HK1023176B - Fuzzy logic liquid level control - Google Patents

Description

Fuzzy logic level control

The present invention is based on U.S. provisional application serial No. 60/015,347 entitled "fuzzy logic level control," 1996, 4/16/1996, which is incorporated herein by reference.

Technical Field

The present invention relates generally to controlling a mechanical liquid refrigeration chiller. In particular, the present invention relates to a fuzzy logic level control for a mechanical liquid refrigeration chiller to control the level of liquid coolant in the condenser to prevent gas from escaping from the condenser.

Background

Coolant liquids are commonly used in mechanical liquid refrigeration chillers. It is desirable to control the flow of the coolant to achieve optimal and efficient operation.

At least some existing systems have used liquid surface sensors to measure the level of coolant. No one is known to the applicant to be able to accurately and optimally control the level of coolant flowing in a coolant chiller system by using a level sensor and fuzzy logic using the measured value as a variable input.

Disclosure of Invention

The present invention provides a cooling system comprising: a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates; a coolant liquid in the condenser; a sensor for measuring the level of liquid coolant in the condenser and providing a measurement signal; said expansion device capable of selectively varying the amount of restriction it applies to the coolant as it flows through the apparatus in response to a control signal; and a microprocessor control means which samples the measurement signal at selected time intervals and applies said sampled measurement signal in a programmable predetermined manner to generate a control signal for the expansion device, wherein said control signal positions the expansion device to control the flow of gaseous coolant to said vaporizer.

The present invention also provides a cooling system comprising: a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates; coolant liquid in the condenser; a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal; said expansion device being disposed between the evaporator and the condenser and being capable of selectively varying the magnitude of its restriction applied to the flow of coolant between the condenser and the evaporator in response to a control signal; and a microprocessor control means receiving the measurement signal from the sensor and generating a control signal to be supplied to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm having a programmable fuzzy logic relationship function to control the flow of gaseous refrigerant to said evaporator.

The present invention also provides a cooling system comprising: a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates; coolant liquid in the condenser; a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal; said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal; and a control means receiving the measurement signal from the sensor and generating said control signal to be supplied to said expansion device to control the flow of refrigerant from the condenser to the vaporizer, wherein the control utilizes a fuzzy logic algorithm to control the flow of gaseous refrigerant to said vaporizer, wherein said control means includes a microprocessor which generates said control signal to selectively open and close said valve to cause the refrigerant level in the condenser to reach a preselected set level, wherein said microprocessor opens the valve relative to the previous position of the valve when the level of refrigerant liquid measured in the condenser is above the preselected set level.

The present invention also provides a cooling system comprising: a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates; coolant liquid in the condenser; a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal; said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal; and a control means receiving the measurement signal from the sensor and generating said control signal to said expansion device to control the flow of refrigerant from the condenser to the vaporizer, wherein the control utilizes a fuzzy logic algorithm to control the flow of gaseous refrigerant to said vaporizer, wherein said control means includes a microprocessor generating said control signal to selectively open and close said valve to cause the refrigerant level in the condenser to reach a preselected set level, wherein said microprocessor closes the valve relative to the previous position of the valve when the refrigerant level measured in the condenser is below the preselected set level.

The present invention also provides a cooling system comprising: a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates; coolant liquid in the condenser; a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal; said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal; and a control means receiving the measurement signal from the sensor and generating said control signal to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm to control the flow of gaseous refrigerant to said evaporator, wherein said control comprises a microprocessor generating said control signal to selectively open and close said valve to cause the refrigerant level in the condenser to reach a preselected set level, wherein said microprocessor periodically samples the level measured by said level sensor, and wherein the microprocessor executes a fuzzy logic algorithm once per sampling period to calculate a level difference equal to the difference between the measured refrigerant liquid level and the preselected set level and applies the level difference to a fuzzy logic algorithm, wherein the microprocessor calculates a rate of change of coolant level by subtracting the coolant level measured at the previous sampling period from the coolant level measured at the current sampling period and applies the rate of change of level to a fuzzy logic algorithm.

The present invention also provides a cooling system comprising: a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates; coolant liquid in the condenser; a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal; said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal and acts as an expansion device; and a control means receiving the measurement signal from the sensor and generating said control signal to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm to minimize or eliminate the flow of gaseous refrigerant to said evaporator, wherein said control comprises a microprocessor generating said control signal to selectively open and close said valve to force the refrigerant level in the condenser to a preselected set level, wherein said microprocessor utilizes a fuzzy logic algorithm to apply a fuzzy min/max method that first performs a "fuzzy and" (min) decision and then a "fuzzy or" (max) decision to derive the closing and opening components.

