JPH10267452A - Prediction device of cooling load of absorption refrigerating machines and device for controlling number of machines - Google Patents

Abstract

PROBLEM TO BE SOLVED: To properly predict a cooling load regardless of the fluctuation in a large cooling load by obtaining a future prediction load based on a load change pattern during the past specific time of a cooling load and at the same time obtaining a weight coefficient from the deviation with an actual cooling load and then calculating a future prediction load. SOLUTION: Cold water obtained from a plurality of absorption refrigerating machines A-C is guided to a cooling load 5 via an outlet assembly pipe 4 and is allowed to flow back through an inlet assembly pipe 6. In this case, the assembly pipes 4 and 6 have temperature sensors 9 and 10, input each detection signal to a processing circuit 11, where a cooling load is detected. Also, a future, first prediction load is estimated based on a load change pattern during the past, first time of the cooling load and at the same time a future, second prediction load is estimated based on the load change pattern during a second time exceeding the past, first time of the cooling load. Then, the weighting coefficient of each prediction load is obtained based on a deviation between the prediction load and an actual cooling load, thus calculating a future, third prediction load.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for predicting the cooling load of an absorption chiller, and to a device for controlling the number of absorption chillers during cooling.

[0002]

2. Description of the Related Art Absorption chillers require a longer time to start and stop compared to compression chillers used mainly for home use, and therefore, unit control is behind the rapid cooling load fluctuation. This causes a state in which the temperature of the chilled water greatly fluctuates. In a large air conditioning system that cools with multiple absorption chillers, such as district cooling and factory cooling, how multiple absorption chillers cooperate in response to disturbance factors such as cooling load changes. It is important to keep the cold water temperature constant.

[0003] A typical prior art is disclosed in JP-B-61-121.
77. In this prior art, the chilled water temperature at the outlet collecting pipe or the inlet collecting pipe of a plurality of absorption chillers is measured, and a gradient of the chilled water temperature within a predetermined time, for example, 5 minutes, is calculated. Stopping.

In this prior art, it is assumed that the cooling load is constant during the predetermined period in which the cooling load is predicted. Therefore, when the cooling load suddenly changes, the prediction system decreases.

[0005] Another prior art is disclosed in Japanese Patent Application Laid-Open No. 5-141806.
Is disclosed. In this prior art, when one absorption refrigerator is stopped due to a decrease in cooling load, a control valve for controlling the heat source heat quantity of a high-temperature regenerator of the absorption refrigerator that continues to operate with a stop command is forcibly opened to reduce the number of units. This prevents rapid fluctuations in refrigeration capacity, thereby preventing fluctuations in chilled water temperature.

[0006] Further, in this prior art, the refrigerating capacity cannot be quickly changed in response to a large fluctuation of the cooling load, and it is difficult to maintain the chilled water temperature at a predetermined set temperature.

[0007]

SUMMARY OF THE INVENTION It is an object of the present invention to provide an absorption chiller which can reliably predict the cooling load with respect to a large fluctuation of the cooling load and can quickly control the number of units. An object of the present invention is to provide a cooling load prediction device and a number control device.

[0008]

According to the present invention, cold water obtained from the evaporators of a plurality of absorption chillers is led to a cooling load via an outlet collecting pipe, and chilled water from the load is returned from the inlet collecting pipe and circulated. In the prediction device of the cooling load of the absorption refrigerator
Means for detecting a cooling load; and first estimating means for responding to an output of the load detecting means and estimating a future first predicted load based on a past load change pattern of the cooling load during a first predetermined time. Second estimating means for responding to the output of the load detecting means and estimating a future second predicted load based on a load change pattern of a cooling load during a predetermined second time exceeding the past first time; In response to the outputs of the first and second estimating means, the first and second predicted loads Qx (t), Qy
Based on the predicted load errors X and Y between (t) and the detected load Qr (t), the first and second predicted loads Qx
Weight coefficient calculating means for calculating weight coefficients Kx and Ky of (t) and Qy (t); and a third predicted load Q in response to outputs of the first and second estimating means and weight coefficient calculating means.
And a prediction load calculating means for calculating the cooling load of the absorption chiller. To quickly control the number of units in response to a change in cooling load, it is very effective to predict a change in cooling load in the future. Therefore, in the present invention, the future cooling load is predicted based on the cooling load from the past to the present, and the future cooling load data thus predicted is used for controlling the number of units. Even when a large change in the cooling load occurs, the temperature of the cold water is kept constant. Claim 1
According to the present invention, in order to predict the future cooling load based on the cooling load up to the present, two types of calculations are performed by the first and second estimating means, and the first and second predictions thus obtained are performed. The load is combined to calculate and estimate a highly accurate future third predicted load Q. In the first estimating means, for example, based on a change tendency of the cooling load data during a predetermined first time in the past, for example, the past 10 minutes to the past 30 minutes, that is, a load change pattern, a first prediction of a future cooling load is made. Estimate the load. According to the first estimating means, a relatively near future cooling load can be accurately estimated. In the second estimating means, based on the change tendency of the cooling load data for one day, that is, the load change pattern during the second time, for example, a period from the present in the middle of the second time to the end of the second time. Is estimated. According to the second estimating means, for example, the concept of the time of the day, which is the second time, is included, so that the daily load change patterns are relatively similar in applications such as industrial cooling. In addition, since the cooling load changes rapidly, this prediction method is very effective. In this manner, the first estimating means performs a short-term prediction of the cooling load based on a load fluctuation pattern that is, for example, local load fluctuation data for the past 10 minutes, and calculates a change in the cooling load that has a precursor to the fluctuation of the cooling load. Prediction is possible with high accuracy. On the other hand, for example, prediction is performed on the basis of a cooling change pattern which is global cooling load change data for one day, and a cooling load change with a fixed time can be predicted with high accuracy. Since the first and second predicted loads by the first and second estimating means change depending on the operation mode and the cooling operation state, the prediction result of either of the first and second estimating means is determined by the actual Whether it is correct or not approximate to the cooling load changes every moment. Therefore, in order to obtain a more reliable cooling prediction load, it is necessary to obtain the obtained first and second cooling loads.
A weight coefficient Kx, Ky is calculated for the predicted load, thereby obtaining a highly accurate third predicted load Q. In this way, it is possible to quickly change the number of units in response to a change in the cooling load, and it is possible to suppress a change in the chilled water temperature due to an excessive or insufficient cooling capacity.

