In one embodiment, the invention relates to absorption heat exchange apparatus for removing heat from a heat load to a heat sink, and particularly to an improved method for varying the capacity of absorption heat exchange apparatus. In another embodiment, the invention relates to a two-stage absorption refrigeration machine and, more particularly, to a control system for detecting and recovering from crystallization within a system heat exchanger. Finally, in a third embodiment, the invention relates more particularly to a control system having a binomial setpoint filter to eliminate overshoot without reducing the response to process disturbances.

Absorption refrigeration, chilling, heat pump, and related apparatus employing a composite refrigerant circulating through a refrigeration loop are well known.

The refrigeration loop includes a generator, a condenser, an evaporator, and an absorber.

A variety of composite refrigerant systems can be used in such apparatus. Two examples are an ammonia/water system and a lithium bromide/water system.

An external source of energy adds heat to the composite refrigerant and absorbent solution in the generator. The generator heats the composite solution sufficiently to distill out a vapor of the more volatile refrigerant (for example, ammonia vapor in the case of the ammonia/water refrigerant and water in the case of the lithium bromide/water system), leaving a less-volatile, concentrated absorbent solution behind.

The concentrated absorbent component is then removed to the absorber.

The condenser receives the vaporized refrigerant from the generator and condenses it to liquid form (also known as a condensate). The heat released by the condensation of the vapor is rejected to a cooling tower, cooling water, some other external heat sink, or another stage of the refrigeration apparatus.

The evaporator withdraws heat from a heat load (i.e. the building air, refrigerator contents, cooling water, or other fluid or object the system is designed to cool) by evaporating the condensed liquid refrigerant in direct or indirect contact with the heat load. The evaporator thus re-vaporizes the volatile refrigerant.

The absorber contacts the refrigerant vapor leaving the evaporator with the concentrated absorbent solution leaving the generator. The contacting process generates heat when the vapor phase is reabsorbed into the less-volatile solution phase. This heat is rejected to a cooling tower, cooling water, another stage of the refrigeration apparatus, or some other heat sink. The original composite refrigerant and absorbent solution is re- formed in the absorber, and then is returned to the generator to complete the cycle.

In one common absorption heat-exchange apparatus, the evaporator and the absorber are located in a single vessel, so the refrigerant vapor generated in the evaporator can pass easily to the absorber for reabsorption. In the typical combined evaporator and absorber, the contacting process involves spraying the concentrated absorbent solution into contact with refrigerant vapor. The presence of concentrated solution at a low temperature produces a saturated vapor pressure in the absorber which is slightly less than the saturation pressure in the evaporator. The pressure imbalance causes refrigerant vapor to flow from the evaporator to the absorber, where it is reabsorbed into solution. The chiller refrigeration capacity is a function of the rate at which the refrigerant evaporates in the evaporator, and thus is directly related to the rate at which the evaporated refrigerant flows from the evaporator to the absorber.

It is desirable to change the refrigeration capacity of absorption chillers to accommodate varying loads which may be imposed upon the chiller. The most common method of controlling refrigeration capacity is to change the concentration of the absorbent solution that is sprayed, at a constant rate, into the absorber. Increasing the concentration of the absorbent solution at the absorber sprays creates a greater pressure imbalance in the absorber, which in turn causes more refrigerant vapor to flow from the evaporator to the absorber, which causes refrigerant to evaporate at a higher rate in the evaporator, which finally increases the refrigeration capacity. Conversely, decreasing the concentration of the absorbent solution at the absorber sprays reduces the refrigeration capacity.

The absorbent solution concentration at the absorber sprays has been varied by varying the flow rate of the concentrated absorbent solution flowing from the generator to the absorber. As the flow from the generator fluctuates, the chiller apparatus maintains a constant flow into the absorber by blending some of the dilute composite solution from the absorber sump with the concentrated absorbent from the generator. The apparatus then passes the blend through the absorber sprays. When the flow rate from the generator is low, for example, the recirculation flow rate will be high and the initially-concentrated absorbent solution entering the absorber will be dilute.

In theory, the chiller capacity would be reduced to zero if the flow of absorbent solution from the generator was reduced to zero (in which case flow to the absorber sprays would consist solely of recirculation from the absorber sump). In practice, however, the apparatus must maintain a minimum flow rate through the generator to prevent stagnation of flow and crystallization in the chiller heat exchangers. To reduce refrigeration capacity beyond a certain point, then, the apparatus must further dilute the absorbent solution flowing from the generator by mixing it with a large amount of excess refrigerant. The use of excess refrigerant before the absorber to further dilute the absorbent solution, however, increases the response time of a chiller and requires a large refrigerant storage tank. For residential, office and industrial heat exchange applications, where users desire absorption chillers that will accommodate low refrigeration loads and yet respond quickly to load changes, another arrangement is needed.

For example, a controller cannot quickly vary an office building chiller system between low and high refrigeration capacities by changing the concentration of absorbent solution at the absorber sprays. If the sun continually fades in and out of the clouds on a partly cloudy day, the system will cycle from low or no refrigeration capacity, when clouds block the sun, to a significant refrigeration capacity when the sun shines through the office windows and increases the inner air temperature. At times of low capacity, the office chiller system will flood the absorbent solution flow with excess refrigerant in order to reduce its concentration. When the next sudden increase in office air temperature requires a significant refrigeration capacity increase, the generator must literally boil the excess refrigerant out of the composite solution to re-form the concentrated absorbent. The absorbent must be re-formed in the generator before the chiller can regain its high capacity. However, by the time refrigeration capacity is restored, the sun may very well be behind a cloud again. At that point, the system must again reduce the refrigeration capacity because the sun is no longer heating up the office.

Thus, although refrigeration capacity control of an absorption chiller by means of varying the absorber solution concentration is well known, existing systems are inefficient when operating in low refrigeration capacity ranges because they have a long response time. In addition, in order to even operate in low refrigeration capacity ranges, such systems must be encumbered with a large refrigerant storage tank.

In another respect, this invention also relates to a two-stage absorption refrigeration machine and, more particularly, to a control system for detecting and recovering from crystallization within a system heat exchanger. During operation of a two-stage absorption refrigeration machine, the occurrence of accidents or malfunctions can cause solidification or crystallization of absorption solution in the flow passages of the machine. One of the most common sites for crystallization is in the concentrated solution passage of the concentrated solution heat exchanger. At this point, the absorption solution has been concentrated by the generators and is being forced back to the absorber.

Between the generators and the absorber, the concentrated solution passes through a heat exchanger, releasing heat to dilute absorbent solution being pumped to the generators from the absorber. If for some reason the absorbent solution becomes too concentrated or it is cooled below its crystallization temperature, the concentrated solution flow passage begins to block and eventually closes completely due to crystallization. This condition can occur over a period of very few minutes and has been known to occur in less than a minute.

A number of conditions can cause crystallization of the concentrated absorbent solution in the heat exchanger. For example, the presence of air or other inert gas in the absorber will prevent dilution of the absorbent solution therein. This will cause the concentration of a concentrated absorbent solution to rise. As the solution becomes supersaturated, it will begin to crystallize. If the condenser water suddenly becomes colder than normal operating temperature, a reduction in the temperature of the dilute absorbent solution leaving the absorber will result. This, in turn, will reduce the temperature of the concentrated absorbent solution in the heat exchanger below the crystallization point and will begin to block the heat exchanger. Overfiring the generator, resulting in supersaturation of the absorbent solution, will also cause crystallization blockage of heat exchange passages.

It is desirable to prevent any of the above conditions from ever occurring.

