Air source heat pump load aggregation modeling method, system and equipment and storage medium

Abstract

The invention relates to the technical field of air source heat pumps, and provides an air source heat pump load aggregation modeling method, system and device and a storage medium. The method comprises the steps that air source heat pump operation data information of a plurality of areas is obtained; classifying the obtained multiple groups of air source heat pump operation data information to obtain multiple data sets; training a preset neural network model based on each data set to obtain a local load prediction model corresponding to each data set; fusing the model parameters of the local load prediction models to obtain fused model parameters, and replacing the model parameters of any local load prediction model with the fused model parameters to obtain a global load prediction model; and determining a prediction loss function of the global load prediction model, and optimizing the global load prediction model based on the prediction loss function to obtain a target load prediction model. According to the method, the universality of the target load prediction model in different regions can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of air source heat pumps, and in particular to a method, system, device and storage medium for load aggregation modeling of air source heat pumps. Background Art

[0002] As an efficient and environmentally friendly heating technology, air source heat pumps have been widely used in the fields of building energy conservation and renewable energy utilization. However, when conducting "air source heat pump load aggregation modeling", researchers face a key challenge: how to balance the contradiction between "climate differences in different regions" and "system versatility" in the model.

[0003] On the one hand, the performance and load characteristics of air source heat pumps are strongly dependent on climatic factors such as ambient temperature and humidity, which vary significantly between different geographical regions. On the other hand, in order to improve the practicality and promotion value of the model, a certain degree of versatility must be ensured so that the model can be applied to air source heat pump systems under different climatic conditions. Summary of the invention

[0004] The present application provides an air source heat pump load aggregation modeling method, system, device and storage medium to solve the problems raised by the above background technology.

[0005] In a first aspect, the present application provides an air source heat pump load aggregation modeling method, comprising: Acquire air source heat pump operation data information of several regions; the air source heat pump operation data information includes climate characteristic information, and a mapping relationship table between air source heat pump load and environmental parameters; Based on a preset classification algorithm, the multiple sets of air source heat pump operation data information obtained are classified to obtain multiple data sets; wherein each of the data sets includes multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data set is greater than a preset similarity; The preset neural network model is trained based on each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is the environmental parameter, and the output information is the air source heat pump load; The model parameters of each of the local load forecasting models are integrated to obtain integrated model parameters, and the model parameters of any of the local load forecasting models are replaced with the integrated model parameters to obtain a global load forecasting model; The prediction loss function of the global load prediction model is determined based on the operation data information of each of the air source heat pumps, and the global load prediction model is optimized based on the prediction loss function to obtain a target load prediction model.

[0006] In a possible implementation, the plurality of sets of air source heat pump operation data information obtained are classified based on a preset classification algorithm to obtain a plurality of data sets, including: Determining a plurality of initial cluster numbers based on the amount of the air source heat pump operation data information; Respectively obtain high-dimensional feature coordinate points of climate feature information corresponding to each of the air source heat pump operation data information; Calculate the distance between each high-dimensional feature coordinate point and the Euclidean origin respectively; Arranging each of the high-dimensional feature coordinate points in sequence based on the distance values ​​corresponding to each of the high-dimensional feature coordinate points to obtain a high-dimensional feature coordinate point sequence; wherein the distance values ​​corresponding to the high-dimensional feature coordinate point sequence increase in sequence; For each of the initial clustering numbers, determining a clustering error corresponding to the initial clustering number based on the high-dimensional feature coordinate point sequence; Determining a target number of clusters based on each of the clustering errors; Based on the target cluster quantity and each of the high-dimensional feature coordinate points, the acquired multiple groups of air source heat pump operation data information are classified to obtain multiple data sets.

