loss and throttling loss in the system. Tlie reduced condensation pressure makes it practical to use environnienully benign refrigerant, such as carbon dioxide or anhydrous ammonia which would o ierwise condense only under a much higher pressure.

At die top end of die column 310, a mixture of ammonia and water vapours in die vapour chamber 301 foπii an upward flow promoted by die sucking force of the compressor 320, which can be of die same type as die compressors in die assembly 1 10. Between die compressor 320 and the theπnoelectric members 1 12, a mutually compensating relationship is formed to encourage the water evaporation by maintaining a low pressure in die chamber 301 , which lowers die boiling temperature of the salt solution in the boiling zone 302 to improve the efficiency of the members 1 12, or vice versa.

Refer to Fig. 23C, when die vapour mixture enters the passage 330. its water content would be the first to condense onto the flexible sheets 332 and give up its heat to die heat storage material 410. Then the condensed water would absorb die ammonia vapour to form an aqueous solution. This process slows down pressure build-up in die passage 330 to nuke room for vapours coming afterwards. By using die flexible sheets 332. die passage 330 can be inflated under the vapour pressure to increase the toul size of the heat exchange surface hence the rate of water condensation and ammonia absoφtion. The ammonia absoφtion is an exothermic process and the heat so evolved is transfened to the heat storage material 410. Finally, die aqueous ammonia fanned in the passage 330 would flow via the comiection tubing 340 into the collector 341 for another cycle of die solvent circulation. To encourage die return flow, the heat unk 400 is tilted towards the outlet end of the passage 330 so the flow is achieved mainly by gravity, however, the actual flow rate is decided by the valve 350 under the control of the unit 500.

When the compressor assembly 1 10 stops operating when the temperature inside die freezer is below a predeteπnined value, and the valve 350 is shut to stop the incoming flow of the aqueous ammonia solvent. At die same time, die compressor 320 is switched to a low-power operation which allows out-flow of die vapour mixture, so d t the coolant solution in the column 310 cools down gradually by further evaporation which causes more salt to precipiute and sink to die salt chamber 305. Similarly, die heat tank 400 would also cool down by dissipating stored heat to ambient air, dierefore causing die vapour inside die passage 330 to condense and dissolve thus to create a low pressure in die passage 330. Due to the continuing operation of the compressor 320. which does not consume much power but acts more like an active one-way valve, low pressure is also established in die vapour chamber 301 , which in tuni causes more water evaporation, dierefore cools the interior of die column further, although die column housing 310 itself provides a good diermal insulation which prevents heat dissipation dirough its wall. Tlie one-way valve 351 and die control valve 350 help to prevent any salt from entering the pipeline 340. At this final status, most of die salt content is accumulated as solid particles in die chamber 305 while most of die solvent, i.e. aqueous ammonia, is accumulated in die passage 330 and die collector 341. Because the temperature in die unk 400 and die collector 341 will eventually approach ambient (e.g. room) temperature, die aqueous ammonia collected in diem would be below saturated concentration, i.e. forming a weak solution ready to absorb ammonia vapour in next session. This is highly beneficial when die compressor assembly 110 works intermittently with relatively short operating periods and long idle periods so dut ammonia is evaporated mainly during die operation while water over all periods.

Fig. 24C shows a second embodiment of die coolant column 310. The differences between this embodiment and that of Fig. 24A are that the cap 115 in Fig. 24A is replaced by a valve assembly formed by a valve member 115', a support member 118 and a spring 119, and a flap valve 319 is attached beneath the collar portion 313 of die flow guide 312 to cover die through-holes 314. In operation, die valve member 1 15' is

biased by die spring 1 19 to keep the imier evaporation zone 304 closed, at die same time die flap valve 319 closes all die through-holes 314. Alternatively, the member 115 ^' can be biased by magnetic force, widi or widiout die spring 1 19. That is to say, the boiling zoue 302 and die vapour chamber 301 are separated from the inner and outer zones 304 and 303, to allow a low pressure to be formed by the effect of the compressor 320. The low pressure and the elevated temperature of die heat exchange member 1 13 in die boiling zone work together to drive water to evaporate so dut a proper proportion of water vapour will enter the unk 400 with die ammonia vapour to form the aqueous ammonia. During this time, a high pressure is built up in die iiuier evaporation zone 304, which would, in combiiution with the low pressure in die vapour chamber 301 , eventually overcome the biasing force on the valve member 1 15' and force it to open. Once diis happens, a bubbling flow will rush into die boiling zone, and at die same time the flap valve 319 will also be opened to allow the high concentration solution in the boiling zone to enter the outer precipitation zone 303 below. In this way an intermittent local circulation of the coolant solution is fanned which balances the amounts of water and ammonia vapours formed in the process. For control puφoses, die average temperature value provided by the sensor 514 is used by the unit 500 to balance diis intermittent operation. Due to this balanced evaporation of different vapour components, the embodiment can be used for continuous operation or for intermittent operation with relatively short idle periods.

