Effect prediction device, effect prediction method, and effect prediction program

Abstract

An effect prediction device (201) predicts an effect of DR control. An air conditioner data acquisition unit (402) acquires air conditioner data including an actual performance set temperature and an actual performance power consumption. An air temperature data acquisition unit (403) acquires air temperature data including an actual performance air temperature. A data region division unit (404) divides the air conditioner data and the air temperature data for each region. A model creation unit (405) creates a power consumption prediction model that predicts the power consumption for each region. A control effect prediction unit (407) predicts the effect of the DR control using the power consumption prediction model for each region.

Description

Technical Field

[0001] This disclosure relates to an effect prediction device, an effect prediction method, and an effect prediction procedure. In particular, it relates to an effect prediction device, an effect prediction method, and an effect prediction procedure for performing demand response control on air conditioning equipment and predicting the effects of demand response control. Background Technology

[0002] Demand response control is known for maintaining a balance between electricity supply and demand. Hereinafter, demand response control will be referred to as DR control. DR control is used to maintain a balance between electricity supply and demand by suppressing or increasing electricity consumption by demanders. Entities such as power companies or aggregators undertake DR control. In DR control, power companies or aggregators implement equipment controls targeting demanders to suppress or increase electricity consumption. Furthermore, power companies or aggregators take measures that enable equipment controls to suppress or increase electricity consumption by demanders. Here, in order to determine the supply and demand control content that can achieve the target electricity reduction, power companies or aggregators need to predict the electricity consumption reduction during supply and demand control, i.e., the effect of DR control. In cases with a large number of demanders subject to DR control, the effect of DR control needs to be estimated for each demander. Therefore, low computational complexity is desired. A high-accuracy prediction of the effect of DR control with low computational complexity is required.

[0003] Patent document 1 discloses a method for predicting power consumption in an air conditioner. Here, power consumption is predicted using regression analysis with indoor temperature or set temperature and outdoor air temperature as explanatory variables.

[0004] Existing technical documents

[0005] Patent documents

[0006] Patent Document 1: Japanese Patent Application Publication No. 2013-120011 Summary of the Invention

[0007] The problem that the invention aims to solve

[0008] It is possible to utilize existing technology to predict the effect of DR control on air conditioners based on the difference between predicted power consumption when DR control is implemented and when DR control is not implemented. However, the state of the air conditioner after DR control differs from its state before DR control. Therefore, there is a problem that if a regression equation is used to represent power consumption, the prediction error will increase.

[0009] The purpose of this disclosure is to improve the predictive accuracy of the effect of DR control during DR control implementation.

[0010] Methods for solving problems

[0011] The effect prediction device disclosed herein predicts the effect of demand response control in an air conditioning system owned by an electricity consumer. The device comprises: an air conditioning data acquisition unit that acquires air conditioning data including the consumer's actual set temperature and actual power consumption; a temperature data acquisition unit that acquires temperature data including the consumer's actual temperature; a data region division unit that divides the air conditioning data and temperature data into regions; a model creation unit that uses the air conditioning data and temperature data divided into regions to create a power consumption prediction model for each region; and a control effect prediction unit that uses the power consumption prediction model for each region to predict the effect of the demand response control.

[0012] Invention Effects

[0013] In the effect prediction device of this disclosure, a different regression equation is used to predict power consumption for each state, and the predicted effect value of DR control is calculated based on the difference between the predicted power consumption when DR control is implemented and when it is not implemented. Therefore, the effect prediction device of this disclosure achieves the effect of reducing the prediction error of power consumption in each state and improving the prediction accuracy of the effect of DR control. Attached Figure Description

[0014] Figure 1 This is a diagram illustrating an example of the overall structure of the air conditioning management system 800 according to Embodiment 1.

[0015] Figure 2 This is a diagram showing the region division of the effect prediction system 200 in Embodiment 1 before and after DR control.

[0016] Figure 3 This is a diagram illustrating an example of the hardware structure of the effect prediction device 201 in Embodiment 1.

[0017] Figure 4 This is a diagram illustrating an example of the functional structure of the effect prediction system 200 of Embodiment 1.

[0018] Figure 5 This is a flowchart illustrating the process of creating a power consumption prediction model in Implementation Method 1.

[0019] Figure 6 This is a diagram illustrating an example of the demand list 412 of implementation method 1.

[0020] Figure 7 This is a diagram showing the detailed operation flow of step S506 in Embodiment 1.

[0021] Figure 8 This is a flowchart illustrating the control effect prediction process of Implementation Method 1.

[0022] Figure 9 This is a diagram showing the detailed operation flow of step S806 in Embodiment 1.

[0023] Figure 10 This is a diagram showing another example of the hardware structure of the effect prediction device 201, a variation of Embodiment 1.

[0024] Figure 11 This is a diagram showing an example of the overall structure of the air conditioning management system 800 according to Embodiment 2.

[0025] Figure 12 This is a diagram showing the detailed operation flow of step S504 in embodiment 5.

[0026] Figure 13 This is a diagram showing the detailed operation flow of step S804 in embodiment 5. Detailed Implementation

[0027] The following description uses the accompanying drawings. In each drawing, the same or equivalent parts are labeled with the same reference numerals. In the description of the embodiment, the same or equivalent parts are appropriately omitted or simplified. Arrows in the drawings primarily indicate data flow or processing flow. Furthermore, in the following drawings, the size relationships of the structural components may sometimes differ from the actual dimensions. Additionally, in the description of the embodiment, orientations or positions such as top, bottom, left, right, front, back, front, and back are sometimes shown. These descriptions are for ease of explanation and do not limit the arrangement, direction, or orientation of the device, appliance, or component.

[0028] Implementation method 1. Structural description

[0029] Figure 1 This is a diagram showing an example of the overall structure of the air conditioning management system 800 according to this embodiment.

[0030] The air conditioning management system 800 includes an effect prediction system 200 as described in this embodiment. The effect prediction system 200 is also called a DR control effect prediction system.

[0031] The effect prediction system 200 is a system that predicts the effect of DR control with high accuracy with relatively low computational cost.

[0032] In addition, air conditioning is short for air conditioning.

