WO2023189446A1

WO2023189446A1 - Heat source machine system, pre-trained model generation method, and pre-trained model

Info

Publication number: WO2023189446A1
Application number: PCT/JP2023/009456
Authority: WO
Inventors: 知行内村; 慎也石原
Original assignee: 荏原冷熱システム株式会社
Priority date: 2022-03-31
Filing date: 2023-03-10
Publication date: 2023-10-05
Also published as: JP2023151008A

Abstract

This heat source machine system comprises: a heat source device for cooling or heating a heat medium to be supplied to a heat-demanding facility; a heat source auxiliary machine; and a control device for adjusting operating states of the heat source device and the heat source auxiliary machine. The control device has a pre-trained control model. The control model is trained with training data-based machine learning processing such that when an operating condition is input, operating states which enable predetermined indices to have values corresponding to the condition are output. The control device controls the heat source device and the heat source auxiliary machine so as to achieve the operating states output by the control model. In the case of this pre-trained model generation method, the training data is generated by means of simulation, and the pre-trained model is generated by implementing machine learning processing by using the training data.

Description

Heat source system, learned model generation method, and learned model

The present disclosure relates to a heat source system, a method for generating a learned model, and a learned model, and particularly relates to a heat source system, a method for generating a learned model, and a learned model that control equipment using a machine-learned model.

In order to supply cold water or hot water to an air conditioner to cool or heat the air, a heat source system is generally constructed by appropriately combining a heat source device and its auxiliary equipment. The heat source device system adjusts the output of the devices that make up the heat source device system and the operating status of the number of devices in operation, depending on the load of the air conditioner. As devices that can be useful for controlling the equipment that makes up the heat source equipment system, there are devices that have a reproduction system that calls and reproduces actual operation data, and a system that can perform simulations using past heat load data and facility environment data ( For example, see Japanese Patent Application Publication No. 2011-163727).

In order to control the equipment that makes up the heat source equipment system with good response, it is preferable to shorten the calculation time to find the appropriate operating state.

In view of the above-mentioned problems, the present disclosure relates to providing a heat source equipment system, a method for generating a trained model, and a trained model that can be operated in an appropriate operating state in response to changes in operating conditions.

A heat source device system according to a first aspect of the present disclosure includes a heat source device that cools or heats a heat medium to be supplied to heat demand equipment, a heat source auxiliary machine that operates in conjunction with the operation of the heat source device, and the heat source device. and a control device that adjusts the operating state of the heat source auxiliary device, the control device having a learned control model, the heat source auxiliary device controlling the heat source to flow the heat medium passing through the heat source device. a medium pump; and a heat source fluid supply device that supplies a heat source fluid that directly or indirectly exchanges heat with the heat medium in the heat source device to the heat source device, and the operating state is the operating status of the heat source device. , a flow rate of the heat medium discharged by the heat medium pump, and a flow rate of the heat source fluid supplied by the heat source fluid supply device, and the control model influences the operating state. Machine learning processing using teacher data is performed so that when operating conditions are input, the operating conditions in which a predetermined index has a value that meets the conditions are output, and the operating conditions are based on the heat demand. The predetermined index includes at least one of the heat demand of the equipment or a physical quantity correlated thereto, and the outside air temperature or a physical quantity correlated thereto, and the predetermined index includes the power consumption of the heat source equipment and the heat source auxiliary equipment, the heat source The control device includes at least one of the operating costs of the equipment and the heat source auxiliary equipment, and the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment, and the control device is configured to be in the operating state outputted by the control model. The heat source equipment and the heat source auxiliary equipment are controlled.

With this configuration, the control model can output an appropriate operating state in response to changes in operating conditions, so the heat source system can be operated with good response.

Further, as a heat source device system according to a second aspect of the present disclosure, in the heat source device system according to the first aspect of the present disclosure, the predetermined index may include power consumption of the heat source device and the heat source auxiliary device; The control model includes a plurality of operating costs of the heat source equipment and the heat source auxiliary equipment, and carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment, and the control model includes a first one of the plurality of predetermined indicators. Selecting a plurality of operating states in which a predetermined index has a value that satisfies a condition, and from among the plurality of selected operating states, a second predetermined index different from the first predetermined index satisfies the condition. The operating state may be configured to output the operating state as a value.

With this configuration, the heat source device system can be operated in an appropriate operating state that is comprehensively determined based on a plurality of predetermined indicators.

Further, as a heat source device system according to a third aspect of the present disclosure, in the heat source device system according to the first aspect or second aspect of the present disclosure, the heat source device, the heat medium pump, and the heat source fluid supply At least one of the devices is composed of a plurality of devices, and the operating state includes the number of operating devices of the heat source device, the heat medium pump, and the heat source fluid supply device, and The control model includes a first control model that outputs the number of operating vehicles as an integer value among the operating states, and an integer value of the number of operating vehicles output by the first control model as one of the operating conditions. and a second control model to be input.

With this configuration, it is possible to avoid the output of the control model regarding the number of operating units from becoming an unrealizable decimal number, and it is possible to operate the heat source equipment system in an appropriate operating state in accordance with actual operation. can.

Moreover, in the heat source device system according to any one of the first to third aspects of the present disclosure, as a heat source device system according to a fourth aspect of the present disclosure, the control device using the output of the control model for some of the operating states of the processing heat amount, the flow rate of the heat medium discharged by the heat medium pump, and the flow rate of the heat source fluid supplied by the heat source fluid supply device, The remaining operating states may be determined by simulation or rule-based.

With this configuration, the time required to output a part of the operating state can be shortened by using the control model, while the rest of the operating state can be output with higher accuracy by using simulation or a rule base.

Further, in the heat source device system according to a fifth aspect of the present disclosure, in the heat source device system according to any one of the first to fourth aspects of the present disclosure, the operating conditions are It may also include the pressure loss coefficient of the equipment.

With this configuration, the heat source system can be operated in an appropriate operating state even when the flow rate of the heat medium changes.

Further, in the heat source device system according to a sixth aspect of the present disclosure, in the heat source device system according to any one of the first to fifth aspects of the present disclosure, the operating conditions are and the cost per unit power consumption of the heat source auxiliary equipment, and the carbon dioxide emissions per unit power consumption of the heat source equipment and the heat source auxiliary equipment.

With this configuration, there is no need to recreate the control model even if the cost per unit power consumption or the amount of carbon dioxide emissions changes.

Further, in the heat source device system according to any one of the first to sixth aspects of the present disclosure as a heat source device system according to a seventh aspect of the present disclosure, the control model is based on the predetermined control model. The index may be one or more of the power consumption of the heat source equipment and the heat source auxiliary equipment, the operating cost of the heat source equipment and the heat source auxiliary equipment, and the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment. The control device may have a plurality of control models, and the control device may use an appropriate control model among the plurality of control models according to the predetermined index desired by the user.

With this configuration, even if the predetermined index is changed by the user, the appropriate operating state of the heat source device system can be maintained.

Further, a method for generating a trained model according to an eighth aspect of the present disclosure includes a heat source device that cools or heats a heat medium supplied to heat demand equipment, and a heat source auxiliary device that operates in conjunction with the operation of the heat source device. A method for generating a learned model used for controlling a heat source device system comprising: By calculating a plurality of driving conditions while changing them through simulation, and defining the driving condition when the predetermined index meets the condition as a relationship with the driving condition, and performing this for a plurality of driving conditions, a step of generating training data by obtaining a plurality of sets of relationships between the driving state and the driving condition when the predetermined index meets the condition; and performing machine learning processing using the training data. a step of generating a trained model in which operating conditions are input and the operating state is output, the heat source auxiliary equipment includes a heat medium pump that flows the heat medium passing through the heat source equipment, and the heat source a heat source fluid supply device that supplies a heat source fluid that directly or indirectly exchanges heat with the heat medium in the equipment to the heat source equipment, and the operating condition is the amount of heat demanded by the heat demand equipment or correlated thereto. The operating state includes at least one of a physical quantity and an outside temperature or a physical quantity correlated thereto, and the operating state includes the operating state of the heat source device, the flow rate of the heat medium discharged by the heat medium pump, and the heat source fluid. The predetermined index includes at least one of the flow rate of the heat source fluid supplied by the supply device, and the predetermined index includes the power consumption of the heat source equipment and the heat source auxiliary equipment, the operating cost of the heat source equipment and the heat source auxiliary equipment, and at least one of the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment.

With this configuration, it is possible to obtain a trained model that allows the heat source device system to operate in an appropriate operating state with good response in response to changes in operating conditions.

Further, as a method for generating a trained model according to a ninth aspect of the present disclosure, in the method for generating a trained model according to the eighth aspect of the present disclosure, the heat source device, the heat medium pump, and the heat source fluid At least one of the supply devices is composed of a plurality of units, and the operating state includes the number of operating units of the heat source device, the heat medium pump, and the heat source fluid supply device, which are composed of a plurality of units, The step of generating the trained model includes the step of generating a first trained model that includes an integer value of the number of operating vehicles in the operating state of the output, and using the integer value of the number of operating vehicles as one of the operating conditions. The method may include a step of generating a second trained model to be input.

With this configuration, it is possible to avoid the output of the control model regarding the number of operating vehicles from becoming an unrealizable decimal number.

Further, as a trained model generation method according to a tenth aspect of the present disclosure, in the trained model generation method according to the ninth aspect of the present disclosure, the first trained model and the second trained The trained model uses a neural network, the first trained model and the second trained model have a common intermediate layer, and the driving state is the output of the first trained model. The items include the driving state item that is the output of the second trained model, and the step of generating the trained model includes first applying machine learning processing to the first learning model to generate the first learning model. A trained model is generated, and then a machine learning process is performed on the model in which the initial value of the weighting coefficient of the intermediate layer of the second learning model is set to the weighting coefficient of the intermediate layer of the first trained model. The second trained model may be generated by performing the following steps.

With this configuration, it is possible to shorten the learning time when generating the second trained model.

Further, as a method for generating a trained model according to an eleventh aspect of the present disclosure, in the method for generating a trained model according to any one of the eighth to tenth aspects of the present disclosure, the In the step of generating data, at least one of the operating conditions and the operating state for performing the simulation may be determined at random.

