Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
REFRIGERATION SYSTEM AND METHOD FOR CONTROLLING A REFRIGERATION SYSTEM
Document Type and Number:
WIPO Patent Application WO/2010/086806
Kind Code:
A2
Abstract:
A refrigeration system comprises a compressor (COMP), a condenser (COND) and at least one multi-stage evaporator (MSE) with at least one first evaporator (FBE) as a first stage and at least one expansion device evaporator (EDE) as at least one further stage downstream the corresponding first stage. Each first evaporator (FBE) is arranged at a respective first compo¬ nent (Cl) for absorbing heat dissipated by the respective first component (Cl). Each expan¬ sion device evaporator (EDE) is arranged at a respective at least one further component (FC) for absorbing heat dissipated by the respective further component (FC). A flow of a working fluid (WF) of the refrigeration system is controlled such that the working fluid (WF) is in a liquid state when entering the first stage and is in a vapor state when leaving a last stage of the respective multi-stage evaporator (MSE).

Inventors:
MICHEL BRUNO (CH)
SMITH BRIAN (US)
Application Number:
PCT/IB2010/050371
Publication Date:
August 05, 2010
Filing Date:
January 27, 2010
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
IBM (US)
MICHEL BRUNO (CH)
SMITH BRIAN (US)
International Classes:
F25B5/04; F25B49/02
Foreign References:
US20020157821A12002-10-31
EP1202003A22002-05-02
US5819554A1998-10-13
US3992898A1976-11-23
US4655051A1987-04-07
Other References:
BRUNO AGOSTINI; MATTEO FABBRI; JUNG E. PARK; LESZEK WOJTAN; JOHN R. THOME; BRUNO MICHEL: "Heat Transfer Engineering", vol. 28, 2007, TAYLOR AND FRANCIS, article "State of the Art of High Heat Flux Cooling Technologies", pages: 258 - 281
Attorney, Agent or Firm:
MEYER, Michael et al. (IBM Research - ZurichIntellectual Property La, Saeumerstrasse 4 Rueschlikon, CH)
Download PDF:
Claims:
CLAIMS

1. Method for controlling a refrigeration system comprising a compressor (COMP), a condenser (COND) and at least one multi-stage evaporator (MSE) with at least one first evapora- tor (FBE) as a first stage, each first evaporator (FBE) being arranged at a respective first component (Cl) for absorbing heat dissipated by the respective first component (Cl), and at least one expansion device evaporator (EDE) as at least one further stage downstream the corresponding first stage, each expansion device evaporator (EDE) being arranged at a respective at least one further component (FC) for absorbing heat dissipated by the respective further component (FC), wherein a flow of a working fluid (WF) of the refrigeration system is controlled such that the working fluid (WF) is in a liquid state when entering the first stage and is in a vapor state when leaving a last stage of the respective multi-stage evaporator (MSE).

2. Method according to claim 1, wherein the flow of the working fluid (WF) is controlled such that the working fluid (WF) is in vapor state when entering the last stage of the respective multi-stage evaporator (MSE).

3. Method according to any one of the preceding claims, wherein the flow of the working fluid (WF) is controlled such that 40 to 60 percent of a mass of the working fluid (WF) leav- ing the first stage is in vapor state.

4. Method according to any one of the preceding claims, wherein at least one temperature and/or pressure of the working fluid (WF) is determined and the flow of the working fluid (WF) is controlled dependent on the at least one temperature and/or pressure.

5. Method according to claim 4, wherein the at least one temperature and/or pressure of the working fluid (WF) is determined for the working fluid (WF) entering one of the stages or leaving one of the stages of the respective multi-stage evaporator (MSE).

6. Method according to any one of the preceding claims, wherein

- a first temperature (TEMPI) is determined of the working fluid (WF) at a first point entering one of the at least one further stages and wherein a second temperature (TEMP2) is deter- mined of the working fluid (WF) at a second point downstream the first point leaving one of the at least one further stages and

- the flow of the working fluid (WF) is controlled dependent on the first and the second temperature (TEMPI, TEMP2) such that the second temperature (TEMP2) exceeds the first tem- perature (TEMPI) by at least a predetermined deviation value (DVAL).