Further, the present invention is characterized in that the first estimating means estimates a first predicted load by constructing a nonlinear function of a neural network for performing a learning procedure. Claim 2
According to the present invention, the cooling load is predicted by the neural network of the first estimating means, and the neural network calculates the first predicted load from various cooling load change patterns by performing a procedure called learning. Can be easily constructed. The first predicted load in the future can be obtained by inputting the detected cooling load change pattern into the constructed nonlinear function. In this way, it is possible to estimate the first predicted load in response to a sudden change in the cooling load.

Further, according to the present invention, the second estimating means includes:
A first memory for storing a plurality of types of predetermined load change patterns in the past over a period of time, a second memory for storing past detected load change patterns in a period less than the second time, and a first and a second memory. In response to the contents stored in the memory, a load change pattern most similar to the load change pattern stored in the second memory is selected from a plurality of types of load change patterns stored in the first memory, and the selected load change pattern is selected. Selecting means for setting the load change pattern as the second predicted load. According to the third aspect of the present invention, in the second estimating means, the change tendency of the cooling load data detected by the cooling load detecting means for, for example, less than one day is stored in the second memory. Comparing the plurality of types of cooling load data of the past day stored in the first memory with the cooling load data of less than one day at each corresponding time, and comparing the plurality of types of cooling load data stored in the first memory. The most similar past cooling load data among the types of cooling load data is used to predict the future cooling load in the remaining time of the day. That is, the load change patterns of the past several cases are compared with the current load change pattern, and the most similar load change pattern is selected.
The second predicted load in the future is estimated based on the selected load change pattern.

Further, according to the present invention, the selection means includes a load Qmk (t) for each time between a load change pattern of each type stored in the first memory and a detected load pattern stored in the second memory. , Qr (t) absolute value | Qr
Calculate the value corresponding to the reciprocal of (t) -Qmk (t) |
A degree of coincidence Ek obtained by adding a value corresponding to the reciprocal of the absolute value | Qr (t) −Qmk (t) | of the difference over a period of less than a second time is calculated, and the plurality of types of load change patterns are calculated. Among them, a load change pattern of the type having the highest degree of coincidence Ek is set as a second predicted load. According to the fourth aspect of the present invention, when selecting the above-described closest load change pattern by the selecting means provided in the second estimating means, the degree of coincidence Ek is calculated and determined.
The degree of coincidence Ek is determined by the actual cooling load Qr (t) detected by the cooling load detecting means at the current time t and the current time tr in each of the above-described types of load change patterns k.
Are based on the cooling load value Qmk (t) in
Over a day, it is a value obtained by adding the reciprocal of the absolute value of the difference at each time from the first midnight to the current time t. The type of load change pattern having the highest degree of coincidence Ek is selected, and the selected type of load change pattern is used as the second predicted load.

According to the present invention, in the weighting coefficient calculating means, the predicted load errors X and Y are as follows: X = | Qx (t) -Qr (t) | Y = | Qy (t) -Qr (t) | The weighting coefficients Kx and Ky are as follows: Kx = Y / (X + Y) Ky = X / (X + Y)
Is characterized in that Q = Kx · Qx (t) + Ky · Qy (t). According to the invention of claim 5,
In order to obtain a more accurate third predicted load Q by weighting the first predicted load and the second predicted load based on the weighting factors Kx and Ky, any one of the predicted values of the first and second predicted loads is obtained. Is more likely to be determined, and the more likely predicted values are weighted heavily. For example, in obtaining the third predicted load Q by combining the obtained two types of first and second predicted loads, the first and second predicted loads do not always have the same value, and in some cases, are completely different. It is conceivable that diametrically opposite predicted values are output from the first and second estimating means, respectively. The completely opposite predicted value corresponds to a case where one of the first and second predicted loads is predicted to increase, but the other predicted load is estimated to decrease. According to the fifth aspect of the present invention, even in such a case, the highly accurate third predicted load Q can be quantitatively obtained.