With this configuration, it is possible to prevent the amount of data to be handled from increasing more than necessary, and additional creation of teacher data can be performed relatively easily.

Further, as a method for generating a trained model according to a twelfth aspect of the present disclosure, in the method for generating a trained model according to any one of the eighth to eleventh aspects of the present disclosure, the teacher In the step of generating data, data of the predetermined index already exists when the operating condition is in a wider range than the operating condition assumed under the operating condition in a wider range than the assumed operating condition. In this case, the predetermined index at the assumed operating state under the assumed operating condition may be extracted from the already existing data and used as the teacher data.

With this configuration, the teacher data can be created using existing data, and the time required to create the teacher data can be shortened.

Further, as a method for generating a trained model according to a thirteenth aspect of the present disclosure, in the method for generating a trained model according to the twelfth aspect of the present disclosure, the step of generating the teacher data is not extracted. The teaching data may be generated by performing a simulation after changing the values of items that do not match the assumed operating conditions and operating states in the data to values that match the operating conditions and operating states. good.

With this configuration, missing teacher data can be efficiently replenished.

Further, as a method for generating a trained model according to a fourteenth aspect of the present disclosure, in the method for generating a trained model according to any one of the eighth to thirteenth aspects of the present disclosure, the driving The conditions include a pressure loss coefficient of the heat demand equipment, and the step of generating the teacher data includes calculating a pressure loss in the heat demand equipment based on the pressure loss coefficient and the flow rate of the heat medium in the simulation. After that, the predetermined index may be determined.

With this configuration, it is possible to generate a trained model that can output an appropriate operating state even when the flow rate of the heat medium changes.

Further, the trained model according to the fifteenth aspect of the present disclosure includes a heat source device that cools or heats a heat medium to be supplied to heat demand equipment, and a heat source auxiliary machine that operates in conjunction with the operation of the heat source device. A trained model installed in a computer used to control a heat source device system, comprising an input layer into which operating conditions of the heat source device system are input, and an output layer into which operating conditions of the heat source device system are output. , an intermediate layer in which parameters are learned using teacher data that inputs the operating condition and outputs the operating condition in which a predetermined index has a value that meets the condition; A heat medium pump that flows the heat medium passing through the device; and a heat source fluid supply device that supplies the heat source fluid to the heat source device, which directly or indirectly exchanges heat with the heat medium in the heat source device; The operating conditions include at least one of the heat demand of the heat demand equipment or a physical quantity correlated thereto, and the outside air temperature or a physical quantity correlated thereto, and the operating state includes the operating status of the heat source equipment; The predetermined index includes at least one of the flow rate of the heat medium discharged by the heat medium pump and the flow rate of the heat source fluid supplied by the heat source fluid supply device, and the predetermined index The operating conditions are input to the input layer, including at least one of the power consumption of the machine, the operating cost of the heat source equipment and the heat source auxiliary equipment, and the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment. Then, the computer is operated to perform calculations in the intermediate layer and output the operating state from the output layer.

With this configuration, the model becomes a trained model that can operate the heat source equipment system in an appropriate operating state and with good response in response to changes in operating conditions.

Further, a heat source device system according to a sixteenth aspect of the present disclosure includes a control device having a trained model according to the fifteenth aspect of the present disclosure, the heat source device, and the heat source auxiliary device, The control device may control the heat source device and the heat source auxiliary device so as to be in the operating state outputted by the learned model.

With this configuration, the heat source device system can be operated in an appropriate operating state with good response in response to changes in operating conditions.

According to the present disclosure, the heat source device system can be operated in an appropriate operating state with good response in response to changes in operating conditions.

FIG. 1 is a schematic diagram of a heat source device system according to an embodiment. It is a flow chart which shows the procedure of creating training data used for machine learning processing of a control model provided in a heat source equipment system concerning one embodiment. FIG. 3 is a diagram illustrating operating conditions assumed when creating training data. It is a figure which illustrates the driving state assumed when creating training data. 12 is a flowchart illustrating a procedure for creating teacher data by reusing some existing data. FIG. 2 is a schematic configuration diagram of a first control model. FIG. 3 is a schematic configuration diagram of a second control model. FIG. 2 is a block diagram showing a calculation procedure of a control device of a heat source device system according to an embodiment. FIG. 1 is a block diagram illustrating the hardware configuration of an exemplary computer.

This application is based on Japanese Patent Application No. 2022-60393 filed in Japan on March 31, 2022, and the contents thereof form a part of the contents of this application.
In addition, the present invention may be more fully understood from the detailed description that follows. Further scope of applicability of the invention will become apparent from the detailed description below. However, the detailed description and specific examples are preferred embodiments of the invention and are presented for illustrative purposes only. From this detailed description, various changes and modifications will be apparent to those skilled in the art within the spirit and scope of the invention.
Applicant does not intend to offer any of the described embodiments to the public, and the applicant does not intend to offer any of the described embodiments to the public, and any disclosed modifications or alternatives that may not literally fall within the scope of the claims are considered equivalents. be part of the invention under discussion.

Hereinafter, embodiments will be described with reference to the drawings. In each figure, members that are the same or correspond to each other are designated by the same or similar reference numerals, and redundant explanations will be omitted.

<Example of configuration of heat source system>
First, with reference to FIG. 1, a heat source device system 1 according to an embodiment will be described. FIG. 1 is a schematic diagram of the heat source device system 1. As shown in FIG. The heat source device system 1 includes three heat source devices 11, 12, and 13, three cold/hot water pumps 21, 22, and 23, three cooling towers 31, 32, and 33, and three cooling water pumps 41. , 42, 43, and a control device 70 as main components. The heat source device system 1 is a system that supplies cold and hot water CH cooled or heated by each of the heat source devices 11, 12, and 13 to the heat demand equipment 99 according to a request. Examples of the heat demand equipment 99 include air conditioning equipment such as an air handling unit and a fan coil unit. The cold/hot water CH is a medium that conveys cold heat or hot heat to the heat demand equipment 99, and corresponds to a heat medium. Cold/hot water CH is a general term for cold water, which is a medium for cold heat, and hot water, which is a medium for heat. Typically, when the heat demand equipment 99 performs cooling, it becomes cold water, and when the heat demand equipment 99 performs heating, it becomes cold water. When the water is drained, the water becomes warm.

The three heat source devices 11, 12, and 13 are devices that cool or heat the cold/hot water CH, and correspond to heat source devices. Hereinafter, in order to distinguish the three heat source devices, they may be referred to as the first heat source device 11, the second heat source device 12, and the third heat source device 13, respectively. Each heat source device 11, 12, 13 can use various types of devices, but in this embodiment, both the first heat source device 11 and the second heat source device 12 are variable speed centrifugal chillers having the same characteristics. The explanation will be given assuming that the third heat source device 13 is a fixed speed centrifugal refrigerator. Each of the heat source devices 11, 12, and 13 is a device that can cool or heat the cold/hot water CH using input electric power. The amount of heat that each heat source device 11, 12, and 13 removes from the cold/hot water CH or gives to the cold/hot water CH is referred to as "processed heat amount." The first heat source device 11 and the second heat source device 12 are capable of adjusting the amount of heat to be processed through capacity control. The third heat source device 13 is operated by switching between an operating state and a stopped state, and the processing heat amount is either rated or zero. The number of operating heat source devices 11, 12, and 13 and the operating capacity of the first heat source device 11 and the second heat source device 12 are referred to as "operating status." The amount of heat processed by each of the heat source devices 11, 12, and 13 is determined depending on the operating status. Each heat source device 11, 12, 13 is capable of detecting a temperature difference and a pressure difference between cold and hot water CH flowing in and out. In addition, each of the heat source devices 11, 12, and 13 gives the heat taken from the cold/hot water CH to the cooling water CD in order to cool the cold/hot water CH, and deprives the cooling water CD of the heat for heating the cold/hot water CH. It is composed of Cooling water CD is a fluid that indirectly exchanges heat with cold/hot water CH in each heat source device 11, 12, 13 via a refrigerant (not shown), and corresponds to a heat source fluid.

The three cold/hot water pumps 21, 22, and 23 are devices that operate in conjunction with the operation of the respective heat source devices 11, 12, and 13, and are a form of heat source auxiliary equipment, and also pump cold and hot water CH. It is a device that causes fluid to flow and corresponds to a heat medium pump. Hereinafter, in order to distinguish the three cold/hot water pumps, they may be referred to as the first cold/hot water pump 21, the second cold/hot water pump 22, and the third cold/hot water pump 23, respectively. Each of the cold and hot water pumps 21, 22, and 23 is a device that can flow cold and hot water CH using input electric power. In this embodiment, each cold/hot water pump 21, 22, 23 has an inverter and is configured to be able to steplessly change the flow rate of the cold/hot water CH to be discharged.

The three cooling towers 31, 32, and 33 are devices that operate in conjunction with the operation of each heat source device 11, 12, and 13, and are a form of heat source auxiliary equipment, and each coolant CD is It is a device that supplies heat source devices 11, 12, and 13, and corresponds to a heat source fluid supply device. Hereinafter, in order to distinguish the three cooling towers, they may be referred to as a first cooling tower 31, a second cooling tower 32, and a third cooling tower 33, respectively. Each of the cooling towers 31, 32, and 33 is a device that can perform heat exchange between the cooling water CD and the atmosphere (typically, outside air) using input electric power. In this embodiment, each cooling tower 31, 32, 33 is operated/stopped in accordance with the operation/stop of each heat source device 11, 12, 13, but the rotation speed of the fan is adjusted steplessly or stepwise. It may be configured such that it can be adjusted.

The three cooling water pumps 41, 42, and 43 are devices that operate in conjunction with the operation of each heat source device 11, 12, and 13, and are a form of heat source auxiliary equipment, and also pump cooling water CD. It is a device that supplies each of the heat source devices 11, 12, and 13, and is one form of a heat source fluid supply device. Hereinafter, in order to distinguish the three cooling water pumps, they may be referred to as a first cooling water pump 41, a second cooling water pump 42, and a third cooling water pump 43, respectively. Each of the cooling water pumps 41, 42, and 43 is a device that can flow the cooling water CD using input electric power. In this embodiment, each of the cooling water pumps 41, 42, and 43 includes an inverter and is configured to be able to steplessly change the flow rate of the cooling water CD to be discharged.