Also shown in Fig. 24C is a second thermoelectric anangement at the lower end of the assembly 1 10, which is fanned by the imier heat exchange member 1 17', diermoelectric members 1 12' and an outer heat exchange member 1 13'. They work in the same way as the members 1 12, 1 13 and 1 17 described above, but in an opposite direction, i.e. die member 113' farms the cold side of the arrangement so it cools the solution in die outer zone 303 to encourage more salt precipiution. Again it is used to maintain die temperature gradient in the column 310, as described above.

Fig. 25 A shows details of die defrosting system 600, which has a dieπnoelectric member 610 atuched to die bottom of the unk 210, and a heat exchange member 620 fitted beneath it. The member 610 is fanned by an anay of Peltier elements sunounded by an insulating member 61 1 for separating die member 620 from die unk 210 and also for absorbing diermal stress between diem. A space 630 is formed below die unk 210 by a hinged member 631 which also serves as an ice collecting member and a separation board 632. Two comer channels 633, also shown in Fig 22B, extend downwards. Two small fans 640 are each fitted to one back comer to draw in air from the front edge of the member 620, dien to drive the air out into die imier space of die freezer. Fig. 25B shows that die member 620 has a base 621 in direct conuct witfi die bottom surface of the thermoelectric elements 610. and a number of fins 622 extending downwards, separating die space between die base 621 and the hinged member 631 into a number of parallel channels. A number of concave cells 623 are formed on the base, each conesponding to one of the diermoelectric elements. An ice detecting system is formed by a light emitting diode (LED) 661 fitted to one side of die heat exchange member 620 with a light sensor 662 fitted opposite to it. A series of snull holes 663 are fomied in die fins 622 allowing a light beam 664 from the LED 620 to pass through and be received by die sensor 662.

The defrosting arrangement 600 can work in two different modes, i.e. the first mode for deep-freezing and the second mode for defrosting/ice collecting.

When the anangement 600 operates in die first mode, an electric cunent is supplied to die thermoelectric element array 610 in a first direction, in which die top surface of each element 610 warms up and gives heat to die evaporator 150, while the bottom surface of each element becomes colder, cooling the member 620 to a temperature lower than dut of the evaporator 150 and die antifreeze liquid in die cold unk 210. That is to say, die heat exchange member 620 becomes die coldest part inside die freezer 200, which can he well below -30°C, i.e. lower dun the operation range when ammonia is used as refrigerant. At die same time, die fans 640 cause an air flow passing dirough die cluiuiel defined by die fins 622, in which die air is dehydrated and its moisture content becomes frost. Then the dry and cold air is driven by the fans 640 down into die comer channels 633 to be released at the bottom into the inner space of die freezer 200. Because die temperature in die concave cells 623 is lower dian at any odier position in die freezer, frost accumulates in diese cells, which helps to keep odier parts of die freezer frost-free. When a certain amount of frost/ice is collected in the cells 623, die light passage 663 will be blocked and diis will be sensed by the sensor 662, so die control unit 500 can switch die operation to the defrosting mode.

In the defrosting mode, a large cunent of reversed direction is supplied to die diermoelectric elements 610, which has the effect of causing diem to reverse dieir operation to absorb heat from die cold storage tank 210 and transferring the same to die heat exchange member 620 to cause a sluφ raise of its temperature. This temperature change of die member 620 causes a thermal expansion of die side walls of die concave ice cells 623 which produces a t eπnal shock to break die accumulated ice, which is brittle. The broken pieces of die ice dien drop onto the ice collecting member 631, forcing it to swing downwards, as shown by die dash line 631 ', to deliver the ice pieces into an ice collecting pocket 650 atuched to die inner surface of d e door 202. After die frost/ice accumulated on the member 620 has been cleared, the sensor 662 would be able to receive light signals from the LED 661, and die operation can be switched back to the first mode.