Further, the present invention relates to an absorption refrigerator in which chilled water obtained from an evaporator of each of a plurality of absorption chillers is led to a cooling load through an outlet manifold, and chilled water from the load is returned from the inlet manifold to circulate. In the number control device, a prediction device that calculates a predicted load of a cooling load, an operating number detection unit that detects the current number of operating absorption chillers, and an absorption chiller that responds to an output of the prediction device and corresponds to the predicted load Responding to the outputs of the first operating number calculating means for calculating the predicted operating number of the vehicle, the operating number detecting means and the first operating number calculating means. Preparation state setting means for setting the absorption chiller in the start preparation state, and when the predicted operation number is less than the current operation number, setting the absorption chiller in operation to the stop preparation state; and Detect Cooling load detecting means, a second operating number calculating means responsive to an output from the cooling load detecting means, and calculating a required operating number of absorption chillers corresponding to the detected cooling load; In response to the output from the second operating number calculating means, when the required operating number exceeds the operating number at the time when the predicted operating number reaches the predicted time, the absorption chiller in the start preparation state is started. And starting / stopping means for stopping the absorption chiller in the stop preparation state when the required operating number is less than the operating number at the time when the predicted operating number is reached. This is a device for controlling the number of absorption refrigerators. Further, according to the present invention, the start / stop unit is configured such that when the operating state in which the predicted number of operating units is equal to the number of operating units at the time of the predicted arrival has elapsed for a predetermined time u or more, the start-up preparation state is set by the preparation state setting unit. The present invention is characterized in that the refrigerator is stopped, or the absorption refrigerator that has been in the stop preparation state is kept running. According to the present invention, for example, the first operating number based on the future third predicted load Q obtained by the cooling load prediction device according to claims 1 to 5 or the predicted load obtained by another method. The predicted number of operating units is calculated by the calculating means, and the preparation state setting means prepares for starting or stopping the absorption refrigerator by the predicted load. Thereby, thereafter, the measured load value of the cooling load detected by the cooling load detecting unit is:
At the point in time when it is as expected, the absorption refrigerator set in the ready state can be immediately started or stopped. The startup preparation state of the absorption refrigerator means that the heat source 28 provided in the regenerator is in an active state, that is, when the heat source is, for example, a burner, the fuel is ignited and heated, and LiBr that has absorbed water as a refrigerant is heated. The absorbent, which is an aqueous solution, is heated to expel the absorbed water from the solution and to suppress the generation of water vapor in the evaporator. In order to suppress the generation of water vapor in the evaporator, the spraying device 19 for injecting water into the heat transfer tube 18 for guiding the cold water is put into a resting state, and for example, water stored in the evaporator is sent to the spraying device. Shut off the guiding pump 20. Therefore, in the absorption refrigerator in the startup preparation state, when a startup command is generated, the above-described pump is driven, for example, to drive water from the spraying device to the cold water heat transfer tube in the evaporator immediately. Thus, quick startup of the absorption refrigerator can be performed. At the time of starting the absorption refrigerator, of course, the heat source of the regenerator remains activated. In the stop preparation state of the absorption refrigerator, the heat source of the regenerator is activated, for example, the burner is in an ignited state, and the water in the evaporator is circulated and injected from the spraying device to the heat transfer pipe of the cold water. Is in a stopped state, and this state is the same as the above-described start preparation state. When setting the absorption refrigerator to the stop state from the stop preparation state by the stop command,
Deactivate the heat source and extinguish the burner, for example.
At this time, the circulation of water to the spraying device of the evaporator is stopped, and for example, the above-described pump is stopped. The cooling load prediction device used in the absorption chiller number control device has the configuration defined by the above-described claims, and the third prediction load enables highly accurate estimation of the cooling load. Become.

Further, according to the present invention, the evaporator provided in the absorption refrigerator has a housing having a storage portion at a lower portion for storing the refrigerant from the condenser, and a housing in the housing above the storage portion, and chilled water is provided. A heat transfer pipe to be guided, a spraying device that injects refrigerant into the heat transfer tube, a circulation pump that supplies water stored in the storage portion to the spraying device, and a storage device without injecting the refrigerant from the condenser into the heat transfer tube. And a guiding conduit. The present invention also provides an absorption refrigerator including an evaporator to which a refrigerant from a condenser is guided, wherein the evaporator has a housing having a storage portion for storing the refrigerant from the condenser at a lower portion, and a storage portion in the housing. A heat transfer pipe through which cold water is introduced, a spraying device that injects refrigerant into the heat transfer tube, a circulating pump that supplies water stored in the storage section to the spraying device, and a heat transfer tube that transfers refrigerant from the condenser. And a pipe leading to the storage section without being injected into the heat pipe. According to the evaporator of the ninth and tenth aspects, it is possible to quickly make a transition from the start preparation state to the start state of the absorption refrigerator, and to make a quick transition from the stop preparation state to the stop state. .

[0015]

FIG. 1 is a block diagram showing the overall configuration of an embodiment of the present invention. Chilled water obtained from a plurality of (three in this embodiment) absorption chillers A, B, and C is supplied from each outlet pipe 1a, 2a, 3a to an outlet collecting pipe 4a.
To a single cooling load 5 or a plurality of cooling loads 5
In parallel, cold water is supplied. The chilled water from the cooling load 5 is returned from the inlet collecting pipe 6 and circulated through the inlet pipes 1b, 2b, 3b for each of the absorption chillers A, B, C.
A circulation pump 8 is provided for circulating the cold water.
In order to detect the cooling load 5, the outlet collecting pipe 4 is provided with a temperature sensor 9 for detecting the temperature of cold water, and the inlet collecting pipe 6 is provided with a temperature sensor 10 for detecting the temperature of cold water. The circulation flow rate of the cold water by the pump 8 is kept constant. Therefore, the cooling load L (unit%) is represented by Expression 1.

L (%) = (T10−T9) / ΔT10 (1) Here, T9 and T10 are the temperatures of the cold water detected by the temperature sensors 9 and 10, for example, T9 = 4 ° C. ΔT10 is a temperature difference between the chilled water detected by the temperature sensors 9 and 10 in the rated load operation state of the three absorption refrigerators A, B and C, and is, for example, ΔT10
= 5 ° C.

