COOLING DISTRIBUTION UNIT Field of Invention

The present invention relates generally to heating and cooling control systems, and more particularly, to a system and method for controlling fluid flow and fluid temperature within a pumped two-phase cooling distribution unit.

Brief Description of the Drawings

One advanced method for cooling heat-generating devices is to use a pumped two-phase cooling system. With reference to Fig. 1 , an exemplary pumped two-phase cooling system 10 is illustrated. The system 10 includes a liquid pump 12, one or more heat generating devices (not shown) cooled by one or more evaporators 14, a condenser 16 and an accumulator tank 18 all in fluid communication via conduits 19a-19d.

In operation, the liquid refrigerant is partially boiled through the one or more evaporators 14 producing a two-phase mixture that is transported to the condenser 16 where it condenses back to a liquid. The evaporator 14 can be in the form of a coiled tube surrounding the heat generating device, a heat sink touching the heat generating device, a liquid to refrigerant heat exchanger removing heat from the heat generating device via an intermediary fluid, an air to refrigerant evaporator cooling heated air from the heat generating device, or another form of removing heat from the heat-generating device. In many industries, including data centers, the portion of the two-phase cooling system that pumps and condenses the fluid is typically located remotely from the evaporator portion. The pumping and condensing portion is often times referred to as the cooling distribution unit (CDU). The CDU also typically contains the control system, which manages the flow and temperature of the fluid.

Summary of the Invention

There are several problems with conventional controllers currently available in CDU's. First, conventional controllers do not have an automatic way to account for varying demand for fluid flow as parallel evaporating loops are added or subtracted. One example is the addition or removal of servers in an IT rack which each contain a fluid evaporating loop, where each fluid loop to each server is plumbed in parallel. In this example, as a server is added more fluid flow is needed from the CDU and as a server is removed less fluid flow is needed from the CDU.

Secondly, conventional controllers do not have a way to automatically adjust the pump speed if cavitation exists. With pumped two-phase systems, the pumps operate with the pump inlet fluid temperature very close to the boiling point of the fluid. During certain conditions such as low heat load, no heat load, or a sudden loss of heat load, the pumps are prone to cavitation, which is damaging to the pumping components.

Thirdly, conventional controllers do not have a method for maintaining temperature control that is both fast enough and stable enough to work with highly transient heat loads. The conventional approach is to use a Proportional- Integral-Derivative (PID) control method to measure the fluid temperature (usually the two-phase return temperature to the CDU) and adjust a water valve or fan to hold that temperature nearly constant. The problem is that if the PID constants are set too high, the system becomes unstable under fast changes in heat load and if they are set too low, the system cannot react fast enough to the same transient conditions.

Lastly, there is a risk when using chilled water to condense the two- phase refrigerant that it may bring the refrigerant temperature lower than the dew point temperature in the room. If this happens, condensation can form on the refrigerant plumbing and drip on to the electrical devices, causing damage to the devices. The conventional approach is to monitor for condensation and alert the user if it occurs.

A system, apparatus and method in accordance with the present disclosure utilize one or more control methodologies to improve performance of a pumped two-phase cooling system. The control methodologies include at least one of flow control based on a pressure differential across the CDU, automatic cavitation control, refrigerant temperature control and dew point control.

According to one aspect of the invention, a control system for a cooling system utilizing a two-phase refrigerant includes: a cooling distribution unit (CDU) including a CDU refrigerant supply line for supplying the refrigerant to at least one evaporator, and a CDU two-phase return line for returning the refrigerant to the CDU; and a controller (22) operatively coupled to the CDU, the controller comprising a first control loop configured to maintain a constant pressure differential across the CDU refrigerant supply line and the CDU two- phase return line.

In one embodiment, the CDU includes: an accumulator tank; a refrigerant pumping portion having at least one pump; and a bypass valve, the accumulator tank, refrigerant pumping portion and bypass valve in fluid communication with each other, wherein the bypass valve is operative to divert refrigerant flow from the at least one pump away from the at least one evaporator and into the accumulator tank, the controller configured to adjust a pump speed of the at least one pump and a position of the bypass valve to maintain the CDU pressure differential.

In one embodiment, when refrigerant demand is below a predetermine threshold, the controller is configured to command the at least one pump to operate at a predetermined minimum speed, and to regulate the CDU pressure differential via the bypass valve.

In one embodiment, when the bypass valve is commanded to bypass a minimum amount of refrigerant the controller is configured to command the at least one pump to vary a speed to regulate the CDU pressure differential.