[0033] The effect prediction system 200 consists of at least one server, for example, configured as a cloud. The server in the effect prediction system 200 receives instructions from the terminal device 210 via the communication network 220 and returns the effect prediction results of DR control in response to the instructions from the terminal device 210.

[0034] The following description will focus on the server as an effect prediction device 201 for predicting the effect of DR control.

[0035] The effect prediction device 201 is connected to multiple users 250 via a communication network 220.

[0036] The electricity consumer 250 has an air conditioning system 255. The air conditioning system 255 includes, for example, an indoor unit 251, an outdoor unit 252, and a data acquisition device 253.

[0037] Furthermore, the effect prediction device 201 obtains data from the temperature information DB (database) 231 of the temperature information collection / prediction system 230 via the communication network 220. Additionally, the effect prediction device 201 obtains data from the air conditioning information DB 241 of the air conditioning information collection system 240 via the communication network 220.

[0038] Temperature Information Collection / Forecasting System 230 collects and forecasts various data related to temperature.

[0039] The air conditioning information collection system 240 collects various data such as set temperature and power consumption from the user 250 via the communication network 220.

[0040] Figure 2 This is a diagram showing the region division of the effect prediction system 200 of this embodiment before DR control and after DR control.

[0041] In the effect prediction system 200 of this embodiment, the objective is to improve the prediction accuracy of the effect of DR control during its implementation. To achieve this objective, the effect prediction system 200 of this embodiment includes, for example... Figure 2 As shown, the states before and after DR control are divided into regions. Then, in the effect prediction system 200, different regression equations are used for each state to predict power consumption, and the predicted effect value of DR control is calculated based on the difference between the predicted power consumption when DR control is implemented and when it is not implemented.

[0042] Therefore, in the effect prediction system 200 of this embodiment, the prediction error of power consumption in each state is reduced and the prediction accuracy of the DR control effect is improved.

[0043] Figure 3This is a diagram illustrating an example of the hardware structure of the effect prediction device 201 of this embodiment.

[0044] The effect prediction device 201 is a computer. The effect prediction device 201 includes at least a processor 910, a memory 921, and an auxiliary storage device 922. Furthermore, the effect prediction device 201 has an input / output interface 930 and a communication interface 950 for communicating with other devices on the communication network 220.

[0045] The communication network 220 is, for example, the Internet. Furthermore, a hardware architecture example will be described in detail later.

[0046] Terminal device 210 is managed by the user of effect prediction system 200. The user sends and receives data between terminal device 210 and effect prediction device 201 within effect prediction system 200 via communication network 220. Communication network 220 is, for example, the Internet. Furthermore, the data sent from terminal device 210 includes DR control content and the date, time, and location of the DR-controlled object. Additionally, the data received by terminal device 210 is the effect prediction value of DR control.

[0047] At least one of the indoor unit 251 and outdoor unit 252 of the air conditioners installed in each user's unit 250 is connected to a data acquisition device 253 for acquiring air conditioner data. The data acquisition device 253 transmits the air conditioner's operating data to the air conditioner information database (DB) 241 within the air conditioner information collection system 240 via a communication network 220. Here, the air conditioner information collection system 240 is configured as a cloud, for example. Furthermore, the communication network 220 is, for example, the Internet. Additionally, the transmitted operating data includes, for example, power consumption and set temperature. Here, the data acquisition device 253 may also be built into either the indoor unit 251 or the outdoor unit 252.

[0048] The effect prediction device 201 within the effect prediction system 200 obtains data required for predicting the effect of DR control from the temperature information DB231 in the temperature information collection / prediction system 230 or the air conditioning information DB241 in the air conditioning information collection system 240 via the communication network 220. Here, the temperature information collection / prediction system 230 is configured as a cloud, for example. Furthermore, the communication network 220 is, for example, the Internet. Additionally, the data obtained by the effect prediction device 201 from the air conditioning information DB241 includes, for example, air conditioning data such as actual power consumption, actual set temperature, and time data. Furthermore, the data obtained by the effect prediction device 201 from the temperature information DB231 includes temperature data such as actual temperature, predicted temperature, and time data.

[0049] Figure 4 This is a diagram illustrating an example of the functional structure of the effect prediction system 200 of this embodiment.

[0050] The effect prediction system 200 is roughly divided into a user effect prediction system 400 that predicts the effect of DR control for each user and a system control unit 401 that controls the effect prediction system 200. The user effect prediction system 400 is also referred to as the user DR control effect prediction system.

[0051] The user effect prediction system 400 also includes an air conditioning data acquisition unit 402, a temperature data acquisition unit 403, a data area division unit 404, a model creation unit 405, a model storage unit 406, and a control effect prediction unit 407.

[0052] The air conditioning data acquisition unit 402 obtains the actual / predicted set temperature, actual power consumption, and time data for any period from the air conditioning information DB241.

[0053] Temperature data acquisition unit 403 obtains actual / predicted temperature and time data from temperature information DB231.

[0054] The data area division unit 404 divides the data obtained by the air conditioning data acquisition unit 402 and the temperature data acquisition unit 403 into a state before DR control and a state after DR control.

[0055] The model making unit 405 uses the actual set temperature, actual power consumption and actual air temperature data after the data region division unit 404 has divided the region to perform regression learning and make a power consumption prediction model.

[0056] The model storage unit 406 stores the power consumption prediction model produced by the model production unit 405.

[0057] The control effect prediction unit 407 predicts the power consumption for each region based on the predicted set temperature, predicted air temperature data, and a power consumption prediction model. The effect of DR control is predicted based on the difference between the predicted power consumption when DR control is implemented and when it is not implemented. The predicted set temperature and predicted air temperature data are obtained from the data region division unit 404 after region division. The power consumption prediction model is obtained from the model storage unit 406.

[0058] Description of the action

[0059] Next, the operation of the effect prediction device 201 of this embodiment will be described. The operation steps of the effect prediction device 201 are equivalent to the effect prediction method. Furthermore, the program that implements the operation of the effect prediction device 201 is equivalent to the effect prediction program.

[0060] In this embodiment, as a method for DR control, a change in the set temperature is considered, for example, an increase in temperature during summer and a decrease in temperature during winter. In this DR control method, the effect of DR control can be adjusted by changing the amount of the set temperature. By changing the set temperature using this DR control method, the effect of DR control can be predicted, and the target amount of power consumption reduction can be achieved.

