JP5544268B2 - Control system - Google Patents

Description

  The present invention relates to a control system using a genetic algorithm.

  In various systems such as an air conditioning system, a control table for the purpose of energy saving is set based on the rating information of each device incorporated in the system at the time of design, and the control values for each device derived from the control table etc. Drive based on. However, there are individual differences in the performance of each device, and the secular change of each device varies depending on the frequency of use, so it is necessary to update the control table etc. as needed after the start of operation in order to optimally control the system . Various methods for updating the control table of the system have been proposed, and there is a method using a genetic algorithm.

  For example, Patent Document 1 discloses a frequency control table using a genetic algorithm so as to conform to dynamic characteristics between a frequency during operation and a detected room temperature in frequency control for an inverter circuit of a compressor motor of an air conditioner. It is described to update. In Patent Document 2, in a refrigeration cycle control device such as an air conditioner, a superheat degree for the valve opening degree of an electronic expansion valve is used by using a genetic algorithm so as to match a dynamic characteristic between the valve opening degree and the superheat degree. It is described that the control table is updated. Patent Document 3 describes that, in a cogeneration system that uses electric power and heat from a generator in a house, the control map is updated using a genetic algorithm based on historical data of energy demand of each house. Yes.

Japanese Patent Laid-Open No. 7-253236 JP-A-8-166169 JP 2006-275473 A

  In the genetic algorithm, initial individuals are randomly generated, and individuals with high fitness (for example, individuals with low energy consumption) are searched from the initial individuals. However, since the initial individuals are randomly generated, if the initial individuals are not appropriate, individuals with high fitness may not be born, and a control table suitable for optimal control may not be obtained. In addition, mutation is also performed as a genetic operation in the genetic algorithm, but if the initial individual is not appropriate, the mutation alone alone has a low probability of producing a highly adaptable individual. As a result, an optimal control value for each device of the controlled system cannot be obtained, and optimal control cannot be performed.

  Therefore, an object of the present invention is to provide a control system capable of obtaining an optimal control value using a genetic algorithm.

A control system according to the present invention is a control system that outputs a control value for controlling a control target system including a plurality of devices, and there is a device characteristic equation for each device, and an optimal control value for the control target system is obtained. An optimum control device for generating, the optimum control device comprising: an execution number setting means for setting an execution number for executing a genetic algorithm; an initial individual generation means for generating an initial individual; and an execution means for executing a genetic algorithm. Until the number of implementations of the genetic algorithm reaches the number of implementations set by the implementation number setting means, and for each device, each time the genetic algorithm is implemented, different initial individuals are generated by the initial individual generation means. look for coefficients of the adaptive high degree equipment characteristic equation to implement a genetic algorithm coefficients equipment characteristic equation as gene execution means using the generated initial population And, if the number of performed genetic algorithm becomes execution count set in execution count setting means, in the genetic algorithm implementations number of times set by the execution count setting means for each device searched device characteristic equation for each select the coefficient of highest fitness is high equipment characteristic equation from the coefficients, generating a control value to be output to the control object system based on the device characteristic equation for each device by using the coefficient for each said selected device Features. The control system according to the present invention is a control system that outputs a control value for controlling a control target system including equipment, and has an optimal control device that generates an optimal control value for the control target system. The optimal control apparatus includes an execution number setting means for setting the number of executions for executing the genetic algorithm, an initial individual generation means for generating an initial individual, and an execution means for executing the genetic algorithm, and executes the genetic algorithm. Each time the genetic algorithm is executed, a different initial individual is generated by the initial individual generation means until the number of executions reaches the number of executions set by the execution number setting means, and the device of the equipment is executed by the execution means using the generated initial individual. The genetic algorithm is executed using the coefficient of the characteristic formula as a gene, the coefficient of the device characteristic formula having high fitness is searched, and the genetic algorithm is executed. The device with the highest fitness among the coefficients of the device characteristic equation respectively searched by the genetic algorithm for the number of executions set by the execution number setting means Select the coefficient of the characteristic equation, generate the control value to be output to the controlled system based on the device characteristic equation using the selected coefficient, and the fitness is calculated by the device characteristic equation using the gene by the genetic algorithm It is a deviation between the measured energy consumption and the measured value of the energy consumption of the device.

  This control system is a system for controlling the system to be controlled, and has an optimal control device that generates a control value for optimally controlling the system to be controlled. In the optimum control device, the number of times of execution of the genetic algorithm is set by the number of times setting means. Until the number of executions of the genetic algorithm reaches the number of executions, in the optimal control device, every time the genetic algorithm is executed, an initial individual is newly generated by the initial individual changing means, and the initial means is used to execute the initial means. In order to search for individuals with high fitness, perform a genetic algorithm. When the number of executions of the genetic algorithm reaches the number of executions, the optimal control device selects the individual with the highest fitness from the individuals searched by the genetic algorithm for the number of executions, and the highest fitness is obtained. A control value to be output to the control target system is generated based on the individual. As described above, by generating the initial individual for the number of times of execution, the genetic algorithm can be implemented by various initial individuals. At this time, even if some initial individuals are inappropriate and individuals with high fitness cannot be obtained, individuals with high fitness can be obtained from other initial individuals, and from among individuals with high fitness An individual with the highest fitness can be obtained, and an optimal control value can be generated. As described above, the control system can obtain an optimum control value by changing the initial individual and repeatedly executing the genetic algorithm.

  In addition, when generating the control value to be output to the control target system based on the individual with the highest fitness, the control value may be generated directly from the individual with the highest fitness, or the individual with the highest fitness. It is also possible to generate a control table, a control characteristic equation, or the like for deriving the control value from the control table, and generate the control value from the control table or the like.

  The control system according to the present invention has a basic control device that generates a basic control value for the control target system, and switches the control value output to the control target system between the optimal control device and the basic control device. The control value of the previous control device may be gradually changed to the control value of the control device after switching.