The present invention also provides a cooling system comprising: a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates; coolant liquid in the condenser; a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal; said expansion device being disposed between the evaporator and the condenser and being capable of selectively varying the magnitude of its restriction applied to the flow of coolant between the condenser and the evaporator in response to a control signal; and a control means receiving the measurement signal from the sensor and generating said control signal to be supplied to said expansion device to control the flow of coolant from the condenser to the vaporizer, wherein the control uses a fuzzy logic algorithm to control the flow of gaseous coolant to said vaporizer, wherein said fuzzy logic algorithm uses as input variables the value of the input variable of the coolant measured by said sensor at a given time and the rate of change of that value at a given time relative to the previous detection time.

Systems and methods consistent with the present invention include a cooling system comprising a vaporizer, a compressor, a condenser, and an expansion device (e.g., a valve), all connected in a closed cooling loop. It is known that coolant flows through this system and often collects in the condenser and evaporator. In the present invention, a level sensor is placed in the condenser to measure the level of the coolant. An expansion device or chamber is disposed between the evaporator and the condenser and includes a variable flow valve disposed therein. One such valve is a butterfly valve that can be selectively opened or closed in sequential steps by a motor, coil or similar actuator connected to or part of the valve. The valve is controlled by a microprocessor via an actuator, which receives one or more output signals from the level sensor and controls the position of the valve to vary the restriction to the coolant to control the flow of the coolant. Preferably, the microprocessor uses a fuzzy logic algorithm based on the level of liquid measured in the condenser and its rate of change to determine the desired position of the valve.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate an embodiment of the invention and together with the description, serve to explain the principles of the invention.

In the drawings:

FIG. 1 is an overall system diagram of a mechanical liquid refrigeration chiller;

FIGS. 2(a) and 2(b) show an example of a fuzzy logic relationship function with level difference (difference between measured level and expected optimal level) and rate of change of level as inputs;

FIG. 3 illustrates an example of a fuzzy logic truth table in accordance with the present invention;

FIG. 4 illustrates a fuzzy logic truth table in accordance with a particular embodiment of the present invention; and

fig. 5A-5E show flow diagrams operated by a microprocessor according to an embodiment of the invention.

Detailed Description

The present invention was developed as a control for a centrifugal chiller system. It will be appreciated that the invention may have other uses.

Reference will now be made in detail to the present preferred embodiments of the invention, an example of which is illustrated in the accompanying drawings. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts.

The use of the liquid level control system of the present invention in a cooling system is illustrated in FIG. 1. The cooling system includes a condenser 100, a vaporizer 110, a centrifugal gas compressor 120 and an expansion device or chamber 130 having a valve 140, all interconnected in a conventional closed cooling circuit. The restriction device is preferably a valve having a plurality of positions, such as a butterfly valve, which varies the resistance to fluid and thus the flow of coolant. Such butterfly valves are readily available in manufacture, such as the D-200 type butterfly valve manufactured and supplied by Norriseal corporation. Valve controller 150 opens and closes valve 140 (relative to its previous position) based on signals received from a controller, such as a microprocessor 160. The valve controller, as is known and available, may be a motor, coil or similar actuator. An example of such an actuator is the Barber Colman Mp-481 or 487 damper controller, which is a readily available motor that rotates whenever there is a pulse (an acceptable ac signal applied for a given period of time). Other types of pulses, such as a dc signal, may be used, as long as it is compatible with the type of actuator used. In a preferred embodiment of the invention, the microprocessor sends a signal in the form of a pulse that causes the actuator to open and close in a direction proportional to the pulse it receives. In another embodiment, other signals and actuators can be used as long as the valve opens and closes in proportion to the received signal.

A coolant sensor 170 is used to detect the amount of liquid coolant in the condenser. The sensor is preferably a level sensor located in the condenser 100. The sensor outputs a signal to microprocessor 160. One preferred example of a sensor is the SHP and SVP level sensor probes, manufactured and supplied by Hansen Technologies Corporation. The sensor is inserted directly into the coolant reservoir, here the condenser, and provides continuous and accurate coolant level measurement. The sensor is a capacitive level sensor. Because liquid coolants inherently have much greater potential for storing electrical energy than gaseous coolants, the capacitance value varies almost proportionally with the level of the coolant liquid.