During die defrosting operation, die frost/ice accumulation is physically expelled by die diermal shock produced by die heat exchange member 620, and to achieve diis effect the amount of energy and the length of time needed are very small because it does not need to melt die ice. as in most conventional anangemeiits. On die odier hand, die ice is formed by pure water, so the pieces themselves are valuable commodities ready for human consumption. In diis sense, nothing is wasted during die operation of die arrangement. Obviously, diis arrangement can also be used for the purposes of ice-making or quick-freezing. To achieve this, a control button can be fitted to die unit 500 so that a user can make ice or quickly freeze fresh goods by putting a tray of water or the goods into die freezer 200, suiubly at die bottom close to die air outlet of die comer channels 633, then starting the deep-freezing mode of the arrangement 600 by pressing the control button. In due course, die defrosting mode would be actuated and a light signal indicates dial the ice pieces are collected in die pocket 650, ready for use.

A method for controlling the operation of die cooling system according to die invention is described hereinbelow with reference to Fig. 21 and Fig. 26. The operation of the control unit 500 is based on consunt evaluation of four temperature values including: tl, sensed by die sensor 511 in the heat tank 400, indicating die temperature of the heat storage material at the ouϋet end of die passage 330; t2, sensed by the sensor 512 in the freezer 200, indicating its internal temperature; t3, sensed by die sensor 513 in the cold tank 210, indicating the temperature of die antifreeze liquid; and t4, sensed by the sensor 514 in the inner evaporation zone, as

described before. In practice, each of die sensors 511 to 514 can be foπiied by a set of sensing elements fitted at different positions of die relevant components, and die values of tl to t4 would be die average values of each set of die sensing elements. The temperature sensor 51 1 m die heat unk can also be replaced by a pressure sensor for die same control purpose, dien die control can be conducted on the basis of the vapour pressure in die passage 330. To faciliute die description, predetermined reference values of a set of control parameters for each parts of the system are marked in Fig. 21 , which are suiuble, e.g. for a standard four star (* * * *( freezer, in which the temperature inside the freezer, i.e. t2, should be controlled below -I4°C. Corresponding to this value, the freezing point of the antifreeze liquid in the cold storage unk 210 is adjusted to about -20°C. that is to say ice crysuls begin to appear in the liquid at -20°C. and it can be considered as fully frozen when below -28°C. For practical puφoses die highest expected room temperature, e.g. in England, is assumed to be below 30°C so this value is selected as the required fusion point of the heat storage material at the outlet of the passage 330 in the heat tank 400. In operation, as long as tl is not beyond this fusion point, one can assume that the unk 400 has not been charged to its full heat storage capacity. Obviously, these values are given as examples, and they should be adjusted when conditions change. From control point of view, the system can operate in any of the following modes:

(a) Full capacity cooling operation (Full cooling) In diis mode, the refrigerant circulating circuit 100, the coolant circuit 300, the brine circuit 200 and die diermoelectric system 600 (in its deep freezing mode) are all operating at dieir respective full capacity. This is achieved by die unit 500 by actuating the compressor assembly 1 10 and the thermoelectric members 1 12, the compressor 320, the control valve 350, the pump 230. die diermoelectric member 610 and the fans 640 (shown in Fig. 25A).

(b) Deep freezing operation (Deep freezing) In diis mode, it is intended to achieve the full capacity cold storage in the cold unk 210. The system operation is basically the same as that of mode (a) except diat the circulating pump 230 is turned off to allow the antifreeze liquid to freeze.

(c) Economic cooling operation (Economic cooling) The difference between diis mode and the above mode (a) is that die system 600 (including die fans 640) does not work so the power consumption is lower.

(d) Air circulating operation (Air circulating) In diis mode, only the fans 640 work to circulate air in die freezer 200. The air is cooled in die space 630 by die frozen antifreeze Iiquid in the cold unk 210.

(e) Defrosting operation (Defrosting) In diis mode, die system 600 is switched to its defrosting mode as described before. It will switch back automatically when die frost/ice accumulation has been cleared. This operation happens during the full cooling mode (a) or the deep freezing mode (b).