The first heat source device 11 is connected to the outgoing header 19 via the first cold/hot water outgoing pipe 15 , and is also connected to the return header 29 via the first cold/hot water return pipe 25 . The first cold/hot water pump 21 is provided in the first cold/hot water return pipe 25 . Further, the first heat source device 11 is connected to the first cooling tower 31 via a first cooling water outgoing pipe 35 and a first cooling water return pipe 45. The first cooling water pump 41 is provided in the first cooling water outgoing pipe 35 . The second heat source device 12 is connected to the outgoing header 19 via the second cold/hot water outgoing pipe 16 , and is also connected to the return header 29 via the second cold/hot water return pipe 26 . The second cold/hot water pump 22 is provided in the second cold/hot water return pipe 26 . Further, the second heat source device 12 is connected to the second cooling tower 32 via a second cooling water outgoing pipe 36 and a second cooling water return pipe 46. The second cooling water pump 42 is provided in the second cooling water outgoing pipe 36. The third heat source device 13 is connected to the outgoing header 19 via the third cold/hot water outgoing pipe 17 , and is also connected to the return header 29 via the third cold/hot water return pipe 27 . The third cold/hot water pump 23 is provided in the third cold/hot water return pipe 27 . Further, the third heat source device 13 is connected to the third cooling tower 33 via a third cooling water outgoing pipe 37 and a third cooling water return pipe 47. The third cooling water pump 43 is provided in the third cooling water outgoing pipe 37.

The forward header 19 is a member that collects cold and hot water CH that has been cooled or heated by each of the heat source devices 11, 12, and 13. The outgoing header 19 is connected to the heat demand equipment 99 via the supply pipe 91, and is configured to be able to guide the cold and hot water CH collected in the outgoing header 19 to the heat demand equipment 99. The heat demand equipment 99 and the return header 29 are connected via a recovery pipe 92. The return header 29 is configured to allow cold and hot water CH whose heat has been utilized in the heat demand equipment 99 to flow in through the recovery pipe 92 . The cold/hot water CH that has flowed into the return header 29 is distributed to each heat source device 11, 12, 13 according to the operating status of each heat source device 11, 12, 13. In this way, the return header 29 is a member that distributes the cold/hot water CH after the cold or hot heat has been utilized to each of the heat source devices 11, 12, and 13.

The control device 70 is a device that controls the operation of the heat source device system 1. The control device 70 is connected to each heat source device 11, 12, 13 by a communication line (wired or wireless). The control device 70 is configured to be able to receive operation information data from each of the heat source devices 11, 12, and 13. Examples of receivable operation data include temperature and pressure at the outlet and temperature and pressure at the inlet of cold and hot water CH, temperature and pressure at the outlet and temperature and pressure at the inlet of cooling water CD, and all or part of power consumption, etc. can be mentioned. Note that in this embodiment, the flow rate can be calculated from the pressure difference between the inflow and outflow of the fluid, but the flow rate may also be directly detected. Further, the control device 70 is configured to be able to adjust the operating status of each heat source device 11, 12, 13 by transmitting a control signal to each heat source device 11, 12, 13. Further, the control device 70 is configured to be able to control the operation of the heat source auxiliary machines via the heat source machines 11, 12, 13 and the auxiliary machine power panel 75. The auxiliary machine power panel 75 is a device that controls the operation of the heat source auxiliary machine. The auxiliary machine power panel 75 is connected to each heat source auxiliary machine by a signal line (wired or wireless). The auxiliary power panel 75 starts, stops, and discharges water by adjusting the power supplied to each cold/hot water pump 21 , 22 , 23 based on the control signal received from each heat source device 11 , 12 , 13 . It is configured to be able to control the flow rate of cold and hot water CH. In addition, the auxiliary power panel 75 controls the start and stop of each cooling tower 31 , 32 , 33 by adjusting the power supplied to each cooling tower 31 , 32 , 33 based on the control signal received from each heat source device 11 , 12 , 13 . It is configured so that it can be In addition, the auxiliary power panel 75 adjusts the power supplied to each of the cooling water pumps 41, 42, and 43 based on the control signals received from each of the heat source devices 11, 12, and 13. It is configured such that the flow rate of the cooling water CD to be discharged can be controlled.

The control device 70 is further configured to receive a temperature information signal via a communication line (wired or wireless) from an outside air thermometer 61 that detects the outside air temperature, so as to be able to grasp the outside air temperature. . In addition, it is preferable that the outside air thermometer 61 is arrange|positioned so that the temperature of the outside air around each cooling tower 31,32,33 may be detected. Further, the control device 70 is configured to be able to grasp the amount of heat demanded in the heat demand equipment 99 by receiving information on the amount of heat demanded in the heat demand equipment 99 as a signal via a communication line (wired or wireless). ing. Further, the control device 70 has a control model 80. The control model 80 is a model that is subjected to machine learning processing on a computer, into which operating conditions are input in order to output operating states. Here, the operating state is the operating state of the heat source equipment and heat source auxiliary equipment that can be controlled by the heat source equipment system 1. As a specific example, first, starting and stopping each heat source equipment 11, 12, 13 (in other words, the number of units in operation) and the amount of heat processed (in other words, the output). In addition, as other examples of the operating state, the discharge flow rate of cold and hot water CH from each cold and hot water pump 21, 22, 23 (flow rate 0 is a stopped state), starting and stopping of each cooling tower 31, 32, 33, and each cooling water pump Examples include the discharge flow rate of the cooling water CD at 41, 42, and 43 (a flow rate of 0 is a stopped state). On the other hand, the operating conditions are conditions that affect the operating state, and are basically conditions that cannot be controlled by the heat source device system 1. The operating conditions include, for example, the amount of heat demanded by the heat demand equipment 99, the outside air temperature, the target temperature of the cold/hot water CH flowing out from the outgoing header 19 toward the heat demand equipment 99, the pressure loss coefficient on the heat demand equipment 99 side, etc. . In addition, when the operating conditions include the unit price of resources (for example, electricity, water, fuel, etc.) consumed during the operation of the heat source equipment system 1, the carbon dioxide emissions per unit consumption of these consumed resources, etc. There is also. The total unit price of various consumption resources used to power the equipment constituting the heat source device system 1 corresponds to the cost per unit consumption power. Further, the total amount of carbon dioxide emissions per unit consumption of various consumption resources used to power the equipment constituting the heat source device system 1 corresponds to the amount of carbon dioxide emissions per unit consumption power.

Note that including the pressure loss coefficient on the heat demand equipment 99 side in the operating conditions has the following advantages. First, as a premise, the pressure loss on the heat demand equipment 99 side changes depending on the total flow rate of cold and hot water CH, which changes depending on the number of operating air conditioners that make up the heat demand equipment 99. Proportional to the square of the flow rate. In this embodiment, this proportionality constant (coefficient) is referred to as a "pressure loss coefficient", and the pressure loss on the heat demand equipment 99 side is calculated by multiplying the pressure loss coefficient by the square of the total flow rate of cold and hot water CH. be able to. Then, based on the determined pressure loss on the heat demand equipment 99 side, the motive power of each cold/hot water pump 21, 22, 23 can be calculated. Therefore, even if the operating state changes and the flow rate of cold/hot water CH on the heat demand equipment 99 side changes, the motive power of each cold/hot water pump 21, 22, 23 can be estimated correctly. The reason why the control device 70 has the control model 80 is based on the following background.

As a common method of conventional system control, there is a method in which simulations are performed assuming various operating conditions for the operating conditions under which the heat source equipment is currently operating, and the optimum operating condition is determined and operated. However, as the number of controllable parameters increases, the computational load increases, making it increasingly difficult to determine the optimal operating state assuming all possible operating states. In practice, this is dealt with by narrowing down the number of variable parameters, reducing the amount of calculation by making various assumptions, or lengthening the calculation cycle. With such limited control, it is difficult to operate under truly optimal operating conditions. Therefore, in the heat source device system 1 according to the present embodiment, a control model 80 subjected to machine learning processing, which will be described later, is used to achieve the optimum operating state as much as possible. An overview of the procedure for generating the control model 80 is to first create training data, and then use the created training data to perform machine learning processing on a computer.

<Creating teaching data>
FIG. 2 is a flowchart showing the procedure for creating teacher data. In the following description, when referring to the configuration of the heat source device system 1, reference will be made to FIG. 1 as appropriate. Further, the heat source devices 11, 12, and 13 may be collectively referred to as heat source devices. Moreover, each cold/hot water pump 21, 22, 23, each cooling tower 31, 32, 33, and each cooling water pump 41, 42, 43 may be collectively referred to as a heat source auxiliary machine. In this embodiment, simulation is used to create teacher data. The simulation in this embodiment is typically performed at a location different from the installation location of the heat source device system 1. The simulation in this embodiment is used to estimate a predetermined index under a given operating condition and an assumed operating state. Here, the predetermined index is a standard that the user of the heat source device system 1 pays attention to when operating the heat source device system 1. Examples of the predetermined index include power consumption, operating cost, carbon dioxide emissions, etc. of the heat source equipment and heat source auxiliary equipment when the heat source equipment system 1 is operated. Note that the simulation calculation of the heat source device system 1 generally requires recursive calculation, and the calculation requires a certain amount of time. The computing devices installed in conventional heat source equipment systems, which use simulation results to directly control equipment, have difficulty adopting high-performance computing devices due to issues such as installation space, power consumption, noise environment, and cost. It tends to be longer. However, since the calculation device used for the simulation in creating the teacher data shown in FIG. 2 does not need to be installed at the site where the heat source device system 1 is constructed, a high-performance calculation device can be employed. Furthermore, since sufficient time is available for simulation calculations, there is an advantage that the optimum operating state can be found by assuming a wide range of operating states.