However, because of malfunction or accident, it is impossible to prevent crystallization in the heat exchanger at all times. When crystallization and heat exchange blockage occur, a practical prior method of clearing the heat exchanger passages had been to heat them with an external heat source and liquify the absorbent solution therein. However, this solution is unacceptable because it requires significant interruption of the absorption machine operation. Other prior crystallization detection and prevention systems incorporated the use of mechanical float valves in the concentrated absorbent solution flow passages that would be activated when flow began to reverse due to crystallization.

However, these mechanical systems have proven to be unreliable and expensive.

Finally, in yet a third respect, the invention is directed generally to control systems and more particularly to a control system having a binomial setpoint filter to eliminate overshoot without reducing the response to process disturbances.

Control systems are available for monitoring and controlling virtually all kinds of systems and machines. Control systems are often utilized because of their economic benefits. For example, the ability to hold a process closer to a desired operating constraint is an advantage. Such controls also increase the safety of the system and also increase the efficiency.

Two basic types of control systems exist. One type of control system is a regulation (regulatory) control system. This type of control system is used primarily to respond to changes and disturbances to the system. Examples of devices controlled by regulation control systems include water chiller machines which are used to provide cooling water for comfort cooling applications.

Another type of control system is a tracking control system. This type of control tracks a change in a setpoint or a related input. Such a control system improves the control of the machine. For example, an initial setpoint is entered to a system or machine, and the control system tracks any deviations from the machine and tries to maintain operation of the machine at the desired setpoint.

Simply stated, a closed loop control system consists of a process, a measurement of the controlled variable, and a controller which compares the actual measurement with the desired value and uses a difference between them to automatically adjust one of the inputs to the process. The physical system to be controlled can be electrical, thermal, hydraulic, pneumatic, gaseous, mechanical or described by any other physical or chemical process.

Generally, a control system is designed to meet one of two objectives.

First, a servomechanism is designed to follow changes in setpoint as closely as possible.

Many electrical or mechanical control systems are servomechanisms. Second, a regulator is designed to keep output constant despite changes in a load or other disturbances.

Regulatory controls are widely used for controlling chemical processes. In general, tracking control systems monitor setpoint changes and make appropriate adjustments.

Regulatory control systems adjust to compensate for process disturbances.

The stability, accuracy and speed of response of a control system are determined by analyzing the steady state and the transient performance. It is desirable to achieve the steady state in the shortest possible time, while maintaining the output within specified limits. Steady-state performance is evaluated in terms of the accuracy with which the output is controlled for a specified input. The transient performance, i.e., the behavior of the output variable as the system changes from one steady-state condition to another, is evaluated in terms of such quantities as maximum overshoot, rise time and response time.

A number of factors affect the quality of control, including the disturbances caused by setpoint changes and process load changes. Both setpoint and process load may be defined in terms of the setting of the final control element to maintain the control variable at the setpoint. Thus, both cause the final control element to reposition. Other disturbances may be variations in inlet process fluid temperature and cooling water temperature in a water chiller, for example.

In many control systems, a step input response results in overshoot.

However, a step input is widely used for analysis for many reasons. First of all, testing is easily implemented. Second, the step input is the most severe disturbance, and the response to a step input shows the maximum error that can occur. Features of transient performance include the existence and magnitude of the maximum overshoot, the frequency of the transient oscillation and the response time.

In certain instances, the output variable overshoots its desired steady state condition and transient oscillation occurs. The first overshoot is the greatest and its effect is a concern to the designer. The primary considerations for limiting this maximum overshoot are (1) to avoid damage to the process or machine due to excessive excursions of the controlled variable beyond that specified by the command signal and (2) to avoid the excessive settling time associated with highly underdamped systems.

As mentioned above, control systems may be used as process regulators or tracking controls. For example, absorption chillers are used in industrial applications. In these applications, the chiller controls may be required to perform more of a tracking control function. Water chiller controls for comfort chilling are mainly process regulators.

The chiller controls evaporator leaving chilled water to a setpoint that may never be changed. For this type of application, it is usually desirable to use a relatively high integral gain in a PID (proportional, integral and derivative) control loop to eliminate error in leaving water temperature caused by a process disturbance. Typically, a higher integral gain is beneficial since it allows the control system to respond faster to load disturbances. However, one of the problems with using a large integral gain is, at startup, the control overshoots its setpoint. Another problem caused by the high integral gain is a shut down on low temperature.

Such an overshoot problem has previously been addressed with a control function called "softloading." During a setpoint change, or at startup, the chiller system experiences an immediate and substantial change. The typical response of the chiller system is to load up to 100% to meet the change. To compensate for this major change, the softloading function limits the commanded output from the controller at startup to slow down the loading on the chiller.

However, the softloading function also has certain problems. For example, the softloading is arranged at the back end of the control system, which makes the function difficult to implement. Since the softloading function limits the commanded output, not the input, the manner in which the output is limited changes with different types of control systems in different machines. For example, to limit a water temperature change, the softloading function must limit the command to control the water temperature.

Also, the manner of limiting the command is a function of what is being controlled. For each system, this requires that the command be tweaked for every system. In addition, a large amount of empirical work is needed to get the response to be well-behaved.

An additional control concern occurs in certain instances in which the leaving water temperature setpoint needs to be changed. Again, with a large integral gain, a large overshoot results. Also, there are comfort cooling applications in which the chilled water setpoint is changed on a daily basis. For example, the temperature may be raised at night and lowered during the day. These regular changes also cause the overshoot problems discussed above.

Therefore, with regulator control, it is desirable to push the integral control higher to make the system respond faster. This increased integral control typically works fine until a setpoint change is required. Then the increased integral gain on the setpoint chance will be excessive and cause overshoot. However, when the same system attempts to perform a tracking type control, overshoot will occur. As a result, a need exists for a control system which eliminates overshoot without reducing the response to process disturbances.

BRIEF SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a faster way to change the refrigeration capacity of a chiller, responsive to load variations, process disturbances, and other reasons for varying the capacity of an absorption chiller.

Another object of the invention is to provide a faster way to change the responsiveness of a chiller when operating at a low refrigeration capacity.

An additional object of the invention is to provide a way to extend the refrigeration operating range of a chiller without sacrificing responsiveness or increasing the size of the chiller apparatus.

One aspect of the invention is a variable capacity absorption chiller. The chiller includes a generator for generating a concentrated absorbent solution and a refrigerant, an absorber that contains a surface contact portion where the concentrated absorbent solution is contacted with the refrigerant, and a reservoir. The chiller also includes at least one conduit (e.g., a tube, pipe, passage, or common vessel) for transferring one flow of concentrated absorbent solution from the generator to the surface contact portion of the absorber and another flow of concentrated absorbent solution from the generator to the reservoir. Finally, the chiller includes a fluid flow regulator (e.g., a valve or a pump) for varying the ratio of the two flows coming from the generator.

Another aspect of the invention is a method of varying the refrigeration capacity of an absorption chiller. The method can be carried out in a generator, an absorber, and a reservoir as described above. A flow of concentrated absorbent solution is provided from the generator to the surface contact portion of the absorber. Another flow of concentrated absorbent solution is provided from the generator to the reservoir. A desired refrigeration capacity is determined for the chiller, and the flow rate ratio of the two flows coming from the generator is varied to provide the desired refrigeration capacity.

The present invention has several advantages. One advantage is that the time to respond to load variation or process disturbance can be greatly reduced. Another advantage is that chiller operating range can be extended into the low capacity regions without sacrificing responsiveness or requiring an enlarged apparatus (such as a large refrigerant reservoir).