[0061] <Power Consumption Prediction Model Creation and Processing>

[0062] This document provides an overview of the process for creating and processing the electricity consumption prediction model.

[0063] Effect prediction device 201 predicts the effect of DR control in air conditioning system 255 owned by electricity demander 250.

[0064] The air conditioning data acquisition unit 402 acquires air conditioning data, which includes the actual set temperature of the user 250 and the actual power consumption.

[0065] Temperature data acquisition department 403 acquires temperature data including actual temperature data of 250 demanders.

[0066] The data region division section 404 divides the air conditioning data and temperature data according to each region.

[0067] Model creation unit 405 uses air conditioning data and temperature data divided into regions to create a power consumption prediction model for each region. Specifically, model creation unit 405 uses air conditioning data and temperature data divided into regions and performs regression learning using regression equations corresponding to the power consumption of each region, thereby creating a power consumption prediction model for each region.

[0068] The control effect prediction unit 407 uses a power consumption prediction model for each area to predict the effect of DR control.

[0069] Figure 5 This is a flowchart illustrating the process of creating a power consumption prediction model according to this embodiment.

[0070] In step S501, the system control unit 401 receives location information of the target for power consumption prediction model creation from the user's terminal device 210 via the communication network 220. However, it is also possible that communication with the terminal device 210 is not required, and the system control unit 401 automatically determines the location of the target for power consumption prediction model creation.

[0071] Figure 6 This is a diagram showing an example of the customer list 412 of this embodiment.

[0072] In step S502, the system control unit 401 obtains a list 412 of demanders within the location where the power consumption prediction model is being created. Here, the system control unit 401 may also obtain the list 412 of demanders from the air conditioning information DB 241 via the communication network 220. Alternatively, the system control unit 401 may obtain the list 412 of demanders from another DB. Furthermore, as... Figure 6 As shown, the demander list 412 contains at least demander ID and demander location information. Figure 6 In the example, City A is used as the location for the electricity consumption forecasting model. However, information that can uniquely identify a consumer can replace the consumer ID, such as a MAC address or the consumer's address.

[0073] If there are unselected customers in the customer list 412, proceed to step S503.

[0074] Once all customers in customer list 412 have been selected, the process ends.

[0075] In step S503, a requester for processing an object is selected from the requester list 412.

[0076] In step S504, the air conditioning data acquisition unit 402 acquires at least the actual set temperature, actual power consumption, and time data for any period of the selected user from the air conditioning information DB241 via the communication network 220. Here, the air conditioning data acquisition unit 402 acquires data including data from past DR control or changes in the user's set temperature. This allows for regression learning of the DR control state in subsequent steps.

[0077] In step S505, the temperature data acquisition unit 403 acquires the actual temperature and time data corresponding to the location of the selected customer or the location of the model making object from the temperature information DB231 via the communication network 220. Here, data from the same period as in step S504 is acquired.

[0078] In step S506, the data obtained in steps S504 and S505 is divided into regions by the data region division unit 404. Here, the data region division unit 404 divides the data into regions according to the region division conditions set by the system user. For example, the data region is divided into a state before DR control and a state after DR control. In the state after DR control, the air conditioner compressor stops operating, so power consumption is significantly reduced. By using different regression equations to predict the power consumption in the state before and after DR control, the prediction accuracy is improved.

[0079] Figure 7 This is a diagram illustrating the detailed operation flow of step S506 in this embodiment.

[0080] Step S506 is further divided into Figure 7 Steps S701 to S703 in step S506. Steps S701 to S703 in step S506 are executed by the data area division unit 404.

[0081] This section provides an overview of the data region partitioning process.

[0082] The data region partitioning unit 404 uses air conditioning data and temperature data to perform regression learning using a regression equation that takes the period of the state after DR control as the target variable, thereby creating a period prediction model. More specifically, the data region partitioning unit 404 uses a regression equation that takes at least one parameter from the air conditioning data and temperature data, including set temperature, set temperature change, indoor temperature, and air temperature, as explanatory variables, and takes the period of the state after DR control as the target variable, thereby creating a period prediction model.

[0083] In step S701, the data region division unit 404 calculates the period ΔT of the post-DR control state at the past DR control implementation point and the set temperature change point. DR The data area partitioning unit 404 determines the period ΔT based on the state after DR control. DR The system divides the area into a pre-DR control state and a post-DR control state. In the post-DR control state, the air conditioning compressor stops operating, and power consumption is reduced. Alternatively, a threshold can be set for actual power consumption, defining the period from the implementation of DR control until the threshold is exceeded as the post-DR control state. Or, the period of the post-DR control state can be determined based on compressor operation information. For example, using compressor operation information, the period from the implementation of DR control until the compressor restarts can be defined as the post-DR control state.

[0084] In step S702, the data region division unit 404 learns the period of the state after DR control by using regression analysis with at least one parameter, including set temperature, set temperature change, indoor temperature, and air temperature data, as explanatory variables and the period of the state after DR control as the target variable. The model after regression learning is, for example, the following mathematical formula 1.

[0085] ΔT DR It is the period after DR control, X i It is an explanatory variable, j i These are the regression coefficients and intercepts obtained through regression learning.

[0086] (Mathematical Formula 1) ΔT DR =j1×X1+j2×X2+…+j n

[0087] In step S703, the period prediction model (j1~j n The status of the demander is registered in the model storage unit 406 to determine the demander's status. For example, it is registered in association with the demander's ID.

[0088] If there are unselected areas, proceed to step S507.

[0089] If all regions have been selected, the system will return a determination of whether there are any users who have not selected their regions.

[0090] Steps S507 to S509 are performed by the model making department 405.

[0091] In step S507, the model making unit 405 selects one area.

[0092] In step S508, the model creation unit 405 performs regression learning on the data of the selected region. By performing regression learning in each region to create a power consumption prediction model, the accuracy is improved, and the prediction accuracy of the DR control effect is also improved. Here, the power consumption prediction model is, for example, the following mathematical formula 2.

[0093] E represents actual electricity consumption, T set It is the actual performance setting temperature, T out The actual temperature is k1, k2, and k3 are the regression coefficients and intercepts obtained through regression learning.