The outputs of the temperature sensors 9 and 10 are provided to a processing circuit 11 realized by a microcomputer or the like, whereby the number of each of the absorption chillers A, B and C is controlled. The memories M1 and M2 are connected to the processing circuit 11. Absorption refrigerators A, B, and C have the same configuration and the same rating in this embodiment, for example.

FIG. 2 is a system diagram showing a specific configuration of the absorption refrigerator A. This absorption refrigerator A uses LiB as an absorbent.
r Use an aqueous solution and use water as a refrigerant. In the absorber 22,
A LiBr aqueous solution is sealed, and a refrigerant liquid (water) 15 is sealed in the evaporator 14.

The evaporator 14 has, at the lower part of the housing 16, a storage section 17 in which water from the condenser 13 is guided and stored via a pipe 17a. Inside the housing 16, a heat transfer tube 18 through which cold water is guided is disposed above the storage unit 17. The heat transfer tube 18 is connected to the outlet tube 1a and the inlet tube 1b described above. Above the heat transfer tube 18, a spraying device 19 is provided in the housing 16. The refrigerant liquid 15 from the spraying device 19 is injected into the heat transfer tube 18.
The circulating pump 20 is configured to store the refrigerant liquid 15 stored in the storage section 17.
Is supplied to the spraying device 19. The pipe 17a is connected to the heat transfer pipe 1
Below 8, the refrigerant liquid from the condenser 13 is supplied to the storage unit 17 from the spraying device 21 which is the outlet. Therefore, the refrigerant liquid from the spraying device 21 is not injected into the heat transfer tube 18.

Therefore, the absorption chiller A can be quickly started by starting the operation by the circulation pump 20, and the absorption chiller A can be quickly stopped by stopping the pump 20. Become. By operating the pump 20 and injecting the refrigerant liquid in the storage unit 17 in the evaporator 14 from the spraying device 19 onto the heat transfer tube 18, the refrigerant liquid in the evaporator 14 evaporates. The cold water flowing through the heat transfer tube 18 is cooled by the latent heat of the evaporation.

The housing 16 of the evaporator 14 includes the absorber 2
2 is connected. LiBr aqueous solution 23 in absorber 22
Absorbs the refrigerant vapor in the container, the absorbed heat is absorbed by the cooling water supplied from the pipe 24, and the solution is cooled.

The low-concentration solution having absorbed the refrigerant liquid in the absorber 22 is sent from a pipe 25 to a regenerator 27 by a solution pump 26. The solution having the reduced concentration is heated in the regenerator 27 by the burner 28 as a heat source. Thereby, the moisture absorbed by the absorber 22 is driven out by the regenerator 27. The solution which has been driven out of water to have a high concentration is returned to the absorber 22 from the pipe line 29 again, and continues the absorbing action.

The condenser 13 cools and condenses the refrigerant vapor expelled from the solution in the regenerator 27 by the cooling water guided through the heat transfer tube 30 and returns the refrigerant vapor to the refrigerant liquid. In the spraying device 21 of the evaporator 14, it is returned to the storage unit 17.

The burner 28 is provided with a flow control valve 31 for controlling the fuel flow. Outlet pipe 1a of absorption refrigerator A
Is provided with a temperature sensor 32. This temperature sensor 3
2, the temperature of the cold water detected by
For example, the processing circuit 11 controls the opening of the flow control valve 31 and controls the rotation speed of the pump 26 so as to maintain the temperature at 4 ° C., thereby controlling the capacity of the absorption refrigerator A alone. In the capacity control, the temperature sensor 3 of the absorption refrigerator A is used.
The amount of heating to the regenerator 27 is controlled so that the chilled water outlet temperature detected by 2 is maintained at the set temperature, and the opening of the flow control valve 31 for the burner 28 is controlled as described above. The circulation amount of the aqueous solution 23 is controlled, and the rotation speed of the pump 26 is controlled.

Further, the burner 28 is provided with an on-off valve 33 for supplying / cutting off fuel. The on-off valve 33 is used for starting and stopping the operation of the absorption refrigerator A to control the number of units. The remaining absorption chillers B and C also have the same configuration as the absorption chiller A. These absorption chillers A, B, and C may be the single-effect absorption chillers described above, may be double-effect absorption chillers, or may be other types of absorption chillers. . These absorption chillers A, B, and C have a long time of about 5 minutes or about 30 minutes from the start of startup until reaching a steady operation state, and also from the start of shutdown to the completion of shutdown. Also requires a long time.

In the number control of the present invention, as shown in FIG. 3, a total of three absorption chillers A, B and C are controlled in this order in accordance with the cooling load L in this order. . The relationship between the cooling load and the number of operating units shown in FIG. 3 is as shown in Table 1.

[0027]

[Table 1]

FIG. 4 is a flowchart of the operation executed by the processing circuit 11. Step a1 to step a5 are:
Predict the cooling load. Step a7 to step a17
Is based on the predicted cooling load,
Control of the number of operating units B and C is performed.

First, a cooling load prediction device will be described with reference to steps a1 to a5. According to the present invention, in order to predict the future cooling load from the cooling load up to the present, the first predicted load Qx (t) is calculated and obtained, and the second predicted load Qx (t) is calculated.
The predicted load Qy (t) is calculated and calculated, and the first and second predicted loads thus obtained are combined to calculate and calculate a highly accurate third predicted load Q.

The process proceeds from step a1 to step a2 in FIG. 4 to calculate and obtain a first predicted load. In order to calculate the first predicted load Qx (t) of the short-term prediction, a method according to an embodiment of the present invention illustrated in FIG. 5 and a method according to another embodiment of the present invention illustrated in FIG. Can be In estimating the first predicted load, for example, a change in cooling load data for a first time equal to the time required for starting from stop and vice versa, for example, the past 10 to 30 minutes,
A first predicted load, which is a future cooling load, is estimated. This makes it possible to accurately estimate the cooling load in the relatively near future.

