Substation building air conditioner control model training method and system

Abstract

The invention discloses a transformer substation building air conditioner control model training method and system, and the method comprises the steps: carrying out the verification and cleaning of each piece of obtained training data and target data, obtaining a plurality of pieces of processed training data and processed target data, inputting each piece of processed training data into a constructed initial control model, and carrying out the verification and cleaning of the obtained training data and target data, the initial control model is used for processing the training data, so that a fuzzification layer of the initial control model carries out fuzzification processing on the processed training data to obtain corresponding membership parameters, according to the membership parameters, a recursion layer is used for processing the time sequence of the processed training data to obtain corresponding feature data, and a decision layer is used for analyzing the feature data to obtain the feature data; and obtaining model temperature data, then obtaining a matching degree according to the model temperature data, and updating parameters in the initial control model according to the matching degree. According to the air conditioner control model obtained through the method, the temperature of each room in the transformer substation can be accurately predicted, and energy wasted by each air conditioner is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of substation air-conditioning control, and particularly to a method and system for training an air-conditioning control model for a substation building. Background Art

[0002] In the power transmission system, the substation, as an important link, undertakes the tasks of power conversion and distribution. And various power equipment in the substation is also constantly updated to respond to the call for energy conservation and emission reduction.

[0003] In the prior art, air conditioners can be installed in each room of the substation, and each air conditioner can start working or stop working according to the operations triggered by the staff or according to the preset time. However, in actual applications, the temperature in the room where the air conditioner is turned on may be too high or too low, resulting in excessive cold or hot air output by the air conditioner, thus causing the problem of energy waste again. Summary of the Invention

[0004] To solve the above technical problems, embodiments of the present invention provide a method and system for training an air-conditioning control model for a substation building to solve the technical problem that the temperature in the room where the air conditioner is turned on may be too high or too low in the prior art, resulting in excessive cold or hot air output by the air conditioner, thus causing energy waste.

[0005] The first aspect of the embodiments of the present invention provides a method for training an air-conditioning control model for a substation building. The method includes: Obtaining a plurality of training data and target data to be trained; Performing verification and cleaning on each training data and target data to obtain a plurality of processed training data and processed target data; Inputting each processed training data into the constructed initial control model, so that the fuzzification layer of the initial control model performs fuzzification processing on each processed training data to obtain membership degree parameters corresponding to each processed training data. According to the membership degree parameters, the recursive layer processes the time series of each processed training data to obtain feature data corresponding to each moment, and the decision layer analyzes each feature data to obtain model temperature data; Comparing the model temperature data with the processed target data to obtain a matching degree, updating the parameters in the initial control model according to the matching degree to obtain an updated initial control model, and continuing to train the updated initial control model until the training times are reached or the matching degree is less than a preset threshold, and then stopping the training to obtain an air-conditioning control model.

[0006] In a possible implementation of the first aspect, the multiple training data includes substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data, and the target data includes indoor temperature sample data.

[0007] In a possible implementation of the first aspect, verify and clean each training data and target data to obtain multiple processed training data, including: Traverse the substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data corresponding to each moment, and determine whether there are duplicate data. If so, delete the duplicate data; Traverse the substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data corresponding to each moment, and determine whether there is missing data. If so, supplement it according to the data corresponding to the previous moment and the data corresponding to the next moment.

[0008] In a possible implementation of the first aspect, after obtaining the model temperature data, it further includes: Obtain the date corresponding to the model temperature data according to the moment corresponding to the model temperature data; Output the air-conditioning working mode according to the date corresponding to the model temperature data.

[0009] In a possible implementation of the first aspect, compare the model temperature data with the processed target data to obtain a matching degree, including: Obtain the temperature difference according to the model temperature data and the corresponding processed target data; Calculate the quotient of the temperature difference and the corresponding processed target data to obtain the matching degree.

[0010] In a possible implementation of the first aspect, the initial control model is constructed based on a BP neural network and combined with a T-S recursive fuzzy neural network.