[0007] In a possible implementation, determining the clustering error corresponding to the initial number of clusters based on the high-dimensional feature coordinate point sequence includes: For each of the initial cluster numbers, determining whether the initial cluster number is 1; If so, determine a center point in the high-dimensional feature coordinate point sequence; when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is an odd number, the center point is a high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point sequence; when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is an even number, the center point is a center point between two high-dimensional feature coordinate points located in the middle of the high-dimensional feature coordinate point sequence; Calculating the square of the Euclidean distance between each high-dimensional feature coordinate point in the high-dimensional feature coordinate point sequence and the center point, and taking the sum of each square value as the clustering error when the initial cluster number is 1; If not, segmenting the high-dimensional feature coordinate point sequence based on the initial cluster number to obtain a plurality of high-dimensional feature coordinate point subsequences; the number of the high-dimensional feature coordinate point subsequences is consistent with the initial cluster number; For each of the high-dimensional feature coordinate point subsequences, a center point is determined in the high-dimensional feature coordinate point subsequence, and the square value of the Euclidean distance between each high-dimensional feature coordinate point in the high-dimensional feature coordinate point subsequence and the center point is calculated respectively, and the sum of each of the square values ​​is used as the intermediate clustering error corresponding to the high-dimensional feature coordinate point subsequence; wherein, when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point subsequence is an odd number, the center point is a high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point subsequence, and when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point subsequence is an even number, the center point is a center point between two high-dimensional feature coordinate points located in the middle of the high-dimensional feature coordinate point subsequence; The sum of the intermediate clustering errors is taken as the clustering error corresponding to the initial number of clusters.

[0008] In a possible implementation manner, determining the target number of clusters based on each of the clustering errors includes: With the initial number of clusters as the horizontal axis and the clustering error as the vertical axis, a mapping curve between the initial number of clusters and the clustering error is drawn; The elbow rule is used to determine the target cluster quantity on the mapping relationship curve.

[0009] In a possible implementation, the multiple sets of air source heat pump operation data information obtained are classified based on the target cluster number and each of the high-dimensional feature coordinate points to obtain multiple data sets, including: Based on the target cluster number, a plurality of initial cluster center points are randomly selected from each of the high-dimensional feature coordinate points; wherein the number of the initial cluster center points is consistent with the target cluster number; For each of the high-dimensional feature coordinate points, the Euclidean distance between the high-dimensional feature coordinate point and each of the initial cluster center points is calculated respectively, and the cluster corresponding to the initial cluster center point corresponding to the shortest Euclidean distance is determined as the initial cluster corresponding to the high-dimensional feature coordinate point; wherein each of the initial clusters includes a plurality of high-dimensional feature coordinate points; For each of the initial clusters, a first intermediate cluster center point is determined among the high-dimensional feature coordinate points in the initial cluster; the first intermediate cluster center point is the cluster center of the initial cluster; For each of the high-dimensional feature coordinate points, the Euclidean distance between the high-dimensional feature coordinate point and each of the first intermediate cluster center points is calculated respectively, and the cluster corresponding to the first intermediate cluster center point corresponding to the shortest Euclidean distance is determined to be the first intermediate cluster corresponding to the high-dimensional feature coordinate point; wherein each of the first intermediate clusters includes multiple high-dimensional feature coordinate points; For each of the first intermediate clusters, determining a second intermediate cluster center point in each high-dimensional feature coordinate point in the first intermediate cluster; the second intermediate cluster center point is the cluster center of the first intermediate cluster; Iterate the steps after determining the first intermediate cluster center point in each high-dimensional feature coordinate point in each initial cluster for each initial cluster until a plurality of target clusters are obtained; wherein each target cluster converges; For each of the target clusters, the air source heat pump operation data information corresponding to each high-dimensional feature coordinate point in the target cluster is combined to obtain a data set corresponding to the target cluster.