In one embodiment, the at least one pump comprises a plurality of pumps, and the controller is configured to place one of the plurality of pumps in a standby mode and use another of the plurality of pumps to regulate the CDU pressure differential.

In one embodiment, the at least one pump comprises a plurality of pumps, and the controller is configured to command each of the plurality of pumps to operate at a reduced speed to regulate the CDU pressure differential.

In one embodiment, the controller is configured to use an output from a single PID controller to command the plurality of fluid pumps.

In one embodiment, the controller further comprises a second control loop configured to detect cavitation at the at least one pump, and automatically vary a speed of the at least one pump upon detection of cavitation.

In one embodiment, the second control loop is configured to detect cavitation by: converting refrigerant pressure at an inlet of the at least one pump to a saturation temperature; comparing the saturation temperature to an actual refrigerant temperature at the inlet of the at least one pump to acquire an amount of subcool in the refrigerant; determining a minimum amount of subcool in the refrigerant that provides cavitation -free operation for a given pump speed; and determining cavitation is present based on a comparison of the amount of subcool in the refrigerant and the minimum amount of subcool.

In one embodiment, the control system includes: a condenser(16) in fluid communication with the two-phase return line; a fluid valve (40) in fluid

communication with the condenser and operative to provide a coolant to the condenser for cooling the refrigerant, wherein the controller includes a third control loop configured to regulate a temperature of the refrigerant by varying a coolant flow through the condenser.

In one embodiment, the system includes: a condenser in fluid

communication with the two-phase return line; a cooling device operative to provide a cooling medium to the condenser for cooling the refrigerant, wherein the controller includes a third control loop configured to regulate a temperature of the refrigerant by varying a flow of the cooling medium through the condenser.

In one embodiment, the third control loop is configured to: generate a refrigerant temperature reference based on a refrigerant temperature setpoint and a tolerance value; convert the refrigerant temperature reference to a corresponding pressure reference; and adjust coolant flow through the fluid valve based on a comparison of the corresponding pressure reference with an actual pressure at the CDU two-phase return line.

In one embodiment, the controller comprises a fourth control loop configured to vary the refrigerant temperature setpoint based on a comparison of an ambient dew point and an actual refrigerant temperature. In one embodiment, the fourth control loop is configured to: calculate the ambient dew point based on ambient temperature and ambient humidity;

calculate a dew point reference based on a predetermined threshold value added to the calculated ambient dew point; and activate a dew point fault when the fluid temperature is below the dew point reference.

In one embodiment, when a dew point fault is active, the fourth control loop is configured to increase the refrigerant temperature setpoint by a predetermined value, and recalculate the ambient dew point.

In one embodiment, when a dew point fault is active and the refrigerant temperature is above the dew point reference, the fourth control loop is configured to decrease the refrigerant temperature setpoint by a predetermined value, and recalculate the ambient dew point.

According to another aspect of the invention, a method for controlling a cooling system utilizing a two-phase refrigerant, the cooling system including a cooling distribution unit (CDU) having a CDU refrigerant supply line for supplying the refrigerant to at least one evaporator, and a CDU two-phase return line for returning the refrigerant to the CDU, the method comprising maintaining a constant pressure differential across the CDU refrigerant supply line and the CDU two-phase return line.

In one embodiment, the CDU includes a refrigerant pumping portion having at least one pump, and a bypass valve is operative to divert refrigerant flow from the at least one pump away from the at least one evaporator, the method comprising adjusting a speed of the at least one pump and a position of the bypass valve to maintain the CDU pressure differential.

In one embodiment, when refrigerant demand is below a predetermine threshold, the method includes operating the at least one pump at a predetermined minimum speed using the bypass valve to regulate the CDU pressure differential.

In one embodiment, the method includes: detecting the occurrence of cavitation at the at least one pump; and automatically varying a speed of the at least one pump upon detection of cavitation, wherein automatically varying the speed includes converting refrigerant pressure at an inlet of the at least one pump to a saturation temperature; comparing the saturation temperature to an actual refrigerant temperature at the inlet of the at least one pump to acquire an amount of subcool in the refrigerant; determining a minimum amount of subcool in the refrigerant that provides cavitation -free operation for a given pump speed; and determining cavitation is present based on a comparison of the amount of subcool in the refrigerant and the minimum amount of subcool.

According to another aspect of the invention, a controller for controlling a two-phase cooling system includes: a processor and memory; and logic stored in the memory and executable by the processor, the logic when executed by the processor configured to cause the processor to carry out the method described herein.