7. Method according to claim 6, wherein the first and second temperature (TEMPI, TEMP2) are determined at the last stage or a stage before the last stage of the respective multi-stage evaporator (MSE).

8. Method according to any one of claims 6 or 7, wherein the predetermined deviation value (DVAL) amounts to at least 0.1 degrees Celsius.

9. Refrigeration system comprising a control device (CD), a compressor (COMP), a con- denser (COND) and at least one multi-stage evaporator (MSE) with at least one first evaporator (FBE) as a first stage, each first evaporator (FBE) being adapted for being arranged at a respective first component (Cl) for absorbing heat dissipated by the respective first component (Cl), and at least one expansion device evaporator (EDE) as at least one further stage downstream the corresponding first stage, each expansion device evaporator (EDE) being adapted for being arranged at a respective at least one further component (FC) for absorbing heat dissipated by the respective further component (FC), wherein the control device (CD) is adapted to control a flow of a working fluid (WF) of the refrigeration system such that the working fluid (WF) is in a liquid state when entering the first stage and is in a vapor state when leaving a last stage of the respective multi-stage evaporator (MSE).

10. Refrigeration system according to claim 9, wherein the at least one first evaporator (FBE) is a micro-channel flow boiler evaporator (MCE).

11. Refrigeration system according to any one of claims 9 to 10, wherein the respective first component (Cl) is a microprocessor (CPU).

12. Refrigeration system according to any one of claims 9 to 11, wherein the respective at least one further component comprises a DC-DC converter (DDC) and/or a memory (MEM) and/or a further electronic component (FEC) dissipating heat during operation.

13. Refrigeration system according to claims 11 and 12, wherein in an order of a designated direction of flow of the working fluid (WF) the first component (Cl) is the microprocessor (CPU), a second component (C2) is the DC-DC converter (DDC), a third component (C3) is the memory (MEM) and a fourth component (C4) is the at least one further electronic component (FEC).

14. Refrigeration system according to claims 11 and 12, wherein in an order of a designated direction of flow of the working fluid (WF) the first component (Cl) is the microprocessor (CPU), a second component (C2) is the memory (MEM), a third component (C3) is the DC- DC converter (DDC) and a fourth component (C4) is the at least one further electronic com- ponent (FEC).

15. Refrigeration system according to any one of claims 9 to 14, wherein

- a respective control valve (CV) is arranged hydraulically between the condenser (COND) and each multi-stage evaporator (MSE), - the working fluid (WF) is supplied from the condenser (COND) to each control valve (CV) and

- the working fluid (WF) is supplied from each multi-stage evaporator (MSE) to the compressor (COMP).

Description:
REFRIGERATION SYSTEM AND METHOD FOR CONTROLLING A REFRIGERATION SYSTEM

TECHNICAL FIELD

The present invention relates to a refrigeration system and a method for controlling the refrigeration system. The present invention particularly relates to a refrigeration system for cooling electronic components.

BACKGROUND OF THE INVENTION

In Bruno Agostini, Matteo Fabbri, Jung E. Park, Leszek Wojtan, John R. Thome and Bruno Michel: "State of the Art of High Heat Flux Cooling Technologies", Heat Transfer Engineering, 28(4):258-281, Taylor and Francis, 2007 different cooling technologies for cooling high heat-flux computer chips are compared. Two-phase flow boiling in micro-channels is identified as a promising approach. A multi-micro-channel heat sink is used. Further, a refrigeration system for computer cooling is disclosed comprising a condenser, a filter/dryer, a thermostatic expansion valve, an evaporator, an accumulator, a dc rotary compressor and a hot-gas bypass valve.

It is a challenge to provide a refrigeration system and a method for controlling a refrigeration system that allow for achieving sufficient cooling with high efficiency.

SUMMARY OF THE INVENTION

According to one aspect of the invention, a method is provided for controlling a refrigeration system. Further, a corresponding refrigeration system is provided. The refrigeration system comprises a compressor, a condenser and at least one multi-stage evaporator. The at least one multi-stage evaporator respectively comprises at least one first evaporator as a first stage. Each first evaporator is adapted to being arranged at a respective first component for absorbing heat dissipated by the respective first component. The at least one multi-stage evaporator comprises at least one expansion device evaporator as at least one further stage downstream the corresponding first stage. Each expansion device evaporator is adapted to being arranged at a respective at least one further component for absorbing heat dissipated by the respective further component. The method comprises controlling a flow of a working fluid of the refrigeration system such that the working fluid is in a liquid state when entering the first stage and is in a vapor state when leaving a last stage of the respective multi-stage evaporator. The re- frigeration system further comprises a control device adapted to perform the method.