In FIG. 5, the time change rate of the cooling load during the time (t1-p) which is p minutes from the current time t1 is calculated and obtained. Based on the cooling load time change rate thus obtained and the cooling load at the current time t1, m
The cooling load at time (t1 + m) minutes later is calculated and obtained according to a linear function. p and m may be, for example, about 10 to 30 minutes as described above.
= M. In the embodiment shown in FIG. 5, it is necessary to assume that the temporal change rate of the cooling load does not change during the predicted time p from the current time t1 to the predicted arrival time (t1 + m). Therefore, the embodiment shown in FIG. 5 has a problem that it cannot cope with a sudden change in cooling load.

This problem is solved by another embodiment of the present invention shown in FIG. In order to cope with a rapid change in the cooling load, it is necessary to consider a change in the time change rate of the cooling load during the predicted time p. For that purpose, it is effective to use a non-linear function that uses the cooling load up to the present time as an input and the future cooling load as an output. In order to construct a non-linear function, a cooling network is predicted and estimated by a neural network. The neural network 35 can easily construct a non-linear function for calculating the load prediction value 37 from various load change patterns 36 by performing a procedure called learning. By inputting the load change pattern into the constructed nonlinear function, a predicted load in the future can be obtained.

The neural network 35 according to the embodiment of the present invention shown in FIG. 6 is a supervised hierarchical neural network. The whole of the neural network 35 is composed of a total of three layers: an input layer 38, a hidden layer 39, and an output layer 40. Each layer 38, 39, 40
Is composed of a nonlinear input / output function called a unit. The units are connected by an information transmitter called a link in each of the layers 38, 39, and 40. Inputs entering the network 35 are transmitted to the units of the input layer 38, further propagate to the hidden layer 39 and the output layer 40, and finally output from the units of the output layer 40.

In the learning of the network 35, a combination of outputs which is a desired correct answer for a certain input is prepared in advance, and the data is used as a teacher to change network internal parameters such as link transmission efficiency and unit output function parameters. I do. That is, the internal parameters are updated so that the output when the teacher input is input to the network approaches the teacher output. By repeating this operation for a plurality of teacher data, a desired multivariable nonlinear input / output function can be obtained.

In still another embodiment of the present invention, a multi-order regression operation is performed to construct a nonlinear function.
The structure which calculates | requires a prediction function and predicts a future cooling load may be sufficient.

With reference to FIGS. 7 and 8, an embodiment of the present invention for calculating and obtaining the second predicted load Qy (t) will be described. In step a3 of FIG. 4, such a second predicted load is calculated and obtained. In FIG. 7, a change pattern of the cooling load at a predetermined second time exceeding the first time, for example, from midnight t0 to the current time t1 in one day, is calculated based on the outputs of the temperature sensors 9 and 10. Ask. The detected cooling load change pattern is stored in the memory M2.

Apart from this, a plurality (four in this embodiment) of load change patterns over the past day are stored in the memory M1 as shown in FIGS. 8 (1) to 8 (4). Keep it. FIG. 8 shows an actual cooling load Qr (t) detected at a certain time t1 in one day.
8 (1) to FIG. 8 (4) by the actually measured cooling load Qmk (t) at the time t during one of the cooling load change patterns k shown in FIG. 8 (4).
The degree of coincidence Ek is calculated and calculated for each cooling load change pattern in (4).

[0038]

(Equation 1)

FIG. 8 (1) to FIG. 8 (4) thus obtained.
After calculating the degree of coincidence Ek of the change pattern of the cooling load for each type shown in (1), the load change pattern having the maximum Ek among the four degrees of coincidence Ek is selected. For example, in the embodiment shown in FIGS.
In the change pattern of the cooling load detected from 0 o'clock t0 to the current time t1 in one day, the degree of coincidence Ek with the change pattern shown in FIG. 8 (4) is the largest as compared with the change pattern shown in FIG.
Thus, the change pattern in FIG. 8A is selected by the pattern matching method of the present invention. Therefore, in accordance with the selected cooling load change pattern shown in FIG. 8A, the second predicted load Qy (t) from the current time t1 to the predicted arrival time (t + m).
Is calculated and estimated. Although Q in the coincidence degree equation of Equation 2 is described as being the absolute value of the cooling load, this Q is “a change amount of the cooling load” or “a dimensionless cooling load (for example, the current cooling load). / Average cooling load of the previous day) ".

In step a4 of FIG. 4, the first predicted load Qx (t) obtained and estimated in step a2 described above.
And the second predicted load Qy (t) obtained in step a3 to obtain a highly accurate third predicted load Q. For this purpose, first, the first and second predicted loads Qx (t) and Qy at the predicted arrival time, which is the current time t1 after m minutes obtained and estimated at the time (tm).
(T) is calculated, and from the cooling load Qr (t) actually measured and detected at the predicted arrival time t1, the predicted load errors X and Y are quantitatively calculated and calculated.

X = | Qx (t) -Qr (t) | (3) Y = | Qy (t) -Qr (t) | (4) From the predicted load errors X and Y thus obtained, Coefficient K
Find x and Ky.

Kx = Y / (X + Y) (5) Ky = X / (X + Y) (6) The weighting factors Kx and Ky are calculated based on the first predicted load Qx (t) and the second predicted load Qy (t). Represents certainty. As a result, the highly accurate third time at the time (t + m) after the next m minutes
The estimated load Q is calculated and estimated.