[0094] (Mathematical formula 2) E=k1×T set +k2×T out +k3

[0095] In step S509, the model creation unit 405 registers the power consumption prediction model (k1, k2, k3) obtained through regression learning in step S508 into the model storage unit 406, in a state that can identify demanders and regions. For example, it registers the model in association with demander ID and region ID. Then, steps S507 to S509 are repeated until no unselected regions remain.

[0096] Finally, repeat steps S503 to 509 until there are no more unselected customers.

[0097] By periodically creating the above power consumption prediction model, regression learning can be performed without following every user instruction, thus reducing computational load. However, regression learning can also be performed according to each user instruction. When the frequency of user instructions is low, performing regression learning according to each instruction reduces computational load.

[0098] <Control Effect Prediction Processing>

[0099] A summary of the control effect prediction process is provided.

[0100] The air conditioning data acquisition unit 402 acquires the predicted set temperature, which represents a prediction of the user 250's set temperature. For example, the air conditioning data acquisition unit 402 acquires the most recent actual set temperature as the predicted set temperature.

[0101] Temperature data acquisition unit 403 obtains the predicted temperature, which represents the predicted temperature for demander 250.

[0102] The data region division unit 404 divides the predicted set temperature and predicted air temperature into regions according to each region. Specifically, the data region division unit 404 uses a periodic prediction model to divide the regions of predicted set temperature and predicted air temperature into a state before DR control and a state after DR control.

[0103] The control effect prediction unit 407 uses a power consumption prediction model for each region, a predicted set temperature, and a predicted air temperature to predict the power consumption of each region. The control effect prediction unit 407 uses the predicted power consumption value for each region to predict the effect of DR control. Specifically, the control effect prediction unit 407 predicts the effect of DR control based on the difference between the predicted power consumption value before DR control and the predicted power consumption value after DR control.

[0104] Figure 8 This is a flowchart illustrating the control effect prediction process of this embodiment.

[0105] In step S801, the system control unit 401 receives the time and location for implementing DR control from the terminal device 210 via the communication network 220.

[0106] In step S802, the system control unit 401 obtains the demand list 412 within the target area of ​​the DR control. Here, the system control unit 401 can obtain the demand list 412 from the air conditioning information DB 241 via the communication network 220, or it can obtain the demand list 412 from other DBs.

[0107] If there are unselected customers in the customer list 412, proceed to step S803.

[0108] If all customers in the customer list 412 have been selected, proceed to step S811.

[0109] In step S803, one user who needs to process an object is selected from the user list 412.

[0110] In step S804, the air conditioning data acquisition unit 402 obtains, via the communication network 220, at least the predicted set temperature of the selected user's DR control time from the air conditioning information DB241. Here, the most recent set temperature is used as the predicted set temperature.

[0111] In step S805, at the DR control time, the temperature data acquisition unit 403 acquires the predicted temperature and time data corresponding to the location of the selected user or the location of the DR control object from the temperature information DB231 via the communication network 220.

[0112] Figure 9 This is a diagram illustrating the detailed operation flow of step S806 in this embodiment.

[0113] In step S806, the data obtained in steps S804 and S805 is divided into regions by the data region division unit 404. Step S806 is further divided into steps S901 and S902. Steps S901 and S902 of step S806 are executed by the data region division unit 404.

[0114] In step S901, the data region partitioning unit 404 obtains from the model storage unit 406 the period prediction model (j1~j2) of the DR control post-state corresponding to the selected demander. n Data area partitioning section 404, for example, uses the requester ID to obtain the data.

[0115] In step S902, the data region division unit 404 predicts the period (j1~j) of the state after DR control based on the period prediction model. n ) and parameters (X^1~X^ （n-1） The period ΔT^ for predicting the post-DR control state is determined as shown in the following mathematical formula 3. DR The period prediction model of the state after DR control (j1~j n The information is obtained in step S901. Furthermore, it is related to the parameters (X^1~X^). （n-1） The parameters corresponding to the explanatory variables during model building. The period ΔT^ based on the determined predicted DR control state. DR The data area is divided into a state before DR control and a state after DR control.

[0116] (Mathematical expression 3) ΔT^ DR =j1×X^1+j2×X^2+…+j n

[0117] If there are unselected areas, proceed to step S807.

[0118] If all regions have been selected, the process proceeds to step S810.

[0119] Steps S807 to S809 are performed by the control effect prediction unit 407 controlled by DR.

[0120] In step S807, the control effect prediction unit 407 selects one region.

[0121] In step S808, the control effect prediction unit 407 retrieves the power consumption prediction model (k1~k2) corresponding to the selected demander and the selected area from the model storage unit 406. n For example, it can be obtained using the demander ID and the region ID.

[0122] In step S809, the control effect prediction unit 407 calculates the predicted power consumption for the selected area based on the predicted set temperature, the predicted air temperature, and the corresponding power consumption prediction model.

[0123] For example, as shown in the following mathematical formula 4, based on the power consumption prediction model (k1~k3), the predicted set temperature T^ set Predicted temperature T^ out The predicted power consumption E^ is calculated. By dividing the state into pre-DR control and post-DR control states, the power consumption prediction accuracy in each region is improved, and the prediction accuracy of the DR control effect is also improved. Steps S807 to S809 are repeated until there are no unselected regions.

[0124] (Mathematical expression 4) E^=k1×T^ set +k2×T^ out +k3

[0125] In step S810, the control effect prediction unit 407 predicts the effect of DR control based on the difference between the predicted power consumption when DR control is implemented and when DR is not implemented. For example, the effect of DR control is predicted based on the difference between the predicted power consumption when the set temperature is changed and when it is not changed.

[0126] Then, repeat steps S803 to S810 until there are no more unselected customers.

[0127] Finally, in step S811, the system control unit 401 sums the predicted DR control effects for all selected users and sends the sum to the user's terminal device 210 via the communication network 220. Alternatively, the control effect prediction unit 407 can calculate the effect prediction separately for multiple users and add the effect predictions calculated separately for multiple users as the effect prediction result.

[0128] Explanation of hardware structure examples

[0129] use Figure 3 An example of the hardware structure of the effect prediction device 201 of this embodiment is shown.