An advantage is that not only heat dissipated by the first component is absorbed but also heat dissipated by the at least one further component is absorbed. By providing both the first evaporator and the at least one expansion device evaporator for absorbing heat from the first and the further components and by controlling the flow of the working fluid such that the working fluid is turned from liquid to vapor during its passage through the multi-stage evaporator high heat flux components can be cooled effectively. Further, this allows for a high temperature level of the working fluid in the refrigeration system, for example 70 degrees Celsius or more when entering the condenser. By this, the absorbed heat can be recycled and can be used for other purposes, for example for district, community, or neighborhood heating. A further advantage is that, because two-phase flow boiling is used in the first stage of the multistage evaporator, the temperature of the working fluid may be rather high when entering the multi-stage evaporator. This allows for a small temperature lift from the working fluid leaving the condenser to the working fluid entering the condenser. The small temperature lift, conse- quently, allows for a small pumping power of the compressor. The refrigeration system can thus be energy efficient and can thus help to reduce carbon dioxide emissions.

The controlling of the working fluid may be performed, for example, by controlling a flow rate of the compressor and/or an inflow of working fluid in the multi-stage evaporator by con- trolling an opening of a control valve. The compressor preferably is designed as a variable flow compressor and the control valve is preferably designed as a proportional valve. In particular, the first and the further components are electronic components. The electronic components, for example, are arranged on a circuit board and dissipate heat during operation. For example, each circuit board has its own multi-stage evaporator. In particular, the first compo- nent is the component with the greatest heat flux of all components on the circuit board. Preferably, all components on the circuit board dissipating a significant amount of heat are considered to represent the at least one first or the at least one further component being cooled by the multi-stage evaporator and contributing to the evaporation or heating of the working fluid. In particular, the circuit board with the electronic components form part of a computer system. The refrigeration system is particularly suited for data centers with many such circuit boards and a high degree of heat dissipation. The heat dissipated by the electronic components may be collected with the multi-stage evaporators and may be recycled and may even be sold. The data center may thus be operated highly energy and cost efficient.

Preferably, the first evaporator is designed for flow boiling of the working fluid. Alternatively or preferably additionally, the first evaporator is designed as a micro-channel evaporator. This allows for effective cooling of the first component.

According to a preferred embodiment, the flow of the working fluid is controlled such that the working fluid is in vapor state when entering the last stage of the respective multi-stage evaporator. This helps to make sure that the working fluid is completely in vapor state when entering the compressor. This helps to achieve a good reliability of the refrigeration system, a long life, and a low energy consumption of the compressor. The compressor may thus be designed to work with vapor only and may thus be inexpensive and efficient. Expensive measures for preventing fluid from entering the compressor may not be necessary.

According to a further preferred embodiment, the flow of the working fluid is controlled such that 40 to 60 percent of a mass of the working fluid leaving the first stage is in vapor state. By this, the first component being a high heat flux component may be cooled very effectively. Further, by not evaporating all of the working fluid in the first stage, drying out of the first evaporator may be prevented and, as a result, overheating of the first component may be prevented. Further, the subsequent further component directly downstream of the first component may also be cooled effectively by a liquid- vapor-mixture of the working fluid. Preferably, the flow of the working fluid is controlled such that 50 percent of the mass of the working fluid leaving the first stage is in vapor state. However, a tolerance may be accepted, preferably up to ten percent of the mass of the working fluid leaving the first stage.

According to a further preferred embodiment, at least one temperature and/or pressure of the working fluid is determined and the flow of the working fluid is controlled dependent on the at least one temperature and/or pressure. The advantage is that this is simple. Temperatures and pressures can be determined easily, for example with a respective temperature sensor or pressure sensor. Preferably, at least two temperatures of the working fluid at different points of the refrigeration system are determined.