Q = Kx · Qx (t) + Ky · Qy (t) (7) In another embodiment of the present invention, the predicted arrival at time t after m minutes estimated at time (tm) The first predicted load Qx (t) and the second predicted load Qy (t) at the time are equally distributed, and the third predicted load Q1 is calculated and obtained.

Q1 = {Qx (t) + Qy (t)} / 2 (8) In still another embodiment of the present invention, the first predicted load Qx
(T) may be prioritized, and the third predicted load Q2 may be estimated.

Q2 = (1−K) · Qx (t) + K · Qy (t) (9) where the coefficient K is 0 <K <1 (10) In still another embodiment of the present invention. , The second predicted load Qy
(T) is prioritized, and the third predicted load Q3 is estimated.

Q3 = K · Qx (t) + (1−K) · Qy (t) (11) In each of the embodiments of the present invention represented by the above-described equations 8 to 11, the first and second equations are used. Predicted load Qx (t), Qy (t)
Is not always the same value, but may be exactly opposite in some cases, that is, one of Qx (t) and Qy (t) predicts that the load will increase, and one of Qx (t) and Qy (t) On the other hand, when the load is predicted to decrease, the predicted loads Q1 to Q3 may be calculated and obtained. In such a case, if one of the predicted loads Qx (t) or Qy (t) was accurate, the third predicted load Q
For 1 to Q3, the prediction accuracy is reduced.

The third predicted load Q obtained by the above-mentioned equations 3 to 7 solves the problems of each of the embodiments of the present invention of the equations 8 to 11 described above.

In step a5 of FIG. 4, third predicted loads Q, Q1 to Q3, for example, Q are calculated and obtained. In step a6, using the third predicted load Q thus obtained,
Based on FIG. 3 and Table 1, the predicted number E1 of absorption chillers A, B, and C is calculated.

Next, with reference to steps a7 to a17 in FIG. 4, a control device for the number of absorption chillers A, B and C will be described. After calculating the predicted number of operating vehicles E1 from the third predicted load Q in step a6 described above, in the next step a7, the current number of operating vehicles E2 is obtained and compared. Cooling load forecast load Q
The number of operating vehicles is increased or decreased from FIG.

Since a plurality of operation steps are required when starting and stopping the absorption refrigerator, it is not preferable to frequently start and stop the absorption refrigerator.
According to the present invention, it is possible to accurately and early estimate a change in the cooling load based on the predicted load Q. If the absorption chiller is started or stopped immediately based on the predicted load Q, if the prediction is incorrect, the stable operation of the absorption chiller is hindered. Therefore, in the embodiment of the present invention, a preparation state for starting or stopping is first made according to the predicted load Q, and thereafter, when the actually detected cooling load reaches a predetermined value as predicted, a start or stop command signal is issued. To actually start or stop.

That is, in step a7, E1> E
When it is determined to be 2, in the next step a8, the stopped absorption refrigerator is instructed to start up, and the stopped absorption refrigerator is set to the startup preparation state. Step a7
In step a9, when it is determined that E1 <E2, the absorption chiller that is performing the cooling start operation is instructed to stop, and the absorption chiller that is performing the cooling operation is brought into a stop preparation state.

When the absorption chiller is ready to start or stop, the on-off valve 33 is open and the burner 28 is ignited, and the water, which is the refrigerant stored in the storage section 17 of the evaporator 14, is sprayed. In this state, the circulating pump 20 for supplying to the pump 19 is stopped. In this preparation state, the absorber 22
The solution pump 26 for supplying the aqueous solution to the regenerator 27 from
You are still driving.

After steps a8 and a9, step a1
Move to 0. When it is determined in step a7 that E1 = E2, the process proceeds to step a10. In step a10, the actual cooling load L at the predicted arrival time (t + m) of the predicted operating number E1 is calculated and detected from the temperatures T9 and T10 detected by the temperature sensors 9 and 10.

In step a11, the above-mentioned step a1
The required number E3 of operating absorption chillers based on FIG. 3 and Table 1 corresponding to the detected cooling load L obtained at 0 is calculated.

In step a12, the required operating number E3 obtained in step a11 is compared with the current operating number E4. If E3> E4, the absorption refrigeration set in the start-up preparation state in step a8 is started. Start the machine. For this start-up, the pump 20 in which the evaporator 14 has been stopped may be brought into the operating state.
The cooling liquid is immediately injected into the heat transfer tube 18 to perform the cooling operation.

When it is determined in step a12 that E3 <E4, the process proceeds to step a14, in which the absorption chiller that has been brought into the stop preparation state in step a9 is stopped. This stop state is immediately achieved by closing the on-off valve 33 and stopping the solution pump 26.

When it is determined in step a12 that E3 = E4, the process proceeds to the next step a15. In step a15, a predetermined time u, for example, 10 minutes from the start of the start-up preparation command in step a8 described above.
It is determined whether or not the minutes have elapsed.
In step 9, it is determined whether the predetermined time u has elapsed from the start of the stop preparation command. The predetermined time u may be, for example, about 10 minutes, and is equal to or longer than the time required for the absorption chillers A, B, and C to transition from the stopped state to the activated state, and is equal to or greater than the time required for the absorption chillers A to transition from the activated state to the stopped state. , For example, 5 minutes or more. The time from the stop state to the start state and the time from the start state to the stop state may be the above-mentioned time m, for example, m = 5 minutes.