[0130] The effect prediction device 201 is a computer. The effect prediction device 201 includes a processor 910, and other hardware such as a memory 921, an auxiliary storage device 922, an input / output interface 930, and a communication interface 950. The processor 910 is connected to and controls this other hardware via signal lines 80.

[0131] As described above, the effect prediction device 201 includes a system control unit 401, an air conditioning data acquisition unit 402, a temperature data acquisition unit 403, a data region division unit 404, a model making unit 405, and a control effect prediction unit 407 as functional elements. The functions of the system control unit 401, the air conditioning data acquisition unit 402, the temperature data acquisition unit 403, the data region division unit 404, the model making unit 405, and the control effect prediction unit 407 are implemented via software. The model storage unit 406 is provided in the memory 921 or the auxiliary storage device 922.

[0132] Sometimes, the functions of the system control unit 401, the air conditioning data acquisition unit 402, the temperature data acquisition unit 403, the data area division unit 404, the model making unit 405, and the control effect prediction unit 407 are referred to as the functions of the effect prediction device 201. Furthermore, the system control unit 401, the air conditioning data acquisition unit 402, the temperature data acquisition unit 403, the data area division unit 404, the model making unit 405, and the control effect prediction unit 407 are sometimes referred to as individual parts of the effect prediction device 201.

[0133] The processor 910 is a device for executing the effect prediction program. The effect prediction program is a program that implements the functions of the effect prediction device 201.

[0134] The processor 910 is an IC that performs computational processing. Specific examples of processor 910 include CPU, DSP, and GPU. IC is short for Integrated Circuit. CPU is short for Central Processing Unit. DSP is short for Digital Signal Processor. GPU is short for Graphics Processing Unit.

[0135] Memory 921 is a storage device for temporarily storing data. Specific examples of memory 921 are SRAM or DRAM. SRAM is short for Static Random Access Memory. DRAM is short for Dynamic Random Access Memory.

[0136] Auxiliary storage device 922 is a storage device for storing data. A specific example of auxiliary storage device 922 is an HDD. Furthermore, auxiliary storage device 922 can also be a removable storage medium such as an SD (registered trademark) memory card, CF card, NAND flash memory, floppy disk, optical disk, compact disc, Blu-ray (registered trademark) disc, or DVD. Additionally, HDD is short for Hard Disk Drive. SD (registered trademark) is short for Secure Digital. CF is short for CompactFlash. DVD is short for Digital Versatile Disk.

[0137] Input / output interface 930 is an interface used to connect input / output devices. Specifically, input / output interface 930 is a USB or HDMI (registered trademark) port. USB is short for Universal Serial Bus. HDMI (registered trademark) is short for High Definition Multimedia Interface.

[0138] Communication interface 950 is an interface used for communicating with external devices. As a specific example, communication interface 950 is an Ethernet (registered trademark) port or a device for wireless communication.

[0139] The effect prediction program is executed in the effect prediction device 201. The effect prediction program is read into the processor 910 and executed by the processor 910. The memory 921 stores not only the effect prediction program but also the operating system (OS). The processor 910 executes the OS while simultaneously executing the effect prediction program. The effect prediction program and the OS can also be stored in the auxiliary storage device 922. The effect prediction program and the OS stored in the auxiliary storage device 922 are loaded into the memory 921 and executed by the processor 910. Alternatively, part or all of the effect prediction program can be embedded within the OS.

[0140] The effect prediction device 201 may also have multiple processors instead of processor 910. These multiple processors share the responsibility of executing the effect prediction program. Each processor is a device that executes the effect prediction program in the same way as processor 910.

[0141] The data, information, signal values, and variable values ​​used, processed, or output by the effect prediction program are stored in registers or cache memory within memory 921, auxiliary storage device 922, or processor 910.

[0142] Alternatively, the term "part" in each section of the effect prediction device 201 can be replaced with "circuit," "process," "step," "process," or "line." The effect prediction program causes a computer to execute each "process" in which the term "part" in each section of the effect prediction device 201 is replaced with "process." Alternatively, the term "process" in each "process" of the effect prediction device 201 can be replaced with "program," "program product," "computer-readable storage medium storing the program," or "computer-readable recording medium recording the program." Furthermore, the effect prediction method is a method performed by executing an effect prediction program by the effect prediction device 201.

[0143] The effect prediction program can also be provided by storing it on a computer-readable recording medium. Furthermore, the effect prediction program can also be provided as a program product.

[0144] Other structures

[0145] <Variation Example>

[0146] In this embodiment, the functions of each part of the effect prediction device 201 are implemented by software. As a variation, the functions of each part of the effect prediction device 201 can also be implemented by hardware.

[0147] Specifically, the effect prediction device 201 has an electronic circuit 909 instead of a processor 910.

[0148] Figure 10 This is a diagram showing another example of the hardware structure of the effect prediction device 201, a variation of this embodiment.

[0149] Electronic circuit 909 is a dedicated electronic circuit that implements the functions of each part of effect prediction device 201. Specifically, electronic circuit 909 can be a single circuit, a composite circuit, a programmable processor, a parallel programmable processor, a logic IC, a GA, an ASIC, or an FPGA. GA is short for Gate Array. ASIC is short for Application Specific Integrated Circuit. FPGA is short for Field-Programmable Gate Array.

[0150] The functions of each part of the effect prediction device 201 can be implemented by a single electronic circuit or distributed among multiple electronic circuits.

[0151] As another variation, some functions of each part of the effect prediction device 201 can be implemented by electronic circuits, while the remaining functions can be implemented by software. Furthermore, some or all functions of each part of the effect prediction device 201 can also be implemented by firmware.

[0152] The processor and electronic circuits are each referred to as processing circuits. That is, the functions of each part of the effect prediction device 201 are realized through processing circuits.

[0153] Explanation of the effects of this implementation method

[0154] As described above, in the effect prediction device of this embodiment, such as Figure 2 As shown, the states before and after DR control are divided into regions. Then, the effect prediction device uses a different regression equation for each state to predict power consumption, and calculates the predicted effect of DR control based on the difference between the predicted power consumption when DR control is implemented and when it is not implemented. Therefore, the effect prediction device according to this embodiment achieves the effect of reducing the prediction error of power consumption in each state and improving the prediction accuracy of the DR control effect.

[0155] Implementation method 2.