In operation, the vaporous refrigerant is compressed in compressor 120 and sent to condenser 100 where it encounters a cooling medium, such as water 180 from a cooling tower (not shown), and is condensed into a liquid refrigerant. The liquid coolant flows to an expansion device or chamber 130 to relieve pressure on the vaporizer 110. The coolant expands as it flows through the expansion chamber 130 to the vaporizer 110. As the coolant flows through the vaporizer 110, circulating water 190 from a building's air conditioning unit exchanges heat with it to be cooled. The refrigerant vaporizes and returns to the suction port of the compressor 120. In this way, water circulating in the air cooling unit of the entire building is cooled in the vaporizer 110. In order to vary the cooling capacity of the building as the cooling demand or load varies, the duty of the compressor 120 is controlled, as is known.

The flow of a liquid or gas through an obstruction, such as a restriction (e.g., expansion chamber 130, valve 140), is dependent on the pressure upstream and downstream of the obstruction and the amount of resistance due to the geometry of the obstruction. Thus, the flow of refrigerant liquid through the expansion device is dependent upon the pressures in the condenser 100 and the vaporizer 110, as well as the geometry and position of the valve 140. By adjusting the position of the valve, the amount of resistance to coolant liquid flow is varied. The control system positions the valve 140 to match the resistance to flow of liquid in the expansion chamber 130 with the desired resistance to minimize, if not eliminate, the amount of vaporous cryogen reaching the vaporizer. In the described embodiment, this effect is achieved by controlling the valve in a manner that seeks to maintain the refrigerant liquid level in the condenser at a preset level. The microprocessor 160 provides two potentially related signals, on and off, preferably based on the output of a fuzzy logic algorithm, to drive the valve actuator 150. The output of the microprocessor causes the valve to open a certain amount relative to its previous position, close a certain amount relative to its previous position, or remain unchanged for any given period of time. As set forth below, the microprocessor makes this determination based on the measured liquid level and the rate of change of the liquid level, preferably according to a fuzzy logic algorithm.

The level sensor 170 measures the level of the coolant liquid in the condenser 100. The sensor is preferably a capacitive level sensor. A preferred sensor comprises two electrodes separated by a coolant liquid. The coolant liquid causes the capacitance of the sensor to change almost proportionally to the level of the coolant liquid. The sensor outputs a voltage that is responsive to the level of the measured liquid. The microprocessor 160 can use the output of the sensor to determine the level of the liquid coolant and a level difference (difference between the actual level and the optimal preselected level), as well as the rate of change of the level.

If the valve 140 is closed too quickly, the liquid in the condenser 100 rises, causing the possibility of an under-boil, and if the valve 140 is opened too quickly, the liquid in the condenser 100 will fall, causing the gas to flow back to the vaporizer 110. When gas recirculation occurs, the compressor 120 must do more work and heat to maintain the flow of gas. This reduces the operating efficiency of the overall freezer. Thus, the microprocessor 160 is programmed with the desired level set point.

The microprocessor 160 stores and uses a fuzzy logic algorithm to control the valve 140. The algorithm determines whether to leave the valve open, closed, or remain the same as desired to achieve the desired level set point. The appropriate level set point is a task of the chiller design and is selected to prevent backflow of gaseous coolant to the vaporizer. The proper coolant level for a given chiller may be best determined by actual experimental observation and testing of the chiller. When gas backflow occurs, the compressor must operate to maintain the flow of the gas. However, no additional cooling capacity is obtained. This reduces the overall operating efficiency of the freezer. The operation efficiency of the freezer is optimized when the valve is placed in a position to prevent gas backflow by maintaining a set liquid level in the condenser.

The fuzzy logic algorithm controls the desired valve position by periodically sampling the actual liquid level measured by the sensor 170 over a preprogrammed period of time. For example, the programmable time period may range from 1 second to 5 seconds. By sampling the output of the sensor 170 and comparing the measured value to a preselected desired level value, and one or more previously stored samples of the measured parameter, the microprocessor can calculate a level difference (lvl _ error) and a level rate of change (lvl _ rate) in accordance with known computer techniques. The ultimate goal of such fuzzy logic algorithms is to bring the liquid level difference close to zero so that little or no gaseous coolant flows into the vaporizer and there is no vaporizer starvation, thereby optimizing chiller operating efficiency. The fuzzy logic algorithm of the microprocessor 160 determines the proportion of the negative, positive and zero relationship associated with each input (the liquid level difference and its rate of change) by weighting each input by a weight between 0 and 100 during each sample period. The fuzzy logic algorithm then determines a few "if, then" rules that incorporate the proportions of the above relationships into the appropriate process taken by the control system.