To the accomplishment of the foregoing and related ends, the invention, then, comprises the features hereinafter fully described and particularly pointed out in the claims. The following description and the annexed drawings set forth in detail certain illustrative embodiments of the invention. These embodiments are indicative, however, of but a few of the various ways in which the principles of the invention may be employed. Other objects, advantages and novel features of the invention will become apparent from the following detailed description of the invention when considered in conjunction with the drawings Brief Description of the Drawings

Many aspects of the invention in accordance with the present disclosure can be better understood with reference to the following drawings. The components in the drawings are not necessarily to scale, emphasis instead being placed upon clearly illustrating the principles in accordance with the present disclosure. Likewise, elements and features depicted in one drawing may be combined with elements and features depicted in additional drawings. Additionally, in the drawings, like reference numerals designate corresponding parts throughout the several views.

Fig. 1 is a schematic diagram illustrating a conventional two-phase cooling system.

Fig. 2 is a schematic diagram illustrating a two-phase cooling system in accordance with the present disclosure.

Fig. 3 is a block diagram illustrating exemplary steps for carrying out a first control methodology in accordance with the present disclosure.

Fig. 4 is graph showing pump speed/valve position vs. PID output for a system in accordance with the present disclosure utilizing one pump.

Fig. 5 is graph showing pump speed/valve position vs. PID output for a system in accordance with the present disclosure utilizing two pumps.

Fig. 6 is graph showing required PID output vs. the number of installed parallel cooling loops.

Fig. 7 is a flow chart illustrating a pump cavitation control methodology in accordance with the present invention.

Fig. 8 is a flow chart illustrating a refrigerant temperature control methodology in accordance with the present invention. Fig. 9 is graph showing water valve position vs. pressure setpoint in accordance with the present disclosure.

Figs. 10A and 10B illustrate steps for carrying out a dew point control methodology in accordance with the present invention.

Detailed Description

A system, apparatus and method in accordance with the present disclosure for controlling a cooling system, such as a two-phase cooling system, solve one or more of the issues listed above. In one embodiment, the system includes at least one and preferably four control loops that are seamlessly integrated together into an integrated control solution.

Referring to Fig. 2, an exemplary system 20 for controlling a two-phase cooling system is illustrated. Like the system 10 in Fig. 1 , the system 20 includes a liquid pump 12, one or more heat generating devices (not shown) cooled by one or more evaporators 14, a condenser 16 and an accumulator tank 18, each in fluid communication via conduits 19a-19d.

A controller 22 executes one or more control algorithms in accordance with the present disclosure to optimize performance of the cooling system. In one embodiment the controller 22 includes a processor and memory communicatively coupled via a system bus, and logic stored in memory and executable by the processor to cause the processor to carry out the method described herein. In another embodiment, the controller 22 comprises an application-specific integrated circuit (ASIC) including logic configured to carry out the method described herein.

A pump inlet refrigerant temperature sensor 24 measures a temperature of the refrigerant entering the pump 12, an inlet pressure sensor 26a measures a pressure of the refrigerant at the inlet of the pump 12, and an outlet pressure sensor 26b measures a pressure at the outlet of the pump 12 (i.e., the pressure exiting the CDU). A return line pressure sensor 27 measures a pressure in the refrigerant in the return line to the CDU. A speed sensor 28 is operatively coupled to the pump 12 to provide data corresponding to the pump speed. Such sensor may be an analog sensor or digital sensor (e.g., tachometer, resolver, encoder, etc.), and may be directly coupled to the pump 12 or indirectly coupled to the pump 12 (e.g., to a motor (not shown) driving the pump). Although not shown, the controller 22 is operatively coupled to a motor that drives the pump 12, and can vary a speed of the motor thereby varying a speed of the pump 12. A bypass valve 29, which is controllable by the controller 22, couples an output of the pump back to the accumulator tank 18 via conduits 19e and 19f .

An ambient temperature sensor 30 and ambient humidity sensor 32 measure the temperature and humidity of the ambient environment, respectively, and communicate the measurements to the controller 22. The ambient environment is the space immediately surrounding the CDU. If the evaporator 14 is located in a space different from the CDU, the ambient environment includes both spaces. One or more setpoint setting devices 34 provide one or more setpoints to the controller 22 for regulating one or more parameters of the system (e.g., refrigerant temperature, dew point safety margin temperature, chilled water temperature, CDU pressure differential, etc.). The temperature setting devices may be any analog or digital device capable of generating signal indicative of a desired temperature, or may be preset values configured in the controller 22 (e.g., stored in memory of the controller). Cooling water is provided from a cooling water source (not shown) to the condenser 16 via a supply and return lines 36 and 38 to regulate a temperature of the refrigerant. A cooling water control valve 40 is operatively coupled to the controller 22 and is operative to control the flow of cooling water through the condenser 16 based on a command from the controller 22.