[0156] In this embodiment, the differences from Embodiment 1 and the points added to Embodiment 1 will be explained.

[0157] In this embodiment, structures having the same functions as in Embodiment 1 are labeled with the same reference numerals, and their descriptions are omitted.

[0158] The effect prediction device 201 is installed in the air conditioning system 255 of the user 250.

[0159] Structural description

[0160] Figure 11 This is a diagram showing an example of the overall structure of the air conditioning management system 800 according to this embodiment.

[0161] In Implementation 1, the DR control effect prediction system 200 is built in the cloud. Furthermore, it is explained that the effect prediction system 200 is connected to the air conditioning information DB241 or the terminal device 210 via a communication network 220.

[0162] In this embodiment, such as Figure 11 As shown, an effect prediction system 200 controlled by DR is formed by effect prediction devices 201 that serve as one or more local servers.

[0163] The effect prediction system 200 consists of effect prediction devices 201 that function as at least one local server. Users directly or indirectly operate and control the effect prediction devices 201 within the effect prediction system 200 through a user interface to obtain the effect prediction results of DR control.

[0164] Each consumer 250 has at least one indoor unit 251 and at least one outdoor unit 252, at least one of which is connected to a data acquisition device 253. The data acquisition device 253 sends air conditioner operation data to the air conditioning information DB 241. For example, it sends power consumption and set temperature data. Here, the data acquisition device 253 may also be built into the indoor unit 251 and the outdoor unit 252.

[0165] When predicting the effect of DR control, the effect prediction system 200 obtains air conditioning data from the air conditioning information DB241. Furthermore, the effect prediction system 200 obtains temperature data from the temperature information DB231 within the temperature information collection / prediction system 230 via the communication network 220. Here, the temperature information collection / prediction system 230 is configured as a cloud, for example, and the communication network 220 is the Internet.

[0166] Description of the action

[0167] The operation of the effect prediction device 201 in this embodiment is the same as that described in Embodiment 1.

[0168] Explanation of the effects of this implementation method

[0169] As described above, in the effect prediction system of this embodiment, buildings such as buildings, factories, or apartments with one or more indoor units and one or more outdoor units are considered as one demander to predict the effect of DR control. By having the system locally, cloud costs can be reduced, and for demanders with many air conditioners, the total cost related to system implementation / maintenance can be reduced.

[0170] Implementation method 3.

[0171] In this embodiment, the differences from embodiments 1 and 2, as well as the points added to embodiments 1 and 2, will be explained.

[0172] In this embodiment, structures that have the same functions as those in embodiments 1 and 2 are labeled with the same reference numerals, and their descriptions are omitted.

[0173] Structural description

[0174] The structure of the effect prediction system 200 in this embodiment is the same as that described in embodiments 1 and 2.

[0175] Description of the action

[0176] Regarding the operation of the effect prediction system 200 in this embodiment, points that differ from those described in embodiments 1 and 2 will be explained.

[0177] In this embodiment, the data region division unit 404 sets a threshold for the following parameters, which uses at least one of the following: set temperature, set temperature change, indoor temperature, air temperature, and elapsed time after demand response control (DR) implementation, from the air conditioning data and air temperature data. The data region division unit 404 defines the period from the implementation of DR control until the threshold is exceeded as the post-DR control state.

[0178] In embodiments 1 and 2, an example is described in which the data region partitioning unit 404 determines the prediction period of the post-DR control state through regression analysis in steps S506 and S806. At this time, the data region partitioning unit 404 uses at least one parameter including the set temperature, the change in set temperature, the indoor temperature, and the air temperature as explanatory variables, and uses the period of the post-DR control state as the target variable.

[0179] In this embodiment, a threshold is set for at least one parameter among the following: set temperature, set temperature change, indoor temperature, air temperature, and elapsed time after DR control is implemented. The period from the implementation of DR control until the threshold is exceeded is defined as the post-DR control state. For example, if the set temperature T is used...set and indoor temperature T in The parameter T set -T in If the threshold is set to 0, then DR control will be implemented until T. set -T in The state after DR control is set up until it exceeds 0.

[0180] Explanation of the effects of this implementation method

[0181] As described above, in the effect prediction system of this embodiment, by using a threshold in the decision-making process during the period after DR control, the computational cost of regression learning can be reduced.

[0182] Implementation method 4.

[0183] In this embodiment, the differences from embodiments 1 to 3 and the points added to embodiments 1 to 3 will be explained.

[0184] In this embodiment, structures that have the same functions as those in embodiments 1 to 3 are labeled with the same reference numerals, and their descriptions are omitted.

[0185] Structural description

[0186] The structure of the effect prediction system 200 in this embodiment is the same as that described in embodiments 1 to 3.

[0187] Description of the action

[0188] Regarding the operation of the effect prediction system 200 in this embodiment, points that differ from those described in embodiments 1 to 3 will be explained.

[0189] In embodiments 1 to 3, when the air conditioning data acquisition unit 402 acquires the predicted set temperature in step S804, the most recent actual set temperature is used as the predicted set temperature. In this embodiment, the predicted set temperature is a value such as the average value, the most frequent value, or the median value of the set temperature over a fixed period. For example, the most frequent value of the set temperature over the most recent two weeks is used as the predicted set temperature.

[0190] As mentioned above, the set temperature before DR control execution can also be set to any one of the average, most frequent, or median values ​​of the set temperature over a fixed period. Alternatively, the set temperature before DR control execution can be determined using at least one of the time-series data of air temperature, humidity, and solar irradiance.

[0191] Explanation of the effects of this implementation method

[0192] As described above, in the effect prediction system of this embodiment, by utilizing values ​​such as average value, most frequent value, or median value, a more reliable prediction set temperature can be set.

[0193] Implementation method 5.

[0194] In this embodiment, the differences from embodiments 1 to 3 and the points added to embodiments 1 to 3 will be explained.

[0195] In this embodiment, structures that have the same functions as those in embodiments 1 to 3 are labeled with the same reference numerals, and their descriptions are omitted.

[0196] Structural description

[0197] The structure of the effect prediction system 200 in this embodiment is the same as that described in embodiments 1 to 3.