Fuzzy logic algorithms use one level difference and the rate of change of the level difference as variable inputs. In a preferred embodiment, three relationship functions are defined for each of the two inputs. Each relationship function determines the proportion of zero, positive or negative, given input in a linear fashion. For example, in FIG. 2(a), a difference of 20% results in a relationship of 50% positive, 50% zero and 0% negative. Similarly, as shown in FIG. 2(b), a rate of change of the liquid level difference equal to-15% yields a relationship function of 60% negative, 40% zero and 0% positive per sample. The relationship functions shown in fig. 2(a) and 2(b) are asymmetric about zero when inputs of the same magnitude are considered, and do not reflect the same relationship proportions for negative and positive values.

In general, the relationship function may or may not be symmetric. These relationship functions are independently programmable and may be changed in the microprocessor 160. Thus, the sensitivity of the level difference and the rate of change of the level difference, both symmetrical and asymmetrical, can be adjusted to optimize system control as desired, with lower amplitude values resulting in higher sensitivity and higher amplitude values resulting in lower sensitivity. In one embodiment, the above-mentioned relationship function is selected to control the valve in the following manner: the valve opens faster to avoid starving the vaporizer of liquid and closes more slowly to prevent overshoot of the set point. In order to achieve flexibility in the regulation control, it is preferable to have a programmable relationship function. A user can thus change the relationship functions applied by the fuzzy logic algorithm described above. The object of the invention is to minimize the above-mentioned flow to an acceptable level to a practical extent.

The table shown in FIG. 3 represents a fuzzy logic truth table that shows how the microprocessor 160 determines fuzzy logic rules according to one embodiment of the present invention. As shown in the table, e N, e Z and e P represent negative, zero and positive liquid level differences, respectively, and d e N, d e Z and d e P represent negative, zero and positive liquid level difference rates of change, respectively. The set of rules (e N, d e N), (eZ, d e N) and (e N, d e Z) cause the pulse to be generated to close the valve 140 to increase the liquid level in the condenser 100, while the set of rules (e P, d e Z), (eP, d e P) and (e Z, d e P) cause the pulse to open the valve 140 to decrease the liquid level in the condenser 100. The remaining three rule sets are not assigned values because they do not generate actions. Thus, a total of six rule sets are judged using a fuzzy judgment minimum/maximum method. This approach means that a minimum value "fuzzy and" decision will be performed first for each of the six rule sets. Then, a "fuzzy or" decision is made wherein the maximum of the three rule sets that produce the valve actuator closing component and the three rule sets that produce the valve actuator opening component are derived, thereby producing two maximum values representing the derived opening and closing values.

The two maxima obtained above need to be combined into a single output determination, i.e. they need to be "defuzzified". The single-element set approach is preferred since the defuzzification centroid (centroid) approach is computationally more exhaustive than required by the present application. In this single element set approach, the maximum value of the valve actuator closing component is subtracted from the maximum value of the valve actuator opening component to yield a single output determination. If this result is less than zero, the output signal of the valve actuator will pulse for a period of time equal to one percent of the sampling time in order to close the valve. If this result is positive, the opening signal of the valve actuator will be pulsed for a period of time equal to one hundredth of the sampling time in order to open the valve. The pulses and time periods are selected based on empirical evidence of flow in the system, including valves and actuators. Finally, these values are best determined by practical empirical operation and testing of the refrigeration unit to which the present invention is to be applied. The result of the calculation may be any value in the range from-100 (valve actuator closing pulse width equal to the sampling period) to 100 (valve actuator opening pulse width equal to the sampling period). The on pulse and the off pulse never occur simultaneously.