(f) Pause In this mode, the control unit 500 turns die system off for a predetermined period, e.g. of 10 minutes. All components are off except die relevant sensors for monitoring condition changes, but the compressor 320 is kept operating at low power as an active one-way valve for at least a part of this period, as explained before. This mode is prolonged when die freezer door is open and stops when die door shuts. Fig. 26 shows die control logic of die control unit 500, which is based on a CPU widi necessary supporting components. In Fig. 26, the first step S000 is die suiting point of a control session and also die point that die program returns i.e. fomi step S006 after each session. Once die process is started, die first diing is to check at step S001 whedier die freezer door is open. If the answer is "Yes", die operation is switched to die mode "Pause" at step S005. As mentioned above, diis mode is maintained as long as die door is kept open

and temiiiuted when the door is shut. Alarming arrangement can be incorporated into the uiύt 500 to alami a user when the door is kept open for too long or not being shut properly. If or once die door is shut, die program goes to die next step S002 to read from die sensors 51 1 to 514 the respective values tl . t2, t3 and t4 Then in step S003, the value of tl is checked to see whether the heat tank 400 is charged to its full capacity. If die answer is "Yes", it goes to die step S005 to pause for a while, allowing die nk to cool down. Odierwise it goes to die step S004 to see whether the low cost electricity is available. This can be done by checking the voluge of different power input teπniiuls. In case of using off-peak electricity, this step is simply to check a tinier in the control unit to see whether it is the right time. If the answer is "Yes", the system is switched to die low-cost electricity dien executes step SI 00. in which it checks whedier t3 is below -28°C to see whedier the cold nk 210 has been charged to its full capacity. If die answer is "Yes", it goes to step S005 to pause e.g. for ten minutes, then returns via step S006 to the surting point S000. If the answer at step SI 00 is "No", i.e. the tank 210 is not fully charged, the control unit executes step S 101 to check the value of t3 to see whether the antifreeze liquid in the unk 210 is partially frozen. If the answer is "Yes", the system starts the "Deep freezing" mode at step S 103. odierwise it goes to step S102 to see whedier t3 is below -14°C. If it is "Yes", it operate in "Full cooling" mode at step S 104 when the antifreeze liquid can circulate. After the operation is surted, the ice detecting system checks at step S105 to actuate the defrosting operation at step S106 if the answer is "Yes", then it returns. If die answer at step S102 is "No", it indicates dut die antifreeze liquid is very waπn, it goes to step S207 for "Economic cooling". This would avoid the system to be overloaded, e.g. during its start-up. Then it returns for next session of the control process. If at the step S004, it is found d t die low-cost electricity is not available, die control unit 500 executes step S200 to see whether the value of t2 is above 0°C, i.e. whedier the interior of die freezer 200 is unaccepubly wann. This may occur when the freezer is suited, its door is kept open for a long time or a large quantity of fresh goods is loaded into it. If the answer is "Yes" , it goes to step S 102. then the following steps would be the same as that mentioned above. If the answer at step S200 is "No" , it checks at step S201 to see whedier die ice-nuking/quick-freezing button is pressed. If the answer is "Yes", it goes to step S 10 I to decide eidier to start full cooling or deep freezing operation according to die value of t3. If the answer at step S201 is "No", i.e. no instruction for deep-freezing, the value of t2 is checked at step S204 to see whether the internal temperature is accepuble. i.e. below -14°C. If die answer is "Yes", the system is switched to "Pause" mode at step S005, then resuπs, odierwise die value of t3 is checked at step S205 to see whedier the antifreeze liquid is still frozen. If it is "Yes", the system surts die air circulation by the fans 640 at step S206. during which the interior of the freezer is cooled by the cold energy stored in the cold nk 210. In practice, this operation would be able to cope widi a temporary change of the internal temperature caused, e.g. by die invasion of wann air when die freezer door is opened briefly. If it is "No" at step S205, die system operates at step S207 in "Economic cooling" mode which is the mode when die low-cost electricity is not available. Once one of die operation modes in step S 103, S104, S206 or S207 is surted, die control unit 500 does not stop the operation until it finds at a later session diat one of die control criteria for stopping die operation exists, i.e. at step S001. S003, S100, or S204.