  In addition to the optimal control device, this control system has a basic control device that generates basic control values for controlling the system to be controlled. In the control system, when switching the control value output to the control target system from the control value of the optimal control device to the control value of the basic control device or from the control value of the basic control device to the control value of the optimal control device, The control value of the control device is gradually changed to the control value of the control device after switching. In this way, the control system can prevent the control value from changing suddenly by gradually changing the control value when switching between the optimal control device and the basic control device, and suppresses steep output in the controlled system. it can. In addition, the control system can perform basic control by the basic control device even when the optimal control device fails and optimization control cannot be performed.

  According to the present invention, an optimal control value can be obtained by changing the initial individual and repeatedly executing the genetic algorithm.

It is a block diagram of the air conditioning system which concerns on this Embodiment. It is a block diagram of the control system in the air conditioning system of FIG. 2 is an example of a database 1; It is an example of the database 2. It is a flowchart which shows the flow of the main program implemented with the personal computer for optimal control of FIG. It is a flowchart which shows the flow of the program A implemented with the personal computer for optimal control of FIG. It is explanatory drawing of the energy consumption calculation by the program B implemented with the personal computer for optimal control of FIG. It is a flowchart which shows the flow of the program C (at the time of ON) implemented with the personal computer for optimal control of FIG. It is a flowchart which shows the flow of the program C (at the time of OFF) implemented with the personal computer for optimal control of FIG. It is a flowchart which shows the flow of the program E (at the time of ON) implemented with the personal computer for optimal control of FIG. It is a flowchart which shows the flow of the program E (at the time of OFF) implemented with the personal computer for optimal control of FIG.

  Hereinafter, embodiments of a control system according to the present invention will be described with reference to the drawings. In addition, the same code | symbol is attached | subjected about the element which is the same or it corresponds in each figure, and the overlapping description is abbreviate | omitted.

In the present embodiment, the control system according to the present invention is applied to a control system for controlling an air conditioning system (heat source system). The control system according to the present embodiment performs basic control for the air conditioning system even when the optimal control function for performing the optimal control for minimizing the energy consumption and the CO ₂ emission amount in the air conditioning system and the optimal control are not performed. It has a basic control function, and the function is switched by central monitoring by the operator.

  First, an air conditioning system 1 according to the present embodiment will be described with reference to FIG. FIG. 1 is a configuration diagram of an air conditioning system according to the present embodiment.

  The air conditioning system 1 includes a plurality of cooling towers 1a, a cooling water pump 1b, a refrigerator 1c, a primary pump (cold water pump) 1d, and the like (in the example of FIG. 1, two cooling towers 1a and a cooling water pump). 1b, only the refrigerator 1c and the primary pump 1d are drawn). The cooling tower 1a cools the high-temperature cooling water that has been returned for cooling by the refrigerator 1c. The cooling water pump 1b circulates cooling water between the cooling tower 1a and the refrigerator 1c. The refrigerator 1c uses the cooling water from the cooling tower 1a to cool again the high temperature water that has been used and returned by the air conditioner (not shown). The primary pump 1d circulates cold water (heat medium) between the refrigerator 1c and the air conditioner. In the present embodiment, the air conditioning system 1 corresponds to a control target system described in the claims.

  The air conditioning system 1 is controlled by the control system 2 and is given control values such as the number of operating units, the heat source capacity, the cooling water temperature, the cooling water flow rate, and the cooling water flow rate for optimal control of the air conditioning system 1. For this control, data such as the temperature and humidity of the outside air, the number of operating units, the heat source capacity, the cooling water temperature, the cooling water flow rate, the cooling water temperature, and the cooling water flow rate are detected at regular intervals during the operation of the air conditioning system 1. Each data is collected by the control system 2 and accumulated as past log data.

  In the air conditioning system 1, the cooling tower 1a, the cooling water pump 1b, the refrigerator 1c, and the primary pump 1d are operated as one pair. For example, when the number of operating refrigerators 1c is three, three refrigerators are operated. The cooling tower 1a, the cooling water pump 1b, and the primary pump 1d that are paired with the machine 1c are also operated. In the air conditioning system 1, when the air conditioning load is at a peak (100%), all the cooling towers 1a, the cooling water pump 1b, the refrigerator 1c, and the primary pump 1d are operated, but when the air conditioning load is not at a peak (partial load). The number of operating units is determined according to the load, and only the cooling tower 1a, the cooling water pump 1b, the refrigerator 1c, and the primary pump 1d corresponding to the operating number are operated.

  Thus, in the air conditioning system 1, not all the devices are always operating, but some devices are operating almost all of the time. Furthermore, even when the vehicle is operating, it is not always operating at 100% of the capacity of the device, but it is operating at a capacity of less than 100% depending on the state of the load almost all of the time. In addition, even with the same number of operating units, the combination of the device to be operated and the device to be stopped changes from time to time, so that the deterioration state due to aging changes with each device. Also, even with the same equipment, there are individual differences and performance is slightly different. When optimal control is performed for the air conditioning system 1 as described above, optimal control is not achieved even if control is performed based on the control table set based on the rating information (catalog information) at the time of design (particularly after the start of operation). As the time elapses), it is necessary to periodically update the control table using the past log data of each operating device.

  The control system 2 according to the present embodiment will be described with reference to FIGS. FIG. 2 is a configuration diagram of the control system. FIG. 3 is an example of the database 1. FIG. 4 is an example of the database 2. FIG. 5 is a flowchart showing the flow of the main program executed on the optimal control personal computer. FIG. 6 is a flowchart showing a flow of the program A executed by the optimal control personal computer. FIG. 7 is an explanatory diagram of energy consumption calculation by the program B executed on the optimal control personal computer. FIG. 8 is a flowchart showing the flow of the program C (when ON) executed on the optimal control personal computer. FIG. 9 is a flowchart showing the flow of program C (when OFF) executed on the optimal control personal computer. FIG. 10 is a flowchart showing a flow of the program E (when ON) executed on the optimal control personal computer. FIG. 11 is a flowchart showing the flow of program E (when OFF) executed on the optimal control personal computer.