[0011] To solve the same technical problem, the second aspect of the embodiments of the present invention provides a substation building air-conditioning control model training system, including an acquisition module, a cleaning module, a training module, and an adjustment module, where The acquisition module is used to acquire multiple training data and target data to be trained; The cleaning module is used to verify and clean each training data and target data to obtain multiple processed training data and processed target data; The training module is used to input each processed training data into the constructed initial control model, so that the fuzzification layer of the initial control model performs fuzzification processing on each processed training data to obtain membership parameters corresponding to each processed training data. According to the membership parameters, the recursive layer is used to process the time series of each processed training data to obtain feature data corresponding to each moment, and the decision layer is used to analyze each feature data to obtain model temperature data; The adjustment module is used to compare the model temperature data with the processed target data to obtain a matching degree, update the parameters in the initial control model according to the matching degree to obtain an updated initial control model, and continue to train the updated initial control model until the training times are reached or the matching degree is less than a preset threshold, and then stop training to obtain an air-conditioning control model.

[0012] In a possible implementation manner of the second aspect, the multiple training data include substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data, and the target data includes indoor temperature sample data.

[0013] The third aspect of the embodiments of the present invention provides a computer device, including: A memory for storing a computer program; A processor for implementing the steps of the substation building air-conditioning control model training method as in the first aspect when executing the computer program.

[0014] The fourth aspect of the embodiments of the present invention provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the substation building air-conditioning control model training method as in the first aspect are implemented.

[0015] The technical solution of the present invention has the following advantages: The substation building air-conditioning control model training method provided by the embodiment of the present invention includes: obtaining a plurality of training data and target data to be trained, verifying and cleaning each piece of training data and target data to obtain a plurality of processed training data and processed target data, inputting each piece of processed training data into the constructed initial control model, so that the fuzzification layer of the initial control model performs fuzzification processing on each piece of processed training data to obtain membership degree parameters corresponding to each piece of processed training data, processing the time series of each piece of processed training data according to the membership degree parameters by using a recursive layer to obtain feature data corresponding to each moment, analyzing each piece of feature data by using a decision layer to obtain model temperature data, then comparing the model temperature data with the processed target data to obtain a matching degree, updating the parameters in the initial control model according to the matching degree to obtain an updated initial control model, and continuing to train the updated initial control model until the training times are reached or the matching degree is less than a preset threshold to stop training, so as to obtain an air-conditioning control model. The air-conditioning control model obtained by the above method can accurately predict the temperature of each room in the substation and reduce the energy wasted by each air conditioner. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0017] Figure 1 is a flowchart of the substation building air-conditioning control model training method in the embodiment of the present invention; Figure 2 is a system block diagram of the substation building air-conditioning control model training system in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0019] The substation building air-conditioning control model training method provided by the embodiment of the present invention is as Figure 1 shown Figure 1 is a flowchart of the substation building air-conditioning control model training method, including steps S101 to S104, and the specific steps are as follows: S101. Obtain multiple pieces of training data and target data to be trained.

[0020] In this embodiment, sample data is obtained. The sample data includes: substation environment data, substation personnel data, air-conditioning control mode, and indoor temperature data. For example, the substation environment data includes: the temperature and humidity corresponding to the environment where the substation is located, the substation personnel data is used to represent the room distribution of the substation, and the number of staff in each room, and the indoor temperature data is used to represent the temperature corresponding to each room of the substation at each moment. The embodiments of the present application do not make specific limitations on the various data in the sample data.

[0021] Moreover, the substation environment data, substation personnel data, air-conditioning control mode, and indoor temperature data all correspond to time parameters. Correspondingly, the indoor temperature data can be respectively corresponding to the substation environment data, substation personnel data, and air-conditioning control mode based on the time parameters.

[0022] In one embodiment, the multiple pieces of training data include substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data, and the target data includes indoor temperature sample data.

[0023] In this embodiment, during the operation of the substation, multiple data such as substation environment data, substation personnel data, air-conditioning control mode, and indoor temperature data can be recorded. Correspondingly, before training the air-conditioning control model, the recorded multiple data can be obtained through a pre-set storage space, so that the obtained various data can be used as sample data, so that in the subsequent steps, the air-conditioning control model can be trained according to the sample data.

[0024] In addition, during the process of inputting the sample data into the initial control model, the sample data can be classified first to obtain training data and target data, and then the training data is input into the initial control model. Among them, the training data can include: substation environment data, substation personnel data, air-conditioning control mode, and the target data can include: indoor temperature data.