By selectively switching between different modes, die system achieves an improved efficiency. For example, if the system is designed to make full use of off-peak electricity at night, by switching die system to

fiill cooling mode at step S104 or deep freezing mode at step S103, die interior of die freezer is cooled to below -28°C. In this case, a very large quantity of cold energy is stored not only in die frozen antifreeze liquid in die cold nk 210 and die channels 220 shown in Figs 2A to 2C, but also in the goods supported on the shelf members 203. The stored cold energy would keep the cooling mechanism idle for a long period after the low cost electricity is no longer available, dierefore reducing the amount of "expensive" electricity used during daytime. In this case, since die cold energy is stored during die period from midnight to early morning, when the ambient temperature is die lowest and also it is lest possible for the freezer door to be opened, the overall cooling efficiency is highest. Industrial Applicability In the above description, a freezer is used to illustrate die inventive concept. Obviously, the concept can be used in other applications, e.g. a fridge widi or widiout a freezer compartment, a refrigerated display case or cabinet, a cold room, a refrigerated vehicle, vessel or aeroplane, or an air-conditioning system. The concept of using off-peak electricity can be equally applied to other kinds of low-cost electricity, such as solar, wind or tidal power, or in case of a mobile application, die ground power in contrast with the electricity supplied by an on-board generator. Odier kinds of power can be used in combination, e.g. using solar power during the day and die off-peak power from national grid at night, combined with wind power whenever it is available. Other evaporable component(s), such as methanol, ethaiiol and/or carbon dioxide can be used in the coolant solution in case ammonia is not suiuble, e.g. when a system has copper parts; the heat storage unk may not be necessary for a system in which heat dissipating is not a problem: and the ice-makiiig/quick-freezing facility would not be needed for air-conditioning or high temperature refrigeration. In die latter case, a cold storage gel of a higher freezing temperature, which is known in die art, can be used to replace the antifreeze for storing cold energy above 0°C, suiuble for air-conditioning.

It should be noted that the embodiments according to the second general object of the invention have the following patentable aspects. 1. An axial flow pump/marine propeller comprising: a hollow body; a sutor in die body defining an inner space; a tubular rotor with propelling means in die imier space; and electromagnetic means for generating a routional nugnetic field in said inner space; wherein die rotor is supported by suspension means for rotational and dirust bearing to retain it at a balanced position in response to its routional and/or axial movements.

2. An apparatus of aspect 1 , wherein die rotor or die sutor has spiral means formed on at least one of the surfaces defining a gap between them, for forming a peripheral flow when die rotor routes.

3. An apparatus of any preceding aspect, further comprising sealing means at each end of die rotor to seal said gap, so that a lubricating and/or cooling liquid can be filled in die sealed gap.

4. An apparatus of aspect 3, further comprising means for supplying and circulating under a suble pressure said lubricating and/or cooling liquid in said gap. 5. An apparatus of any preceding aspect, wherein amiular members of high magnetic resisunce are fitted to conespouding positions on the sutor and rotor to fomi a magnetic registration mechanism therebetween. 6. An apparatus of any preceding aspect, wherein said bearing means include magnetic suspension anangement comprising: an electromagnet with tow annular poles and a permanent magnet with matching poles so diat repelling or attracting forces can be controlled by die electric cunent to die electromagnet.

7. An apparatus of aspect 6, wherein the poles of die electromagnet and permanent magnet are ananged to form complemeiiury cylindrical and/or comcal surfaces so as to provide roury and/or thrust suspension.

8. An apparatus of aspect 6 or 7, further comprising a control unit and sensing means for adjusting die electric cunents supplied to the suspension anangement to maintain the rotor's axial position. 9. An apparatus of any preceding aspect, wherein a throttle ring is fitted to an axial end of said rotor, which ring restricts the peripheral flow when the rotor moves towards d t end.

10. An apparatus of any preceding aspect, wherein the suspension bearing includes flow dividing means for forming two opposite peripheral flows, so as to keep the rotor's position self-balanced.

1 1. An apparatus of any of die preceding aspects, wherein said propelling means includes one or more flexible spiral blades which are compressible in response to changes of power input or working load.

12. An apparatus of aspect 11. wherein said blades are biased elastically to keep them axially expanded.

1 . An apparatus of any preceding aspect, wherein said propelling means includes two coaxial spiral means in opposite spiral directions so as to cancel each other's swirling effects and produce high pressure ouφut.

14. An apparatus of aspect 13. further comprising a second set of a sutor, a rotor and electromagnetic means to route in an opposite direction.

16. An apparatus of aspect 13 or 14. wherein die central one of said two spiral means is supported by a pivot bearing.

17. An apparatus of any of the preceding aspects, wherein said propelling means includes a conical impeller.

It should be noted that the embodiments according to the third general object of the invention have the following patentable aspects.