As shown in FIG. 2, when creating teacher data, first, fixed conditions are prepared (S1). Fixed conditions are elements that do not change, such as the energy consumption characteristics and pressure loss coefficient (different from the pressure loss coefficient on the heat demand equipment 99 side) of each heat source equipment and heat source auxiliary equipment. It is unique. The fixed conditions can generally be prepared based on characteristic data of each device and various formulas for the fluid (for example, a relational formula between flow rate and pressure drop). Once the fixed conditions are prepared, operating conditions are assumed (S2). Operating conditions can be assumed within the range of expected applications. The expected range of application (assumed operating conditions) is intended to exclude, for example, conditions such as an outside temperature of 70° C., which is unlikely to occur in reality. As a specific example of the assumed operating conditions, as shown in operating condition "1" in FIG. Examples include a pressure loss coefficient of 200 kPa/1000 LPM on the demand equipment 99 side. In addition, specific examples of the assumed operating conditions may include electricity charges of 15 yen/kWh and water charges of 150 yen/m3. Further, although not shown, the carbon dioxide emission coefficient for electricity and the carbon dioxide emission coefficient for water supply may be included.

Once the operating conditions are assumed, the operating state is assumed (S3). Operating conditions can also be assumed to the extent that the application is envisaged. The expected range of application (assumed operating state) is intended to exclude, for example, a situation where the number of heat source devices is three and an unrealizable operation of five heat source devices. As a specific example of the assumed operating state, as shown in operating state "1" in FIG. An example of this is to set the flow rate of cold/hot water CH to 100%. In the operating state "1", in addition, all three cooling towers 31, 32, and 33 are operated, and the flow rate of cooling water CD in the three cooling water pumps 41, 42, and 43 is set to 100%. I'm assuming. After assuming the operating condition, a predetermined index under the assumed operating condition and the assumed operating condition is calculated by simulation (S4). Calculation of this predetermined index can be performed, for example, in the following manner.

First, the total flow rate of cold and hot water CH can be determined from the number of operating heat source devices 11, 12, and 13 under assumed operating conditions and the flow rate of cold and hot water CH of each cold and hot water pump 21, 22, and 23. . Then, the return temperature of the cold/hot water CH (the temperature of the cold/hot water CH entering the heat source device) can be determined from the total flow rate of the cold/hot water CH, the target temperature of the cold/hot water CH under the assumed operating conditions, and the required heat amount. Next, the load heat amount of each heat source device 11, 12, 13 can be calculated from the temperature difference between the input and output of cold/hot water CH to each heat source device 11, 12, 13 and the flow rate of cold/hot water CH. Then, assuming the efficiency (COP) of each heat source device 11, 12, 13, the amount of heat to be processed is determined. Then, the temperature of the cooling water CD entering each heat source device 11, 12, 13 is determined from the amount of heat to be processed, the flow rate of the cooling water CD in each cooling water pump 41, 42, 43, and the outside air temperature. The efficiency of each heat source device 11, 12, 13 is determined from the inlet temperature of the cooling water CD, the flow rate of the cooling water CD, and the conditions on the cold/hot water CH side. Here, iterative calculations (convergence calculations) are performed for each of the heat source devices 11, 12, and 13 so that the calculated efficiency matches the previously assumed efficiency. Thereby, the efficiency of each heat source device 11, 12, 13 is determined, and thereby the power consumption of each heat source device 11, 12, 13, the amount of supplementary water, etc. can be determined.

Regarding the power of the heat source auxiliary equipment, for example, the power consumption of each cold/hot water pump 21, 22, 23 can be determined as follows. First, the pressure loss on the heat demand equipment 99 side is determined from the flow rate of cold/hot water CH and the pressure loss coefficient on the heat demand equipment 99 side, and then the pressure of each heat source equipment 11, 12, 13 defined as a fixed condition is determined. The pressure loss of each heat source device 11, 12, and 13 is determined from the loss data. By adding up both pressure losses, the required head (pressure) of each cold/hot water pump 21, 22, 23 can be determined, and from this and the flow rate of cold/hot water CH, the power consumption of each cold/hot water pump 21, 22, 23 can be calculated. It can be estimated. Similarly, by estimating the power consumption of other heat source auxiliary machines and adding these together, the power consumption of the entire heat source device system 1 can be determined. Furthermore, the amount of water consumed can be calculated from the amount of evaporated water obtained by dividing the amount of heat processed in each of the heat source devices 11, 12, and 13 by the latent heat of vaporization of water, and the target concentration ratio. Note that, as is generally done, the water consumption amount may be determined by multiplying the flow rate of the cooling water CD by a certain coefficient (about 5%). In this way, if you can calculate power consumption, water consumption, etc., you can calculate the operating cost by multiplying it by the cost per unit quantity and adding it up, and you can calculate the carbon dioxide emissions by multiplying it by the carbon dioxide emission coefficient and adding it up. You can ask for it.

After calculating the predetermined index under the assumed operating condition and the assumed operating state in the manner described above, it is determined whether the calculated number of predetermined indexes under the operating condition is sufficient (S5). Here, whether the calculated number of predetermined indicators is sufficient is to be determined by comparing and examining multiple predetermined indicators to determine what kind of operating state should be achieved under the relevant operating conditions. However, the question is whether a predetermined index exists that is sufficient to determine a preferable operating state. This means that under the relevant operating conditions, the above-mentioned simulation in which the predetermined index is calculated by changing the operating state is performed multiple times. In the explanation of the procedure for creating teacher data up to this point, the predetermined index has been calculated only in one driving state under this operating condition, so the number of predetermined indexes calculated is not sufficient. In the step (S5) of determining whether the calculated number of predetermined indexes under the operating condition is sufficient, if the number is not satisfied, an unused operating state is assumed under this operating condition (S6 ).

As an example of an unused operating state, as shown in operating state "2" in FIG. An example of this is to change the flow rate of cold/hot water CH in 21 and the second cold/hot water pump 22 to 50%. In addition, as shown in the operating state "2" in FIG. One example of this is to set the flow rate of the cooling water CD in the pump 42 to 50%. Alternatively, as another example of an unused operating state, as shown in operating state "3" in FIG. 4, the third heat source device 13 and the third cooling tower 33 are stopped for operating state "1". At the same time, the third cold/hot water pump 23 and the third cooling water pump 43 may be stopped (flow rate is 0). In addition, it is preferable that the flow rates of the cold/hot water CH and the cooling water CD flowing through the first heat source device 11 and the second heat source device 12, which are variable speed centrifugal refrigerators, are the same for both heat source devices 11 and 12, respectively. The reason for this is to reduce the calculation load, since considering the symmetry, it is extremely unlikely that energy can be saved by operating both heat source devices 11 and 12 under different conditions.

In this way, when changing the operating states under the relevant operating conditions, it is preferable to apply each operating state in combination (brute force) in order to avoid overlooking a certain operating state. At this time, in this embodiment, each of the heat source devices 11, 12, 13 and each cooling tower 31, 32, 33 has a choice of operating or stopping, so the combinations thereof are limited. On the other hand, the flow rate of the cold/hot water CH of each of the cold/hot water pumps 21, 22, 23 and the flow rate of the cooling water CD of each of the cooling water pumps 41, 42, 43 can be changed steplessly, so there may be a large number of combinations. . In this case, for example, for each flow rate, the number of combinations can be suppressed by setting the range of change other than when stopped (0%) between 50% and 100% in 10% increments or 5% increments. However, before performing the simulation, obviously inappropriate driving conditions should be excluded before or during the simulation. For example, if the total maximum output of the heat source devices in operation does not satisfy the heat demand for the operating conditions, the simulation may not be performed. In addition, the demand heat amount, the supply temperature of cold and hot water CH (the temperature of cold and hot water CH leaving the heat source equipment), and the return temperature of cold and hot water CH derived from the flow rate of cold and hot water CH (the exit temperature of the heat demand equipment 99) are based on the specifications. This can also be excluded if the temperature exceeds . In addition, the case where the temperature of the cooling water CD is equal to or higher than a certain level and variable flow rate control of the flow rate of the cooling water CD is unnecessary can also be excluded.

Once the unused operating state is assumed in the manner described above, the process returns to the step of calculating a predetermined index (S4) and the procedure described above is followed. Then, a plurality of predetermined indexes are calculated, and in the step (S5) of determining whether or not the number of predetermined indexes calculated under the relevant operating conditions is sufficient, if the number is sufficient, the predetermined number under the relevant operating conditions is determined. An operating state in which the index has a value that meets the conditions is identified (S7). Here, the value for which the predetermined index meets the conditions is a value that is worth selecting rationally and is typically an optimal value. For example, when the predetermined index is the operating cost, the smaller the index is, the more preferable it is, when the predetermined index is the amount of carbon dioxide emissions, the lower the index is, and the higher the energy saving index is. There are several possible ways to determine the optimal value of the predetermined index. For example, if the predetermined index is operating cost, you may simply select the minimum one, and the heat source Other aspects such as a small number of operating equipment and cooling towers may also be considered. Alternatively, a weighted average may be applied to a plurality of values within an appropriate range of a predetermined index, and the driving state at the value of the predetermined index may be adopted. Alternatively, the average of the driving states corresponding to a plurality of values (for example, the top three) within an appropriate range of the predetermined index may be adopted as the driving state.

Once a set of operating conditions and operating states that meet the conditions of a predetermined index is identified, it is determined whether the number of sets satisfies the required number (S8). Here, the required number of sets corresponds to the number of teacher data required to generate the control model 80, and is generally about several hundred to several thousand sets, depending on the configuration of the model. In the step of determining whether the number of sets satisfies the required number (S8), if the number is not sufficient, an unused operating condition is assumed (S9). An unused operating condition is an operating condition that has not been used to identify a combination of an operating condition and an operating state for which a predetermined index satisfies the condition. In this embodiment, several hundred to several thousand operating conditions are ultimately assumed in order to ultimately identify several hundred to several thousand pairs of operating conditions and operating conditions. becomes.

When assuming a relatively large number of operating conditions in this way, it is sufficient to be able to assume all operating conditions, but as the number of operating conditions increases, the number of operating conditions increases by multiplication, so the number becomes enormous. In addition, since the operating conditions are analog values, the number of operating conditions to be generated is determined by the separation step, so if sufficient machine learning cannot be performed with the training data once generated, the operating conditions to be machine learned again. It is troublesome to create. Therefore, in this embodiment, the computer that performs the simulation randomly determines the operating conditions (creates them using random numbers). That is, each operating condition is set using random numbers within a variable range. This is repeated every time the step (S9) of assuming unused operating conditions is performed. In this way, each item (parameter) is evenly distributed even though the number of operating conditions is smaller than when specified by combination (round robin). Therefore, when setting operating conditions using random numbers, training data that can create a highly accurate control model 80 even with less data (data creation time) than when specifying by combination (brute force) is used. Obtainable. Furthermore, when new operating conditions are assumed in order to add training data after the fact, data may be created again using random numbers as many times as necessary.