The control system 2 performs optimal control so that the energy consumption and CO ₂ emission amount in the air conditioning system 1 are minimized for the purpose of optimal operation of the air conditioning system 1, and if the optimal control is not performed, the air conditioning is performed. Basic control is performed on the air conditioning system 1 in order to ensure control over the system 1. For this purpose, the control system 2 includes an intelligent controller 2a, a BEMS [Building Energy Management System] 2b, a heat source group management PLC [Programmable Logic Controller] 2c, an optimal control PLC 2d, an optimal control personal computer (hereinafter referred to as an optimal control personal computer). 2e, a control system controller 2j is provided. The optimal control is performed by the optimal control PLC 2d and the optimal control personal computer 2e, the basic control is performed by the control system controller 2j, and the control value by the optimal control personal computer 2e is obtained when the operator turns on the optimal control in the central monitoring Output through the optimal control PLC 2d, and output the control value of the control system controller 2j when the operator is OFF in the central monitoring, and output the control value that gradually changes during the transition period when ON and OFF are switched To do. In this embodiment, the optimal control PLC 2d and the optimal control personal computer 2e correspond to the optimal control device described in the scope of patent control, and the control system controller 2j corresponds to the basic control device described in the claims. To do.

  Communication between the intelligent controller 2a, the BEMS 2b, and the optimal control personal computer 2e is performed over an IPv [Internet Protocol Version] 4 or IPv6 network 2f. Communication between the intelligent controller 2a and the heat source group management PLC 2c is performed by the coaxial cable 2g. Communication between the heat source group management PLC 2c and the optimum control PLC 2d is performed by the coaxial cable 2h. Communication between the heat source group management PLC 2c and the control system controller 2j is performed by the coaxial cable 2k. Communication between the optimal control PLC 2d and the optimal control personal computer 2e is performed by Ethernet (registered trademark) 2i.

  The intelligent controller 2a collects and manages data detected from the outside air, and transmits the data to the optimal control personal computer 2e. Examples of data to be collected include outside air temperature and outside air humidity. The optimal control PLC 2d collects and manages data detected by the air conditioning system 1, and transmits the data to the optimal control personal computer 2e. The collected data includes, for example, the number of operating air conditioning systems 1 and the operating state of each device (cooling water outlet temperature and inlet temperature, cooling water flow rate, cooling water outlet temperature and inlet temperature, cooling water flow rate, heat source capacity, etc.). is there.

  The BEMS 2b collects and manages energy data of the air conditioning system 1 and transmits the data to the optimal control personal computer 2e. The data to be collected includes, for example, the power consumption (energy consumption) of each device of the air conditioning system 1 and the load factor (heat consumption rate) of the air conditioner.

  Each time the heat source group management PLC 2c receives a control value for the air conditioning system 1 from the optimal control PLC 2d, it transmits the control value to the control system controller 2j. Each time the control system controller 2j receives a control value for the air conditioning system 1 from the optimal control PLC 2d via the heat source group management PLC 2c, it outputs a command to each device of the air conditioning system 1 based on the control value. As commands for each device, for example, a cooling water temperature control command for each cooling tower 1a, an inverter frequency control command based on a cooling water flow rate ratio for each cooling water pump 1b, and an inverter frequency based on a primary pump flow rate ratio for each primary pump 1d There are control instructions. At this time, since the number of operating units is determined according to the air conditioning load state, energization is stopped and control is not performed for the devices whose operation is stopped. The control system controller 2j performs basic control on the air conditioning system 1 based on the current outside air temperature, outside air humidity, the operating state of each device of the air conditioning system 1, etc., and obtains the control value for the air conditioning system 1 to obtain the optimum When the control is OFF, the control is performed. A conventional method is applied to this basic control. Note that a plurality of heat source group management PLCs 2c may be configured.

  Each time the optimal control PLC 2d receives a control value for the air conditioning system 1 from the optimal control personal computer 2e, the optimal control PLC 2d transmits the control value to the heat source group control PLC 2c.

The optimal control personal computer 2e is the best for minimizing energy consumption and CO ₂ emission based on data such as the operating state and power consumption of each device of the air-conditioning system 1 during operation after the air-conditioning system 1 starts operating. Take control. For this purpose, the optimal control personal computer 2e holds a log database for storing past logs of the air conditioning system 1, and the outside air temperature and the outside air humidity from the intelligent controller 2a and the air conditioning system 1 from the optimal control PLC 2d. The operation state of each device, the power consumption (energy consumption) of each device of the air conditioning system 1 from the BEMS 2b, and the load factor (heat consumption rate) of the air conditioner are accumulated in the log database. The optimal control personal computer 2e holds the database 1 and the database 2 as control tables for setting control values of each device, and is stored in the log database at regular intervals (for example, every month). The database 1 and the database 2 are updated in order to perform optimum control as a whole of the air conditioning system 1 by using the obtained data. At this time, the optimal control personal computer 2e uses a genetic algorithm to search for the optimum value of the coefficient of the characteristic formula for calculating the energy consumption of each device.

  For the optimum operation of the air conditioning system 1 as a whole, the system efficiency is maximized by linking the controls for each device. As energy saving control for each device, increasing the cooling water temperature of the cooling tower 1a reduces the power consumption of the cooling tower 1a, and reducing the cooling water flow rate by the cooling water pump 1b reduces the power consumption of the cooling water pump 1b. If the flow rate of the chilled water by the primary pump 1d is reduced, the power consumption of the primary pump 1d is reduced. Therefore, the system efficiency is maximized by linking these controls. For this purpose, the power consumption (Energy consumption) of the entire air conditioning system 1 for all combinations of control values for each device is obtained for each number of operating units, and the combination of control values that minimizes the power consumption of the entire air conditioning system 1 To extract.

Before describing specific processing in the optimal control personal computer 2e, the database 1 and the database 2 used in the optimal control personal computer 2e will be described. First, the database 1 will be described with reference to FIG. The database 1 is a database for determining the number of operating air conditioning systems 1. The database 1 is a database that defines the number of operating units determined by the outside air enthalpy (physical quantity determined by the outside air temperature and outside air humidity) and the secondary device load factor (heat consumption rate at the air conditioner). The database 1 shown in FIG. 3 is for the case where the maximum number of operating units of the air conditioning system 1 is four. For example, when the outside air enthalpy is 56 and the secondary device load factor is 0.40, the number of operating units is three (3 Since the three refrigerators 1c are operated, the cooling tower 1a, the cooling water pump 1b, and the primary pump 1d that are paired with the three refrigerators 1c are also operated). When the operation of the air conditioning system 1 is started, the database 1 is set so that the energy consumption and the CO ₂ emission amount are minimized by simulation based on the catalog value of each device at the time of design.