[0025] S102. Check and clean each piece of training data and target data to obtain multiple processed training data and processed target data.

[0026] In this embodiment, in order to improve the accuracy of training the air-conditioning control model, the sample data can be cleaned to remove redundant data in the sample data and avoid problems during the process of training the air-conditioning control model. Specifically, the sample data is cleaned, duplicate data in the sample data is deleted, and the sample data is checked to complete the missing data in the sample data.

[0027] In one embodiment, checksum verification and cleaning are performed on each piece of training data and target data to obtain multiple pieces of processed training data, including: Traverse the substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data corresponding to each moment, and determine whether there are duplicate data. If there are, delete the duplicate data; Traverse the substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data corresponding to each moment, and determine whether there is missing data. If there is, supplement it according to the data corresponding to the previous moment and the data corresponding to the next moment.

[0028] In this embodiment, according to the time parameter, traverse the substation environment data, substation personnel data, air-conditioning control mode, and indoor temperature data corresponding to each moment to determine whether there are duplicate data. If there are duplicate data, the duplicate data can be deleted to complete the cleaning of the sample data.

[0029] During the process of cleaning the sample data, the missing data in the sample data can also be supplemented, so that the air-conditioning control model can be trained based on the complete sample data.

[0030] Similarly, according to the time parameter, traverse the substation environment data, substation personnel data, air-conditioning control mode, and indoor temperature data corresponding to each moment to determine whether any data corresponding to any moment is complete. If the data is complete, there is no need to supplement the sample data. If the data is incomplete, it can be supplemented according to the data corresponding to the previous moment and the data corresponding to the next moment. For example, the average value between the data corresponding to the previous moment and the data corresponding to the next moment can be selected as the supplementary data. Of course, it can also be supplemented in other ways, and the embodiments of the present application do not make specific limitations on this.

[0031] S103. Input each piece of processed training data into the constructed initial control model, so that the fuzzification layer of the initial control model performs fuzzification processing on each piece of processed training data to obtain the membership degree parameters corresponding to each piece of processed training data. According to the membership degree parameters, use the recursive layer to process the time series of each piece of processed training data to obtain the feature data corresponding to each moment, and use the decision layer to analyze each piece of feature data to obtain the model temperature data.

[0032] In this embodiment, after obtaining the processed training data, the processed training data can be input into a preset initial control model, so that in subsequent steps, the initial control model can be continuously trained with sample data to obtain an air conditioner control model. Specifically, the processed training data is input into the input layer of the initial control model, and the training data of each dimension (such as substation environment data, substation personnel data, and air conditioner control mode) is input into different membership functions in the fuzzification layer to obtain the membership parameters corresponding to each dimension. Then, through the data processing and analysis of the recursive layer and the decision layer, the predicted target temperature is obtained.

[0033] After receiving the input training data, the initial control model can output model temperature data similar to the target data, so that the initial control model can be adjusted based on the model temperature data and the actually collected indoor temperature data to obtain an air conditioner control model.

[0034] Specifically, through the neural network in the initial control model, feature extraction is performed on each item of data in the training data to obtain feature data. There are multiple layers of neural networks set in the initial control model. Through the feature extraction of the training data by the multiple layers of neural networks, the feature data corresponding to each item of data in the training data can be obtained, so that in subsequent steps, the model temperature data can be output according to the feature data.

[0035] In practical applications, based on the time parameters corresponding to each item of data, feature extraction can be performed on the training data to obtain the feature data corresponding to each moment. That is, for the training data corresponding to each moment, according to the time parameters, through the neural network in the initial control model, feature extraction is performed on the training data to obtain the feature data corresponding to the moment. For example, the training data in the sample data can be input into the input layer of the initial control model, and the training data is fuzzified through the fuzzification layer of the initial control model to obtain the membership parameters corresponding to the training data of different dimensions. Then, in the recursive layer, a Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), or Gate Recurrent Unit (GRU) is used to process the time series in the training data, so that the feature data corresponding to each moment can be obtained.