After assuming an unused operating condition in the manner described above, the process returns to the step of assuming an operating state (S3), and the above-described procedure is followed thereafter. Then, in the step (S8) of determining whether or not a plurality of combinations of operating conditions and operating states for which a predetermined index satisfies the conditions is identified, and the number of the combinations satisfies the required number (S8), if the number of combinations is satisfied; , finish creating the training data.

Teacher data can be created in this way, but if there is data used in another system and that data can be used, it is possible to save labor in creating teacher data by doing the following, for example. I can do it. The outline is that the difference between the training data that has been created for the existing heat source equipment system and the training data that is to be created for the new heat source equipment system 1 is that there are some differences in the range of operating conditions. In some cases, the remaining common parts may be reused. A specific example will be described below.

FIG. 5 is a flowchart illustrating a procedure for creating teacher data by reusing some existing data. In this specific example, the newly created training data (hereinafter referred to as "new data") is based on operating conditions and conditions in which the maximum number of operating heat source units is 3, the heat demand is 200 to 6000 kW, and predetermined indicators meet the conditions. It is assumed that there are 9000 pairs with the state (hereinafter referred to as "set data"). Assume that the existing teacher data (hereinafter referred to as "existing data") has a maximum operating number of heat source devices of 5, a heat demand of 200 to 10,000 kW, and 9,000 data sets. First, from the existing data, those whose heat demand matches the content of the new data are extracted (S11). As a result, from the existing data, data in which the maximum number of operating heat source devices is 5 and the heat demand is 200 to 6000 kW is extracted, and in this example, the number of data is 6000. Next, the existing data extracted based on the heat demand is sorted based on the number of operating heat source devices (S12). As a result, the required amount of heat is 200 to 6000 kW, but these can be divided into those with three or less heat source machines in operation and those with more than three heat source machines in operation. In this example, it is assumed that there are 5000 cases in which the number of operating heat source devices is 3 or less, and 1000 cases in which the number of operating heat source devices is 3 or less. Among these, those in which the number of operating heat source devices is more than three do not match the new data in which the number of heat source devices is three, and therefore cannot be used as new data as is. However, in this example, rather than excluding this condition unconditionally, the operating conditions are inherited, the operating conditions are changed to values that match the new data, and a simulation is performed to create set data (S13). As a result, 1,000 pieces of data are created using existing data as a reference, with a maximum number of operating heat source units of 3 and a heat demand of 200 to 6,000 kW. Next, the 5000 pieces of data extracted from the existing data and the 1000 pieces of data created with reference to the existing data are totaled (S14). As a result, it is possible to obtain data for 6,000 cases where the maximum number of operating heat source devices is 3 and the required heat amount is 200 to 6,000 kW. At this point, there is a shortage of 3,000 new data items. Therefore, 3000 new sets of data are created in which the maximum number of operating heat source devices is 3 and the heat demand is 200 to 6000 kW (S15). When creating these 3000 new pieces of data, it is preferable to randomly set the operating conditions as described above. Finally, the newly created 3000 pieces of data are added to the 6000 pieces of data obtained based on the existing data (S16). As a result, 9000 sets of data are obtained, with the maximum number of operating heat source devices being 3 and the heat demand being 200 to 6000 kW. This may be used as teacher data for applying to the control model 80 of the new heat source device system 1. In this way, the burden of creating new data can be reduced.

It goes without saying that the training data explained above can be created automatically using a general-purpose computer by creating a program.

<Generation of control model (trained model)>
Once the teacher data is created, the control model 80 is created by subjecting the computer to machine learning processing using the teacher data. In this embodiment, for each set of operating conditions and operating states that meet the conditions of a predetermined index, which have been created in large numbers (about several hundred to several thousand sets) as training data, the operating conditions are input and the operating states are determined. As an output, machine learning processing is performed on the computer to create a control model 80. The control model 80 can use any appropriate one among a wide variety of proposed methods in terms of type and machine learning method, but in this embodiment, it will be described as one using a neural network. Neural networks generally have perceptrons (operators) in an input layer, a middle layer, and an output layer, and the output of the perceptron in the previous stage is multiplied by a weighting coefficient by the perceptron in the subsequent stage, and the result is a new output through an activation function. The calculation is performed by doing this. The control model 80 is a model that derives a large number of numerical values from a large number of numerical inputs. Here, since the output layer of the neural network is generally expressed as a probability value, it will be output as a real number (decimal number). Then, among the operating states that are output, the number of operating vehicles that should originally be an integer is also output as a real number (decimal) instead of an integer. Therefore, in order to determine the actual operating state, it is necessary to perform integer processing such as rounding or rounding up, but this will result in the number of operating units being different from the output of the control model 80, so this state is optimal. It becomes impossible to say. Therefore, in the present embodiment, the control model 80 is a first control model that outputs the number of operating vehicles as an integer value among the operating states to be output, and an integer value of the number of operating vehicles output by the first control model as an operating condition. The second control model is input as one of the control models.

FIG. 6 is a schematic configuration diagram of the first control model (hereinafter referred to as "first control model 81"). FIG. 7 is a schematic configuration diagram of the second control model (hereinafter referred to as "second control model 82"). As shown in FIG. 6, the first control model 81 includes an input layer, an intermediate layer, and an output layer. Operating conditions are input to the input layer, and operating conditions are output from the output layer. It is configured as follows. The first control model 81 includes, as the operating conditions input to the input layer, each item illustrated in FIG. Contains the water supply carbon dioxide emission factor. Note that the electric power carbon dioxide emission coefficient is the amount of carbon dioxide emitted per unit power consumption, and the water supply carbon dioxide emission coefficient is the amount of carbon dioxide emitted per unit amount of water consumption. By including the unit price of consumed resources and the amount of carbon dioxide emissions per unit consumption in the operating conditions input to the input layer, there are the following advantages. The premise is that these variables generally fluctuate over a period of several months to several years, and are different from heat demand and outside air temperature, which fluctuate over several minutes to several hours at the most. Therefore, in conventional control using simulation, it is reasonable to treat these as fixed conditions and change them when they change. However, if these are treated as fixed conditions in this embodiment, the control model 80 will have to be recreated (the machine learning process will be performed again) if they change. Therefore, in this embodiment, these operating conditions are used so that the control model 80 does not need to be recreated even if the operating conditions change. This will also be of great benefit to users whose energy costs change on a daily basis due to the introduction of a variable pricing system for electricity, which has been under consideration in recent years, or due to diversification of electricity procurement.

In addition, as shown in FIG. 6, the first control model 81 includes the number of operating fixed-speed centrifugal chillers and the number of variable-speed centrifugal chillers in operation, as well as the number of operating units of these various chillers, as the operating states output to the output layer. It includes the flow rate of each cold/hot water CH and the flow rate of cooling water CD. As described above, the third heat source device 13 in FIG. 1 corresponds to the fixed speed centrifugal chiller, and the first heat source device 11 and the second heat source device 12 correspond to the variable speed centrifugal chiller. Therefore, the number of fixed-speed centrifugal chillers in operation is 0 or 1, and the number of variable-speed centrifugal chillers in operation is 0, 1, or 2, all of which are integer values. Further, since the number of operating cooling towers 31, 32, and 33 matches the number of operating heat source devices 11, 12, and 13, the cooling towers are not shown in FIG. In other words, the number of operating heat source devices 11, 12, 13 and the number of operating cooling towers 31, 32, 33 are the same integer value. Further, the fixed-speed centrifugal chiller operating cold/hot water flow rate of the output layer shown in FIG. value (continuous value). Similarly, the variable speed centrifugal chiller operation cold/hot water flow rate is the total flow rate of the first cold/hot water pump 21 and the second cold/hot water pump 22. Further, the fixed speed centrifugal chiller operation cooling water flow rate is the flow rate of the third cooling water pump 43, and the variable speed centrifugal chiller operation cooling water flow rate is the total flow rate of the first cooling water pump 41 and the second cooling water pump 42. be. In this embodiment, the range of the flow rate of each of these pumps is 0% or any value (continuous value) from 50 to 100%.

Here, in light of the above-mentioned purpose, the output layer of the first control model 81 only needs to be able to output only the number of operating units, which does not fit in with the output being a decimal value, so flow rates, etc. other than the number of operating units are output. It may not be included in the output to the layer. However, in reality, there is a close relationship between the number of operating heat source units and the flow rate of cold and hot water (in other words, the load for each heat source unit), so it is better to add the flow rate of cold and hot water, etc. to the output, so that the middle layer can handle this. In some cases, a perceptron that is closely related to the above occurs, and learning progresses efficiently. For these, the same training teacher data may be given to a plurality of control models to learn them, and the optimal model may be determined from the learning results. Note that if the flow rate or the like is added to the output for learning, the learning time of the second control model 82 may be shortened as described later. In the case of a neural network, model learning means adjusting the weighting coefficients of the perceptron in the intermediate layer so that the input data and output data of the teacher data match. The procedure is to first separate the training data into "learning" and "verification", use the training data to adjust the weighting coefficients using a method called backpropagation, and then use the verification data to adjust the weighting coefficients. Generally, the accuracy is checked, and if it is insufficient, the weighting coefficients are adjusted again. If the required accuracy is not obtained even after repeated learning, as mentioned above, it is possible to generate new training data by randomly assuming driving conditions and perform additional learning using the new training data. can.