As will be described in more detail below, the controller 22 executes one or more control methodologies in accordance with the present disclosure to control the two-phase cooling system. More particularly, the controller 22 may execute a first control methodology for automatically controlling flow through the system based on a pressure differential between a liquid supply line and a two- phase return line. A second control methodology can prevent pump cavitation, while a third control methodology controls a temperature of the refrigerant by adjusting water flow rate via a water control valve. Finally, a fourth control methodology prevents the occurrence of condensation.

Pressure differential flow control

A first control methodology that may be executed by the controller 22 is automatic flow control via CDU pressure differential control. The flow control loop works to maintain a constant pressure differential between a CDU liquid supply connection 19b and the CDU two-phase return connection 19c. This is achieved by adjusting the speed of the pump 12 (or pumps) as well as adjusting the position of the pump bypass valve 29, which when open diverts flow from the pump 12 back to the tank 18 within the CDU, thus bypassing the rest of the system.

The control loop is designed to control from zero flow up to full rated flow of the system. Because the system may use positive displacement pumps, it is not feasible to block all flow while the pump 12 is running, as this would cause a large spike in pressure and could be damaging to the system. Thus the bypass valve 29 may be needed during operating conditions when little or no flow is needed.

One example of why the system needs to operate with little or no flow is when most or all of the cooling loops are disconnected from the system. Such scenario may occur, for example, if removable servers with integrated cooling loops are utilized. During initial installation or during a period of maintenance, some or all of the servers may be removed while the cooling system is still running. As each cooling loop is removed, less flow is needed from the CDU, therefore the flow output of the CDU must be dynamic.

The control loop works off of a PID feedback loop with CDU dP being the feedback signal. The dP setpoint is user selectable over a broad range (e.g., 20 psid to 100 psid). The control output is a 0-100% command that is then converted to a pump speed command and a bypass valve position command through a series of equations. Fig. 3 illustrates exemplary flow chart 50 in accordance with the present disclosure for implementing dP control across the CDU. The steps illustrated in Fig. 3 may be implemented by the controller 22.

Beginning at step 52, the pump outlet pressure sensor 26b obtains a pressure reading for the liquid supply line leaving the CDU 19b and communicates the pressure reading to the controller 22, and at step 54 the pressure sensor 27 obtains a pressure reading for the two-phase return line and communicates the pressure reading to the controller 22. Next at step 56 the controller 22 calculates a pressure differential dP based on a difference between the measured supply line pressure and measured return line pressure. The calculated pressure difference is used by the controller as the differential feedback value dP. At step 58, the controller 22 obtains a differential pressure setpoint value. Such setpoint value may be obtained, for example, via an operator input device, communicated to the controller 22 via another device, calculated based on ambient and/or system conditions, or preset in the controller 22.

Moving to step 60, the controller 22 calculates an error signal by computing the difference between the dP setpoint value as obtained in step 58 and the dP feedback value as obtained at step 56. The error signal then is provided to a control algorithm, such as a PID (proportional, integral and derivative) controller or the like, as indicated at step 62. The control algorithm generates an output that attempts to cause the dP feedback value to approach the dP setpoint value.

More particularly, and as indicated at step 64, the control algorithm can calculate two different outputs. A first output corresponds to a setting for the bypass valve 29, e.g., between a minimum opening and a maximum opening of the bypass value. A second output corresponds to the speed of the pump 12. An exemplary relationship between the command value corresponding to pump speed and a command value corresponding to bypass valve position is shown in Figs. 4 and 5.

As can be seen in Figs. 4 and 5, at the lowest flow condition, zero flow leaving the CDU, the pump 12 (or pumps) must still run to maintain CDU differential pressure dP. However, due to mechanical and electrical constraints with the pump 12, it may not be feasible to slow the pump down to near zero flow so it is commanded to run at an idle speed, which is still within the pump's safe operating range. This sets the lowest value the pump 12 can run at while the system is running. At this condition, the bypass valve 29 must be at least partially open to allow a path for the fluid pumped by the pump 12 to flow. This value can be 100% open or may be less depending on the design of the valve 29. In the example above, there is little change in the effect of the valve 29 between 60% and 100%, so the maximum valve position is set to 60%.