[0036] After obtaining the feature data, the initial control model can output the model temperature data corresponding to the feature data in combination with the preset neural network. Among them, the model temperature data is the temperature data output by the initial control model, which is used to indicate the target temperature that the air conditioner corresponding to each room in the substation needs to output.

[0037] As an example of this embodiment, after obtaining the feature data, the decision-making layer of the initial control model can analyze the feature data to obtain the corresponding model temperature data, and then output the model temperature data through the output layer, that is, output the target temperature corresponding to the air conditioner in the current environment.

[0038] In one embodiment, the initial control model is constructed based on a BP neural network in combination with a T-S recursive fuzzy neural network.

[0039] In this embodiment, the initial control model is constructed based on a BP neural network in combination with a Takagi-Sugeno Recurrent Fuzzy Neural Network, and can output a prediction for the sample data to obtain the target temperature. The neural network adopted by the initial control model in the embodiments of the present application is not specifically limited.

[0040] In one embodiment, after obtaining the model temperature data, it further includes: Obtain the date corresponding to the model temperature data according to the time corresponding to the model temperature data; Output the air conditioner working mode according to the date corresponding to the model temperature data.

[0041] In this embodiment, in practical applications, affected by the temperatures in different seasons, the air conditioner can also be adjusted to cool or heat according to the date corresponding to the sample data. Therefore, after the initial control model outputs the model temperature data according to the feature data, the date corresponding to the model temperature data can be obtained according to the time corresponding to the model temperature data, and then the air conditioner working mode can be output according to the date corresponding to the model temperature data.

[0042] Specifically, after obtaining the model temperature data, the working mode of the air conditioner can also be determined according to the time corresponding to the model temperature data. Correspondingly, the time corresponding to the model temperature data can be determined through the initial control model, and the corresponding date can be extracted based on this time, so as to determine the corresponding season in the subsequent steps. The air conditioner working mode is used to represent that the air conditioner works in a cooling or heating manner.

[0043] After obtaining the date corresponding to the model temperature data, the corresponding season can be determined according to this date, and thus the working mode corresponding to the air conditioner can be determined according to this season. For example, if the corresponding season is summer, the output air conditioner working mode is the cooling mode; if the corresponding season is winter, the output air conditioner working mode is the heating mode.

[0044] It should be noted that in practical applications, affected by different geographical locations, the air-conditioning working modes required by substations in different positions cannot be accurately determined according to seasons. Therefore, during the process of outputting the air-conditioning working mode, it can also be adjusted according to the position where the substation is located. The embodiments of the present application do not make specific limitations on this.

[0045] S104. Compare the model temperature data with the processed target data to obtain a matching degree. Update the parameters in the initial control model according to the matching degree to obtain an updated initial control model, and continue to train the updated initial control model until the training times are reached or the matching degree is less than a preset threshold, and then stop the training to obtain an air-conditioning control model.

[0046] In this embodiment, after obtaining the model temperature data, the model temperature data can be compared with the indoor temperature data in the target data to obtain a comparison result therebetween, so that in subsequent steps, the initial control model can be adjusted according to the comparison result to obtain an air-conditioning control model. Specifically, according to the comparison result and in combination with the indoor temperature data in the target data, the matching degree between the model temperature data and the target data is determined, where the matching degree is used to represent the accuracy of the output of the initial control model, and the air-conditioning control model is trained.

[0047] In one embodiment, comparing the model temperature data with the processed target data to obtain a matching degree includes: Obtain a temperature difference according to the model temperature data and the corresponding processed target data; Calculate a quotient value of the temperature difference and the corresponding processed target data to obtain a matching degree.

[0048] In this embodiment, during the process of comparing the model temperature data with the indoor temperature data, the model temperature moment corresponding to the model temperature data can be determined first, and then according to the model temperature moment, the indoor temperature data at the same moment is searched from the target data, so that the model temperature data can be compared with the indoor temperature data corresponding to the model temperature moment to obtain a comparison result.

[0049] The comparison result is the temperature difference between the model temperature data and the indoor temperature data, that is, subtracting the model temperature data from the target data to obtain the difference between the two. Moreover, the comparison result can also include the magnitude relationship between the model temperature data and the indoor temperature data. Further, in practical applications, the matching degree between the model temperature data and the target data can be determined according to the temperature difference in combination with the indoor temperature data in the target data, where the matching degree is used to represent the accuracy of the output of the initial control model. For example, after obtaining the temperature difference, the control device can divide the temperature difference by the corresponding indoor temperature data to obtain the quotient between the two, so that the quotient can be used as the matching degree between the model temperature data and the target data.