As shown in FIG. 7, the second control model 82 also includes an input layer, an intermediate layer, and an output layer, in which operating conditions are input to the input layer, and operating conditions are output from the output layer. It is configured as follows. The second control model 82 includes, as operating conditions input to the input layer, each item input to the input layer of the first control model 81 (see FIG. 6), as well as information output to the output layer of the first control model 81. This includes the number of operating heat source units (and associated cooling towers). The value of the number of operating heat source devices, etc. input to the input layer of the second control model 82 is It becomes an integer value with the decimal number rounded up. On the other hand, the operating state output to the output layer of the second control model 82 does not include the number of fixed-speed centrifugal chillers in operation and the number of variable-speed centrifugal chillers in operation; These are the flow rate of cold/hot water CH and the flow rate of cooling water CD of each centrifugal chiller. These values output to the output layer of the second control model 82 are also output to the output layer of the first control model 81 in this embodiment; may differ from the corresponding value. This is because, as described above, the second control model 82 limits the value of the number of operating heat source devices, etc., input to the input layer, to an integer value. For this reason, the second control model 82 is a cold/hot water CH that more closely matches the conditions of the predetermined index in a realistic state than when the number of operating units of each heat source device, etc. is output as a real number (including decimal numbers). This means that the operating status such as the flow rate of the cooling water CD and the flow rate of the cooling water CD can be output.

As mentioned above, when the output of the first control model 81 includes the content of the output of the second control model 82, the middle layer of both models 81 and 82 has the same configuration, and the content of the output of the first control model 81 is The coefficients of the intermediate layer may be used as initial values of the coefficients of the intermediate layer of the second control model 82. This is a type of so-called "transfer learning", and the only difference between the first control model 81 and the second control model 82 is whether or not there is the number of operating vehicles in the input, so the middle layer has similar coefficients in principle. will have. Therefore, as described above, by making the intermediate layers of both models 81 and 82 the same configuration and using the coefficients of the intermediate layer of the first control model 81 as the initial values of the coefficients of the intermediate layer of the second control model 82, In some cases, the learning time when generating the second control model 82 can be significantly shortened. Note that the first control model 81 and the second control model 82 are conceptually distinct, and may be physically configured separately or integrally.

It is preferable that the control model 80 (including the case where it is divided into a first control model 81 and a second control model 82) is created for each predetermined index that is desired to be optimized. For example, a model that minimizes operating costs, a model that minimizes carbon dioxide emissions, etc. may be created. In this way, the control model 80 can be switched and used depending on the situation, and the heat source system 1 can be operated more appropriately.

The control model 80 configured (generated) as described above (including the case where it is divided into the first control model 81 and the second control model 82) is installed in the control device 70 of the heat source device system 1. . The control device 70 equipped with the control model 80 is configured to control the operations of the heat source devices and heat source auxiliary devices that constitute the heat source device system 1 so as to achieve the operating state outputted by the control model 80. Below, the operation of the heat source device system 1 including the control of the control device 70 will be explained.

<Operation of the heat source system>
FIG. 8 is a block diagram showing the calculation procedure of the control device 70 of the heat source device system 1. Hereinafter, the operation of the heat source device system 1 will be described with reference mainly to FIGS. 1 and 8, and with appropriate reference to FIGS. 2 to 7. While the heat source device system 1 is in operation, the necessary heat source devices 11, 12, and 13 are in operation, and in conjunction with this, each of the cold and hot water pumps 21, 22, and 23, and each of the cooling towers 31, 32, and 33 are operated. , the necessary cooling water pumps 41, 42, and 43 are operated. For the time being, in order to avoid complicating the situation, the description will be made assuming that each of the heat source machines 11, 12, 13 and their associated heat source auxiliary machines are all in operation.

By the operation of each cold/hot water pump 21 , 22 , 23 , cold/hot water CH flows from the return header 29 to each heat source device 11 , 12 , 13 via each cold/hot water return pipe 25 , 26 , 27 . The cold and hot water CH flowing into each of the heat source devices 11, 12, and 13 is cooled (during cooling) or heated (during heating), and typically the temperature is adjusted to increase the difference from the ambient environment temperature. The cold and hot water CH whose temperature has been adjusted in each of the heat source devices 11, 12, and 13 is conveyed to the outgoing header 19 via each of the cold and hot water outgoing pipes 15, 16, and 17. The cold and hot water CH that has flowed into the outgoing header 19 flows through the supply pipe 91 and is supplied to the heat demand equipment 99 by a secondary pump (not shown), and then flows through the recovery pipe 92 and flows into the return header 29. Flow. The cold/hot water CH supplied to the heat demand equipment 99 is used for heat load processing, so that the temperature changes in a direction in which the difference from the ambient environment temperature becomes smaller.

On the other hand, the operation of each cooling water pump 41, 42, 43 causes cooling water CD to flow from each cooling tower 31, 32, 33 to each heat source device 11, 12, 13 via each cooling water outgoing pipe 35, 36, 37. Inflow. The cooling water CD flowing into each heat source device 11, 12, 13 exchanges heat with the refrigerant in each heat source device 11, 12, 13, and then passes through each cooling water return pipe 45, 46, 47 to each cooling tower 31. , 32, 33. Regarding the operation of each device constituting the heat source device system 1, as described above, the control device 70 controls starting and stopping of each of the heat source devices 11, 12, and 13. Based on the operation of each heat source device 11, 12, 13, the auxiliary equipment power panel 75 starts and stops each cooling tower 31, 32, 33, each cold/hot water pump 21, 22, 23, and each cooling water pump 41. , 42 and 43 are controlled. When the heat source device system 1 is operating in this manner, the control device 70 performs control in the following manner so that the operating state is appropriate (typically, optimal).

The control device 70 receives temperature information from the outside air thermometer 61. Further, the control device 70 receives temperature and pressure information from each heat source device 11, 12, and 13, and measures the amount of heat demanded of the heat demand equipment 99, the flow rate of cold/hot water CH, and the pressure loss of the heat demand equipment 99. . Then, the control device 70 calculates the pressure loss coefficient of the heat demand equipment 99 from the pressure loss of the heat demand equipment 99 and the flow rate of cold/hot water CH. Further, the control device 70 receives information regarding the target temperature of the cold/hot water CH from the heat demand equipment 99. On the other hand, the electric power unit price, water rate, electric power carbon dioxide emission coefficient, and water supply carbon dioxide emission coefficient are input and held in the control device 70 as set values. As described above, these setting values generally do not change in a short period of time, so it is often sufficient to correct them when they change. The control device 70 inputs these specified values into the control model 80 as operating conditions.

When the operating conditions are input, the control model 80 performs processing based on a learned algorithm and outputs an appropriate operating state. In general, learning a mathematical model using a neural network (backpropagation calculation) in particular requires very high computing power, such as a computational element such as a GPU, but inference using a trained model (forward propagation calculation) (Calculation) does not require such high computing power. Therefore, in inference using a trained model, even if a dedicated calculation element or the like is used, sufficient calculations can be performed using a calculation device that is relatively inexpensive and consumes little power. In light of this embodiment, the control model 80 that performs inference based on a learned algorithm has a smaller computational load than a simulation performed when creating teacher data, and is suitable for installation in the field. .

When the control model 80 receives the operating conditions, it first inputs the operating conditions into the first control model 81 (see FIG. 6). The first control model 81 to which the operating conditions are input performs processing based on a learned algorithm, converts the output into an integer, and determines the number of operating heat source devices 11, 12, and 13. In addition, at this time, in order to improve the accuracy of the output, it is preferable to output the flow rate of the cold/hot water CH and the flow rate of the cooling water CD as well, as described above. However, the flow rate of cold/hot water CH and the flow rate of cooling water CD output by the first control model 81 are not intended to be used directly for controlling the heat source device system 1 . After obtaining the output of the first control model 81, the control model 80 applies the operating conditions input to the first control model 81 and the operating conditions output from the first control model 81 to the second control model 82 (see FIG. 7). Input the integer value of the number of operating heat source devices 11, 12, and 13. The second control model 82 into which the integer value of the number of operating units and the operating conditions are input performs processing based on a learned algorithm, and outputs the flow rate of cold/hot water CH and the flow rate of cooling water CD. In this embodiment, since the number of operating vehicles is output as an integer value in the first control model 81, the output of the integer value can be directly applied to actual control. Furthermore, since the integer value of the number of operating vehicles is included in the input of the second control model 82, the accuracy of the output of the second control model 82 can be improved.

The control device 70 controls the operation of each heat source device 11 , 12 , 13 so that the number of operating units output by the first control model 81 is achieved. Further, the control device 70 controls each cold/hot water pump 21, 22, 23, each cooling water pump 41, 42, The discharge flow rate of 43 is controlled. By controlling each device constituting the heat source device system 1 according to the output of the control model 80, the predetermined indexes set by the user, such as power consumption, operating cost, and carbon dioxide emissions, are set to meet the conditions. (e.g. to a minimum). Control of each device based on calculations using such a control model 80 and its output is performed by, for example, changing the number of units in stages or gradually changing the flow rate command value in order to avoid sudden changes in operating conditions. It may be changed. Further, it is preferable that the calculation in the control model 80 and the adjustment of the number of operating units and the flow rate based on the output thereof be performed at predetermined intervals. The predetermined interval can be determined as appropriate depending on the situation, taking into consideration the calculation time (or calculation load) in the control model 80 and the accuracy of control of the heat source device system 1. Examples of predetermined intervals include 3 minutes, 5 minutes, 10 minutes, 15 minutes, 30 minutes, etc. However, if, for example, the amount of heat demanded changes by more than the preset value or if other operating conditions change significantly, the calculation using the control model 80 may be performed at a timing other than the preset predetermined interval. Control calculations for each device may be performed based on the output.

Note that the control device 70 may be equipped with a plurality of control models 80 according to the type of predetermined index, and may be appropriately switched to an appropriate control model 80 according to user settings and instructions. For example, the control model 80 may be set as a predetermined index that meets the conditions, such as one with the minimum power consumption, one with the minimum operating cost, and one with the minimum carbon dioxide emissions, and the control models 80 may be switched depending on the situation. It may also be a thing. Further, the predetermined index when generating one control model 80 is not limited to one type, and for example, the control model 80 with the lowest carbon dioxide emission within the range of 5% from the lowest operating cost. The control device 70 may have the following. In this case, the operating cost corresponds to the first predetermined index (the index with the highest priority), and the amount of carbon dioxide emissions corresponds to the second predetermined index (the index with the second highest priority).