[0010] In a possible implementation, the step of determining the prediction loss function of the global load prediction model based on the operation data information of each of the air source heat pumps includes: Inputting each environmental parameter in all the air source data information into the global load prediction model respectively to obtain the air source heat pump predicted load corresponding to each environmental parameter; For each of the environmental parameters, calculating the difference between the predicted load of the air source heat pump corresponding to the environmental parameter and the load of the air source heat pump; Constructing an environmental parameter-difference mapping relationship table based on the difference values ​​corresponding to each of the environmental parameters; The prediction loss function is generated based on the environmental parameter-difference mapping relationship table.

[0011] In a second aspect, the present application provides an air source heat pump load aggregation modeling system, comprising: An acquisition module is used to acquire air source heat pump operation data information of several regions; the air source heat pump operation data information includes climate characteristic information, a mapping relationship table between air source heat pump load and environmental parameters; A classification module, for classifying the acquired multiple groups of air source heat pump operation data information based on a preset classification algorithm to obtain multiple data sets; wherein each of the data sets includes multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data set is greater than a preset similarity; A training module, used to train a preset neural network model based on each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is environmental parameters, and the output information is the air source heat pump load; A parameter fusion module, used to fuse the model parameters of each of the local load forecasting models to obtain a fusion model parameter, and use the fusion model parameter to replace the model parameter of any of the local load forecasting models to obtain a global load forecasting model; The model optimization module determines the prediction loss function of the global load prediction model based on the operation data information of each of the air source heat pumps, and optimizes the global load prediction model based on the prediction loss function to obtain a target load prediction model.

[0012] In a third aspect, the present application provides a terminal device, characterized in that it includes a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, it implements the air source heat pump load aggregation modeling method as described in any of the above items.

[0013] In a fourth aspect, the present application provides a computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, wherein when the computer program is executed by a processor, the air source heat pump load aggregation modeling method as described in any of the above items is implemented.

[0014] The present application provides an air source heat pump load aggregation modeling method, system, device and storage medium, wherein the method comprises: obtaining air source heat pump operation data information of several regions; the air source heat pump operation data information comprises climate characteristic information, a mapping relationship table between air source heat pump load and environmental parameters; classifying the obtained multiple groups of air source heat pump operation data information based on a preset classification algorithm to obtain multiple data sets; wherein each of the data sets comprises multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data set is greater than a preset similarity; respectively based on each The preset neural network model is trained with each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is the environmental parameters, and the output information is the air source heat pump load; the model parameters of each of the local load prediction models are fused to obtain fused model parameters, and the fused model parameters are used to replace the model parameters of any of the local load prediction models to obtain a global load prediction model; the prediction loss function of the global load prediction model is determined based on the operating data information of each of the air source heat pumps, and the global load prediction model is optimized based on the prediction loss function to obtain a target load prediction model. The method first constructs each of the local load prediction models, then fuses the model parameters of each of the local load prediction models to obtain fused model parameters, and uses the fused model parameters to replace the model parameters of any of the local load prediction models to obtain a global load prediction model. The global load prediction model trained in this way not only retains the advantages of the local model, but also helps to achieve global management of the air source heat pump load in each area. On the other hand, by determining the prediction loss function of the global load prediction model based on the operating data information of each of the air source heat pumps, and optimizing the global load prediction model based on the prediction loss function, a target load prediction model is obtained, which helps to further improve the prediction accuracy and versatility of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0016] Figure 1 A schematic diagram of a flow chart of an air source heat pump load aggregation modeling method provided in an embodiment of the present application; Figure 2 A schematic block diagram of the structure of the air source heat pump load aggregation modeling system provided in an embodiment of the present application; Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change based on actual conditions.

[0019] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0020] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0021] As an efficient and environmentally friendly heating technology, air source heat pumps have been widely used in the fields of building energy conservation and renewable energy utilization. However, when conducting "air source heat pump load aggregation modeling", researchers face a key challenge: how to balance the contradiction between "climate differences in different regions" and "system versatility" in the model.