[0050] After obtaining the comparison result, the accuracy of the output of the initial control model can be determined according to the comparison result, so that the loss function in the initial control model can be adjusted according to the accuracy, so that the adjusted initial control model can output more accurate model temperature data, and then an air-conditioning control model can be obtained.

[0051] In the process of adjusting the initial control model, the loss function in the initial control model can be adjusted first according to the comparison result until the initial control model meets the pre-set training conditions, that is, an air-conditioning control model is obtained. The training conditions can be conditions related to the number of training times or conditions related to the matching degree.

[0052] Specifically, during the process of training the initial control model, the control device can count the number of times of training the initial control model. When the number of training times of the initial control model reaches the pre-set training threshold, the training can be stopped, indicating that the training conditions are met, and an air-conditioning control model is obtained.

[0053] Alternatively, during the process of training the initial control model, the control device can obtain the model temperature data output by the initial control model, and obtain the matching degree between the model temperature data and the target data according to the model temperature data and the indoor temperature data in the target data. When the matching degree between the model temperature data and the target data is less than or equal to the pre-set matching threshold, the training of the initial control model is also stopped to obtain an air-conditioning control model. Among them, the training threshold can be determined according to the amount of data required for training the air-conditioning control model, and the matching threshold can be set according to the required accuracy of the air-conditioning control model.

[0054] It should be noted that in the process of adjusting the initial control model through the loss function, the comparison result can be combined with the different weights corresponding to the fuzzification layer, the recursive layer and the decision layer of the initial control model to adjust the various parameters in the fuzzification layer, the recursive layer and the decision layer, so that the adjusted initial control model can output more accurate model temperature data.

[0055] This method obtains sample data and inputs the sample data into a preset initial control model. Through the sample data, the initial control model can be trained to obtain an air-conditioning control model. The obtained air-conditioning control model can accurately predict the temperature of each room in the substation according to the environmental data of the substation and the distribution data of the internal staff in the substation. Therefore, according to the accurately predicted temperature, the air conditioners in the substation can be controlled in advance, thereby reducing the power consumption of each air conditioner, reducing the wasted energy of each air conditioner, and improving the energy conservation and emission reduction effect of the substation.

[0056] The substation building air-conditioning control model training system provided by the embodiment of the present invention, as Figure 2 shown, Figure 2 is a system block diagram of the substation building air-conditioning control model training system 200, including an acquisition module 201, a cleaning module 202, a training module 203, and an adjustment module 204. Among them, The acquisition module 201 is used to acquire a plurality of training data and target data to be trained; The cleaning module 202 is used to check and clean each training data and target data to obtain a plurality of processed training data and processed target data; The training module 203 is used to input each processed training data into the constructed initial control model, so that the fuzzification layer of the initial control model performs fuzzification processing on each processed training data to obtain the membership degree parameters corresponding to each processed training data. According to the membership degree parameters, the recursive layer is used to process the time series of each processed training data to obtain the feature data corresponding to each moment, and the decision layer is used to analyze each feature data to obtain the model temperature data; The adjustment module 204 is used to compare the model temperature data with the processed target data to obtain a matching degree, update the parameters in the initial control model according to the matching degree to obtain an updated initial control model, and continue to train the updated initial control model until the training times are reached or the matching degree is less than a preset threshold, and then stop training to obtain an air-conditioning control model.

[0057] In one embodiment, the plurality of training data includes substation environment sample data, substation personnel sample data, and air-conditioning control mode sample data, and the target data includes indoor temperature sample data.

[0058] The specific implementation manner of the system for training the substation building air-conditioning control model is basically the same as the specific embodiment of the above-mentioned substation building air-conditioning control model training method, and will not be repeated here.

[0059] In an embodiment of the present application, a computer device is provided. The computer device includes a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the above steps are implemented. For the computer device provided in this embodiment, its implementation principle and technical effects are similar to those of the above method embodiment, and will not be elaborated here.