The fuzzy decision of the present invention will be further clarified by the following example, which is merely exemplary of the present invention, which is illustrated in the truth table of fig. 4. As shown in this table, the quantities of relationship assigned to ∈ Z (50) and to d ∈ N (60) are combined by performing a minimum value fuzzy determination, i.e., a fuzzy and routine. The fuzzy AND routine produces a minimum value 50 that is assigned to the first closing component C1. The same fuzzy and procedure applies to the second and third closing components C2 and C3, each resulting in a value of 0, while simultaneously resulting in values for the first, second and third opening components O1, O2 and O3. After the minimum fuzzy decision, a maximum fuzzy decision is performed, i.e., a fuzzy or routine is applied to C1, C2 and C3, or MAX (50, 0, 0), which yields a maximum value of 50 that is assigned as the combined actuator closure component. The same maximum fuzzy decision is applied to the opening components O1, O2 and O3, or MAX (40, 0, 0), which yields a maximum value of 40 that is assigned as the combined actuator opening component.

The next step in the fuzzy logic routine is to combine or "defuzzify" the resulting close and open components so that there is a single output determination. The defuzzification centroid method, well known in the fuzzy logic art, can be used to obtain a single output. However, the single-element set approach is preferred since the centroid method is computationally more exhaustive than required by the present application. The off component (50) is subtracted from the on component (40) as is well known in the fuzzy logic art for the set of singlets, yielding the value-10. Since this value is less than zero, the valve actuator closure output signal is pulsed for a period of time equal to ten percent of the sample time. If the sampling time is 4 seconds, the off signal is pulsed for 0.4 seconds.

The steps of the microprocessor 160 executing the fuzzy logic program described above are shown in the flow diagrams of fig. 5A-5E. When a cycle timer in the microprocessor 160 expires, the fuzzy logic program begins execution (step 500). To restart the timer, the next sampling CYCLE is initiated (step 501), and the CYCLE timer CYCLE _ TMR is set equal to the variable LEVEL PERIOD LEVEL _ PERIOD set in the microprocessor. The LEVEL difference ERROR is derived from subtracting a programmable desired LEVEL LEVEL _ SETP of between 20% and 80% from a measured LEVEL LEVEL _ CONV of between 5% and 100% based on input from the LEVEL sensor 170 (step 501). If the coolant LEVEL LEVEL _ LAST for the LAST cycle time was equal to zero (step 502), then the LEVEL of the LAST cycle is set to the LEVEL actually measured for the current cycle (step 503). This is used to prevent error rate calculations from occurring if the fuzzy logic program is first re-entered after the system has been dormant for any period of time.

The RATE of change of liquid level difference RATE is calculated by subtracting the liquid level measured in the previous cycle from the liquid level measured in the current cycle (step 504). However, the present invention also contemplates a system that takes as input the derivative of the measured liquid level. Thus, the last variable is set equal to the level currently measured for the next cycle of the fuzzy logic program (step 504).

The next step of the program determines whether the head of liquid ERROR is between +/-3% or equal to +/-3% (step 505). If the level difference ERROR is in this range, the level difference ERROR is set equal to zero (step 506), otherwise the program determines whether the RATE of change of the level difference RATE is between +/-1% or equal to +/-1% (step 507). If so, then the rate of change of the liquid level difference is set equal to zero (step 508). Otherwise, the program compares the level difference to a variable PROPORTIONLIM CLOSE, which is independently programmable from 10% to 50% (step 509). If the level difference is less than or equal to the variable PROPORTION _ LIM _ CLOSE, the program sets the negative level difference (E N) ERROR _ NEG to 100, the zero level difference (E Z) ERROR _ ZER and the positive level difference (E P) ERROR _ POS to zero (step 510), and invokes subroutine B, otherwise invokes subroutine A.

As shown in fig. 5(B), subroutine a begins by determining whether the level difference ERROR is less than zero, i.e., negative (step 511). If the level difference ERROR is negative, the program sets the level difference ERROR _ NEG equal to- (100-ERROR) divided by the variable PROPORTION _ LIM _ CLOSE; setting the level difference ERROR _ ZER equal to 100 minus ERROR _ NEG; the level difference ERROR POS is set equal to zero (step 512). If ERROR in step 511 is greater than or equal to zero, the program compares the ERROR value to an independently programmable, between 10% and 50% variable PROPORTION _ LIM _ OPEN (step 513). When ERROR is greater than PROPORTION _ LIM _ OPEN, the program sets the level differences ERROR _ NEG and ERROR _ ZER to zero and ERROR _ POS to 100 (step 514). Otherwise, the program sets ERROR _ NEG to zero; let ERROR _ POS equal to (100-ERROR) divided by the variable priority _ LIM _ OPEN; let ERROR _ ZER equal 100 minus ERROR _ POS (step 515).