  Next, the database 2 will be described with reference to FIG. The database 2 is a database for determining control values when operating the air conditioning system 1 with the number of operating units determined in the database 1. The database 2 is a database including the power consumption (energy consumption) of the entire air conditioning system 1 for each combination of control values and the power consumption of each device, and the power consumption of the entire air conditioning system 1 for each number of operating units is the minimum. This is a database showing combinations of control values. In addition, about the database 2, the data for every operation number may be shown, or the data of all the operation numbers may be shown.

  The database 2 shown in FIG. 4 is for the case where the maximum number of operating units of the air conditioning system 1 is four. 1-NO. 4 cooling tower 1a, cooling water pump 1b, refrigerator 1c, and primary pump 1d. In particular, FIG. 4 shows a part of data when the number of operating units in the database 2 is three. As the control value, the device No. 1-NO. 4, the water supply temperature to the air conditioner, the flow rate ratio of the primary pump 1d, the cooling water set temperature in the cooling tower 1a, and the flow rate ratio of the cooling water pump 1b. In addition, as the power consumption, the power consumption of the entire air conditioning system and the equipment No. 1-NO. 4, the power consumption of the refrigerator 1c, the power consumption of the primary pump 1d, the power consumption of the cooling water pump 1b, and the power consumption of the fan of the cooling tower 1a. Note that the operation of the device whose control value or power consumption is “0” (device No. 4 in the example of FIG. 4) is stopped.

  The combinations of the control values are all combinations in which the water supply temperature, the primary pump flow rate ratio, the cooling tower set temperature, and the cooling water pump flow rate ratio are changed with a constant step size within the respective fluctuation ranges. In the example of FIG. 4, the flow rate ratio is changed with a step size of “0.05”, and the temperature is changed with a step size of “1”. The power consumption of each device of the air conditioning system 1 is calculated for all combinations of the control values, the system power consumption is calculated from the power consumption of each device, and the system power consumption is calculated for all combinations of the control values. The minimum system power consumption is extracted from the amount, and the combination of control values at the extracted system power consumption is the optimal control.

The example shown in FIG. 4 is a case where the number of operating units is 3, and the device No. 1-NO. 3 and the equipment No. 4 is shown, and the control value is the equipment number. 4 are all “0”, and the power consumption is also the device No. All the values of 4 are “0”. In the case of the example shown in FIG. 4, “647.4” is minimized as the system power consumption, and the control value combination C when the system power consumption is “647.4” is extracted. The extracted combination C of control values is stored in a specific location of the database 2. Therefore, when the number of operating units is determined to be “3” in the database 1, the combination of control values with the minimum system power consumption is extracted from a specific location in the database 2, thereby minimizing energy consumption and CO ₂ emissions. It becomes the optimal combination of control values.

  Processing in the optimal control personal computer 2e will be described. In the optimal control personal computer 2e, each program (main program, program A, program B, program C (ON time), program C (OFF time), program D, program E for optimal control of the air conditioning system 1 in the memory device. (When ON), program E (when OFF), etc.) are held, and each program is loaded into the RAM and executed by the CPU. In the present embodiment, each process in the program A corresponds to an execution number setting unit, an initial individual generation unit, and an execution unit described in the claims.

  First, the flow of processing of the main program in the optimal control personal computer 2e will be described with reference to the flowchart of FIG. The optimal control personal computer 2e accumulates the past log data of the air conditioning system 1 in the log database at regular intervals (the same cycle as the control cycle, for example, every 5 minutes) (S1). The data stored in the log database includes, for example, the outside air temperature, the outside air humidity, and the operation state of each device of the air conditioning system 1 (cooling water outlet temperature and inlet temperature, cooling water flow rate, cooling water at every time the data is collected. There are combinations of outlet temperature and inlet temperature, cold water flow rate, etc., energy consumption (power consumption) of each device, operation mode, and control value.

  In the optimal control personal computer 2e, the program A is started, and the coefficient of the device characteristic formula for calculating the energy consumption (power consumption) for each device of the air conditioning system 1 is inherited using the past log data of the log database. Calculation is performed using a genetic algorithm (S2). Then, the optimal control personal computer 2e creates a coefficient file of the device characteristic formula in the CSV [Comma Separated Values] format for each device of the air conditioning system 1 (S3). This file is a coefficient sequence of the device characteristic formula for each device.

In the optimal control personal computer 2e, the program B is started, and for each case (combination of all control values for each number of operating units), a device characteristic formula for each device (equipped with the coefficient value of the CSV file in the device characteristic formula) Is used to calculate the energy consumption, and the energy consumption of the entire air conditioning system 1 is calculated (S4). Then, the optimal control personal computer 2e rewrites the database 2 with the calculated energy consumption (power consumption) of each device and the energy consumption (system power consumption) of the entire air conditioning system for each case (S5). . In response to the rewriting of the database 2, the optimal control personal computer 2e also rewrites the database 1. This rewriting (updating) of the database is performed every designated period (for example, once a month). Therefore, in the optimal control, an optimal control value that minimizes the energy consumption amount and the CO ₂ emission amount is set by using the database 1 and the database 2 that are updated every specified period.

  In the optimal control personal computer 2e, in response to ON / OFF for the optimal control by the central monitoring, the program C (when ON) is started when ON, and the program C (when OFF) is started when OFF, The control value is transferred to the control system controller 2j via the optimal control PLC 2d and the heat source group management PLC 2c (S6). When the optimum control is switched from OFF to ON, the control value output by the program C (when ON) is used for a predetermined period. When the optimum control is completely turned on, the control value by the database 2 (the optimum control value) is obtained. Used as is. On the other hand, when the optimum control is switched from ON to OFF, the control value output by the program C (when OFF) is used for a predetermined period. When the optimum control is completely turned OFF, the control value (the basic control value) of the control system controller 2j is used. Control value) is used as it is.