Furthermore, instead of determining all of the operating states for controlling the devices constituting the heat source device system 1 based on the output of the control model 80, the control device 70 determines some of the operating states based on the output of the control model 80. It may also be controlled by In this case, the remaining operating states may be controlled based on the simulation results performed by the control device 70 or based on a rule base predefined in the control device 70. For example, the number of operating heat source devices 11, 12, 13 and the flow rate of cold/hot water CH in each cold/hot water pump 21, 22, 23 are controlled based on the output of the control model 80, and each cooling water pump 41, 42, 43 is The flow rate of the cooling water CD may be controlled based on a rule.

As explained above, according to the heat source device system 1 according to the present embodiment, the component devices are controlled based on the output of the control model 80, so that the appropriate operating state is maintained according to changes in operating conditions. It allows for responsive driving. Furthermore, since it includes a first control model 81 that outputs the number of operating vehicles as an integer value and a second control model 82 that inputs the integer value, it is possible to maintain an appropriate operating state in accordance with actual operation. can. Further, when the operating conditions include the pressure loss coefficient on the heat demand equipment 99 side, appropriate operation can be performed even when the flow rate of cold/hot water CH supplied to the heat demand equipment 99 changes.

<Others>
In the above description, as heat source devices, the first heat source device 11 and the second heat source device 12 are variable speed centrifugal refrigerators, and the third heat source device 13 is a fixed speed centrifugal refrigerator. However, various heat source devices other than the centrifugal refrigerator, such as an absorption refrigerator, a cold/hot water generator, a heat pump, etc., can be used as the heat source device depending on the purpose. Furthermore, in the above explanation, the first heat source device 11 and the second heat source device 12 are both the same type of equipment with the same characteristics, but they may be the same type of equipment with different characteristics, or they may be different types of equipment. There may be. In addition, in the above explanation, three heat source devices 11, 12, and 13 are provided as heat source devices, but the total number of heat source devices is not limited to three, and can be more than three depending on the purpose. It may be less. For example, more than three devices of different types may be provided, or a plurality of types (for example, two or three devices) of the same type may be provided.

In the above description, each of the heat source devices 11, 12, and 13 is configured to be able to detect the temperature difference and pressure difference between the inflowing and outflowing cold and hot water CH. However, instead of each heat source device 11, 12, 13 being equipped with a meter for detecting temperature and pressure, a meter for detecting temperature and pressure may be provided in the nearby piping.

In the above description, the cooling towers 31, 32, and 33 are provided as the heat source fluid supply device, and the heat source fluid is the cooling water CD. However, the heat source fluid supply device may be an air-cooled heat pump chiller, and the heat source fluid may be air. In this case, the air-cooled heat pump chiller will serve as both the heat source device and the heat source fluid supply device. In other words, the heat source device and the heat source fluid supply device are typically physically integrated (housed in one housing). ) will be configured.

In the above description, the control device 70 indirectly controls the heat source auxiliary device via the auxiliary device power panel 75, but the control device 70 may directly control the heat source auxiliary device.

In the above explanation, when creating the training data, it is preferable to combine (brute force) the driving states to be changed based on the assumed driving conditions, but it is preferable to set the driving states to be changed at random. Good too. However, as described above, in order to avoid overlooking a specific operating state, it is preferable to set the operating states to be changed in combination (brute force).

In the above explanation, fixed conditions are prepared before assuming operating conditions, but the items listed as examples of fixed conditions may be treated as operating conditions. In this case, the step (S1) of preparing fixed conditions in the flowchart shown in FIG. 2 is omitted.

In the above explanation, the items included in the operating conditions input to the input layer are the amount of heat demanded, the outside air temperature, the target temperature of cold and hot water CH, the pressure loss coefficient on the heat demand equipment 99 side, the unit price of consumed resources, and the unit carbon dioxide. Although it is assumed to be the amount of emissions, the input items may be increased or decreased as appropriate. For example, when a heat source device (for example, an absorption refrigerator) that obtains a heat source by burning fuel such as oil or gas is provided, the fuel unit price may be included in the input items. On the contrary, the pressure loss coefficient and others on the heat demand equipment 99 side may be excluded from the input items. However, if the pressure loss coefficient on the heat demand equipment 99 side is included in the input items, even if the flow rate of cold/hot water CH on the heat demand equipment 99 side changes, the power of each cold/hot water pump 21, 22, 23 cannot be estimated correctly. It has the advantage of being able to Moreover, instead of the required heat amount inputted as the operating condition, a physical quantity correlated to the demanded heat amount may be inputted as the operating condition. As a physical quantity correlated with the amount of heat demanded, for example, the return temperature of the cold/hot water CH (the temperature of the cold/hot water CH flowing into each heat source device 11, 12, 13), etc. can be mentioned. Further, instead of the outside air temperature input as the operating condition, a physical quantity correlated to the outside air temperature may be input as the operating condition. Examples of physical quantities correlated with the outside air temperature include the inlet temperature of the cooling water CD (the temperature of the cooling water CD flowing into each heat source device 11, 12, and 13), the temperature of the lower water tanks of the cooling towers 31, 32, and 33, etc. Can be mentioned.

In the above description, it is assumed that the integer value output by the first control model 81 is used to control the heat source device system 1 for the number of operating heat source devices 11, 12, and 13. However, if each cold/hot water pump 21, 22, 23 and/or each cooling water pump 41, 42, 43 is configured to control the number of units, and output as an integer value is required as the operating state of these pumps, these The output of the first control model 81 may be used as the operating state. On the other hand, if each of the heat source devices 11, 12, and 13 is configured to perform capacity control with the output adjustable (for example, can be operated at 50% to 100% of the maximum output), the first control The model 81 may be omitted and the output of the second control model 82 may be used for control.

For example, the hardware configuration of each of the arithmetic device (computer) that creates the teacher data, the control device 70, and the control model 80 (first control model 81 and second control model 82) described above is as follows. A computer can be used.

FIG. 9 is a block diagram of an exemplary computer 100. Computer 100 may be used to provide computational functionality associated with the algorithms, methods, functions, processes, and procedures described in this disclosure.

Computer 100 may be a server, desktop computer, embedded computer, laptop/notebook computer, smartphone, tablet computer device, or one or more processors therein (including physical instances, virtual instances, or both). ) may include any computing device. Computer 100 can include input devices, such as a keypad, keyboard, and touch screen, that can accept information entered by a user. Computer 100 may also include output devices that convey information associated with the operation of computer 100. This information may include digital data, visual data, audio information, or a combination of these information. This information can be displayed on a graphical user interface (GUI).

Computer 100 may serve as a client, network component, server, database, persistence, or component of a computer system to perform the processes, procedures, etc. described in this disclosure. Exemplary computer 100 is communicatively coupled to network 120. In some implementations, one or more components of computer 100 are configured to operate within an environment including a cloud computing-based environment, a local environment, a global environment, and a combination of environments. be able to. Computer 100 can receive requests via network 120, for example, from a client application running on another computer. Computer 100 may be configured to respond to incoming requests by processing the incoming requests using software applications.

The computer 100 typically includes a processor 102, a first memory 104, a second memory 106, and an interface 108 as components.

The processor 102 processes various types of information in the computer 100. Processor 102 may execute instructions (programs) and manipulate data to perform operations of computer 100, including operations using any of the algorithms, methods, functions, processes, and procedures described in this disclosure. can. Processor 102 may be a single processor or two or more processors. Processor 102 may include a central processing unit (CPU), graphics processing unit (GPU), microprocessor, controller card, circuit board, or other electrical circuitry.

The first memory 104 temporarily or permanently stores programs and/or data used for information processing in the computer 100. First memory 104 may store any data consistent with this disclosure. The first memory 104 may be a single memory, or may be two or more memories. The first memory 104 may include volatile memory such as RAM and cache, and nonvolatile memory such as ROM.

The second memory 106 typically holds data used by the computer 100, but may also hold data used outside the computer 100. The second memory 106 may hold programs, including an operating system, that are executable on the computer 100 or other equipment. The second memory 106 may be a single memory or two or more memories. The second memory 106 may include a hard disk drive (HDD), a solid state drive (SSD), a flash memory, and the like.

Interface 108 is used by computer 100 to communicate with other systems connected to network 120 (whether shown or not) in a distributed environment. Interface 108 may include logic encoded in software or hardware operable to communicate with network 120. Interface 108 may include software that supports one or more communication protocols associated with communication. In this manner, network 120 or interface hardware may be operable to send and receive signals in and out of computer 100. When communicating with other devices, the interface 108 uses, for example, a wired communication standard such as Ethernet (registered trademark), and/or a wireless communication standard such as 4G, 5G, or Wi-Fi (registered trademark). be able to. Interface 108 may be a single interface or may have two or more interfaces.

Each component of the computer 100 can communicate using a system bus. In some embodiments, any or all components of computer 100 (including hardware or software components) communicate (interface) with each other or with interface 108 (or a combination of the two) over a system bus. be able to.

Additionally, the computer 100 has a power supply 110. Power source 110 typically includes a power plug that receives power from a utility or other power source. Power source 110 may include a replaceable or non-replaceable battery, and the battery may be rechargeable by receiving power from a utility or other power source.

In the computer 100, the processor 102 reads programs stored in the first memory 104 and/or the second memory 106 and executes calculations to execute processes and procedures according to the purpose. When the computer 100 is used as an arithmetic device that creates teacher data, the processor 102 reads out programs and/or data stored in the first memory 104 and/or the second memory 106 and creates the control model 80 (the 1 control model 81 and second control model 82)). When the computer 100 is used as the control device 70, the processor 102 reads out programs and/or data stored in the first memory 104 and/or the second memory 106, and controls the heat source device system 1 by the control device 70. Operation instructions (control) are given. When the computer 100 is used for the control models 80 (the first control model 81 and the second control model 82), the processor 102 executes programs and/or data stored in the first memory 104 and/or the second memory 106. It reads out and outputs the appropriate operating state of the heat source device system 1.

In the above description of the computer 100, the programs and/or data used for information processing in the computer 100, which are stored in the first memory 104 and/or the second memory 106, are stored in a non-transitory computer-readable medium. It may be stored. The non-transitory computer-readable medium stores computer-readable instructions and/or data utilized to perform a computer-implemented method. Computer-readable media include magneto-optical disks and optical memory devices, as well as digital video discs (DVD), CD-ROM, DVD+/-R, DVD-RAM, DVD-ROM, HD-DVD, and BLURAY®. etc. can be included. Computer readable media can also include magnetic devices such as tapes, cartridges, cassettes, and removable disks. Each computer program is a set of computer program instructions encoded on a tangible, non-transitory computer-readable medium for execution by, or to control the operation of, an information processing device, including computer 100. It can include one or more modules. Furthermore, the program and/or data may be downloaded from an external device via a network.