Therefore, it is an object of the present invention to provide a reliable and inexpensive apparatus and method for detecting crystallization of concentrated absorbent solution in a heat exchanger of a two-stage absorption refrigeration machine.

It is a further object of the present invention to provide a method and apparatus for recovering from crystallization of the concentrated absorbent solution, once detected, without utilizing external heat sources.

In another respect, the present invention provides a method for detecting crystallization in a two-stage absorption refrigeration machine including a controller, an absorber, an evaporator, a high temperature generator, a low temperature generator, a condenser, a low temperature heat exchanger for placing concentrated absorbent solution from the high temperature and low temperature generators and dilute absorbent solution from the absorber in a heat exchange relationship, a first passage directing concentrated absorbent solution through the low temperature heat exchanger, a second passage directing concentrated absorbent solution from the low temperature generator to the first passage, a third passage for directing concentrated absorbent solution from the high temperature generator to the first passage, a fourth passage for directing concentrated absorbent solution from the low temperature generator to the high temperature generator, a second passage temperature sensor for sensing the temperature of the concentrated absorbent solution in the second passage, a third passage temperature sensor for sensing the temperature of the concentrated absorbent solution in the third passage, and a fourth passage temperature sensor for sensing the temperature of the concentrated absorbent solution in the fourth passage.

The absorption refrigeration machine also includes an evaporator spray pump delivering dilute refrigerant from an evaporator refrigerant collector to at least one evaporator spray nozzle, an absorber spray pump delivering concentrated absorbent solution to at least one absorber spray nozzle, a fifth passage directing concentrated absorbent solution from the low temperature heat exchanger to the absorber, a sixth passage directing dilute absorbent solution from the collector to the absorber spray pump, and a valve controlled by the controller disposed in the sixth passage controlling flow of dilute absorbent solution in the sixth passage. A fifth passage temperature sensor that senses the temperature of the concentrated absorbent solution in the fifth passage between the low temperature heat exchanger and the absorber.

In normal operation, the temperature sensed by the fourth passage temperature sensor is substantially equal to the temperature sensed by the second passage temperature sensor, and the valve in the sixth passage is closed. When crystallization begins to block the first passage, the temperature sensed by the second passage temperature sensor begins to exceed the temperature sensed by the fourth temperature sensor. According to the present crystallization detection method, a crystallization alert is issued when the temperature sensed by the second passage temperature sensor is substantially equal to the average of the temperature sensed by the fourth temperature sensor and the temperature sensed by the third passage temperature sensor. This temperature is called the trip temperature. When the temperature sensed by the second passage temperature sensor meets or exceeds the trip temperature, i.e., crystallization is detected in the low temperature heat exchanger, then the control system goes into crystallization recovery mode.

According to this aspect of the present invention, during the crystallization recovery mode, the control system completes the following steps: 1. concentrating of absorbent solution in both the low temperature and high temperature generators is temporarily halted by deactivating heat to the source; 2. the circulation of absorbent solution is temporarily halted by deactivating all system pumps; 3. the valve in the sixth passage between the evaporator and the absorber spray pump is opened to allow flow of dilute absorbent solution from the collector to the absorber spray pump; 4. after about 3 minutes of deactivation, the low temperature generator and high temperature generator pumps are reactivated for approximately 5 minutes to partially flush the high concentration absorbent solution that caused crystallization; 5. the low temperature generator and high temperature generator pumps are again deactivated for approximately 3 minutes to counteract any recrystallization that may have occurred during flushing; 6. all system pumps are reactivated, the valve is closed, and the control system adjusts the heat input at both the high temperature and low temperature generators such that the temperature of the concentrated absorbent solution between the low temperature heat exchanger and the absorber is maintained at a level such that the margin between the crystallization temperature of the concentrated solution in this region and the actual temperature is increased by 5"F. (about 3"C.) over the previous margin.

The control system will automatically go through the recovery cycle twice.

When crystallization is detected for the third time, all operation of the absorption refrigeration machine will be stopped, as this indicates a systemic problem that must be corrected.

To this end, yet another embodiment of the present invention provides a control system which eliminates overshoot without reducing the response to process disturbances. In particular, an embodiment of the control system includes a binomial setpoint filter for filtering setpoint changes to provide a more gradual response that consequently eliminates overshoot.

One aspect of the invention is a method of controlling a water chiller to provide a supply of chilled water at a preselected temperature representing the nominal setpoint of the system. The water chiller has a control system having an input. The method is carried out as follows. A nominal setpoint temperature is selected. The setpoint temperature is filtered, using a binomial filter, to provide a filtered setpoint temperature. The filtered setpoint temperature is then provided to the input of the control system. The filtered setpoint temperature is a function of time and temperature.

Preferably, the filtered setpoint temperature is initially the current temperature of the chilled water, then changes as a function of time to asymptotically approach the nominal setpoint temperature.

A more general aspect of the invention is a control system comprising a system input, a feedback loop, and a binomial filter having a setpoint input and a filtered setpoint output. The feedback loop is connected to the system input via a summing node.

The binomial filter is arranged to receive a nominal setpoint at its setpoint input and provide in response a filtered setpoint output to the system input via the summing node.

Another aspect of the invention is a method of reducing overshoot in a control system. A control system is provided having an input. A nominal setpoint is selected. The nominal setpoint is filtered, using a binomial filter, to provide a filtered setpoint. The filtered setpoint is provided to the input of the control system. The filtered setpoint approaches the nominal setpoint asymptotically thus reducing or eliminating overshoot of the nominal setpoint.

An advantage of an embodiment of the control system having a binomial setpoint filter is that using the filtered setpoint allows using more integral gain in the control system to respond faster to load disturbances, while responding to startups and setpoint changes without causing overshoot.

Also, an advantage of an embodiment of the control system having a binomial setpoint filter is that using the filtered setpoint for softloading is relatively simple to implement. For example, instead of limiting the output command, the filter works on the front end to drive the control response to follow a desired trajectory. In an embodiment, a critically damped second order response is the desired trajectory which is obtained using a second order binomial filter. Such a binomial filter provides a gradual rise to the desired setpoint instead of an abrupt one, as is the case without a filter or with a first order filter.

Another advantage of an embodiment of the control system having a binomial setpoint filter is that by filtering the setpoint, the control system does not see a step input when the temperature setpoint is changed. In contrast, the response is more like a series of small process disturbances.

A further advantage of an embodiment of the control system having a binomial setpoint filter is that it uses closed loop controls already in place for leaving water temperature control, since the binomial setpoint filtering is performed at the front end, or input side, of the control system.

One or more of the preceding objects, or one or more other objects which will become plain upon consideration of the present specification, are satisfied in whole or in part by the invention described herein.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of an absorption refrigeration apparatus according to the present invention.

FIG. 2 is a schematic view of a portion of an absorption refrigeration apparatus according to one embodiment of the present invention.

FIG. 3 is a schematic view of a portion of an absorption refrigeration apparatus according to a second embodiment of the present invention.

FIG. 4 is a schematic illustration of a two-stage absorption refrigeration machine employing an embodiment of the present invention.

FIG. 5 is a flow chart illustrating the recovery control system according to the present invention.

FIG. 6 is a block diagram of an embodiment of a control system of the present invention illustrating a second order binomial filter for filtering a setpoint in the control system.

FIG. 7 is a graph illustrating responses of first and second order filters to a step input.

FIG. 8 is a graph illustrating chilled water leaving temperature with respect to time relative to a binomially filtered setpoint temperature in a control system, at startup, operating in accordance with the apparatus and method of the present invention.