[0198] Description of the action

[0199] In embodiments 1 to 3, when the air conditioning data acquisition unit 402 acquires the predicted set temperature in step S804, the most recent actual set temperature is used as the predicted set temperature. In this embodiment, the predicted set temperature is calculated based on regression learning with at least one of air temperature, humidity, and solar radiation as explanatory variables and the set temperature as the target variable. By predicting the set temperature in this way, a predicted set temperature that takes into account the user's setting preferences can be calculated.

[0200] Figure 12 This is a diagram illustrating the detailed operation flow of step S504 in this embodiment.

[0201] In this embodiment, such as Figure 12 As shown, step S504 is divided into steps S1101 to S1103.

[0202] In step S1101, the air conditioning data acquisition unit 402 acquires the selected user's actual set temperature, actual power consumption, and time data from the air conditioning information DB241 in the air conditioning information collection system 240 via the communication network 220. Simultaneously, it acquires at least one of the actual temperature, actual humidity, and actual sunshine duration data. Here, the actual temperature, humidity, and sunshine duration data can be acquired from the temperature information DB231 in the temperature information collection / prediction system 230 via the communication network 220, or they can be acquired via the temperature data acquisition unit 403.

[0203] In step S1102, the air conditioning data acquisition unit 402 performs regression learning with at least one of temperature, humidity, and sunshine duration as explanatory variables and the set temperature as the target variable. The model after regression learning is, for example, the following mathematical formula 5. X i To represent explanatory variables, l i This represents the regression coefficient and intercept.

[0204] (Mathematical expression 5)T set =l1×X1+l2×X2+…+l n

[0205] In step S1103, the air conditioning data acquisition unit 402, when able to determine the user's needs, generates a prediction set temperature model (l i ~l n Registered in model storage unit 406. For example, registered in association with the demander ID.

[0206] Figure 13 This is a diagram illustrating the detailed operation flow of step S804 in this embodiment.

[0207] In this embodiment, step S804 is divided into Figure 13 Steps S1201 to S1203 are shown.

[0208] In step S1201, the air conditioning data acquisition unit 402 acquires at least one of the predicted temperature, predicted humidity, and predicted sunshine amount data of the selected user, as well as the time data. Here, the predicted temperature, predicted humidity, and predicted sunshine amount data can be obtained from the temperature information DB231 in the temperature information collection / prediction system 230 via the communication network 220, or they can be obtained via the temperature data acquisition unit 403.

[0209] In step S1202, the air conditioning data acquisition unit 402 retrieves the predicted set temperature model (l_1~l_n) corresponding to the selected user from the model storage unit 406. For example, it retrieves the model using the user ID.

[0210] Next, in step S1203, the air conditioning data acquisition unit 402 sets the temperature according to the predicted temperature model (l1~l n ) and the parameters (X^1~X^) corresponding to the explanatory variables during model creation. （n-1） The predicted setpoint temperature T^ is determined as shown in the following mathematical formula 6. set .

[0211] (Mathematical expression 6) T^ set =l1×X^1+l2×X^2+…+l n Explanation of the effects of this implementation method

[0212] As described above, in the effect prediction system of this embodiment, the predicted set temperature is calculated based on regression learning with at least one of temperature, humidity, and sunshine as explanatory variables and the set temperature as the target variable. By predicting the set temperature in this way, a predicted set temperature that takes into account the setting preferences of the users can be calculated.

[0213] In the effect prediction system of the above embodiments 1 to 5, a method for predicting the reduction in power consumption of air conditioner control related to the demand response of an air conditioner consisting of the following elements is explained.

[0214] The reduction in electricity consumption is calculated by comparing the electricity consumption before and after the execution of the control measures involved in demand response.

[0215] The calculations are performed using a predictive model relating to power consumption prior to the execution of the control involved in the demand response and a predictive model relating to power consumption after the execution of the control involved in the demand response.

[0216] The predictive model related to power consumption prior to the execution of the demand response control is learned based on air conditioner operating data or meteorological information prior to the execution of the same control as the demand response control. The predictive model related to power consumption after the execution of the demand response control is learned based on air conditioner operating data or meteorological information after the execution of the same control as the demand response control.

[0217] Based on the prediction model corresponding to the power consumption before the execution of the demand response control and the predicted air conditioner operation data or predicted meteorological information at the time point of the demand response control, the power consumption before the execution of the demand response control is predicted.

[0218] Based on the prediction model corresponding to the power consumption after the execution of the demand response control, the predicted air conditioner operation data or predicted meteorological information at the time point of the demand response control, and the content of the demand response control, the power consumption after the execution of the demand response control is predicted.

[0219] Furthermore, in the above prediction method, the demand response control content of the air conditioner is the change of the set temperature of the air conditioner.

[0220] The predictive model for power consumption before control execution in the demand response is learned by using time series data of power consumption before the change of the air conditioner's set temperature, set temperature, and air temperature.

[0221] The predictive model for power consumption after control execution involved in the demand response is learned by time series data of power consumption after changes in the set temperature of the air conditioner, the set temperature, and the air temperature.

[0222] Furthermore, in the above prediction method, the period after the control execution involved in the demand response is learned and determined based on at least one of the following: the period of the state after the control execution involved in the demand response, the amount of change in the set temperature, the set temperature of the air conditioner before the set temperature change, the indoor temperature and the air temperature, the set temperature of the air conditioner after the set temperature change, and the time series data of the indoor temperature and the air temperature.

[0223] Furthermore, in the above prediction method, for the period after the execution of the control involved in the demand response, a threshold is set for at least one of the following parameters: the set temperature of the air conditioner before the set temperature change, the indoor temperature and air temperature, the set temperature of the air conditioner after the set temperature change, and the elapsed time from the execution of the demand response control. The period from the execution of the control involved in the demand response until the parameter becomes above the threshold is set as the period of the state after the execution of the control involved in the demand response.

[0224] In embodiments 1 to 5 above, each part of the effect prediction system was described as an independent functional block. However, the structure of the effect prediction system may not be as described in the above embodiments. The functional blocks of the effect prediction system can have any structure as long as they can perform the functions described in the above embodiments. Furthermore, the effect prediction system may not be a single device, but a system composed of multiple devices.

[0225] Furthermore, multiple parts of embodiments 1 to 5 can be combined to implement the embodiments. Alternatively, one part of these embodiments can be implemented. Furthermore, these embodiments can be implemented in any combination, either as a whole or in part.