If it is determined in step a15 that the predetermined time u has elapsed since the start of the preparation state, the process proceeds to step a16, in which the start preparation command or the stop preparation command is canceled. That is, the start preparation state is canceled and the operation is returned to the stop state, or the stop preparation state is returned to the operation state as the start state. In this way, in step a15, the absorption chillers A, B, and C are not maintained for a long time in the preparation state for starting or stopping, or frequent start and stop are prevented. Step a
If the predetermined time u has not elapsed at 15, the process proceeds to step a17, and the startup operation is continued with the current number of operating absorption chillers A, B, C.

In step a13, if E3> E4 and more than the number of units in the ready state have to be started, the stopped absorption chiller is further started. Similarly, in step a14, when E3 <E4 and the number of units requiring a stop exceeds the number of units in the stop preparation state, the absorption chiller during the start-up operation is set to the stop state. Even in such an operation, the fluctuation of the chilled water temperature T9 in the outlet collecting pipe 4 can be suppressed as much as possible.

[0060]

According to the first aspect of the present invention, the cooling load can be predicted with high accuracy, and generally, the follow-up of the cooling load fluctuation can be performed as compared with the compression refrigerator. And to be able to quickly control the number of absorption refrigerators with poor response characteristics at start and stop,
Therefore, the temperature of the chilled water supplied to the cooling load can be kept constant irrespective of the fluctuation of the cooling load. Thus, it is possible to suppress the fluctuation of the chilled water temperature due to the excess or deficiency of the cooling capacity.

According to the second aspect of the present invention, it is possible to easily construct a nonlinear function for calculating the first predicted load from various cooling load change patterns. This makes it possible to accurately obtain future first and third predicted loads.

According to the third aspect of the present invention, it is possible to accurately predict a change in the cooling load with a fixed time.

According to the fourth aspect of the present invention, the selecting means provided in the second estimating means quantitatively obtains the degree of coincidence Ek, and determines the degree of coincidence Ek among the plurality of types of load changing patterns. It is possible to reliably obtain an approximate type of load change pattern, whereby the second predicted load can be obtained with high accuracy.

According to the fifth aspect of the present invention, it is possible to obtain a highly accurate predicted load Q by weighting the first predicted load and the second predicted load.

According to the sixth aspect of the present invention, when the number of units is increased or decreased according to the predicted load calculated by the prediction device, the absorption chiller is set to a start / stop preparation state, and then the actual detection is performed. Since the start or stop can be performed at the time when the determined cooling load reaches a predetermined value as predicted, quick start and stop corresponding to the fluctuation of the cooling load can be performed.

According to the present invention, the start preparation state or the stop preparation state is determined for a predetermined time u, for example, about 1 hour.
When 0 minutes or more have passed, cancel the preparation state,
Stop the absorption chiller that is ready to start or leave the absorption chiller that is ready to stop running, thus preventing frequent repetition of starting and stopping the absorption chiller, and ensuring a stable operating condition. Can be achieved.

According to the present invention, the first to third predicted loads of the cooling load can be obtained with high accuracy.

According to the ninth and tenth aspects of the present invention, it is possible to realize an absorption refrigerator capable of rapidly shifting from the start preparation state to the start state and from the stop preparation state to the stop state. it can.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an overall configuration of an embodiment of the present invention.

FIG. 2 is a system diagram showing a specific configuration of the absorption refrigerator A.

FIG. 3 is a diagram illustrating a relationship between a cooling load and the number of operating units.

FIG. 4 is a flowchart of an operation executed by the processing circuit 11;

FIG. 5 shows a first predicted load Qx according to the embodiment of the present invention.
FIG. 9 is a diagram illustrating a method for obtaining (t).

FIG. 6 is a diagram showing a method for obtaining a first predicted load Qx (t) according to another embodiment of the present invention.

FIG. 7 is a diagram showing a change pattern of a cooling load up to a current time in one day for calculating and obtaining a second predicted load Qy (t).

FIG. 8 is a diagram showing a one-day load change pattern stored in advance in a memory M1 used for calculating a degree of coincidence Ek.

[Explanation of symbols]

 1a, 2a, 3a Outlet pipe 1b, 2b, 3b Inlet pipe 4 Outlet collecting pipe 5 Cooling load 6 Inlet collecting pipe 8, 20 Circulation pump 9, 10, 32 Temperature sensor 11 Processing circuit 13 Condenser 14 Evaporator 15 Refrigerant liquid 16 Housing 17 Reservoir 17a Pipeline 18, 30 Heat transfer tube 19, 21 Spray device 22 Absorber 23 LiBr aqueous solution 24, 25 Pipeline 26 Solution pump 27 Regenerator 28 Burner 29 Pipeline 31 Flow control valve 33 Open / close valve 35 Neural network 36 Load change pattern A, B, C Absorption refrigerator M1, M2 Memory

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Susumu Hashidera 1-1, Kawasaki-cho, Akashi-shi, Hyogo Kawasaki Heavy Industries, Ltd. Inside Akashi Plant (72) Inventor Yoshinobu Mori 1-1-1, Kawasaki-cho, Akashi-shi, Hyogo Kawasaki Inside Heavy Industries, Ltd.Akashi Factory (72) Inventor Teruo Tanabe 1000, Aochi-cho, Kusatsu-shi, Shiga Prefecture Inside Kawagei-Hiroshi Kogyo Co., Ltd. In-house (72) Inventor Shigeru Saegusa 1000 Aochi-cho, Kusatsu-shi, Shiga Prefecture

Claims (10)