  And in the control system 2, each apparatus of the air conditioning system 1 is controlled by the command from the control system controller 2j. While the air conditioning system 1 is in operation, the outside air data is collected by the intelligent controller 2a, the operation state data of each device of the air conditioning system 1 is collected by the optimal control PLC 2d and managed as log data, and the air conditioning system by the BEMS 2b. The power consumption (energy consumption) of each device is collected and managed as log data. Then, the optimal control personal computer 2e starts the program D and receives the log data managed by the intelligent controller 2a and the optimal control PLC 2d and the log data managed by the BEMS 2b at regular intervals (S7). The received log data and the control value at that time are accumulated in the log database (S1).

The program A will be described along the flowchart of FIG. Before specifically describing the program A, the device characteristic formula will be described. As an example of the device characteristic formula, a device characteristic formula (1) for calculating the energy consumption f ₁ of the refrigerator is shown. In formula (1), x ₁ (t) to x ₅ (t) are past log data stored in the log database, x ₁ (t) is a load factor, and x ₂ (t) is a cooling water temperature. Yes, x ₃ (t) is the cooling water flow rate, x ₄ (t) is the primary cooling water flow rate, and x ₅ (t) is the water supply temperature. t is a time parameter when log data stored in the log database is collected. In formula (1), c ₁ (t) to c ₅ (t) are influence levels. c ₁ (x ₁ (t), x ₂ (t)) is obtained by Expression (2), and is an influence degree of the load factor and the cooling water temperature. c ₂ (x ₃ (t)) is obtained by the equation (3) and is an influence degree of the cooling water flow rate. c ₃ (x ₄ (t)) is 1.0 as shown in the equation (4), which is the degree of influence of the primary cold water flow rate. c ₄ (x ₅ (t)) is obtained by the equation (5) and is an influence degree of the water supply temperature. In the expressions (2) to (5), u ₁ (i) to u ₁₃ (i) are coefficients of the device characteristic expression. i is the number of individuals. As described above, the energy consumption f ₁ of the refrigerator is expressed by the coefficient of u ₁ (i) to u ₁₃ (i) using x ₁ (t) to x ₅ (t) as variables using the equation (1). You can ask for it. In the program A, the optimum value of each coefficient u of the device characteristic formula for each device is obtained using the past log data stored in the log database.

  Now, the program A will be described. The program A is executed for each device of the air conditioning system 1 and calculates the optimum coefficient of the device characteristic formula for all devices. Here, each device uses a genetic algorithm to calculate energy consumption sequentially using the device characteristic equation while sequentially changing the combination of coefficients, and the energy consumption calculated value and energy consumption measurement based on the device characteristic equation are calculated. The combination of coefficients that minimizes the deviation from the value (past log data) is determined.

  The optimal control personal computer 2e acquires the initial condition data of the program and the specific data of each device to be calculated (S10). As the data of the initial condition, for example, the number of data in the log database (for example, the number of data collected in a 5-minute period for the past month), the maximum number of steps for repeatedly executing the genetic algorithm (for example, 5 times, 10 times, 20 times), there is a maximum number of generations of genetic algorithms. Specific data of each device includes, for example, device rating information (for example, rated capacity, rated power consumption), device reference coefficient, device influence, and initial individual fitness. Further, the optimal control personal computer 2e acquires past log data from the log database (S11). For example, in the case of a refrigerator, the past log data includes a cold water flow rate, a cold water outlet temperature, a cold water inlet temperature, a cooling water flow rate, a cooling water outlet temperature, a cooling water inlet temperature, and power consumption. As the past log data, data at the time of system start / stop may fluctuate, so data from 30 minutes after system start up to 30 minutes before stop is used.

  In the optimal control personal computer 2e, each coefficient of the device characteristic formula is regarded as a gene, and initial individuals are randomly generated (S12). Then, implement a genetic algorithm. Specifically, the optimal control personal computer 2e determines a living individual (S13), performs gene crossover (S14), and performs mutation (S15). In the optimal control personal computer 2e, the deviation between the energy consumption calculated by the device characteristic equation based on the generated gene (each coefficient of the device characteristic equation) and the actual measured value of energy consumption (past log data) ( (Corresponding to fitness) is calculated (S16). Furthermore, in the optimal control personal computer 2e, the energy consumption (fitness) calculated based on each gene is sorted in descending order (S17). Then, in the optimal control personal computer 2e, it is determined whether or not the gene operation for the maximum number of generations has been repeated (S18). If it is determined in S18 that the maximum number of generations of genetic manipulations have not been repeated, the optimal control personal computer 2e returns to the processing in S13 and performs genetic manipulations again.

  When it is determined in S18 that the genetic operation for the maximum number of generations has been repeated, the optimal control personal computer 2e determines whether the genetic algorithm for the maximum number of steps has been repeated (S18). If it is determined in S18 that the genetic algorithm for the maximum number of steps has not been repeated, the optimal control personal computer 2e returns to the processing in S12 to regenerate an initial individual at random, and inherits based on a different initial individual. Implement a genetic algorithm. Here, the genetic algorithm is repeatedly performed by sequentially changing the initial individuals. There may be cases where the initial individuals are not appropriate and individuals with high fitness cannot be obtained, but since there are several initial individuals, individuals with high fitness must be searched from those initial individuals. The individual with the highest fitness can be obtained from the individuals with the highest fitness.

  If it is determined in S18 that the genetic algorithm for the maximum number of steps has been repeated, the optimal control personal computer 2e uses the energy consumption (fitness) from all the solids that have executed the genetic algorithm for the maximum number of steps. Is selected (S20). Then, the optimal control personal computer 2e outputs the gene of the selected individual as each coefficient of the device characteristic equation (S21). The above processing is performed for all the devices of the air conditioning system 1, and the optimum coefficient of the device characteristic equation for all the devices is obtained.

  In this way, the genetic algorithm is not performed only once as in the prior art, but the genetic algorithm is performed a predetermined number of times while changing the initial individual. Since the genetic algorithm is implemented by various initial individuals, even if there are inappropriate initial individuals that do not become highly adaptable among the initial individuals, it is ensured that individuals with high fitness are obtained by other initial individuals Can do.