Sub-routine B is entered to determine the amount of relationship of the rate of change of the liquid level difference. The program determines if the RATE of change of liquid level difference RATE is less than a programmable, variable RATE _ LIM _ CLOSE between 10% and 50% (step 516), and if so, RATE _ NEG is set to 100 and RATE _ ZER and RATE _ POS are set to zero (step 517). Otherwise, the routine determines whether the RATE of change of liquid level difference RATE is less than zero, i.e., negative (step 518). If the RATE of change of liquid level difference RATE is negative, the program sets the RATE of change of liquid level difference RATE NEG equal to- (100-RATE) divided by the variable RATE _ LIM _ CLOSE; setting the liquid level difference change RATE _ ZER as 100 minus RATE _ NEG; the RATE of change of liquid level difference RATE _ POS is set equal to zero (step 519). If the RATE in step 518 is greater than or equal to zero, the program compares the RATE value to an independently programmable variable RATE _ LIM _ OPEN that is between 10% and 50% (step 520). When the RATE is greater than RATE _ LIM _ OPEN, the program sets the RATE of change of level difference RATE _ NEG and RATE _ ZER to zero and RATE _ POS to 100 (step 521). Otherwise, the program sets RATE _ NEG equal to zero; rate _ POS equals (100-Rate) divided by Rate _ LIM _ CLOSE; RATE _ ZER equals 100 minus RATE _ POS (step 522).

In FIG. 5(C), subroutine C represents the minimum fuzzy valve closure determination technique described above. Here, it determines whether the negative liquid level difference change RATE _ NEG is less than or equal to the negative liquid level difference ERROR _ NEG (step 523). If so, the valve closing component CLOSE is set equal to the negative RATE of change of head RATE _ NEG, which is the minimum value (step 524). Otherwise, the valve closing component CLOSE is set equal to the negative head difference ERROR _ NEG (step 525). At this point, the valve closing contribution CLOSE is equal to the second valve closing contribution C2.

The program then determines whether the negative level difference change RATE _ NEG is less than or equal to the zero level difference ERROR _ ZER (step 526), and if so, a dummy variable TEMP is set equal to the negative level difference change RATE _ NEG (step 527). Otherwise the dummy variable TEMP is set equal to zero level difference ERROR _ ZER (step 528). At this time, the dummy variable TEMP represents the first valve-closing component C1.

The program determines if the valve closing component CLOSE is less than the dummy variable TEMP (step 529) and if so, the valve closing component CLOSE is set equal to the dummy variable TEMP (step 530). This operation is equivalent to finding the maximum valve closing component between the first and second valve closing components C1, C2. If the program finds that the zero level difference change RATE _ ZER is less than or equal to the negative level difference ERROR _ NEG (step 531), the dummy variable TEMP is set equal to the zero level difference change RATE _ ZER (step 532). Otherwise the dummy variable TEMP is set equal to the negative level difference ERROR NEG (step 533). Subroutine D is then called. The dummy variable TEMP now represents the third valve closing contribution C3.

As shown in fig. 5(D), the subroutine D starts by determining whether the CLOSE component CLOSE representing the maximum value between the first and second CLOSE components C1, C2 is less than the dummy variable TEMP representing the third CLOSE component C3 (step 534). If so, the CLOSE component CLOSE is set equal to the dummy variable TEMP (step 535). After such operation, the maximum value of the shutdown component is determined and stored as CLOSE.

The remainder of subroutine D represents the minimum fuzzy valve opening determination technique as previously described and begins by determining whether the zero head RATE of change RATE _ ZER is less than or equal to the positive head ERROR _ POS (step 536). If so, the opening component OPEN is set equal to the zero level difference change RATE RATE _ ZER, which is the minimum value (step 537). Otherwise, the opening component OPEN is set equal to the positive level difference ERROR _ POS (step 538). At this time, the opening component OPEN is equal to the first opening component O1.

The program determines whether the positive head change RATE _ POS is less than or equal to the zero head ERROR _ ZER (step 539). If so, the dummy variable TEMP is set equal to the positive level difference change RATE RATE _ POS (step 540). Otherwise, the dummy variable TEMP is set equal to zero level difference ERROR _ ZER (step 541). At this time, the virtual variable TEMP represents the third opening component O3. If the OPEN contribution OPEN is less than the dummy variable TEMP (step 542), the OPEN contribution OPEN is set equal to the dummy variable TEMP (step 543). This operation is equivalent to finding the maximum value of the opening component between the first and third opening components O1, O3. If the positive liquid level difference change RATE _ POS is less than or equal to the positive liquid level difference ERROR _ POS (step 544), the dummy variable TEMP is set equal to the positive liquid level difference change RATE _ POS (step 545) and subroutine F is invoked, otherwise subroutine E is invoked.