All references, including publications, patent applications, and patents, cited herein are hereby incorporated by reference in their entirety. To the same extent, they are incorporated herein by reference.

The use of nouns and similar referents used in connection with the description of the invention (particularly in connection with the following claims) shall be used herein unless otherwise indicated or clearly contradicted by the context. , shall be construed as extending both in the singular and in the plural. The words "comprising," "having," "including," and "including" are to be interpreted as open-ended terms (ie, meaning "including, but not limited to"), unless otherwise specified. The recitation of numerical ranges herein is intended solely to serve as shorthand for individually referring to each value falling within the range, unless otherwise indicated herein. and each value is incorporated herein as if individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or clearly contradicted by context. Any examples or exemplary language (e.g., "etc.") used herein, unless specifically stated otherwise, are intended merely to better explain the invention and are intended to place no limitations on the scope of the invention. isn't it. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the invention.

Preferred embodiments of this invention are described herein, including the best mode known to the inventors for carrying out the invention. Modifications of these preferred embodiments will be apparent to those skilled in the art upon reading the above description. The inventors anticipate that those skilled in the art will apply such modifications as appropriate, and the inventors contemplate that the invention may be practiced otherwise than as specifically described herein. . Accordingly, this invention includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Furthermore, any combination of the above-described elements in all variations is encompassed by the present invention, unless otherwise indicated herein or clearly contradicted by context.

Claims

A heat source device that cools or heats a heat medium supplied to heat demand equipment;
a heat source auxiliary machine that operates in conjunction with the operation of the heat source equipment;
A control device that adjusts the operating state of the heat source equipment and the heat source auxiliary equipment, the control device having a learned control model,
The heat source auxiliary equipment includes a heat medium pump that flows the heat medium passing through the heat source device, and a heat source that supplies the heat source fluid to the heat source device, which directly or indirectly exchanges heat with the heat medium in the heat source device. a fluid supply device;
The operating state includes at least one of the operating status of the heat source device, the flow rate of the heat medium discharged by the heat medium pump, and the flow rate of the heat source fluid supplied by the heat source fluid supply device,
The control model is subjected to machine learning processing using training data so that when operating conditions that affect the operating condition are input, the operating condition is outputted in which a predetermined index has a value that satisfies the condition. has been
The operating conditions include at least one of the heat demand of the heat demand equipment or a physical quantity correlated thereto, and the outside temperature or a physical quantity correlated thereto,
The predetermined index is at least one of the power consumption of the heat source equipment and the heat source auxiliary equipment, the operating cost of the heat source equipment and the heat source auxiliary equipment, and the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment. including one
The control device controls the heat source equipment and the heat source auxiliary equipment so that they are in the operating state output by the control model.
Heat source machine system.
The predetermined index may include a plurality of power consumption of the heat source equipment and the heat source auxiliary equipment, operating costs of the heat source equipment and the heat source auxiliary equipment, and carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment. including,
The control model selects a plurality of operating states in which a first predetermined index among the plurality of predetermined indicators satisfies a condition, and selects the first predetermined index from among the plurality of selected operating states. configured to output the operating state in which a second predetermined index different from the predetermined index has a value that satisfies the condition;
The heat source device system according to claim 1.
At least one of the heat source device, the heat medium pump, and the heat source fluid supply device is composed of a plurality of units,
The operating state includes the number of operating units of a plurality of the heat source equipment, the heat medium pump, and the heat source fluid supply device,
The control model includes a first control model that outputs the number of operating vehicles as an integer value among the operating states, and an integer value of the number of operating vehicles output by the first control model as one of the operating conditions. a second control model input as;
The heat source device system according to claim 1 or claim 2.
The control device is configured to control a part of the operating state of the amount of heat processed by the heat source device, the flow rate of the heat medium discharged by the heat medium pump, and the flow rate of the heat source fluid supplied by the heat source fluid supply device. using the output of the control model to determine the remaining operating state by simulation or rule-based;
The heat source device system according to any one of claims 1 to 3.
The operating conditions include a pressure loss coefficient of the heat demand equipment,
The heat source device system according to any one of claims 1 to 4.
The operating conditions include at least one of the cost per unit power consumption of the heat source equipment and the heat source auxiliary equipment, and the amount of carbon dioxide emissions per unit power consumption of the heat source equipment and the heat source auxiliary equipment.
The heat source device system according to any one of claims 1 to 5.
The control model sets the predetermined index to the power consumption of the heat source equipment and the heat source auxiliary equipment, the operating cost of the heat source equipment and the heat source auxiliary equipment, and the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment, having a plurality of control models with one or more of the following;
The control device utilizes an appropriate control model among the plurality of control models according to the predetermined index desired by the user.
The heat source device system according to any one of claims 1 to 6.
A method for generating a trained model used for controlling a heat source equipment system comprising a heat source device that cools or heats a heat medium supplied to heat demand equipment, and a heat source auxiliary machine that operates in conjunction with the operation of the heat source equipment. And,
Under assumed operating conditions, a plurality of predetermined indices for the assumed operating state of the heat source device system are determined by simulation while changing the operating state, and the predetermined index satisfies the conditions. By defining the operating state at the time as a relationship with the operating condition and performing this for multiple operating conditions, a set of relationships between the operating state and the operating condition when the predetermined index meets the condition can be determined. A step of generating training data by obtaining multiple pieces of data,
A step of generating a trained model in which the operating conditions are input and the operating state is output, by performing machine learning processing using the teacher data,
The heat source auxiliary equipment includes a heat medium pump that flows the heat medium passing through the heat source device, and a heat source that supplies the heat source fluid to the heat source device, which directly or indirectly exchanges heat with the heat medium in the heat source device. a fluid supply device;
The operating conditions include at least one of the heat demand of the heat demand equipment or a physical quantity correlated thereto, and the outside temperature or a physical quantity correlated thereto,
The operating state includes at least one of the operating status of the heat source device, the flow rate of the heat medium discharged by the heat medium pump, and the flow rate of the heat source fluid supplied by the heat source fluid supply device,
The predetermined index is at least one of the power consumption of the heat source equipment and the heat source auxiliary equipment, the operating cost of the heat source equipment and the heat source auxiliary equipment, and the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment. including one
How to generate a trained model.
At least one of the heat source device, the heat medium pump, and the heat source fluid supply device is composed of a plurality of units,
The operating state includes the number of operating units of a plurality of the heat source equipment, the heat medium pump, and the heat source fluid supply device,
The step of generating the trained model includes the step of generating a first trained model that includes an integer value of the number of operating vehicles in the operating state of the output, and using the integer value of the number of operating vehicles as one of the operating conditions. generating a second trained model to be input;
The method for generating a trained model according to claim 8.
The first trained model and the second trained model use a neural network, and the first trained model and the second trained model have a common intermediate layer,
The driving state item that is the output of the first trained model includes the driving state item that is the output of the second trained model,
The step of generating the trained model includes first performing machine learning processing on the first learning model to generate the first trained model, and then calculating the weight coefficients of the middle layer of the second learning model. generating the second trained model by performing machine learning processing on the weighting coefficient of the intermediate layer of the first trained model with an initial value set;
The method for generating a trained model according to claim 9.
The step of generating the teacher data includes randomly determining at least one of the operating conditions and the operating state when performing the simulation.
The method for generating a trained model according to any one of claims 8 to 10.
In the step of generating the teacher data, data of the predetermined index under the operating conditions in a wider range than the assumed operating conditions and in the operating conditions in a wider range than the assumed operating conditions is generated. If it already exists, extracting the predetermined index for the operating state assumed under the operating condition assumed from the already existing data and using it as the teacher data;
The method for generating a trained model according to any one of claims 8 to 11.
The step of generating the teacher data includes changing the values of items that are not compatible with the assumed driving conditions and driving states of the driving conditions and driving states in the data that have not been extracted to values that are compatible with the driving conditions and the driving states. generate the training data by performing a simulation with
The method for generating a trained model according to claim 12.
The operating conditions include a pressure loss coefficient of the heat demand equipment,
The step of generating the teacher data includes calculating the pressure loss in the heat demand equipment based on the pressure loss coefficient and the flow rate of the heat medium in the simulation, and then determining the predetermined index.
The method for generating a trained model according to any one of claims 8 to 13.
A trained computer installed in a computer used to control a heat source system that includes a heat source device that cools or heats a heat medium supplied to heat demand equipment, and a heat source auxiliary device that operates in conjunction with the operation of the heat source device. A model,
an input layer into which operating conditions of the heat source device system are input;
an output layer in which the operating state of the heat source device system is output;
an intermediate layer in which parameters are learned using teacher data that inputs the driving conditions and outputs the driving conditions in which a predetermined index has a value that satisfies the conditions;
The heat source auxiliary equipment includes a heat medium pump that flows the heat medium passing through the heat source device, and a heat source that supplies the heat source fluid to the heat source device, which directly or indirectly exchanges heat with the heat medium in the heat source device. a fluid supply device;
The operating conditions include at least one of the heat demand of the heat demand equipment or a physical quantity correlated thereto, and the outside temperature or a physical quantity correlated thereto,
The operating state includes at least one of the operating status of the heat source device, the flow rate of the heat medium discharged by the heat medium pump, and the flow rate of the heat source fluid supplied by the heat source fluid supply device,
The predetermined index is at least one of the power consumption of the heat source equipment and the heat source auxiliary equipment, the operating cost of the heat source equipment and the heat source auxiliary equipment, and the carbon dioxide emissions of the heat source equipment and the heat source auxiliary equipment. including one
inputting the operating conditions into the input layer, calculating the operating conditions in the intermediate layer, and causing the computer to function so as to output the operating conditions from the output layer;
Trained model.
A control device having the trained model according to claim 15;
The heat source device;
and the heat source auxiliary machine,
The control device controls the heat source equipment and the heat source auxiliary equipment so as to be in the operating state outputted by the learned model.
Heat source machine system.