1. A cooling system comprising: a primary mechanism for transferring heat from a cold-generating member to a heat-rejecting member, and a circuit thermally engaged with said heat-rejecting member of said mechanism; wherein a coolant with evaporable component is circulated in said circuit which lus means for promoting evaporation of said component to improve efficiency of said primary mechanism by said evaporation. 2. A system of aspect 1 , wherein said circuit has endothermic salt for cooling by its endothermic dissolution.

3. A system of aspect 1 or 2, wherein said coolant includes aqueous ammonia or carbonated water.

4. A system of any preceding aspect, wherein said promoting means further comprises a diermoelectric member ananged with its cold-side engaging said heat-rejecting member and its hot-side engaging said coolant solution to provide an elevated temperature for increasing said evaporation. 5. A system of any preceding aspect, wherein said promoting means further comprises a vapour pump.

6. A system of any preceding aspect, wherein said circuit further comprises: a heat absorbing portion thermally coupled with said heat-rejecting member and a heat dissipating portion in fluid communication widi said heat absorbing portion; wherein said heat absorbing portion lus an upper part widi a coolant outlet, an lower part widi a coolant inlet, and an intermediate zone engaging said heat-rejecting member, said zone has baffle means for subilising a temperature gradient between said upper and lower parts so diat said evaporable component of said coolant can evaporate in said upper part and flow to and condense in said heat dissipating portion.

7. A system of aspect 6 in combination with aspect 2, wherein said heat absorbing portion further comprising a flow guide means which separate said intermediate zone into a first pan defining an evaporation zone in diermal conuct widi said heat-rejecting member in which said coolant component evaporates, and a second part defining

a precipiution zone separated from said heat-rejecting member but in fluid communication with said first part so that said endothermic salt in the coolant can precipiute and settle to said lower part.

8. A system of aspect 7, wherein a salt chamber is formed at said lower part of die heat absorbing portion to ave said endotheπiiic cooling effect to said heat-rejecting member. 9. A system of any preceding aspect, further comprising: a brine circuit theπnally engaged with said cold- generating member for improving its heat-exchange efficiency.

10. A gas compression assembly for a system of any preceding aspect, comprising: a plurality of compressors serially connected so that a refrigerant can be progressively compressed and supplied to a refrigeration circuit: wherein said compressors are arranged as aid heat-rejecting member engaging said coolant circuit so that the heat generated therein by them is absorbed and carried away by circulating said coolant.

1 1. An assembly of aspect 10. further comprising condenser means connected to the downstream end of said gas co pression chain, and being thermally coupled with said coolant circuit.

12. A heat storage tank far a cooling system comprising a theπually conductive casing, a fluid passage with a thermally conductive and flexible wall in said casing, and at least one heat storage material filled between said casing and said fluid passage for absorbing heat from fluid passing through said passage: wherein said casing is divided into chambers of different heat storage materials, and wherein the chamber at the up-stream end of said fluid passage has a phase-change material of a fusion point higher than that of the material in the chamber at the down-stream end of said fluid passage.

1 . A defrosting arrangement comprising a theniioelectric member with a theπnal pole coupled to a cold- generating member of a cooling system and the other thermal pole to a heat exchange member, and a control unit for changing electric supply to said diermoelectric member to reverse the heat transfer direction, therefore selectively changing said heat exchange member from a frost accumulating mode to a defrosting mode.

14. An anangement of aspect 13, further comprising means for sensing the amount of ice accumulated onto said heat exchange member and indicating the same to said control unit. 15. An arrangement of aspect 13 or 14. further comprising means for collecting ice expelled during its defrosting operation, including a hinged mechanism for delivering ice to a pocket.

16. A niediod of operating a cooling system having an antifreeze circuit, comprising steps of: a) setting a mode control unit for selecting one of two operatioiul modes in response to the availability of low cost electricity: b) operating in a first mode when said low cost electricity is not available, to cool said antifreeze to a temperature under which it can be circulated; and c) operating in a second mode when said low cost electricity is available, to freeze said antifreeze so as to store cold energy dierein.

17. A method of aspect 16. further comprising a step of: d) causing an air circulation widiin a cooled space during the operation of said second mode; or e) causing frost accumulation on a heat exchange member associated widi a thermoelectric member; and f) causing defrosting by reversing die operation of said thermoelectric member.