  The program B will be described with reference to FIG. Program B calculates the energy consumption (power consumption) of each device by the device characteristic formula incorporating the coefficient of each device obtained in Program A, and the energy consumption of the entire air conditioning system from the energy consumption of each device Calculate the quantity. This calculation of energy consumption is performed for all combinations of control values that are changed by a predetermined step size for each number of operating units, and the data in the database 2 is rewritten. Even when the number of operating units is the same, it is performed for all combinations in which the operating device and the stopping device are changed.

  First, the optimal control personal computer 2e acquires the definition of mode information and boundary information. The boundary information includes, for example, external conditions, secondary heat quantity (load factor), and the number of operating units.

  Next, the optimal control personal computer 2e acquires connection information and attribute information of each device when the calculation time is zero. Examples of the connection information include a water information connection number with each device, an air information connection number, a mode information connection number, an operation information connection number, a calculation order, and a device type. As attribute information, for example, in the case of a refrigerator, there are rated capacity, rated input, rated auxiliary machine input, rated cold water flow rate, rated cooling water flow rate, and coefficient. Further, the optimal control personal computer 2e acquires control values and operation mode information of each device. In this program B, a control value (for example, a flow rate by the inverter frequency, a cooling water temperature) is given as a boundary condition. Then, the optimal control personal computer 2e calculates the energy consumption of the device (object) using the acquired information. Until the flow rate and the amount of heat change, the combination of control values is changed and the above processing is repeated to calculate the energy consumption.

  FIG. 7 shows a model for calculating the energy consumption of an object. In the object O (in the case of the refrigerator), the attribute information and the device coefficient information of the device characteristic equation obtained by the program A are input, and the device characteristic equation including the device coefficient is incorporated. The object O has a relationship of water connection and air connection with other objects based on the acquired connection information. The operation information, mode information, water information (cold water), and water information (cooling water) are assigned to the nodes N1 to N4. ) Is input, water information (cold water), water information (cooling water), and energy information (power information) are output from each of the nodes N5 to N7.

  The optimal control personal computer 2e rewrites the database 2 every time the energy consumption (power consumption) of the device is calculated for any combination of control values. In the optimal control personal computer 2e, when the energy consumption (power consumption) of all devices is calculated for any combination of control values, the energy consumption (power) of the entire air conditioning system is calculated from the energy consumption of all devices. (Consumption) is calculated, and the database 2 is rewritten. In the optimal control personal computer 2e, the above processing is performed for all the numbers of operating units (all combinations of devices to be operated and devices to be stopped) and all combinations of control values, and the database 2 is completely rewritten. Finally, a combination of control values that minimizes the energy consumption of the entire air conditioning system 1 is selected for each number of operating units, and the database 2 is rewritten.

  The program C (when ON) will be described with reference to the flowchart of FIG. Program C (when ON) is executed when the operator turns on the optimal control during central monitoring. The program C (when ON) switches from the control value of the control system controller 2j to the control value of the optimum control when the optimum control is turned on. Therefore, the control for gradually switching the control value so that the control value does not change suddenly. The value is output, and when there is no deviation in the control value, the control value is completely switched to the optimal control value.

  When the optimum control is manually turned on by the central monitoring (S30), the optimum control personal computer 2e at every constant control period (for example, every five minutes), the temperature and humidity of the outside air, and the operating state of each device Or energy consumption is acquired (S31). Then, the optimal control personal computer 2e uses the database 1 to determine the number of operating units from the outside air enthalpy and the heat source load factor (secondary device load factor) (S32). Further, the optimal control personal computer 2e extracts the control value (control value of optimal control) of each device corresponding to the determined number of operating units from the database 2 (S33).

  In the optimal control personal computer 2e, the program E (when ON) is started, and the control value is changed from the past data and the immediately preceding data, the operation state of each device, and the like (S34). Here, when the optimal control is turned on, the control of the optimal control is controlled from the basic control value so that the control value does not change suddenly when switching from the basic control value of the control system controller 2j to the control value of the optimal control. A control value that gradually changes to a value is obtained. At this time, the balance between the cooling water pump 1b and the primary pump 1d is taken into consideration.

  In the optimal control personal computer 2e, it is determined whether or not the deviation between the control value of the optimal control by the database 2 and the control value by the program E (when ON) has become zero (S35). If it is determined in S35 that the deviation has become zero (S36), the optimum control is completely turned on (S36), and the optimum control personal computer 2e obtains the control value of the optimum control from the database 2 via the optimum control PLC 2d. Output as is. On the other hand, when it is determined in S35 that the deviation is not zero, the optimal control personal computer 2e outputs the control value output by the program E (when ON) to the control system controller 2j via the optimal control PLC 2d (S37). ).

  In the optimal control personal computer 2e, it is determined whether or not the optimal control is manually turned off by central monitoring (S38). If it is determined in S38 that the optimum control is not turned off, the optimum control personal computer 2e returns to the process of S31. On the other hand, if it is determined in S38 that the optimum control is turned off, the optimum control personal computer 2e shifts to the program C (when turned off) (S39).

  The program E (when ON) will be described with reference to the flowchart of FIG. The program E (when ON) outputs a control value that gradually switches from the basic control value of the control system controller 2j to the optimal control value. When the optimal control is turned on and the control value of the optimal control is close to the control value of the previous step, the control value of the optimal control is output as it is without gradually switching the control value.

  When the program E (when ON) is activated, the optimal control personal computer 2e acquires the temperature and humidity of the outside air in the previous step, the operating state and energy consumption of each device, the operation mode, and the time of the current step (S34a). Further, the optimal control personal computer 2e holds control values for the past one hour from the present time (S34b). For example, if the control cycle is a 5-minute cycle, control values for 12 times are held in one hour. Here, at least the control value of the previous step and the control value of the previous step may be held.

  In the optimal control personal computer 2e, the control value (Yco, n) of the optimal control by the database 2 is acquired using the data acquired in S34a, similarly to the processing of S32 and S33 of the program D (S34c). Note that n is the total number of control values.