[0060] In an embodiment of the present application, a computer-readable storage medium is provided. A computer program is stored thereon. When the computer program is executed by a processor, the above steps are implemented. For the computer-readable storage medium provided in this embodiment, its implementation principle and technical effects are similar to those of the above method embodiment, and will not be elaborated here.

[0061] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered that the scope described in this specification is covered.

[0062] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only for the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. In particular, it is pointed out that for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A training method for a substation building air conditioning control model, characterized in that: include: Acquire multiple training data and target data to be trained; Verifying and cleaning each of the training data and the target data to obtain a plurality of processed training data and processed target data; Input each of the processed training data into the constructed initial control model, so that the fuzzification layer of the initial control model performs fuzzification processing on each of the processed training data to obtain the membership parameter corresponding to each of the processed training data, and according to the membership parameter, use the recursive layer to process the time series of each of the processed training data to obtain the feature data corresponding to each moment, and use the decision layer to analyze each of the feature data to obtain the model temperature data; The model temperature data is compared with the processed target data to obtain a matching degree, and the parameters in the initial control model are updated according to the matching degree to obtain an updated initial control model. The updated initial control model is continuously trained until the training times are reached or the matching degree is less than a preset threshold, and the training is stopped to obtain an air conditioning control model.

2. The substation building air conditioning control model training method according to claim 1, characterized in that: The plurality of training data include substation environment sample data, substation personnel sample data and air conditioning control mode sample data, and the target data include indoor temperature sample data.

3. The substation building air conditioning control model training method according to any one of claims 1 to 2, characterized in that: The checking and cleaning of each of the training data and the target data to obtain a plurality of processed training data includes: Traversing the substation environment sample data, the substation personnel sample data and the air conditioning control mode sample data corresponding to each moment, determining whether there is duplicate data, and if so, deleting the duplicate data; The substation environment sample data, the substation personnel sample data and the air conditioning control mode sample data corresponding to each moment are traversed to determine whether there is missing data. If so, they are supplemented according to the data corresponding to the previous moment and the data corresponding to the next moment.

4. The substation building air conditioning control model training method according to claim 1, characterized in that: After obtaining the model temperature data, it also includes: According to the time corresponding to the model temperature data, obtaining the date corresponding to the model temperature data; The air conditioning operating mode is output according to the date corresponding to the model temperature data.

5. The substation building air conditioning control model training method according to claim 1, characterized in that: The comparing the model temperature data with the processed target data to obtain a matching degree includes: Obtaining a temperature difference value according to the model temperature data and the corresponding processed target data; The quotient of the temperature difference and the corresponding processed target data is calculated to obtain a matching degree.

6. The substation building air conditioning control model training method according to claim 1, characterized in that: The initial control model is constructed based on BP neural network combined with TS recursive fuzzy neural network.

7. A substation building air conditioning control model training system, characterized in that: It includes acquisition module, cleaning module, training module and adjustment module, among which, The acquisition module is used to acquire a plurality of training data and target data to be trained; The cleaning module is used to verify and clean each of the training data and the target data to obtain a plurality of processed training data and processed target data; The training module is used to input each of the processed training data into the constructed initial control model, so that the fuzzification layer of the initial control model performs fuzzification processing on each of the processed training data to obtain the membership parameters corresponding to each of the processed training data, and according to the membership parameters, the time series of each of the processed training data is processed by the recursive layer to obtain the feature data corresponding to each moment, and the feature data is analyzed by the decision layer to obtain the model temperature data; The adjustment module is used to compare the model temperature data with the processed target data to obtain a matching degree, update the parameters in the initial control model according to the matching degree to obtain an updated initial control model, continue to train the updated initial control model until the training number is reached or the matching degree is less than a preset threshold, and stop training to obtain an air conditioning control model.

8. The substation building air conditioning control model training system according to claim 7, characterized in that: The plurality of training data include substation environment sample data, substation personnel sample data and air conditioning control mode sample data, and the target data include indoor temperature sample data.

9. A computer device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the substation building air conditioning control model training method as described in any one of claims 1 to 6 when executing the computer program.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the substation building air conditioning control model training method according to any one of claims 1 to 6 are implemented.

Images