In this respect, it is advantageous if the at least one temperature and/or pressure of the work- ing fluid is determined for the working fluid entering one of the stages or leaving one of the stages of the respective multi-stage evaporator. By this, a reliable operation of the refrigeration system may be achieved with low cost.

According to a further preferred embodiment, a first temperature is determined of the working fluid at a first point entering one of the at least one further stages and a second temperature is determined of the working fluid at a second point downstream the first point leaving one of the at least one further stages. The flow of the working fluid is controlled dependent on the first and the second temperature such that the second temperature exceeds the first temperature by at least a predetermined deviation value. The advantage is, that by this it is possible to detect reliably if the working fluid is completely in vapor state. As long as there is liquid working fluid, the heat is absorbed by evaporating the liquid and the temperature of the working fluid drops slightly due to the reduced pressures in these further stages. After evaporation of all liquid working fluid heat absorption results in the rising of the temperature of the vapor working fluid, which can be detected easily.

In this respect, it is advantageous if the first and second temperatures are both determined at the last stage or a stage before the last stage of the respective multi-stage evaporator. By this it can reliably be made sure that no liquid working fluid leaves the multi-stage evaporator and that no liquid working fluid enters the compressor.

In this respect, it is further advantageous if the predetermined deviation value amounts to at least 0.1 degrees Celsius. This is a temperature rise which is sufficient for reliably detecting the all vapor state of the working fluid. Preferably, the predetermined deviation value amounts to at least 0.5 degrees Celsius. This is a temperature rise which is significantly greater than the uncertainty of the measurement. This allows for a particularly reliable detection of the all vapor state of the working fluid. According to a further preferred embodiment, the at least one first evaporator is a micro- channel flow boiler evaporator. This allows for particularly effective cooling of the first component, which may be a very high heat flux component. The first component may thus be prevented reliably from overheating.

According to a further preferred embodiment, the respective first component is a microprocessor. The microprocessor as a high heat flux component may be cooled effectively and reliably. A temperature of the microprocessor may reliably be kept below a temperature limit of, for example, 85 degrees Celsius.

According to a further preferred embodiment, the respective at least one further component comprises a DC-DC converter and/or a memory and/or a further electronic component dissipating heat during operation. DC-DC converters and memories typically dissipate a lot of heat. It is advantageous to use this heat for evaporating the liquid working fluid or raising the temperature of the vapor working fluid. This heat may then be recycled and used for other purposes. Because DC-DC converters have a higher allowed operation temperature and memories typically have a smaller heat flux than, for example, the microprocessor, the DC- DC converter and/or memory are considered as further components cooled by the expansion device evaporator, which may be less efficient than the first evaporator.

In this respect, it is advantageous if, in an order of a designated direction of flow of the working fluid, the first component is the microprocessor, a second component is the DC-DC converter, a third component is the memory and a fourth component is the at least one further electronic component. On a typical circuit board of a server this order may represent an order of heat flux of the respective components. The components with the greatest heat flux are cooled first, preferably with liquid working fluid or a mixture of liquid and vapor working fluid. Components with smaller heat flux may be cooled with vapor only working fluid. By this, most heat dissipated by the components may be absorbed and all components may be cooled sufficiently to allow for reliable operation.

Alternatively, it is advantageous if, in an order of a designated direction of flow of the working fluid, the first component is the microprocessor, the second component is the memory, the third component is the DC-DC converter and the fourth component is the at least one further electronic component. On a typical circuit board of a server this order may represent an order of heat flux and/or temperature sensitivity of the respective components. The components with the greatest heat flux and lowest required operation temperature are cooled first, preferably with liquid working fluid or a mixture of liquid and vapor working fluid. Components with smaller heat flux or higher operation temperatures may be cooled with vapor only working fluid. By this, most heat dissipated by the components may be absorbed and all components may be cooled sufficiently to allow for reliable operation.

In a further embodiment a flow path of the working fluid branches downstream the first com- ponent into at least two parallel flow paths. Each of these parallel flow paths comprise at least one expansion device evaporator adapted for being arranged at at least one of the further components. By this, most heat dissipated by the components may be absorbed and all components may be cooled sufficiently to allow for reliable operation. For this, however, the power dissipation ratio of the components in the parallel flow paths preferably is constant.