[0022] On the one hand, the performance and load characteristics of air source heat pumps are strongly dependent on climatic factors such as ambient temperature and humidity, which vary significantly between different geographical regions. On the other hand, in order to improve the practicality and promotion value of the model, it is necessary to ensure a certain degree of versatility so that the model can be applied to air source heat pump systems under different climatic conditions. To this end, the embodiments of the present application provide an air source heat pump load aggregation modeling method, system, device and storage medium to solve the above problems.

[0023] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0024] See also Figure 1 , Figure 1 A flow chart of the air source heat pump load aggregation modeling method provided in the embodiment of the present application is shown in FIG. Figure 1 As shown, the air source heat pump load aggregation modeling method provided in the embodiment of the present application includes steps S1 to S5.

[0025] Step S1, obtaining air source heat pump operation data information of several areas; the air source heat pump operation data information includes climate characteristic information, and a mapping relationship table between air source heat pump load and environmental parameters.

[0026] Specifically, the air source heat pump operation data information corresponding to each area is obtained from the air source heat pump database corresponding to each of the areas.

[0027] Step S2, classify the multiple groups of air source heat pump operation data information obtained based on a preset classification algorithm to obtain multiple data sets; wherein each of the data sets includes multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data sets is greater than the preset similarity.

[0028] Specifically, step S2 includes the following steps: Determine multiple initial cluster numbers based on the number of air source heat pump operation data information; specifically, calculate the square root corresponding to the number, and determine an integer not greater than the square root as the initial cluster number. For example, the number is 1000, and since the square root of 1000 is approximately 31.62, the initial cluster number is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31; Respectively obtain high-dimensional feature coordinate points of climate feature information corresponding to each of the air source heat pump operation data information; specifically, respectively input each of the climate feature information into a preset high-dimensional feature extraction model to obtain high-dimensional feature coordinate points corresponding to each of the climate feature information; wherein the high-dimensional feature extraction model is a pre-trained deep learning model, and the high-dimensional feature extraction model includes an input layer, a feature extraction layer, a high-dimensional feature point generation layer and an output layer; Calculate the distance between each of the high-dimensional feature coordinate points and the Euclidean origin respectively; it can be understood that the Euclidean origin is the coordinate origin in the Euclidean space (0, 0, 0, 0, 0 ...); Arranging each of the high-dimensional feature coordinate points in sequence based on the distance values ​​corresponding to each of the high-dimensional feature coordinate points to obtain a high-dimensional feature coordinate point sequence; wherein the distance values ​​corresponding to the high-dimensional feature coordinate point sequence increase in sequence; For each of the initial clustering numbers, determining a clustering error corresponding to the initial clustering number based on the high-dimensional feature coordinate point sequence; Determining a target number of clusters based on each of the clustering errors; Based on the target cluster quantity and each of the high-dimensional feature coordinate points, the acquired multiple groups of air source heat pump operation data information are classified to obtain multiple data sets.

[0029] Furthermore, the determining the clustering error corresponding to the initial clustering quantity based on the high-dimensional feature coordinate point sequence comprises the following steps: For each of the initial cluster numbers, determining whether the initial cluster number is 1; If so, determine a center point in the high-dimensional feature coordinate point sequence; when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is an odd number, the center point is a high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point sequence; when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is an even number, the center point is a center point between two high-dimensional feature coordinate points located in the middle of the high-dimensional feature coordinate point sequence; Calculating the square of the Euclidean distance between each high-dimensional feature coordinate point in the high-dimensional feature coordinate point sequence and the center point, and taking the sum of each square value as the clustering error when the initial cluster number is 1; If not, segmenting the high-dimensional feature coordinate point sequence based on the initial cluster number to obtain a plurality of high-dimensional feature coordinate point subsequences; the number of the high-dimensional feature coordinate point subsequences is consistent with the initial cluster number; For each of the high-dimensional feature coordinate point subsequences, a center point is determined in the high-dimensional feature coordinate point subsequence, and the square value of the Euclidean distance between each high-dimensional feature coordinate point in the high-dimensional feature coordinate point subsequence and the center point is calculated respectively, and the sum of each of the square values ​​is used as the intermediate clustering error corresponding to the high-dimensional feature coordinate point subsequence; wherein, when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point subsequence is an odd number, the center point is a high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point subsequence, and when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point subsequence is an even number, the center point is a center point between two high-dimensional feature coordinate points located in the middle of the high-dimensional feature coordinate point subsequence; The sum of the intermediate clustering errors is taken as the clustering error corresponding to the initial number of clusters.