  In the optimal control personal computer 2e, the control value (Yco, n) of the optimal control by the database 2 is used as a target value, and the PID is calculated from the deviation between the control value used in the previous step and the control value used in the previous step. A control value (Yff, n) is calculated by the [Proportional Integral Derivative] theory (S34d). This control value (Yff, n) is a control value that gradually approaches the control value (Yco, n) of the optimum control, and is a control value in which the amount of change from the control value Y ′ of the previous step is limited. The extent to which the control value is approximated can be set by a control parameter in the PID theory, and a person in the room is uncomfortable due to a change in the operating state of the air conditioning system 1 according to a change in the control value of the preceding and following steps. Not to the extent. Here, the control value Yff is calculated for each of n control values.

Then, the optimum control personal computer 2e calculates the difference | ΔY1 | between the control value Y ′ adopted in the previous step and the control value (Yco, n) of the optimum control by the database 2 by the equation (6) (S34e). ). Further, the optimal control personal computer 2e calculates the difference | ΔY3 | between the control value Y ′ adopted in the previous step and the control value (Yff, n) based on the PID theory by the equation (7) (S34e). Then, the optimal control personal computer 2e uses the | ΔY1 | and | ΔY3 | and the control value Y ′ of the previous step to calculate the control value Y at the current step according to the equation (8) (S34e). Here, until | ΔY1 | becomes equal to or smaller than | ΔY3 |, the control value (Yff, n) according to the PID theory is adopted as the control value Y at the current step. When | ΔY1 | As the control value Y, the control value (Yco, n) of optimum control is adopted. The control value Y of the current step is set so that the difference between the control value Y of the current step and the control value Y ′ of the previous step is at most 10% of the maximum manipulated variable (maximum fluctuation range of the control value). Set. Then, the optimal control personal computer 2e outputs the control value Y (S34f).

  In the transition period from OFF to ON of the optimal control, a control value that gradually approaches the control value of the optimal control from the control value of the basic control of the control system controller 2j is set as a control value that is actually output, and the control that is actually output When the deviation between the value and the control value of the optimum control disappears, the optimum control is completely turned on and the control value of the optimum control is output.

  The program C (when OFF) will be described with reference to the flowchart of FIG. Program C (when OFF) is executed when the operator turns off the optimal control in the central monitoring. The program C (when OFF) switches from the control value of the optimal control to the control value of the control system controller 2j when the optimal control is turned OFF. Therefore, the control for gradually switching the control value so that the control value does not change suddenly. When the value is output and there is no deviation in the control value, the control value is completely switched to the control value of the control system controller 2j.

  When the optimum control is manually turned off by the central monitoring (S50), the optimum control personal computer 2e starts the program E (at the time of OFF) every fixed control cycle (for example, every 5 minutes). The control value is changed from the data, the immediately preceding data, the operating state of each device, etc. (S51).

  The optimal control personal computer 2e determines whether or not the deviation between the control value by the control system controller 2j and the control value by the program E (when OFF) has become zero (S52). If it is determined in S52 that the deviation has become zero, the optimum control is completely turned OFF (S53), and the control value of the control system controller 2j is used. On the other hand, if it is determined in S52 that the deviation is not zero, the optimal control personal computer 2e outputs the control value output by the program E (when OFF) to the control system controller 2j (S54). The operation with the optimum control OFF is executed within 120 minutes at the maximum. In the case of more than that, it is forcibly switched to the control value of the control system controller 2j.

  The program E (at the time of OFF) is demonstrated along the flowchart of FIG. The program E (when OFF) outputs a control value that gradually switches from the control value of the optimum control to the control value of the basic control of the control system controller 2j. If the control value of the basic control is close to the control value of the previous step when the optimum control is turned off, the control value of the basic control is output as it is without gradually switching the control value.

  When the program E (when OFF) is activated, the optimum control personal computer 2e acquires the temperature and humidity of the outside air in the previous step, the operating state and energy consumption of each device, the operation mode, and the time of the current step (S51a). Further, the optimal control personal computer 2e holds control values for the past one hour from the present time (S51b).

  The optimal control personal computer 2e acquires the control value (Ym, n) of the basic control of the control system controller 2j based on the data acquired in S51a (S51c).

  In the optimal control personal computer 2e, the control value (Ym, n) of the basic control of the control system controller 2j is set as the target value, and the control value adopted in the previous step and the control value adopted in the previous step are The control value (Yff, n) is calculated from the deviation by the PID theory (S51d). The control value (Yff, n) is a control value that gradually approaches the control value (Ym, n) of the basic control of the control system controller 2j, and the amount of change from the control value Y ′ of the previous step is limited. Value. Here, the control value Yff is calculated for each of n control values.

Then, the optimum control personal computer 2e calculates the difference | ΔY2 | between the control value Y ′ adopted in the previous step and the control value (Ym, n) of the basic control of the control system controller 2j by the equation (9). (S51e). Further, the optimal control personal computer 2e calculates the difference | ΔY3 | between the control value Y ′ adopted in the previous step and the control value (Yff, n) based on the PID theory by the equation (10) (S51e). Then, the optimal control personal computer 2e uses the | ΔY2 | and | ΔY3 | and the control value Y ′ of the previous step to calculate the control value Y at the current step according to the equation (11) (S51e). Here, until | ΔY2 | becomes equal to or smaller than | ΔY3 |, the control value (Yff, n) based on the PID theory is adopted as the control value Y in the current step. When | ΔY2 | As the control value Y, the control value (Ym, n) of the basic control is adopted. Note that the control value Y of the current step is set so that the difference between the control value Y of the current step and the control value Y ′ of the previous step is 10% of the maximum manipulated variable of the control value. Then, the optimal control personal computer 2e outputs the control value Y (S51f).

  In the transition period from ON to OFF of the optimal control, a control value that gradually approaches the control value of the basic control of the control system controller 2j from the control value of the optimal control is set as a control value that is actually output, and the control that is actually output When the deviation between the value and the control value of the basic control disappears, the optimal control is completely turned off and the control value of the basic control is output.