According to a further preferred embodiment, a respective control valve is arranged hydrauli- cally between the condenser and each multi-stage evaporator. The working fluid is supplied from the condenser to each control valve. The working fluid is supplied from each multi-stage evaporator to the compressor. The advantage is, that more than one multi-stage evaporator can be operated independently in parallel. By controlling the opening of the control valve, a flow of the working fluid through the respective multi-stage evaporator may be controlled individually. By this, a reliable operation of the refrigeration system and of all components may be achieved.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will be more fully appreciated by reference to the following detailed description of presently preferred but nonetheless illustrative embodiments in accordance with the present invention when taken in conjunction with the accompanying drawings.

The figures are illustrating:

FIG. 1, a refrigeration system and FIG. 2, a flow diagram for controlling the refrigeration system.

DETAILED DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a refrigeration system comprising a compressor COMP, a condenser COND and at least one multi-stage evaporator MSE. Preferably, a respective control valve CV is arranged hydraulically between the condenser COND and each multi-stage evaporator MSE.

In FIG. 1, only a single multi-stage evaporator MSE is shown. Preferably, more than one multi-stage evaporator MSE is provided. Such further multi-stage evaporators FMSE may be arranged in parallel to the multi-stage evaporator MSE shown in FIG. 1 and a flow of a working fluid WF through each of the multi-stage evaporator MSE and further multi-stage evaporators FMSE may be controlled individually by controlling an opening of the respective con- trol valve CV corresponding to the respective multi-stage evaporator MSE and further multistage evaporator FMSE. In the following, the refrigeration system is described only with respect to the multi-stage evaporator MSE shown in FIG. 1. However, the description does also apply accordingly to the further multi-stage evaporators FMSE. Working fluid WF is supplied from the condenser COND to each control valve CV and working fluid WF leaving each multi-stage evaporator MSE and further multi-stage evaporator FMSE is supplied to the compressor COMP. The compressor COMP pumps the working fluid WF into the condenser COND and by this raises the pressure and temperature of the working fluid WF supplied to the condenser COND.

The refrigeration system further comprises a control device CD and may further comprise at least one sensor, particularly at least one temperature sensor and/or pressure sensor. Preferably, the compressor COMP is designed as a variable speed compressor allowing for adjusting a flow rate of the working fluid in the refrigeration system. The at least one control valve CV preferably is designed as a proportional valve. The compressor COMP and/or the at least one control valve CV are coupled with the control device CD and may be controlled by the control device CD, that is, the flow rate of the compressor COMP and/or the opening of the respective control valve CV may be controlled by the control device CD. The multi-stage evaporator MSE comprises at least one first evaporator FBE as a first stage and at least one expansion device evaporator EDE as at least one further stage. The at least one first evaporator FBE preferably is designed as a micro-channel evaporator. Preferably, the first evaporator FBE is designed for flow boiling. The at least one first evaporator FBE thus is preferably designed as a micro-channel flow boiler evaporator MCE. The micro-channel flow boiler evaporator MCE preferably comprises multiple micro-channels for conducting the working fluid WF. Preferably, each micro-channel comprises a constriction and a flow deflection portion at its entrance to allow for a smooth and steady boiling without superheating and maldistribution of the working fluid WF in the respective micro-channel. The at least one ex- pansion device evaporator EDE preferably is designed as a single meandering tube. Preferably, a diameter of the tube increases in a designated direction of the flow of the working fluid WF. Preferably, at least two stages of the expansion device evaporator EDE are provided. In FIG. 1, three stages of the expansion device evaporator EDE are shown.

Preferably, each multi-stage evaporator MSE is used for absorbing heat dissipated by components and particularly electronic components of one circuit board. Each circuit board with its electronic components may form part of a computer system and particularly of a server. Each circuit board, for example, may be a processor board of the computer system or server. On the circuit board, a first component Cl is arranged. The first component Cl preferably is the component with the greatest heat flux of all components on the circuit board. The first component C 1 preferably is a microprocessor CPU. The first evaporator FBE is arranged on the first component Cl such that at least part of the heat dissipated by the first component Cl is absorbed by the first evaporator FBE when the working fluid WF flows through the first evaporator FBE. The respective expansion device evaporator EDE is arranged on at least one further component FC on the circuit board such that at least part of the heat dissipated by the respective further component FC is absorbed by the expansion device evaporator EDE when the working fluid WF flows through the expansion device evaporator EDE. The at least one further component FC preferably comprises a DC-DC converter DDC and/or a memory MEM and/or at least one further electronic component FEC on the circuit board dissipating heat during operation.