As shown in fig. 5(E), subroutine F is identical to subroutine E except that it bypasses step 546. In step 546, the dummy variable TEMP is set equal to the positive level difference ERROR _ POS. The dummy variable TEMP now represents the second opening component O2. As further shown in fig. 5(E), step 547 determines whether the opening component OPEN, representing the maximum value of the opening component between the first and third opening components O1, O3, is less than the dummy variable TEMP, now representing the second opening component O2. If so, the OPEN component OPEN is set equal to the dummy variable TEMP (step 548). After such operation, the maximum value of the opening component is determined and stored in the microprocessor 160 as OPEN.

Subroutine E continues to step 549 in which dummy variable TEMP is set equal to the difference between the maximum on component OPEN and the maximum off component CLOSE. This run is equivalent to the method of defuzzifying the set of single elements. The routine determines whether the value of TEMP is greater than 2 (step 550) and if so, the valve is pulsed OPEN for a PERIOD of time equal to TEMP multiplied by LEVEL _ PERIOD divided by 100 as represented by the variable OPEN _ LEVEL _ TMR (step 551). CLOSE _ LEVEL _ TMR is set to zero (step 551) because a CLOSE pulse is not expected to occur.

If the routine determines that the value of TEMP is less than-2 (step 552), then CLOSE _ LEVEL _ TMR is set to TEMP times LEVEL _ PERIOD divided by 100, representing the amount of time the valve is pulsed off, and OPEN _ LEVEL _ TMR is set to zero, representing the valve is not pulsed on (step 553). If TEMP is not greater than 2 or not less than-2, then CLOSE _ LEVEL _ TMR and OPEN _ LEVEL _ TMR are both set to zero (step 554), indicating that no valve pulse needs to be generated.

As can be seen from the foregoing, the present invention utilizes a level sensor to measure the level of the coolant liquid in the condenser, which is then used in a fuzzy logic control algorithm to quickly and accurately control the level of the coolant liquid in the condenser. In this way, the cooling system can be controlled to operate at the most efficient operating point.

Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

Claims (20)

1. A cooling system comprising:

a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates;

a coolant liquid in the condenser;

a sensor for measuring the level of liquid coolant in the condenser and providing a measurement signal;

said expansion device capable of selectively varying the amount of restriction it applies to the coolant as it flows through the apparatus in response to a control signal; and

a microprocessor control means which samples a measurement signal at selected time intervals and applies said sampled measurement signal in a programmable predetermined manner to generate a control signal for an expansion device, wherein said control signal positions the expansion device to control the flow of gaseous coolant to said vaporizer.

2. The chiller system of claim 1 wherein said expansion device comprises a multi-position valve.

3. The chiller system of claim 2 wherein said control generates a control signal pulse to proportionally open and close said valve in response to said pulsed control signal to cause the liquid level in the condenser to reach a preselected set level.

4. The cooling system of claim 3 wherein said microprocessor applies the sampled measurement signal to a fuzzy logic algorithm.

5. A cooling system comprising:

a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates;

coolant liquid in the condenser;

a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal;

said expansion device being disposed between the evaporator and the condenser and being capable of selectively varying the magnitude of its restriction applied to the flow of coolant between the condenser and the evaporator in response to a control signal; and

a microprocessor control means receiving measurement signals from the sensor and generating control signals to be supplied to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm having a programmable fuzzy logic relationship function to control the flow of gaseous refrigerant to said evaporator.

6. The chiller system of claim 5 wherein said sensor is a level sensor for measuring the level of liquid coolant in the condenser.

7. The cooling system of claim 6 wherein said level sensor is a capacitive level sensor.

8. The chiller system of claim 5 wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal.

9. The chiller system of claim 8 wherein said microprocessor generates a control signal which selectively opens and closes said valve to cause the refrigerant level in the condenser to reach a preselected set level.