[Claims] 1. A cooling load of an absorption refrigerator which circulates cold water obtained from an evaporator of each of a plurality of absorption chillers through an outlet collecting pipe to a cooling load, and returns cold water from the load from the inlet collecting pipe to circulate. Means for detecting a cooling load; and estimating a future first predicted load based on a load change pattern of the cooling load during a first predetermined time in the past in response to an output of the load detecting means. A second estimating means for estimating a future second predicted load based on a load change pattern during a predetermined second time period exceeding a past first time period of the cooling load in response to an output of the load detecting means; Estimating means, and a load Q detected as first and second predicted loads Qx (t) and Qy (t) in response to outputs of the first and second estimating means.
A weight coefficient K for the first and second predicted loads Qx (t) and Qy (t) based on the predicted load errors X and Y with respect to r (t).
a weighting coefficient calculating means for calculating x and Ky; and a prediction load calculating means for calculating a future third predicted load Q in response to outputs of the first and second estimating means and the weighting coefficient calculating means. A cooling load prediction device for an absorption refrigerator. 2. The apparatus for predicting a cooling load of an absorption refrigerator according to claim 1, wherein the first estimating means estimates a first predicted load by constructing a nonlinear function of a neural network performing a learning procedure. . 3. A second memory for storing a plurality of types of predetermined load change patterns in the past over the second time, a change in a detected load in the past less than the second time. A second memory for storing a pattern, and a load change pattern stored in the second memory among a plurality of types of load change patterns stored in the first memory in response to the contents stored in the first and second memories. Selecting means for selecting a load change pattern most similar to the above, and selecting the selected load change pattern as a second predicted load. 3. The apparatus for predicting a cooling load of an absorption refrigerator according to claim 1 or 2, wherein . 4. The load control system according to claim 1, wherein each of the load change patterns of each type stored in the first memory and the detected load patterns stored in the second memory has a load Qmk (t), Qr ( a value corresponding to the reciprocal of the absolute value | Qr (t) -Qmk (t) | of the difference t), and a value corresponding to the reciprocal of the absolute value | Qr (t) -Qmk (t) | of the difference Is calculated over a second period of time, and the degree of coincidence Ek is calculated.
4. The apparatus for predicting a cooling load of an absorption chiller according to claim 3, wherein a load change pattern of a kind in which k is the maximum is used as the second predicted load. 5. The weighting factor calculation means, wherein the predicted load errors X and Y are: X = | Qx (t) -Qr (t) | Y = | Qy (t) -Qr (t) | Kx = Ky = Kx = Y / (X + Y) Ky = X / (X + Y) In the predictive load calculating means, the third predictive load Q is: Q = Kx · Qx (t) + Ky · Qy (t) The apparatus for predicting a cooling load of an absorption refrigerator according to any one of claims 1 to 4, characterized in that: 6. A number control device for the number of absorption chillers that guides chilled water obtained from the evaporators of each of a plurality of absorption chillers to a cooling load via an outlet collecting pipe, and returns and circulates chilled water from the load from the inlet collecting pipe. A predicting device for calculating a predicted load of a cooling load, an operating number detecting means for detecting a current operating number of the absorption chillers, and a predictive operation of the absorption chiller corresponding to the predicted load in response to an output of the predicting device. A first operating number calculating means for calculating the number of operating units, and a stopped absorption chiller responsive to an output of the operating number detecting means and the first operating number calculating means, wherein the predicted operating number exceeds the current operating number. When the predicted number of operating units is less than the current number of operating units, a preparation state setting unit for setting the operating absorption chillers to a stop preparation state, and detecting a cooling load at the time when the predicted number of operating units reaches the predicted state. Air conditioning Load detecting means, second operating number calculating means responsive to the output from the cooling load detecting means, and calculating the required number of operating chillers corresponding to the detected cooling load; In response to the output of the operating number calculation means, when the required operating number exceeds the operating number at the time of the predicted arrival of the predicted operating number, start the absorption refrigerator in the start preparation state, When the required number of operating units is less than the number of operating units at the time when the predicted number of operating units reaches the predicted number, start / stop means for stopping the absorption refrigerator in the stop preparation state is included. Control device for the number of machines. 7. The starting / stopping means, when the operating state in which the predicted number of operating vehicles is equal to the number of operating vehicles at the time of prediction reaching has passed for a predetermined time u or more,
7. The absorption refrigerator according to claim 6, wherein the absorption refrigerator set in the start preparation state by the preparation state setting means is stopped, or the absorption refrigerator set in the stop preparation state is kept activated. Unit control device. 8. The apparatus according to claim 6, wherein the prediction device is one of claims 1 to 5. 9. An evaporator provided in an absorption refrigerator has a housing having a storage portion at a lower portion for storing a refrigerant from a condenser, and a transmission portion disposed above the storage portion in the housing to guide cold water. A heat pipe, a spraying device that injects refrigerant into the heat transfer tube, a circulation pump that supplies water stored in the storage portion to the spraying device, and a conduit that guides the refrigerant from the condenser to the storage portion without injecting the refrigerant into the heat transfer tube. The number control device of absorption chillers according to any one of claims 6 to 8, comprising: 10. An absorption refrigerator including an evaporator through which a refrigerant from a condenser is guided, wherein the evaporator has a lower portion having a storage portion for storing the refrigerant from the condenser, and a storage portion in the housing. A heat transfer tube that is located above the cooling unit and guides cold water, a spraying device that injects refrigerant into the heat transfer tube, a circulation pump that supplies the water stored in the storage unit to the spraying device, and a refrigerant that is transferred from the condenser. An absorption refrigerator comprising: a conduit for guiding a heat pipe to a storage section without spraying the heat pipe.