Preferably, as shown in FIG. 1, the at least one further component FC comprises a second component C2, a third component C3 and a fourth component C4 in the designated direction of flow of the working fluid WF. In a first preferred embodiment, the second component C2 is the DC-DC converter DDC, the third component C3 is the memory MEM and the fourth component C4 is the at least one further electronic component FEC. In a second preferred embodiment, which is not shown, the second component C2 is the memory MEM, the third component C3 is the DC-DC converter DDC and the fourth component C4 is the at least one further electronic component FEC. It may be necessary in this second embodiment to provide a heat spreader to the DC-DC converter DDC to reduce the heat flux and to be sufficiently cooled as the third component C3.

Preferably, a temperature one Tl and/or a temperature two T2 and/or a temperature three T3 and/or a temperature four T4 and/or a temperature five T5 of the working fluid WF are determined. Preferably, additionally or alternatively, a pressure one Pl and/or a pressure two P2 and/or a pressure three P3 and/or a pressure four P4 and/or a pressure five P5 of the working fluid WF are determined. Further, a vapor quality one ql and/or a vapor quality two q2 and/or a vapor quality three q3 and/or a vapor quality four q4 and/or a vapor quality five q5 of the working fluid WF may be determined. A vapor quality of zero represents an all liquid working fluid WF and a vapor quality of one represents an all vapor working fluid WF. Vapor quality values between zero and one represent a ratio of a mass of the working fluid WF in vapor state with respect to a total mass of the working fluid WF at the respective point in the refrigeration system. The determining of the temperatures, pressures or vapor qualities may be performed by measurement with an appropriate sensor or by calculation according to a model of the refrigeration system.

The temperature one Tl, the pressure one Pl and the vapor quality one ql correspond to a point of the refrigeration system with the working fluid WF entering the first evaporator FBE, that is hydraulically between the control valve CV and the corresponding first evaporator FBE. The temperature two T2, the pressure two P2 and the vapor quality two q2 correspond to a point of the refrigeration system with the working fluid WF leaving the first evaporator FBE, that is hydraulically between the first evaporator FBE and the expansion device evapo- rator EDE of the second component C2. The temperature three T3, the pressure three P3 and the vapor quality three q3 correspond to a point of the refrigeration system with the working fluid WF leaving the expansion device evaporator EDE of the second component C2, that is hydraulically between the expansion device evaporator EDE of the second component C2 and the expansion device evaporator EDE of the third component C3. The temperature four T4, the pressure four P4 and the vapor quality four q4 correspond to a point of the refrigeration system with the working fluid WF leaving the expansion device evaporator EDE of the third component C3, that is hydraulically between the expansion device evaporator EDE of the third component C3 and the expansion device evaporator EDE of the fourth component C4. The temperature five T5, the pressure five P5 and the vapor quality five q5 correspond to a point of the refrigeration system with the working fluid WF leaving the expansion device evaporator EDE of the fourth component C4, that is hydraulically between the expansion device evaporator EDE of the fourth component C4 and the compressor COMP.

FIG. 2 shows a flow diagram of a program for controlling the refrigeration system. The control device CD preferably is adapted to execute the program and to control the flow of the working fluid WF in the refrigeration system accordingly, for example by controlling the flow rate of the compressor COMP and/or by controlling the opening of the respective control valve CV. The program begins in a step Sl. In a step S2, at least one temperature and/or pressure of the working fluid WF is determined. Preferably, the temperature one Tl and/or the temperature two T2 and/or the temperature three T3 and/or the temperature four T4 and/or the temperature five T5 and/or the pressure one Pl and/or the pressure two P2 and/or the pressure three P3 and/or the pressure four P4 and/or the pressure five P5 of the working fluid WF are determined. In particular, a first temperature TEMPI and a second temperature TEMP2 are determined. Preferably, the first temperature TEMPI is the temperature three T3 or the temperature four T4. Preferably, the second temperature TEMP2 is the temperature four T4 if the first temperature TEMPI is the temperature three T3 or the second temperature TEMP2 is the temperature five T5. However, the first and second temperature TEMPI, TEMP2 may be de- fined differently. Preferably, the first and second temperature TEMPI, TEMP2 are determined by measurement.