FIG. 9 is a graph illustrating chilled water leaving temperature with respect to time relative to a binomially filtered setpoint temperature as a result of a setpoint change in a control system operating in accordance with the apparatus and method of the present invention.

FIG. 10 is a block diagram of an embodiment of a process control system illustrating an absorber chiller having a second order binomial filter for filtering a setpoint in the control system operating in accordance with the apparatus and method of the present invention.

DETAILED DESCRIPTION OF THE INVENTION While the invention will be described in connection with one or more embodiments, it will be understood that the invention is not limited to those embodiments.

On the contrary, the invention includes all alternatives, modifications, and equivalents as may be included within the spirit and scope of the appended claims.

In the present description no particular refrigerant will be referred to. One of ordinary skill in the art is aware of refrigerant systems useful in the apparatus which utilize the present method. The same or different refrigerant systems may be used in relevant apparatus.

This description refers generically to the components of a typical absorption refrigeration solution, which are a more volatile refrigerant or refrigerant vapor (which, in liquid form, is sometimes referred to as a condensed vapor) and a less volatile absorbent component. These components may coexist as a solution, they may be separated by applying heat to the solution, thus distilling the more volatile refrigerant away, and they may be re-combined to reconstitute the solution and reject heat. The vapor may also be condensed to reject heat or vaporized to accept heat. Absorption refrigerant solutions which operate in a different manner, but which may be used in comparable apparatus, are also contemplated for use herein.

Referring first to FIG. 1, the block diagram illustrates the heat and refrigerant transfers of an absorption refrigeration system.

The system 10 is used to transfer heat from the heat load 12 to the heat sink 14. As is known, this heat transfer can be carried out whether the heat load 12 is at a higher temperature than, a lower temperature than, or the same temperature as the heat sink 14.

Heat from the load 12 enters the evaporator 16 of the apparatus via the path 18. (All heat transfers to or from one of the elements are represented in FIG. 1 by the letter Q next to an arrow indicating the direction of transfer.) Either the evaporator 16 is in direct heat-transfer contact with the heat load 12 or heat exchangers connect the evaporator 16 and the heat load 12 to accomplish this heat transfer.

The heat Q entering the evaporator 16 evaporates the condensed refrigerant vapor which has entered the evaporator 16 via the path 20. The effluent of the evaporator 16, which traverses the path 22, is refrigerant vapor which bears the heat from the heat load 12.

The contact portion of the absorber 24a receives the refrigerant vapor via the path 22 and contacts it with the less-volatile concentrated absorbent solution received from the generator 26 via the paths 28 and 28a. The resulting absorption of the refrigerant vapor into the less-volatile absorbent liquid both condenses the vapor, releasing its heat of vaporization, and releases heat of dissolution as the result of the absorption process. The resulting heat Q is rejected via the path 30 to the heat sink 14. The reconstituted composite refrigerant solution is released via the path 24b into the reservoir 24c, where it mingles with concentrated absorbent solution received from the generator 26 via the paths 28 and 28b. In the typical absorption refrigeration apparatus, the reservoir 24c will be the absorber sump. Any other suitable vessel, container, or apparatus for containing solution, however, will also operate effectively in the described chiller. The reservoir solution is passed via the paths 34 through the heat exchanger 35 to the generator 26. The heat exchanger 35 preheats the composite refrigerant traversing the paths 34 before it enters the generator 26, using heat which otherwise would escape from the generator via the concentrated absorbent solution paths 28.

A temperature sensor 116 senses the temperature of the heat-exchange fluid flowing along the path 18 downstream of the evaporator 16. A control line 13 connects the temperature sensor 116 to the controller 118. Another control line, 15, connects the controller 118 and the flow regulator 111. Having received the input from the temperature sensor 116, the controller 118 controls the flow regulator 111 accordingly, regulating the ratio of concentrated absorbent solution from the generator 26 that flows to the contact portion of the absorber 24a via the paths 28 and 28a with that flowing to the reservoir 24c via the paths 28 and 28b. The controller 118 may reduce the flow rate ratio to zero by directing the flow regulator 111 to divert all concentrated absorbent solution flowing from the generator 26 via the path 28 into the reservoir 24c via the path 28b. The fluid flow regulator 111 may consist of a pump, a valve, a series of pumps or valves, or any other apparatus which can be adjusted to vary the flows along paths 28a and 28b as described.

In the generator 26, the heater 36 heats the composite refrigerant solution sufficiently to distill away the more volatile refrigerant vapor, leaving the less volatile concentrated absorbent solution behind. The refrigerant vapor is delivered via the path 38 to the condenser 40. The concentrated absorbent solution goes to the contact portion of the absorber 24a and the reservoir 24c via the paths 28, 28a, and 28b (as previously described).

The condenser 40 condenses the refrigerant vapor entering via the path 38.

The heat of condensation Q is rejected, and follows the path 42 to the heat sink 14. The condensed refrigerant vapor then exits the condenser 40 via the path 20 and returns to the evaporator 16 to complete the cycle. Thus, heat from the heat load 12 and the heater 36 enters the loop, and heat leaves the loop from the absorber 24 and condenser 40. Apart from any waste heat which is lost, all the heat taken from the heat load 12 and the heater 36 goes to the heat sink 14.

Referring now to FIG. 2, apparatus and methods are disclosed which will function in conjunction with or as an integral part of apparatus as illustrated in FIG. 1.

Certain parts of FIG. 2 correspond to those in FIG. 1, and thus share identical reference characters.

FIG. 2 shows a particular embodiment of the absorption refrigeration system 10 of the present invention.

The system 10 comprises an evaporator 16, an absorber 24, a generator 26, a condenser 40, a heat load 12, a heat sink 14, and a heat exchanger 35. The generator 26 has a dilute absorbent inlet 75, a refrigerant vapor outlet 55, and a concentrated absorbent outlet 85.

An evaporator-absorber shell 23 combines portions of the evaporator 16 and the absorber 24. The evaporator 16 includes refrigerant sprays 21, a heat load coil 18, an evaporator pan 17, a refrigerant storage tank 99, and a pump 102. The absorber 24 includes absorber sprays 101, a heat-exchange coil "X", a contact portion 24a, and an absorber sump 24c. The contact portion 24a of the absorber 24 is the volume and surfaces inside the evaporator-absorber shell 23 where refrigerant vapor is contacted with absorbent solution. In this embodiment, the contact area is principally the volume covered by the sprayers 101 and the surfaces of the heat-exchange coil "X". The evaporator- absorber shell also has a composite refrigerant outlet 107 communicating with the path 34, an absorbent port 130 communicating with the path 28b, a refrigerant storage inlet 97 for receiving refrigerant from the outlet 93 of the evaporator pan 17 via the path 95, a refrigerant storage outlet 100 for transferring refrigerant out of the refrigerant storage tank 99 to the refrigerant sprays via the path 96 and the pump 102, and an inlet 144 for receiving condensed refrigerant vapor from the path 20.

The system 10 provides three sensors -- 115, 116, and 117 -- as well as the controller 118 and the adjustable frequency drive 120. The system 10 also contains three pumps. The pump 102 is connected to the refrigerant storage tank 99 as previously described. The pump 103 is connected to the absorber sump via the line 34 and the composite refrigerant outlet 107, and the pump 111 is connected to the junction 131 between the paths 28, 28a, and 28b.

In the generator 26, the composite refrigerant solution enters at the dilute absorbent inlet 75 and is heated sufficiently to distill away the more volatile refrigerant vapor, leaving the less volatile concentrated absorbent solution behind. The refrigerant vapor is delivered through the refrigerant vapor outlet 55 via the line 38 to the condenser 40, where it is condensed. The concentrated absorbent solution is delivered through the concentrated absorbent outlet 85 via the lines 28 to the absorber 24. As the concentrated absorbent solution travels toward the absorber 24 via the line 28, it passes through the heat exchanger 35, where it is cooled by transferring heat to the composite refrigerant solution flowing through the line 34.