As more flow is needed, for example when more parallel loops are installed, the CDU pressure differential dP will drop since there are more parallel loops requiring flow. As the actual dP drops, the control scheme implemented by the controller at step 64 (e.g., a PID control loop) will increase the output in an attempt to return the dP value back to the dP setpoint. As the output increases, the bypass valve 29 position will begin to close, forcing less flow to bypass and more flow to leave the CDU. In the example shown above, this is achieved via a linear equation, however a non-linear equation could be used.

As more flow is required and the control algorithm continues to increase towards 100%, at some point the valve 29 will come to its minimum value. The minimum value can be set at 0% open, or in the case of the example above, it can be set at some value which is still partially open (e.g., 30% open). One reason for not fully closing the valve 29 is to always have some bypass flow to help buffer large changes in flow required. The point at which the bypass valve 29 has reached its minimum value can be referred to as the switch point. At the switch point, as more flow is demanded by the algorithm output, the pump 12 starts to speed up from idle speed. In the case of a single pump 12, it will ramp from idle speed to 100% of its rated flow as the algorithm output continues to climb up to 100%. In the case of two pumps, each pump is set to ramp from its idle speed to 50% speed at 100% algorithm output. The maximum speed with two running pumps is reduced to half so that the total flow is approximately the same as with one running pump. In the case of more than two pumps, a transfer equation can be implemented and tuned for the total number of pumps (e.g., there may be "n" transfer equations for "n" pumps), where the maximum speed of the pumps is limited to 100%/m, where m is the number of pumps running at the time.

One advantage to having two pumps is system redundancy. If one pump

(e.g., first) were to fail, the other (e.g., second) pump can turn on and/or change speed to maintain full cooling capability until the first pump can be serviced. One method for controlling redundant pumps is to have one pump on and the other in standby, where the standby pump is ready to turn on and provide the same flow as the first pump should the first pump fail. The system can cycle between the two pumps within a given time interval. Another method is to have both pumps running at a reduced speed and quickly ramp one pump up to a higher speed should the other pump fail. This has several advantages, which include reducing wear on both pumps by running them slower and having a higher degree of certainty that the backup pump is functioning correctly prior to the primary pump failing.

However, when two pumps are running at a reduced rate and one pump suddenly fails, the other pump needs to adjust very quickly to achieve the same CDU flow output and pressure differential. Since the control algorithm outputs a value to set the pump speed and bypass valve based on CDU dP, it is ideal for the control algorithm to output a value indeterminate of how many pumps are running. The "1 Pump" and "2 Pumps" curves as seen in Figs. 4 and 5 have been tuned so that any change from one pump to two pumps or two pumps to one pump along the algorithm output spectrum will result in negligible difference in CDU dP. This can be seen in Fig. 6, which shows that as more cooling loops are installed the PID output for a one-pump system and two-pump system are approximately the same. The switch points can be tuned by utilizing a computer simulation of the system or by testing. In the case of testing, the required PID output for a given set of pump speed and bypass valve constraints can be found over a broad range of flow demand from zero flow to max flow. A computer optimization routine can then be used to find the optimal switch point values such that the difference in PID output from 1 pump and 2 pump modes for the same flow demand is minimized.

Remaining at the same algorithm output regardless of whether one or two pumps are running is advantageous because the algorithm output does not need to suddenly change when the system switches between two pumps and one or vice versa, only the transfer functions from algorithm output to pump speed and bypass valve command change. This allows the control system to essentially have advance knowledge of how to adjust to a pump failure without disrupting cooling operation to the electrical devices being cooled. Minor differences in algorithm output can be adjusted via the algorithm control loop with minimal impact on flow output.

One further improvement on this flow control method is to incorporate a ramp rate in the dP setpoint any time it is changed, including when the system is first enabled. The rate, which can be hard-coded or user selectable, changes the setpoint at a fixed amount in a fixed time, for example one psid per second. This prevents large step changes in the setpoint, which can cause instability in the control loop. This also allows the system to gradually ramp up to the operating point from an initial value of 0 psid during an initial enable event. Automatic Cavitation Control

The second control methodology in accordance with the present disclosure is automatic cavitation control. Automatic cavitation control includes both detecting when cavitation occurs and reacting to the cavitation event in order to prevent damage to the pump 12 and return to an optimal pump speed as soon as possible.

In cavitation control mode the system overrides a pump speed value commanded via the pressure differential flow control loop (the first control loop) detailed above. Cavitation control mode will reduce the pump speed by a specified amount, for example 1 %, then wait for a specified amount of time, for example one second, and then repeat the loop by checking again for cavitation. To prevent running a pump too slow, the pump speed may be limited to idle speed, even if cavitation is still detected.