[0226] That is, in embodiments 1 to 5, each embodiment can be freely combined, or any constituent element of each embodiment can be modified, or any constituent element can be omitted in each embodiment.

[0227] Furthermore, the above-described embodiments are essentially preferred examples and are not intended to limit the scope of this disclosure, its applications, or its uses. The above-described embodiments can be modified as needed. For example, the steps described using flowcharts or sequence diagrams can be appropriately altered.

[0228] Label Explanation

[0229] 80 Signal line, 201 Effect prediction device, 200 Effect prediction system, 210 Terminal device, 220 Communication network, 230 Temperature information collection / prediction system, 231 Temperature information database, 240 Air conditioning information collection system, 241 Air conditioning information database, 250 Customer, 251 Indoor unit, 252 Outdoor unit, 253 Data acquisition device, 255 Air conditioning system, 400 Effect prediction system for each customer, 401 System control unit, 402 Air conditioning data acquisition unit, 403 Temperature data acquisition unit, 404 Data area division unit, 405 Model making unit, 406 Model storage unit, 407 Control effect prediction unit, 412 Customer list, 800 Air conditioning management system, 909 Electronic circuit, 910 Processor, 921 Memory, 922 Auxiliary storage device, 930 Input / output interface, 950 Communication interface, 950 Communication device.

Claims

1. An effect prediction device for predicting the effect of demand response control on an air conditioning system owned by an electricity consumer, wherein, The effect prediction device includes: The air conditioning data acquisition unit acquires air conditioning data, which includes the actual set temperature indicating the user's set temperature and the actual power consumption indicating the user's power consumption. The temperature data acquisition unit acquires temperature data, which includes the actual temperature of the user's expressed temperature. The data region division department divides the air conditioning data and the temperature data according to each region; The model making department uses the air conditioning data and temperature data divided according to each of the regions to make a power consumption prediction model for each region. as well as The control effect prediction unit uses the power consumption prediction model for each of the regions to predict the effect of the demand response control.

2. The effect prediction device according to claim 1, wherein, The model-making department uses the air conditioning data and temperature data divided according to each region, and performs regression learning using regression equations corresponding to the power consumption of each region, thereby creating a power consumption prediction model for each region.

3. The effect prediction device according to claim 1 or 2, wherein, The data region division unit uses the air conditioning data and the temperature data to perform regression learning using a regression equation that takes the period of the demand response control state as the target variable, thereby creating a period prediction model.

4. The effect prediction device according to claim 3, wherein, The data region division unit uses the following regression equation for regression learning, which takes at least one of the following as explanatory variables: set temperature, set temperature change, indoor temperature, and air temperature from the air conditioning data and the air temperature data, and takes the period of the demand response control state as the target variable.

5. The effect prediction device according to claim 3 or 4, wherein, The air conditioning data acquisition unit acquires the predicted set temperature, which represents a prediction of the user's desired temperature. The temperature data acquisition unit acquires the predicted temperature, which represents a forecast of the temperature for the user. The data region division unit divides the predicted set temperature and the predicted air temperature into regions according to each region. The control effect prediction unit uses the power consumption prediction model, the predicted set temperature, and the predicted air temperature for each region to predict the predicted power consumption value for each region, and uses the predicted power consumption value for each region to predict the effect of the demand response control.

6. The effect prediction device according to claim 5, wherein, The data region division unit uses the period prediction model to divide the predicted set temperature and the predicted air temperature into a pre-demand response control state and a post-demand response control state.

7. The effect prediction device according to claim 5, wherein, The control effect prediction unit predicts the effect of the demand response control based on the difference between the predicted power consumption value in the state before the demand response control and the predicted power consumption value in the state after the demand response control.

8. The effect prediction device according to claim 1 or 2, wherein, The data region division unit sets a threshold for the following parameters, which uses at least one of the following: set temperature, set temperature change, indoor temperature, air temperature, and elapsed time after the implementation of demand response control in the air conditioning data and the air temperature data. The data region division unit defines the period from the implementation of demand response control to the point where the threshold is exceeded as the post-demand response control state.

9. The effect prediction device according to claim 5, wherein, The air conditioning data acquisition unit obtains the most recent actual set temperature as the predicted set temperature.

10. The effect prediction device according to claim 5, wherein, The air conditioning data acquisition unit acquires the average, most frequent, or central value of the set temperature over a fixed period as the predicted set temperature.

11. The effect prediction device according to claim 5, wherein, The air conditioning data acquisition unit calculates the predicted set temperature based on regression learning that uses at least one of temperature, humidity, and sunshine as explanatory variables and the set temperature as the target variable.

12. The effect prediction device according to any one of claims 1 to 11, wherein The effect prediction device is connected to multiple users via a communication network. The control effect prediction unit calculates the effect prediction for each of the multiple users, and adds the effect predictions calculated for each user together to output the effect prediction result.

13. The effect prediction device according to any one of claims 1 to 11, wherein, The effect prediction device is installed in the air conditioning system of the user.

14. An effect prediction method for predicting the effect of demand response control in an air conditioning system owned by an electricity consumer, wherein, The computer obtains air conditioning data, which includes the user's actual set temperature and actual power consumption. The computer acquires temperature data, which includes the actual temperature readings of the user. The computer divides the air conditioning data and the temperature data according to each region. The computer uses the air conditioning data and temperature data, divided according to each of the said regions, to create a power consumption prediction model for each of the said regions. The computer uses the power consumption prediction model for each of the regions to predict the effectiveness of the demand response control.

15. An effect prediction program that predicts the effect of demand response control on an air conditioning system owned by an electricity consumer, wherein, The effect prediction program causes the computer to perform the following processes: Air conditioning data acquisition and processing: acquiring air conditioning data, which includes the actual set temperature indicating the user's set temperature and the actual power consumption indicating the user's power consumption. Temperature data acquisition and processing: acquire temperature data, which includes the actual temperature of the user's expressed temperature. Data region segmentation processing is performed to divide the air conditioning data and the temperature data into each region; The model creation process uses the air conditioning data and temperature data divided according to each region to create a power consumption prediction model for each region. as well as The control effect prediction process uses the power consumption prediction model for each region to predict the effect of the demand response control.

Images