Condensed refrigerant vapor from the condenser 40 traverses the line 20 and enters the evaporator pan 17 through the inlet 144. The condensed refrigerant vapor flows from the evaporator pan 17 through the outlet 93 via the line 95 through the refrigerant storage inlet 97 into the refrigerant storage tank 99. The pump 102 withdraws the refrigerant present in the refrigerant storage tank 99 through the refrigerant storage outlet 100, and then forces the refrigerant to the refrigerant sprays 21 via the line 96. The condensed refrigerant vapor is sprayed out of the refrigerant sprays 21 over the heat load line 18. The residual spray which remains in liquid form is collected in to the evaporator pan 17, where it mingles with additional condensed refrigerant vapor that enters through the inlet 144 via the line 20. The condensed refrigerant vapor in the evaporator pan 17 is again withdrawn to the refrigerant storage tank 99, and the cycle described above is repeated.

The heat entering the evaporator 16 comes from the heat load 12 and traverses the heat load line 18. The heat traversing the heat load line 18 comes into heat- exchange contact with the evaporator 16 and evaporates the condensed refrigerant which is sprayed into the evaporator 16 via the refrigerant sprays 21 through the line 96. The effluent of the evaporator 16 is refrigerant vapor contained in the evaporator-absorber shell 23, which flows into the absorber 24 bearing the heat from the heat load 12.

Concentrated absorbent solution enters the contact portion 24a of the absorber 24 through the absorber sprays 101, and enters the absorber sump 24c through the absorbent port 130. In this embodiment, the ratio of concentrated absorbent solution entering the contact portion 24a versus the absorber sump 24c is regulated as follows.

The stream of concentrated absorbent solution traversing the line 28 enters the junction 131, where the line 28a connected through the pump 111 and the line 28b connected to the absorber sump 24c also converge. The pump 111 is a variable capacity pump which meters concentrated absorbent solution to the absorber sprays 101.

At times when the pump 111 is operating at a higher capacity than the flow rate through the path 28, sump solution from the absorbent port 130 via the line 28b combines at the junction 131 with the concentrated absorbent solution from the line 28.

The combined solution is drawn into the pump 111, where it is then forced through the line 28a to the absorber sprays 101. At times when the pump 111 is dormant or operating at a lower capacity than the flow rate through the path 28, concentrated absorbent solution traversing the line 28 enters the junction 131 and, to the extent that the pump 111 drawing concentrated absorbent solution through the line 28a cannot accommodate the entire flow, travels through the line 28b and into the absorber sump 24c via the absorbent port 130.

When the pump 111 is completely dormant, the flow of absorbent solution to the absorber sprays 101 via the line 28a is halted, thereby reducing the flow rate ratio to zero. An alternative method of regulating the ratio of concentrated absorbent solution entering the contact portion 24a versus the absorber sump 24c is to provide a pump in the line 28b rather than in the line 28a, or to provide separate pumps in each line. By varying pump capacities, flow to the absorber is regulated in a manner similar to that described above.

The refrigerant vapor from the evaporator 16 contacts the concentrated absorbent solution as it leaves the absorber sprays 101. The resulting absorption of the refrigerant vapor into the less-volatile liquid both condenses the vapor, releasing its heat of vaporization, and releases heat of dissolution as the result of the absorption process.

The resulting heat is rejected via the line 30 to the heat sink 14.

The pump 103 withdraws the reconstituted composite refrigerant through the composite refrigerant outlet 107 and forces it to the generator 26 via the line 34. The heat exchanger 35 preheats the composite refrigerant traversing the line 34 before it enters the generator 26, using heat which otherwise would escape from the generator via the concentrated absorbent solution line 28.

Heat from the load 12 enters the evaporator 16 of the apparatus via the heat load line 18. The evaporator 16 is in heat-transfer contact with the heat load 12 to accomplish this heat transfer. The temperature sensor 115 senses the temperature of the fluid traversing the heat load line 18 as the fluid flows from the heat load 12. The temperature sensor 116 senses the temperature of the fluid traversing the heat load line 18 as the fluid flows to the heat load 12. The control line 133 and the control line 135 respectively connect the sensor 115 and the sensor 116 to the controller 118. The control line 137 connects the controller 118 to an adjustable frequency drive 120. The control line 139 connects the adjustable frequency drive 120 to the pump 111. The adjustable frequency drive 120 controls the pumping rate of the pump 111 according to the frequency of the AC electrical power supplied to the drive 120.

The controller 118 is also connected to the sensor 117 by a control line 140. The temperature sensor 117 senses the temperature in the line 30 through which heat is transferred from the absorber 24 to the heat sink 14. The controller 118 also uses the temperature in the line 30 sensed by the temperature sensor 117 to control the adjustable frequency drive 120.

When the controller 118 senses a change in the necessary heat load via the temperature sensors 115, 116, and 117, it modulates the adjustable frequency drive 120, and thus the pump 111, accordingly. If the controller 118 detects an increase in the heat load, for example, it increases the frequency of the adjustable frequency drive 120, which increases the speed of the pump 111, which in turn increases the flow of concentrated absorbent solution to the absorber sprays 101, which then increases the pressure differential in the absorber 24, which causes more refrigerant vapor to flow from the evaporator 16 to the absorber 24, which causes refrigerant to evaporate at a higher rate in the evaporator 16, which promptly increases the refrigeration capacity of the system 10. Conversely, when the controller 118 detects a decrease in the heat load, it decreases the speed of the pump 111 to decrease the refrigeration capacity of the system 10. If the controller 118 decreases the speed of the pump 111 and thus the flow rate ratio to zero, it stops all flow to the absorber sprays 101, thereby effectively reducing refrigeration capacity to zero.

The pump 103, which withdraws reconstituted composite refrigerant from the absorber sump 24c and forces it to the generator 26, may be a variable capacity pump that further varies the refrigeration capacity of the apparatus by varying the total flow rate of absorbent solution and composite refrigerant flowing to and from the absorber 24. An alternate method of further varying refrigeration capacity in this manner is to provide a variable capacity pump in the line 28 rather than in the line 34, or to provide separate pumps in each line.

FIG. 3 shows an alternate embodiment of the invention. In FIG. 3, the controller 118 is connected to a flow regulator valve 111 via a control line 139. The flow regulator valve 111 divides the relatively constant flow of concentrated absorbent solution from the generator 26 via the line 28 into a flow to the absorber sprays 101 via the line 28a and a flow to the absorber sump 24c via the line 28b and the absorbent port 130.

The controller 118 controls the flow regulator valve 111, varying the ratio of concentrated absorbent solution flowing from the generator 26 to the absorber sprays 101 with that flowing from the generator 26 to the absorber sump 24c. The flow regulator valve 111 can be a proportioning valve, a simple valve in either of the lines 28a and 28c, or separate valves in both lines.

When the controller 118 in this embodiment senses a change in the heat load, it adjusts the flow regulator valve 111 accordingly. If the controller 118 detects a decrease in the heat load, for example, it adjusts the flow regulator valve 111 to divert some or all of the concentrated absorbent solution traversing the line 28 into the absorber sump 24c, which in turn decreases the flow of concentrated absorbent solution to the absorber sprays 101, which then decreases the pressure differential in the absorber 24, which causes less refrigerant vapor to flow from the evaporator 16 to the absorber 24, which causes refrigerant to evaporate at a lower rate in the evaporator 16, which finally allows for a smaller refrigeration capacity.