A step S3 may be provided for determining the vapor quality one ql and/or the vapor quality two q2 and/or the vapor quality three q3 and/or the vapor quality four q4 and/or the vapor quality five q5 of the working fluid WF, preferably by calculation dependent on at least one of the temperatures and/or pressures determined in step S2. In a step S4 it is determined if the current flow of the working fluid WF is appropriate for sufficiently cooling the components. Further, an amount of energy required for driving the compressor COMP should be small to allow for an efficient operation of the refrigeration system. If the current flow of the working fluid WF is considered appropriate, the current flow rate is maintained in a step S5. The program then ends in a step S6 or, preferably, is continued in step S2. If the current flow of the working fluid WF is considered not appropriate, that is too high or too low, the flow is adapted appropriately in a step S7. The program then ends in step S6 or, preferably, is continued in step S2.

Preferably, the flow of the working fluid is controlled such that the vapor quality one ql equals zero or is approximately zero. This allows for effective cooling of high heat flux components such as the first component Cl. The flow of the working fluid is further controlled such that the vapor quality five q5 is always one and preferably also the vapor quality four q4 equals one or is approximately one. By this, liquid working fluid WF is prevented from enter- ing and destroying the compressor COMP. In particular, this can be achieved by controlling the flow of the working fluid WF such that the second temperature TEMP2 exceeds the first temperature TEMPI by at least a deviation value DVAL. The deviation value DVAL is greater than zero and preferably amounts to at least +0.1 degrees Celsius. Preferably, the deviation value DVAL amounts to at least +0.5 degrees Celsius. Further, preferably, the flow of the working fluid WF is controlled such that the vapor quality two q2 is in a range between 0.4 and 0.6. By this, drying out of the first evaporator FBE and overheating of the first component Cl may be prevented. For a larger vapor quality there is a risk for local dryout with the consequence of strongly reduced heat transfer efficiency. Since a high heat transfer coefficient has to be guaranteed with a large safety margin for sufficient cooling of the first component the vapor quality of the working fluid WF preferably is kept as low as possible. The vapor quality cannot be too low, however, since otherwise the subsequent further components cannot convert the rest of the fluid to vapor and guarantee that no fluid enters the compressor. The lowest vapor quality may for example be calculated by dividing the power of the processor by the total power of the electronic board and adding a safety margin of at least 10%.

Further, preferably, the flow of the working fluid WF is controlled such that a system pressure and/or a system temperature level of the working fluid WF in the refrigeration system is in a respective predetermined range. The system pressure and/or the system temperature level may be determined dependent on the temperature one Tl and/or the temperature two T2 and/or the temperature three T3 and/or the temperature four T4 and/or the temperature five T5 and/or the pressure one Pl and/or the pressure two P2 and/or the pressure three P3 and/or the pressure four P4 and/or the pressure five P5 as determined in step S2.

Typically, a small negative temperature gradient is created, for example about three degree Celsius, along the first evaporator FBE mainly due to a pressure drop of approximately 0.5 bar, that is for example T2 = Tl - 3° C and P2 = Pl - 0.5 bar. There are two choices for the DC-DC converter DDC cooling, as described above: either they are cooled with a compact cooler directly after the microprocessor CPU with a vapor quality lower than one or they are cooled after the memory MEM preferably with a vapor quality of one and a heat spreader to distribute the relative high power density of these chips over a greater surface. Both approaches are safe since DC-DC converter chips are much more temperature tolerant than microprocessor chips: they typically can survive temperatures up to 125 degrees Celsius com- pared to microprocessors CPU that typically have to be kept below 85 degrees Celsius. In the configuration shown in FIG. 1, the DC-DC converter cooler may, for example, increase the vapor quality from 0.6, that is q2 = 0.6, to 0.75, that is q3 = 0.75, with only a little pressure drop and temperature increase such that approximately T3 = T2 and P3 = P2.