Once the system has sufficient subcool to be deemed free of cavitation, the control loop begins ramping the pump speed back up to its commanded speed determined via the pressure differential flow control loop. If at any point during the process of ramping the pump speed back to the correct value cavitation is detected, the system will enter back into cavitation control mode and the control loop activate again. Fig. 7 illustrates a flow chart 80 for implementing automatic cavitation control in accordance with the present disclosure.

The automatic cavitation control algorithm 80 illustrated in Fig. 7 detects when cavitation occurs by first acquiring the refrigerant pressure at the pump inlet, for example, via pressure transducer 26a located at the pump inlet as indicated at step 82. At step 84 this measured inlet pressure is converted to a corresponding saturation temperature based on a known refrigerant saturation curve. At step 86 the actual refrigerant temperature at the pump inlet is measured using, for example, pump inlet temperature sensor 24, and at step 88 a subcool in the refrigerant is determined based on a comparison of the refrigerant saturation temperature obtained at step 84 and the actual refrigerant inlet temperature obtained at step 86.

Pumps, based on their speed, need a different amount of subcool to operate free of cavitation. At step 90 the required amount of subcool for the actual pump speed (which may be obtained from speed sensor 28) is determined. The required subcool may be determined, for example, using a transfer function that calculates a required subcool based on actual pump speed. The transfer equation can be a linear equation where the required subcool is determined based on pump speed and a minimum required subcool (slope and intercept) based on the design of the pump. A non-linear equation can also be used, as well as a more complex transfer function that uses additional parameters such as the pump inlet temperature.

At step 92 the required subcool value as determined in step 90 is compared to the actual subcool value as determined at step 88. If the actual subcool value is not greater than the required subcool value a message is output at step 94 to provide indication that the system is in cavitation control mode. The method then moves to step 96 where the actual pump speed as determined from the speed sensor 28 (or a pump speed setpoint) is compared to a preset speed, e.g., idle speed. If the actual pump speed (or pump speed setpoint) is greater than the preset idle value then at step 98 a command is provided to a pump speed controller (not shown) to reduce the pump speed, e.g., reduce the speed by a predetermined percentage, step, etc. A purpose of step 98 is to bring the pump speed down to a level that requires a subcool less than the actual subcool at the pump inlet (thereby eliminating cavitation). The method then moves to step 106 where a time delay is introduced and then the method moves back to step 82 and repeats. Moving back to step 96, if the pump speed (or pump speed setpoint) is not greater than the preset speed (minimum) then the pump speed cannot be further reduced and the method moves to step 106 as described above.

Moving back to step 92, if the actual subcool value is greater than the required subcool value, then cavitation is not present and the method moves to step 100 to determine if the pump speed had previously been reduced due to cavitation control. If the pump speed had previously been reduced, the method moves to step 102 where the pump speed is increased by a predetermined percentage or step, and the method moves to step 106 as discussed above. If at step 100 the pump speed had not previously been reduced (or is at the desired setpoint speed), the method moves to step 104 where cavitation mode is deactivated and the method moves to step 106.

Although not shown, a hysteresis can be built into step 92 such that the subcool must be above the threshold by a certain amount before the system can be deemed no longer cavitating.

It should be mentioned that other methods of cavitation detection can also be used, such as determining the NPSHa (actual net positive suction head) instead of subcool, or detecting the onset of cavitation via pressure fluctuations, increase in sound or vibration on or around the pump, change in power required by the pump, etc. These alternative methods would replace steps 82-92, however the rest of the logic would remain the same. Refrigerant temperature control

The third control methodology in accordance with the present disclosure regulates the temperature of the refrigerant flowing through the cooling system. More particularly, the refrigerant temperature (i.e., the two-phase return fluid to the CDU) is regulated by varying a water flow rate through the condenser 16 via the water control valve 40.

Referring to Fig. 8, a flow chart 150 illustrating exemplary steps for controlling the refrigerant temperature is illustrated. The control methodology begins at step 152 by obtaining a desired temperature setpoint for the refrigerant. Such temperature setpoint may be obtained, for example, via operator entry, calculated by the controller 22 based on ambient conditions, or stored (preset) in memory of the controller 22. Next at step 154 a temperature setpoint tolerance is obtained via operator input or by retrieving a preset value (e.g., 3 degrees C) from memory of the controller 22. At step 156, a high temperature setpoint and a low temperature setpoint are calculated based on the temperature setpoint obtained at step 152 and the temperature setpoint tolerance obtained at step 154. For example, the tolerance may be added to the temperature setpoint to obtain the high temperature setpoint, and the tolerance may be subtracted from the temperature setpoint value to obtain the low temperature setpoint.