Thus, capacity control of absorption refrigeration apparatus has been shown which is faster in responding to load changes than previous systems and methods, particularly when the apparatus has been operating at a low refrigeration capacity. In addition, the described apparatus will function in an extended operating range without sacrificing responsiveness or requiring an enlarged refrigerant storage tank. To reach this range of low capacity operation without modulating flow to the absorber sprays as described, a comparable system would require a refrigerant storage tank that is more than three times as large. This apparatus and method, one particular embodiment of which has been built, will respond faster than prior apparatus and methods without requiring such a significant increase in reservoir size.

Referring now to Fig. 4, a second aspect of the present invention is disclosed. A two-stage absorption refrigeration machine 400 includes a low temperature generator 401 and a condenser 402 which are enclosed in a first fluid tight shell 403. A second fluid tight shell 404 contains an evaporator 405 and an absorber 406. A high temperature generator 407 is enclosed in a third fluid tight shell 408. The absorber 406 contains a heat exchanger 409 which is supplied with cooling fluid through a passage 410, which also passes through the condenser 402, from a cooling fluid source (not shown).

The cooling fluid leaves the absorber heat exchanger 409 via the passage 410 and enters condenser heat exchanger 411 and is returned to the cooling fluid source (not shown).

Various suitable types of refrigerants and absorbents may be used in the present two-stage absorption machine. A solution of lithium bromide absorbent in a refrigerant such as water is satisfactory. The term "concentrated solution" as used herein means a solution which is concentrated in absorbent. A "dilute solution" is one which is dilute in absorbent.

Steam flows from a source such as a boiler (not shown), through the high temperature generator heat exchanger 412 of the high temperature generator 407 via a steam passage 413. The steam passage 413 returns condensate to the steam source through a condensate heat exchanger 414. It is, of course, understood that other suitable sources of heat can be used to concentrate absorbent solution in the high temperature generator 407. (For example, the high temperature generator might be directly heated by a burner). Heat from condensing steam in the high temperature generator heat exchanger 412 causes the refrigerant solution in the high temperature generator 407 to boil, thereby producing refrigerant vapor and concentrating the absorbent solution.

Refrigerant vapor produced in the high temperature generator 407 is directed to the low temperature generator heat exchanger 456 in heat the low temperature generator 401 through a refrigerant vapor passage 415 and is then condensed in the condenser 402. The dilute solution in the low temperature generator 401 is boiled through heat exchange with the refrigerant vapor in the refrigerant vapor passage 415 and is also condensed in the condenser 402. At least a portion of the concentrated solution generated in the low temperature generator 401 is delivered through a fourth passage 416 to a high temperature generator pump 417 and pumped through a passage 418 to a high temperature heat exchanger 419. In the high temperature heat exchanger 419 at least a portion of the concentrated solution in passage 418 is preheated on its way to the high temperature generator 407 through heat exchange with the high temperature concentrated solution flowing in a passage 420. A portion of the concentrated solution flowing in the passage 418 is directed through a passage 421 to the condensate heat exchanger 414 where it is brought into a heat exchange relationship with the condensate in the condensate passage 413 before rejoining the solution in the passage 418 and being delivered to the high temperature generator 407.

The high temperature concentrated solution is directed from the high temperature generator 407 through the high temperature heat exchanger 419 to a high temperature concentrated solution accumulator 422 via the passage 420. The high temperature concentrated solution from the accumulator 422 is directed through a third passage 423 to be joined with low temperature concentrated solution leaving the low temperature generator 401 via a second passage 424 at a mixing point 425. From the mixing point 425, the combined concentrated solution is directed to the low temperature heat exchanger 426 via a first passage 427, and subsequently delivered to the absorber 406 through a fifth passage 428, an absorber spray pump 429, and a passage 430.

Liquid refrigerant from the condenser 402 passes through a passage 431 to the evaporator 405. The liquid refrigerant is vaporized in the evaporator 405, thus removing heat from a chilled fluid flowing in a passage 432 through the evaporator heat exchanger 433. This chilled fluid is circulated to a heat load, such as a building requiring cooling.

Since the absorber 406 is in vapor communication with the evaporator 405, the absorbent solution can absorb refrigerant vapor from the evaporator 405, thus removing heat from the evaporator section. At least a portion of the refrigerant liquid dropping from the evaporator heat exchanger 433 is collected in a collector 434. The refrigerant liquid flows from the collector 434 through a passage 435 to a storage vessel 436. Via a passage 437, refrigerant liquid is delivered from the storage vessel 436 to the evaporator spray pump 438 which delivers the refrigerant liquid through a passage 439 to be sprayed in the evaporator 405 through nozzles 440.

Dilute solution from the absorber 406 flows through a passage 441, a low temperature generator pump 442, a passage 443, the low temperature heat exchanger 426 and a passage 444 to the low temperature generator 401, where it is concentrated. In the low temperature heat exchanger 426 the concentrated solution is brought into a heat exchange relationship with the dilute solution from the absorber 406 which is being delivered to the low temperature generator 401, whereby the dilute solution is preheated.

From the low temperature heat exchanger 426 the concentrated solution flows through fifth passage 428 to the absorber spray pump 429. The concentrated solution is forced by the absorber spray pump 429 through the passage 430 and is discharged into the absorber 406 through the absorber spray nozzles 445. A passage 446 is disposed between the storage vessel 436 and the absorber spray pump 429. Flow between the storage vessel 436 and the absorber spray pump 429 is controlled by a valve 447 which is normally closed.

When crystallization occurs in the concentrated solution in the low temperature heat exchanger 426, then the flow of concentrated solution in the first passage 427 is reversed due to crystallization blockage. This effect makes it possible to detect crystallization by monitoring the temperatures of certain solution streams.

A fourth passage temperature sensor 448 senses the temperature of the concentrated solution between the low temperature generator 401 and the high temperature generator pump 417 in the fourth passage 416. A second passage temperature sensor 449 senses the temperature of the concentrated solution in the second passage 424 between the low temperature generator 401 and the mixing point 425. A third passage temperature sensor 450 senses the temperature of the high temperature concentrated solution in the third passage 423. A fifth passage temperature sensor 451 senses the temperature of the concentrated solution in the fifth passage 428.

Operation of the absorption refrigeration machine is typically controller 453 having processing circuitry, for example, a microprocessor. The controller 453 may be of the feedback type that incorporates an input signal receiver 454 and an output signal generator 455. Output control signals are generated by the signal generator 455 in response to input signals received by the input signal receiver 454.

During normal steady state operation, the temperature sensed by the fourth passage temperature sensor 448 is substantially equal to the temperature sensed by the second passage temperature sensor 449, and the control system modulates the heat input into the high temperature generator 407 such that the temperature of the concentrated solution in the fifth passage 428 sensed by the fifth passage temperature sensor 451 is maintained at about 15OF. higher than the concentrated solution's crystallization temperature.

When crystallization of the concentrated solution in the low temperature heat exchanger 426 occurs, the flow through the first passage 427 begins to reverse due to blockage. Accordingly, the temperature sensed by the second passage temperature sensor 449 begins to exceed that sensed by fourth passage temperature sensor 448.