[0030] Further, the determining the target number of clusters based on each of the clustering errors comprises the following steps: With the initial number of clusters as the horizontal axis and the clustering error as the vertical axis, a mapping curve between the initial number of clusters and the clustering error is drawn; The elbow rule is used to determine the target number of clusters on the mapping relationship curve; wherein the target number of clusters is one of the initial numbers of clusters, and after the coordinate point corresponding to the target number of clusters, the speed of reducing the clustering error will be significantly slowed down.

[0031] Understandably, when the number of clusters is too small, the trained model will be too simple and unable to accurately predict the air source load in different areas. When the number of clusters is too large, the complexity of the model will increase, leading to overfitting. Using the elbow rule to determine the target number of clusters can help prevent these two situations and improve the effectiveness of the model.

[0032] Furthermore, the method of classifying the acquired multiple groups of air source heat pump operation data information based on the target cluster quantity and each of the high-dimensional feature coordinate points to obtain multiple data sets includes the following steps: Based on the target cluster number, a plurality of initial cluster center points are randomly selected from each of the high-dimensional feature coordinate points; wherein the number of the initial cluster center points is consistent with the target cluster number; For each of the high-dimensional feature coordinate points, the Euclidean distance between the high-dimensional feature coordinate point and each of the initial cluster center points is calculated respectively, and the cluster corresponding to the initial cluster center point corresponding to the shortest Euclidean distance is determined as the initial cluster corresponding to the high-dimensional feature coordinate point; wherein each of the initial clusters includes a plurality of high-dimensional feature coordinate points; For each of the initial clusters, a first intermediate cluster center point is determined among each of the high-dimensional feature coordinate points in the initial cluster; the first intermediate cluster center point is the cluster center of the initial cluster; specifically, the distance values ​​between each of the high-dimensional feature coordinate points and the Euclidean origin are calculated respectively, and each of the high-dimensional feature coordinate points is arranged in sequence based on the size of the distance value corresponding to each of the high-dimensional feature coordinate points to obtain a high-dimensional feature coordinate point sequence, when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is a cardinality, the high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point sequence is determined as the cluster center, and when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is an even number, the high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point sequence that is closer to the Euclidean origin is determined as the cluster center; For each of the high-dimensional feature coordinate points, the Euclidean distance between the high-dimensional feature coordinate point and each of the first intermediate cluster center points is calculated respectively, and the cluster corresponding to the first intermediate cluster center point corresponding to the shortest Euclidean distance is determined to be the first intermediate cluster corresponding to the high-dimensional feature coordinate point; wherein each of the first intermediate clusters includes multiple high-dimensional feature coordinate points; For each of the first intermediate clusters, determining a second intermediate cluster center point in each high-dimensional feature coordinate point in the first intermediate cluster; the second intermediate cluster center point is the cluster center of the first intermediate cluster; Iterate the steps after determining the first intermediate cluster center point in each high-dimensional feature coordinate point in each initial cluster for each initial cluster until multiple target clusters are obtained; wherein each target cluster converges; wherein each target cluster converges means that the steps after determining the first intermediate cluster center point in each high-dimensional feature coordinate point in each initial cluster for each target cluster are performed on each target cluster, and each target cluster remains unchanged; For each of the target clusters, the air source heat pump operation data information corresponding to each high-dimensional feature coordinate point in the target cluster is combined to obtain a data set corresponding to the target cluster.