Next at step 158, the temperature setpoint, high temperature setpoint and low temperature setpoint are converted to a corresponding pressure based on a known refrigerant saturation curve for the refrigerant used in the system. Pressure can be used interchangeably with temperature because the returning refrigerant is a two-phase mixture of liquid and gas, or is a liquid that is nearly two-phase in the case of no heat rejection to the refrigerant to cause boiling. At step 160 an equation (e.g., a linear equation) is derived from the low pressure setpoint and the high pressure setpoint such that the water valve 40 is at a low limit (e.g., 0% open) at the low saturation pressure and at a high limit (e.g., 100% open) at the high saturation pressure. Fig. 9 illustrates an exemplary linear equation for the water valve position based on the low and high saturation pressure setpoints (which correspond to the low and high setpoints).

Next at step 162 the actual pressure at the two-phase CDU return is acquired using, for example, pressure sensor 27. Based on the actual return line pressure as obtained at step 162 and the equation derived at step 160, the controller 22 determines a position for the water valve 40 and commands the valve to the desired position as indicated at steps 164 and 166. The valve 40 then moves to this position, and at step 168 a delay is introduced to allow the refrigerant temperature to change. A hysteresis can be added to prevent continuous movement of the water valve 40, which could lead to premature wear. At step 170 it is determined if a new temperature setpoint has been entered into the system (e.g., an operator changed the setpoint). If a new temperature setpoint has not been entered, the method moves to step 162 and repeats. However, if a new temperature setpoint has been entered, the method moves back to step 152 and repeats.

Although described above with respect to a water-cooled CDU utilizing a proportional water valve, the same logic can be applied to a CDU in which the heat is rejected by an air-cooled condenser with variable speed fans. In this regard, the speed of the fans can be adjusted based on the required amount of cooling instead of the position of the water valve. One advantage of the control method of Fig. 8 is that it is virtually instantaneous, with the only lag time being the water valve speed. Because this method reacts very quickly to changes, there is little to no overshoot of the refrigerant temperature. Just as the system reacts quickly to sudden heat inputs, it also reacts very quickly to a sudden reduction in heat load. By quickly reducing water flow when the heat load is reduced, the system is better able to prevent over cooling of the refrigerant (which can lead to pump cavitation). The control loop uses pressure instead of temperature because the fluid pressure reacts much quicker to changes in the system compared to the fluid temperature, which has thermal mass and can be slow to react to changes.

One improvement on this control methodology, similarly to the automatic flow control method, is to add a ramp rate to the temperature setpoint such that the controlling temperature cannot change too quickly. The rate, which can be preset or user selectable, changes the setpoint at a fixed amount in a fixed time, for example 1 degree C per minute. This ensures that a change in the temperature setpoint does not cause a sudden change in the water valve, which could lead to system cavitation and loss of performance.

Dew point control

The fourth control methodology in accordance with the present disclosure automatically adjusts the refrigerant temperature setpoint mentioned above if the system detects the dew point of the room has become too close to the current refrigerant temperature. The flow chart shown in Figs. 10A and 10B provide exemplary steps 200a and 200b for carrying out dew point control in accordance with the present disclosure. Beginning at step 202, the ambient air temperature for the area is acquired using, for example, ambient air temperature sensor 30, and at step 204 the relative humidity of the area is obtained using ambient humidity sensor 32. Next at step 206 the ambient dew point for the area is calculated based on the ambient temperature and the ambient humidity as obtained at steps 202 and 204. In calculating the dew point, a margin, for example 2 degrees C, may be added to the dew point calculated at step 208 to provide dew point threshold setpoint that has a built-in a safety margin between the dew point of the room and the temperature of the refrigerant. The margin may be a preset value stored in memory of the controller or it may be based on user entry.

Next at step 210 it is determined if a dew point fault condition is already active. If a dew point fault condition is not active, the method moves to step 212 and obtains the refrigerant temperature at the pump inlet, which may be obtained via temperature sensor 24. At step 214 the dew point temperature threshold as obtained in step 208 is compared to the refrigerant temperature obtained at step 212. If the refrigerant temperature at the pump inlet is not less than the dew point temperature threshold the method moves back to step 202. However, if the refrigerant temperature at the pump inlet is less than the dew point temperature threshold, the method moves to step 216 and activates a dew point fault condition, and at step 218 the current status of automatic dew point control is obtained. Dew point control status can be obtained, for example, by reading an operator input (e.g. a selector switch or the like) or provided by a supervisory control system.