According the present invention, a crystallization alert is issued and corrective action is indicated when the temperature sensed by the second passage temperature sensor 449 meets or exceeds the mathematical average of the temperature sensed by the fourth passage temperature sensor 448 and the temperature sensed by the third passage temperature sensor 450 as determined by the following formula: T3+T4 TTRIP=- 2 Where: T3 is the temperature sensed by the third passage temperature sensor; and T4 is the temperature sensed by the fourth passage temperature sensor; This temperature value is called the "trip temperature." If the temperature sensed by the second passage temperature sensor 449 meets or exceeds the trip temperature, then the control system begins to take action to recover from crystallization of the concentrated solution in the low temperature heat exchanger 426.

In the recovery mode, the control system takes the following steps: 1. heat sources to the low temperature generator 401 and the high temperature generator 407 are deactivated to stop producing concentrated solution. In the embodiment of FIG. 4, this would be achieved by interrupting the steam supply to the high temperature generator 407 by closing a steam valve 452; 2. the circulation of absorbent solution is stopped by deactivating the high temperature generator pump 417, the low temperature generator pump 442, the absorber spray pump 429 and the evaporator spray pump 438; 3. the solution in the absorber is diluted with dilute solution by opening the valve 447 to allow dilute solution from the storage vessel 436 to flow to the absorber spray pump 429; 4. after about 3 minutes, the concentrated solution that in crystallized is flushed by reactivating the low temperature generator pump 442 and the high temperature generator pump 417 for about 5 minutes; 5. the low temperature generator pump 442 and the high temperature generator pump 417 are again deactivated for about 3 minutes (this is done because the reactivation of these pumps according to step 4 may result in temporary recrystallization); 6. the heat source to the high temperature generator 407 is reactivated.

However, the control system adjusts the heat input to the high temperature generator 407 and the low temperature generator 401 such that the temperature of the concentrated solution leaving the low temperature heat exchanger 426 via the fifth passage 428 is maintained at a level such that the margin between the crystallization temperature of the concentrated solution in this region and the actual temperature is increased by about 5"F. (3"C.) over the previous control margin; and 7. after the 3 minute period according to step 5 has elapsed, all pumps are reactivated.

The control system is programmed to allow the system to go through the recovery sequence a predetermined number of times, e.g., twice. If crystallization is detected more times than this predetermined number, the control system will shut down all operation of the absorption machine so that required maintenance can be performed to correct the recurring crystallization.

FIG. 5 is a flowchart illustration of the recovery sequence of the present invention described above.

The following describes another aspect to the present invention. The main function of a control system applied to a water chiller, for example, is process regulation.

Process regulation involves maintaining chilled leaving water at a desired temperature.

Thus, the control system must respond quickly to process disturbances to minimize the magnitude and the duration of differences between the leaving water temperature and the desired setpoint. To minimize these differences, the control response can be adjusted to favor the integral action.

As set forth above, problems may be encountered in certain instances. For example, when setpoint changes are made, and during setup, high integral gains cause the system to overshoot the setpoint significantly. By filtering setpoint changes, overshoot can be eliminated without reducing the response to process disturbances.

Therefore, the setpoint filter of the present invention has another functional advantage. During the prestart sequence of the chiller, the filtered setpoint is initialized to the current leaving water temperature. When closed loop control takes over, the initial error in leaving water temperature is zero. As the filtered setpoint approaches the desired setpoint, a small error will be detected by closed loop control. The control system will track the filtered setpoint as it changes over time. The time it takes the filtered setpoint to reach the desired setpoint is the settling time. By allowing the settling time to be an adjustable input, the filtered setpoint will replace what is known as a "softloading" function. The advantage of using the filtered setpoint for softloading on the front end of the control system is that it uses the closed loop controls that are already in place for the leaving water temperature control. This simplifies implementation and verification of the function.

An embodiment of a second order binomial filter designed to prefilter the control setpoint is shown by the block diagram in FIG. 6. A control system 500 is illustrated. The control system 500 includes a feedback loop 505 with a summing node 508. FIG. 6 also illustrates a binomial filter referenced generally at 510. The binomial filter 510 is connected and arranged to receive a setpoint input 515. The binomial setpoint filter 510 subsequently provides an output of a filtered setpoint referenced 520.

The filtered setpoint 520 is connected via the summing node 508 as an input to the control system 500 and feedback loop 505. As shown in FIG. 6, the binomial setpoint filter 510 is arranged at the input side of the control system 500.

FIG. 7 illustrates responses to a step input. For example, the ideal response of the control system to the step input is a critically-damped to slightly underdamped second order function (see FIG. 7). The filter receives a step input (in a typical method, the leaving water temperature setpoint is changed) and outputs a critically-damped second order output. If the settling time of the filter is small enough that the cut-off frequency is within the bandwidth of the open loop system, prefiltering the setpoint does not increase the settling time.

As shown in FIG. 7, the input is a step input indicated by reference letter I. A response curve of a first order filter is indicated by a reference letter F. The first order filter response F has a steep slope near the origin which causes an abrupt discontinuity at the beginning of the step input. In contrast to the first order response, a second order response curve indicated by reference letter S is also shown. As indicated in FIG. 7, the second order response curve S has a gradual slope at the beginning of the curve to provide a smoother transition in response to the step input. As can be seen from FIG. 7, the first order response F converges to the step input I slightly sooner than the second order response S. However, the second order response S is within acceptable response time limits.

FIG. 8 is a graph illustrating a series of curves. The curves illustrate the operation of the binomial setpoint filter 510 in an absorption water chiller operating in accordance with the apparatus and method of the present invention. The graph of FIG.

8 indicates temperature on the vertical axis and elapsed time on the horizontal axis.

As illustrated, the first few minutes involve starting the burner and the pumps in the chiller. For the first 12 minutes, the chiller is being preheated and the system is run for the time period indicated. At approximately the 12 minute point, the controls are released and the binomial setpoint filter 510 is initialized. to be the same as the leaving water temperature. Once the controls are released, it is preferred that the system follow the filtered setpoint as indicated in FIG. 8.

In FIG. 8, the water begins at approximately 82OF. but the filter setpoint indicated by FS comes down to 440F. The chilled water leaving the evaporator curve referenced CwL also nearly follows the filtered setpoint FS downward as indicated in FIG. 8. The error between the filtered setpoint FS and the chiller water leaving CWL is the feedback to the control system. FIG. 8 shows that the error is not as large as if the chiller had just started at the 44" setpoint and the 80-83° chilled water temperature. Thus, the binomial setpoint filtering within the embodiment of the present invention eliminates the occurrence of a large error at the initial startup. Binomial filtering of the setpoint brings the chilled water temperature down at a more gradual, smoother rate. In addition, as shown in FIG. 8, there is no overshoot of the chilled water leaving evaporator curve CWL versus the filtered setpoint curve FS. Chilled water entering the evaporator curve CWE is also shown. The chilled water entering evaporator curve CWE illustrates that a loop controller tries to maintain the entering temperature once it reaches approximately the 51° level as indicated in FIG. 8. FIG. 8 shows a slight bit of overshoot in the chilled water entering the evaporator. Then a load is added as the water temperature comes down.

Since the binomial setpoint filter 510 is arranged at the front end or input of the control system (see FIG. 6), as opposed to at the output, the response is gradual as shown in FIG. 7. Without the binomial filter 510, the absorber chiller would have immediately tried to load up to 100% due to the 44" setpoint differential. As the water temperature came at a fast rate, the chiller would then try to limit the loading. In certain slower systems, that sequence does not work well because the chiller receives a full load before the water temperature would change. Thus the prior manner of softloading is more reactionary and difficult to implement. In contrast however, the binomial setpoint prefiltering at the input stage of the control system overcomes these problems as discussed above.

FIG. 9 graphically illustrates the situation in which an absorption chiller, for example, has a constant load and is then subjected to a drop in the setpoint. In FIG.