10. A cooling system comprising:

a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates;

coolant liquid in the condenser;

a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal;

said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal; and

a control means receiving a measurement signal from the sensor and generating said control signal to be supplied to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm to control the flow of gaseous refrigerant to said evaporator, wherein said control comprises a microprocessor generating said control signal to selectively open and close said valve to cause the refrigerant level in the condenser to reach a preselected set level, wherein said microprocessor opens the valve relative to the previous position of the valve when the level of refrigerant liquid measured in the condenser is above the preselected set level.

11. A cooling system comprising:

a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates;

coolant liquid in the condenser;

a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal;

said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal; and

a control means receiving a measurement signal from the sensor and generating said control signal to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm to control the flow of gaseous refrigerant to said evaporator, wherein said control means includes a microprocessor generating said control signal to selectively open and close said valve to cause the refrigerant level in the condenser to reach a preselected set level, wherein said microprocessor closes the valve relative to its previous position when the refrigerant liquid level measured in the condenser is below the preselected set level.

12. The cooling system of claim 11 wherein said valve functions as an expansion device.

13. The chiller system of claim 11 wherein said microprocessor periodically samples the level of liquid measured by said liquid level sensor and wherein said microprocessor executes a fuzzy logic algorithm once per sampling period.

14. The cooling system of claim 15 wherein said microprocessor calculates a level difference equal to the difference between the measured level of the coolant liquid and the preselected setpoint level and applies the level difference to a fuzzy logic algorithm.

15. A cooling system comprising:

a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates;

coolant liquid in the condenser;

a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal;

said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal; and

a control means receiving the measurement signal from the sensor and generating said control signal to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm to control the flow of gaseous refrigerant to said evaporator, wherein said control comprises a microprocessor generating said control signal to selectively open and close said valve to cause the refrigerant level in the condenser to reach a preselected set level, wherein said microprocessor periodically samples the level measured by said level sensor, and wherein the microprocessor executes a fuzzy logic algorithm once per sampling period to calculate a level difference equal to the difference between the measured refrigerant liquid level and the preselected set level and applies the level difference to a fuzzy logic algorithm, wherein the microprocessor calculates a rate of change of coolant level by subtracting the coolant level measured at the previous sampling period from the coolant level measured at the current sampling period and applies the rate of change of level to a fuzzy logic algorithm.

16. The chiller system of claim 15 wherein said microprocessor uses fuzzy logic algorithms to determine the amount of relationship (negative, zero, or positive) associated with each rate of change of level and level difference value, respectively, by assigning a preselected weight to each calculation.

17. The cooling system of claim 16, wherein the asymmetric relationship function determines the relationship quantity.

18. The cooling system of claim 11, wherein the microprocessor uses a fuzzy logic algorithm to determine the rule based on the calculated relationship quantity to open and close the valve.

19. A cooling system comprising:

a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates;

coolant liquid in the condenser;

a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal;

said expansion device being positioned between the evaporator and the condenser and being capable of selectively varying the amount of restriction it imposes on the flow of coolant between the condenser and the evaporator in response to a control signal, wherein said expansion device includes a multi-position valve that opens and closes in response to said control signal and acts as an expansion device; and

a control means receiving a measurement signal from a sensor and generating said control signal to said expansion device to control the flow of refrigerant from the condenser to the evaporator, wherein the control utilizes a fuzzy logic algorithm to control the flow of gaseous refrigerant to said evaporator, wherein said control comprises a microprocessor generating said control signal to selectively open and close said valve to cause the refrigerant level in the condenser to reach a preselected set level, wherein said microprocessor utilizes a fuzzy logic algorithm to apply a fuzzy minimum/maximum method which first performs a "fuzzy and" (minimum) decision and then a "fuzzy or" (maximum) decision to derive the closing and opening components.

20. A cooling system comprising:

a vaporizer, a compressor, a condenser and an expansion device, all connected in a closed cooling circuit along which a coolant circulates;

coolant liquid in the condenser;

a sensor for measuring the refrigerant liquid in the condenser and providing a measurement signal;

said expansion device being disposed between the evaporator and the condenser and being capable of selectively varying the magnitude of its restriction applied to the flow of coolant between the condenser and the evaporator in response to a control signal; and

a control means receiving a measurement signal from a sensor and generating said control signal to be supplied to said expansion device to control the flow of coolant from the condenser to the evaporator, wherein the control uses a fuzzy logic algorithm to control the flow of gaseous coolant to said evaporator, wherein said fuzzy logic algorithm uses as input variables the value of an input variable of coolant measured by said sensor at a given time and the rate of change of that value at a given time relative to a previous detection time.