At step 220 it is determined if dew point control is enabled or disabled based on the dew point control status obtained at step 218. If automatic dew point control is not enabled, the method moves back to step 202. Disabling automatic dew point control may be desirable when the cooling system needs to maintain a set temperature and the user does not want the system to deviate from the temperature, even if dew point is detected.

If automatic dew point control is enabled, the method moves to step 222 where a refrigerant temperature setpoint is obtained (e.g., from operator input) and at step 224 a temporary refrigerant temperature setpoint is generated by adding a prescribed value, e.g., 0.1 degrees C, to refrigerant temperature setpoint using a prescribed ramp rate (e.g., 1 degree C per minute). Next at step 226 a delay is introduced to allow the temporary refrigerant temperature setpoint to ramp to the new value and/or to allow the system to react to the new setpoint, and then the method moves back to step 202.

Moving back to step 210, if a dew point fault is active the method moves to step 228 and obtains the current status of automatic dew point control (e.g., as described above with respect to step 218). At step 230 the controller 22 determines if automatic dew point control is enabled or disabled. If automatic dew point control is disabled, the method moves to step 232 and obtains the refrigerant temperature at the pump inlet via temperature sensor 24, and at step 232 the refrigerant temperature at the pump inlet is compared to the dew point temperature threshold as derived at step 208. If the refrigerant temperature at the pump inlet is less than the dew point temperature threshold the method moves back to step 202. However, if the refrigerant temperature at the pump inlet is not less than the dew point temperature threshold, the method moves to step 236 and deactivates the dew point fault condition, and then moves back to step 202.

Moving back to step 230, if dew point control is enabled the method moves to step 238 and obtains the refrigerant temperature at the pump inlet (e.g., via temperature sensor 24), and at step 240 the refrigerant temperature at the pump inlet is compared to the dew point temperature threshold as derived at step 208. If the refrigerant temperature at the pump inlet is less than the dew point temperature threshold the method moves to step 242 where the temporary refrigerant temperature setpoint is increased by a prescribed value, e.g., 0.1 degrees C, at a prescribed ramp rate (e.g., 1 degree C per minute). Next at step 244 a delay is introduced to allow the temporary temperature setpoint to ramp to the new value, and then the method moves back to step 202.

Moving back to step 240, if the refrigerant temperature at the pump inlet is not less than the dew point temperature threshold, the method moves to step 246 where the refrigerant temperature setpoint is obtained, and at step 248 the refrigerant temperature setpoint is compared to the temporary refrigerant temperature septoint. If the temporary refrigerant temperature setpoint is not greater than the refrigerant temperature setpoint, the method moves to step 236 where the dew point fault condition is deactivated and then the method moves back to step 202. However, if the temporary refrigerant temperature setpoint is greater than the refrigerant temperature setpoint, the method moves to step 250 where the refrigerant temperature at the pump inlet, which may be determined from temperature sensor 24, is compared to the dew point temperature threshold. If the refrigerant temperature at the pump inlet is equal to the dew point temperature threshold, the method moves back to step 202. However, if the refrigerant temperature at the pump inlet is not equal to the dew point temperature threshold, then the method moves to step 252 where the temporary refrigerant temperature setpoint is decreased by a prescribed value (e.g., 0.1 degrees C) at a prescribed ramp rate (e.g., 1 degree C per minute). The method then moves to step 254 where a delay is introduced to allow the temporary temperature setpoint to ramp to the desired value and/or the system to adjust to the new setpoint, and then the method moves back to step 202.

By using one or more of the four control methodologies described above, the CDU control system provides a highly stable, customizable control system that is very robust to changing flow requirements, heat loads and external factors such as changing ambient conditions.

The scope of the present invention is not limited to only the control methodologies described above, but also includes other variations of the above mentioned control methods that have not been specifically stated above, but would be obvious to someone knowledgeable in the art.

Although the invention has been shown and described with respect to a certain embodiment or embodiments, it is obvious that equivalent alterations and modifications will occur to others skilled in the art upon the reading and understanding of this specification and the annexed drawings. In particular regard to the various functions performed by the above described elements (components, assemblies, devices, compositions, etc.), the terms (including a reference to a "means") used to describe such elements are intended to correspond, unless otherwise indicated, to any element which performs the specified function of the described element (i.e., that is functionally equivalent), even though not structurally equivalent to the disclosed structure which performs the function in the herein illustrated exemplary embodiment or embodiments of the invention. In addition, while a particular feature of the invention may have been described above with respect to only one or more of several illustrated embodiments, such feature may be combined with one or more other features of the other embodiments, as may be desired and advantageous for any given or particular application.