According to the control system 2, an optimal control value can be obtained from the control table by changing the initial individual and repeatedly executing the genetic algorithm to update the control table. By generating the initial individual a predetermined number of times, the genetic algorithm can be implemented by various initial individuals. As a result, individuals with high fitness can be obtained from appropriate initial individuals among them, individuals with the highest fitness can be obtained from individuals with high fitness, and individuals with the highest fitness can be obtained. The optimal device coefficient for updating the control table can be acquired. Moreover, the energy efficiency of the air conditioning system 1 can be optimized by operating the air conditioning system 1 using an optimal control value. As a result, the energy consumption and CO ₂ emission amount of the entire system can be minimized, and energy saving and CO ₂ saving can be achieved.

  Furthermore, according to the control system 2, the control table is periodically updated (for example, every month) based on past log data such as the operation state and power consumption of each device of the air conditioning system 1. Optimal control can be performed with a control table reflecting the secular change of each device, and an optimal control value can be obtained.

  Further, according to the control system 2, when the control value of the optimum control by the database 2 and the control value of the basic control of the control system controller 2j are switched, the control value is changed abruptly so that the control value changes rapidly. This can prevent the steep change in the operating state of the air conditioning system 1. Further, in the control system 2, even when a failure occurs in the optimal control personal computer 2e or the optimal control PLC 2d and optimal control cannot be performed, basic control is possible by the control system controller 2j, and control of the air conditioning system 1 can be guaranteed.

  As mentioned above, although embodiment which concerns on this invention was described, this invention is implemented in various forms, without being limited to the said embodiment.

  For example, in the present embodiment, the present invention is applied to an air conditioning system (heat source system), but can be applied to other systems such as various plants.

  In the present embodiment, the control system is configured such that the control system controller performs basic control to ensure the minimum operation control for the air conditioning system 1. However, the basic control is not performed and optimal control is always performed. May be. Moreover, although an example of the hardware configuration for realizing the present invention has been described in the present embodiment, the hardware configuration for realizing the present invention may be configured in any other form.

In the present embodiment, the optimum control for the purpose of saving energy and reducing CO ₂ is used. However, the optimum control can be applied to other purposes.

  In the present embodiment, the control value for optimum control is obtained using the two control tables of database 1 and database 2. However, the control value for optimum control may be obtained by other methods. For example, the control value may be obtained from only one database as the control table. Further, instead of using a control table, a control value may be obtained using a control characteristic equation or the like. The control coefficient searched by the genetic algorithm changes according to such an optimal control method.

  In this embodiment, an example of a method of gradually switching from one control value to the other when the optimal control is turned from ON to OFF or from OFF to ON is shown. You may make it switch a value gradually. For example, the maximum allowable change amount (fixed value) from the control value of the previous step is prepared in advance for each control value, and the control value of the current step is obtained from the control value of the previous step with the maximum allowable change amount as a limit. . When the optimal control is switched from OFF to ON, if the difference between the control value of the optimal control and the control value of the previous step is greater than the maximum allowable change amount, the maximum allowable change amount is added to the control value of the previous step and the current When the difference between the control value of the optimal control and the control value of the previous step is equal to or less than the maximum allowable change amount, the control value of the optimal control is set as the control value of the current step.

  DESCRIPTION OF SYMBOLS 1 ... Air conditioning system, 1a ... Cooling tower, 1b ... Cooling water pump, 1c ... Refrigerator, 1d ... Primary pump, 2 ... Control system, 2a ... Intelligent controller, 2b ... BEMS, 2c ... PLC for heat source group management, 2d ... PLC for optimal control, 2e ... PC for optimal control, 2f ... network, 2g, 2h ... coaxial cable, 2i ... Ethernet, 2j ... control system controller, 2k ... coaxial cable.

Claims (3)

A control system that outputs a control value for controlling a control target system including a plurality of devices ,
There is a device characteristic formula for each device,
An optimal control device for generating an optimal control value for the system to be controlled;
The optimal control device is:
Implementation number setting means for setting the number of executions of the genetic algorithm;
An initial individual generation means for generating an initial individual;
An implementation means for implementing a genetic algorithm,
Up to the number of times it was performed genetic algorithm is performed the number of times set in the execution count setting means, for each of the devices, the generated different initial individual in early ontogeny means each for implementing a genetic algorithm, the generated By using the initial individual and performing the genetic algorithm using the coefficient of the device characteristic equation as a gene in the implementation means to search for the coefficient of the device characteristic equation having high fitness,
If the number embodying the genetic algorithm has become the practice number of times set by the execution count setting means, the device characteristics are searched respectively genetic algorithm implementation number of times set by the execution count setting means for each of the devices The coefficient of the device characteristic equation having the highest fitness is selected from among the coefficients of the equation , and output to the control target system based on the device characteristic equation for each device using the coefficient for the selected device. A control system for generating a control value to be generated. A control system that outputs a control value for controlling a control target system including equipment ,
An optimal control device for generating an optimal control value for the system to be controlled;
The optimal control device is:
Implementation number setting means for setting the number of executions of the genetic algorithm;
An initial individual generation means for generating an initial individual;
An implementation means for implementing a genetic algorithm,
Each time the genetic algorithm is executed, a different initial individual is generated by the initial individual generation means until the number of executions of the genetic algorithm reaches the execution number set by the execution number setting means, and the generated initial individual is used. The implementation means searches for the coefficient of the device characteristic equation having a high fitness by performing a genetic algorithm using the coefficient of the device characteristic equation of the device as a gene .
When the number of executions of the genetic algorithm is the number of executions set by the execution number setting means, the coefficients of the device characteristic formulas respectively searched by the genetic algorithm for the number of executions set by the execution number setting means Select a coefficient of the device characteristic formula having the highest fitness from among them, and generate a control value to be output to the control target system based on the device characteristic formula using the selected coefficient .
The fitness is a deviation between an energy consumption calculated by the device characteristic equation using the gene by a genetic algorithm and an actual measurement value of the energy consumption of the device . A basic control device for generating basic control values for the controlled system;
When the control value output to the control target system is switched between the optimum control device and the basic control device, the control value of the control device before switching is gradually changed to the control value of the control device after switching. The control system according to claim 1 or 2 .

Images