9, the drop is from approximately 550 down to 49.50. Also, the chilled water entering the evaporator according to the curve CWE is illustrated. FIG. 9 is similar to the response illustrated in FIG. 7, only in an inverted manner.

FIG. 9 illustrates a situation where an operator resets the water temperature in the morning, for example, after having the chiller temperature set for comfort cooling at a higher temperature setpoint during the night. Thus, the user desires to lower the cooling water temperature during the day.

Thus, FIG. 9 shows that the setpoint has decreased from 55OF. down to 49.5OF. The second order response of the binomial setpoint filter similar to that of FIG.

7 is shown. The water temperature comes down at almost a steady rate and gradually comes into the setpoint. The dotted line represents the chilled water leaving the evaporator. Without the binomial setpoint filtering, the initial decrease in water temperature would cause a 60OF. temperature error all at one time. However, the filtering of the setpoint using the binomial setpoint filter of the present invention increases the load only slightly represented as a series of smaller changes instead of pulling the chiller to full load immediately. Thus, the present invention acts more like a series of process disturbances all in a row instead of one large jump that kicks the chiller into a 100% operative state. As a result, the present invention avoids loading up the chiller to 100% and then having the temperature come down so quickly that the chiller backs off in a rapid manner in response.

The actual filter setpoint will converge in the settling time, but nothing is done to reset it until the machine is turned off. The chiller is then set to the actual leaving water temperature. Otherwise, the chiller just follows the actual setpoint.

An additional advantage of the present invention is that in the opposite situation, for example, changing the setpoint from 49OF. to 550F. the present invention operates equally beneficially. For example, on most known chillers there is something known as a differential stop. If one tries to raise the setpoint greater than a certain amount representing the differential stop, the machine shuts off immediately because of a violation of the differential stop maximum. However, because of the gradual nature of the binomial setpoint filter as explained above, the chiller operating in accordance with the principles of an embodiment of the present invention would gradually increase the load and raise the leaving water temperature without shutting off.

FIG. 10 is another embodiment of the present invention. FIG. 10 illustrates an absorber chiller as part of a larger process. For example, the absorber indicated in FIG. 10 incorporates a binomial setpoint filter 510 as explained above. In this application, the absorption chiller is just part of the overall system of a larger scale. In this application, chiller controls may be required to perform more of a tracking control function as indicated in FIG. 10. In this type of application, the response of the chiller is defined by the binomial setpoint filter 510. This makes the design of the process controls illustrated in FIG. 10 easier because the dynamics of the chiller are already well defined.

Thus, the present invention makes use of filtering a setpoint to avoid overshoot while maintaining response to process disturbances. As discussed above, a first order filter could be used; however, comparison of the responses of a first and a second order filter with reference to FIG. 7 shows that the second order filter has a smoother initial response. In contrast, the initial response of the first order filter is rather abrupt.

A first order filter or a second order filter can be implemented digitally.

This is beneficial since the digital implementation and mathematical representation of the filters can be easily programmable using a computer. Further, such programmability allows a microprocessor to be used in the control system. The following methodology can be used to develop a discrete representation of a first order digital filter. As set forth above, the result is in a form that is easily programmable using a microprocessor, or the like.

To begin with, a first order Laplace transfer function is used: G(s) = a where a is the cut off frequency of the filter. s+a Next, the impulse transfer function is calculated from the following: residues os F(p)(1/1-e-(s-p).T) F(p) = (1-sSt/s)(a/s+a)s=p p where F(p) is the Laplace transform of a zero order hold with s replaced by p.

The poles of F(p) are 0 and -a. Therefore, the impulse transfer function is: <BR> <BR> <BR> <BR> #(s)=[(1-e-s.T/p)(a)(1/1-e-(s-p).T)]p=-a+[(1-e-s.T)(a/p+a)(1 /1-e-(s-p).T)]p=0<BR> <BR> <BR> <BR> <BR> <BR> <BR> #(s)=[(1-e-s.T)(-1/1-e-(s+a).T)]=[(1-e-s.T)(1/1-e-s.T)] #(s)=e-s-.T(1-e-a.T)/1-s-s.T . e-a.T) Conversion to the Z domain is performed by substituting as follows: z-1=e-s.T #(z)=(z-1)(1-e-a-.T)/1-(z-1).e-a.T Finally, conversion to programmable form is performed by: r(z) = O(z)/I(z) O(z)-[1-(z-1).e-a.T]=I(z).[(z-1).(1-e-a.T)] O(z) [1-e-a.T) 1n.i + e-a.T. °n-l n = 2...10 step:= 10...100 Let Settling time:=5-60 minutes O:= 0 In-1:= 0 #c:= 5 Settling time O1:= 0 Istep:= 10 T:= 5 sec Timen:= n.T/60 <BR> <BR> <BR> <BR> <BR> <BR> α:=#c<BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> On:= (1-e-α.T) . In-1 + e-α.T . On-1 Thus, the above equation can be implemented digitally in a programmable form on a computer or microprocessor. Similarly, a binomial filter can be represented in programmable form. A binomial filter is defined as having identical and real pole locations. A binomial filter also has the features of slow response with no overshoot.

The programmable form of a binomial filter is determined in the same basic manner set forth above with respect to the first order filter. The main equations are set forth below.

Laplace form: G(s) = #c²/(s+#c)² Z transform: G(z) = z . (1-e-T.#c)-#c . T . -e-T.#c) + e-T.#c . (e-T.#c-1+#c . T)/(z-e-#c.T)² An algorithm for an embodiment of the second order filter 510 follows: Cutoff frequency = 5 Settling Time *60 a = eAt*Cutoff frequency Coeff 1 = 1 - a - Cutoff frequency*At*a Coeff 2 = -a *(l - a - Cutoff frequency*#t) Coeff 3 = 2*α Coeff 4 = l-Coeff 1 - Coeff 2 - Coeff 3 where At = Cycle Time The calculation of coefficient (a can be approximated by a series expansion. The first 3 terms of the expansion yield adequate results: a = 1-A2t*Cutoff frequency + Qt*At*Cutoff frequency*Cutoff frequency/2.

Thus, the binomial filter represented in programmable form is as follows: Filtered Setpointn = Coeff 1*Setpointn-1 + Coeff 2*Setpointn2 + Coeff 3*Filtered Setpointn, + Coeff 4*Filtered Setpoint.2 Also, a first order plant with a PID controller will behave as a second order function. Thus, the system will naturally follow the second order setpoint. Response of a second order function can be characterized by specifying the natural frequency and damping for the function. By choosing a binomial function (two identical poles) the response is critically damped, which means that the response is damped as fast as possible without overshoot. Thus, only one parameter needs to be set, the settling time.

By prefiltering the setpoint, the control system does not see a sudden, large error in the leaving water temperature when the setpoint is changed. When a step change is made to the leaving water temperature setpoint, the filtered setpoint changes a small fraction of the step every control cycle, thus the control system sees only a small error in the leaving water temperature and reacts accordingly. As the filtered setpoint continues to change, the control system will see a small but persistent error term and will continue to change the leaving water temperature. The filtered setpoint will start to approach the actual setpoint asymptotically. Because the filtered setpoint slowly approaches the actual setpoint, the control will react to prevent the leaving water temperature from overshooting the desired setpoint (or at least will minimize the overshoot).

While particular elements, embodiments and applications of the present invention have been shown and described, it will be understood, of course, that the invention is not limited thereto since modifications may be made by those skilled in the art, particularly in light of the foregoing teachings. It is therefore contemplated by the appended claims to cover such modifications as incorporate those features which come within the spirit and scope of the invention.