When a memory cooler is added to this configuration another 30 percent of the working fluid WF may be evaporated although at a much lower areal heat flux. Such a heat flux can be handled by a cooler where a main heat exchange path is not solid to fluid but directly solid-togas. This cooler preferably first increases the vapor quality to one and then adds a small temperature gradient. This temperature gradient is small, however, since the memory cooler has the function of an expansion device with a larger pressure gradient due to the long tube-length and the small diameter. For example, T4 = T3 - 3° C and P4 = P3 - 0.3 bar.

The remainder of the further components FC, that is the further electronic components FEC, is then cooled with the working fluid vapor leading to a temperature increase and also allow- ing for a safety margin that guarantees that the output is under all load conditions 100 percent vapor and no liquid. This cooler preferably has the function of an expansion device with a large pressure gradient and a defined increase of temperature, for example T5 = T4 + 0.5° C and P5 = P4 - 0.2 bar, which is required for the control loop that defines the mass flow of the working fluid WF through the refrigeration system.

An effective hydraulic radius of the expansion device evaporator EDE of the DC-DC con- verter DDC preferably is larger than that of the first evaporator FBE, particularly that of the micro-channel flow boiler evaporator MCE, to allow larger volume fluxes with the increased vapor quality. A power density of the memory MEM typically is smaller than that of the DC- DC converter DDC. The memory MEM may thus be cooled safely by direct heat transfer from solid to the working fluid vapor. An optimum geometry of the multi-stage evaporator MSE may be based on a balance between pressure drop in the fluid loop and the ability to transfer heat from components to the working fluid.

There is a feedback loop that controls the working fluid flux so that preferably the temperature increase in the cooler for the at least one further electronic component FEC is equal or larger than +0.1 degrees Celsius and that the temperature drop of the working fluid entering and leaving the expansion device evaporator EDE of the DC-DC converter DDC preferably is greater than zero but less than two degrees Celsius. This ensures that all refrigerant is evaporated and also that the vapor quality at the exit of the DC-DC converter cooler is still below one. With the other condition for the feedback loop being a small temperature drop in the memory cooler a vapor quality two q2 of less than 0.6 can be guaranteed. The flow control is done preferably by a proportional valve, that is the control valve CV, hydraulically arranged upstream of each multi-stage evaporator MSE. Along the sequence of coolers is a continuous reduction of pressure. This reduces the boiling temperature in the memory cooler and the following coolers and ensures that there is no overheating despite the needed gradient in the last component to carry out the feedback.

Having a controlled inflow into each circuit board allows operation of more than one multistage evaporator MSE using the same compressor COMP and the same condenser COND for all circuit boards of the computer system or server is possible. To save energy in the compres- sor COMP the control of the flux of the working fluid WF preferably is accomplished by the variable speed compressor COMP. The speed of the compressor COMP preferably is controlled by a setup that tries to control the control valves CV to a desired average opening value of about 70 percent and to keep the opening value of the maximally open control valve CV below 95 percent.

The invention is used to collect most energy dissipated from a typical processor board and to deliver it via the condenser COND to a remote heating network for community heating with a temperature level preferably between 60 and 70 degrees Celsius. The temperature of the working fluid entering the multi-stage evaporator MSE preferably is maintained at a quite high level, for example at 40-50° C, so that the required compressor energy is kept small. One bar pressure drop along all components with a system pressure of 4 bar, for example, leads to a temperature lift of 68 degrees Celsius while the temperature lift is only 34 degrees Celsius at a system pressure of 8 bar. The pumping energy in the latter case is only half since the mass flow is twofold smaller. Higher system pressures are advantageous since mass flow and compressor power are smaller. The overall pressure drop preferably is minimized so that not too much energy is needed for the compressor COMP.

Preferably, during shutdown all liquid working fluid WF is pumped into the condenser COND and the control valves CV are kept completely closed. This ensures that in the whole loop the vapor quality is one. This ensures that during startup the compressor COMP has vapor at its input. To ensure that this condition remains during the entire startup phase the control valves CV preferably are carefully opened until the control loop can take over and the vapor qualities at each control point are as specified.