[0033] The above method for classifying multiple groups of air source heat pump operation data information uses Euclidean distance sorting and iterative clustering process, which not only improves the classification efficiency but also ensures the classification quality, which helps to improve the effectiveness of the modeling method.

[0034] Step S3: training the preset neural network model based on each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is the environmental parameters, and the output information is the air source heat pump load.

[0035] Specifically, for each of the data sets, the environmental parameters in the data set are used as the input of the neural network model, and the air source heat pump load corresponding to the environmental parameters is used as the output of the neural network model to train the neural network model.

[0036] Step S4: Fusing the model parameters of each of the local load prediction models to obtain a fused model parameter, and using the fused model parameter to replace the model parameter of any of the local load prediction models to obtain a global load prediction model.

[0037] Specifically, the model parameters of each local load prediction model are fused by using a weighted average method to obtain fused model parameters.

[0038] Step S5: determining a prediction loss function of the global load prediction model based on the operation data information of each of the air source heat pumps, and optimizing the global load prediction model based on the prediction loss function to obtain a target load prediction model.

[0039] Specifically, determining the prediction loss function of the global load prediction model based on the operation data information of each of the air source heat pumps includes the following steps: Inputting each environmental parameter in all the air source data information into the global load prediction model respectively to obtain the air source heat pump predicted load corresponding to each environmental parameter; For each of the environmental parameters, calculating the difference between the predicted load of the air source heat pump corresponding to the environmental parameter and the load of the air source heat pump; specifically, subtracting the load of the air source heat pump from the predicted load of the air source heat pump to obtain the difference; Constructing an environmental parameter-difference mapping relationship table based on the difference values ​​corresponding to each of the environmental parameters; The prediction loss function is generated based on the environmental parameter-difference mapping relationship table.

[0040] The method provided in this embodiment, on the one hand, first constructs each of the local load prediction models, then fuses the model parameters of each of the local load prediction models to obtain fused model parameters, and uses the fused model parameters to replace the model parameters of any of the local load prediction models to obtain a global load prediction model. The global load prediction model trained in this way not only retains the advantages of the local model, but also helps to achieve global management of the air source heat pump load in each area. On the other hand, the prediction loss function of the global load prediction model is determined based on the operating data information of each of the air source heat pumps, and the global load prediction model is optimized based on the prediction loss function to obtain a target load prediction model, which helps to further improve the prediction accuracy and versatility of the model.

[0041] See also Figure 2 , Figure 2 The structure schematic block diagram of the air source heat pump load aggregation modeling system 100 provided in the embodiment of the present application is as follows: Figure 2 As shown, the air source heat pump load aggregation modeling system 100 includes: The acquisition module 110 is used to acquire air source heat pump operation data information of several regions; the air source heat pump operation data information includes climate characteristic information and a mapping relationship table between air source heat pump load and environmental parameters.

[0042] The classification module 120 is used to classify the multiple groups of air source heat pump operation data information obtained based on a preset classification algorithm to obtain multiple data sets; wherein each of the data sets includes multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data set is greater than the preset similarity.

[0043] The training module 130 is used to train the preset neural network model based on each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is the environmental parameters, and the output information is the air source heat pump load.

[0044] The parameter fusion module 140 is used to fuse the model parameters of each of the local load prediction models to obtain fused model parameters, and use the fused model parameters to replace the model parameters of any of the local load prediction models to obtain a global load prediction model.

[0045] The model optimization module 150 determines the prediction loss function of the global load prediction model based on the operation data information of each of the air source heat pumps, and optimizes the global load prediction model based on the prediction loss function to obtain a target load prediction model.

[0046] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned air source heat pump load aggregation modeling method embodiment, and will not be repeated here.

[0047] The air source heat pump load aggregation modeling system 100 provided in the above embodiment can be implemented in the form of a computer program. The computer program can be used in Figure 3 The system is run on the terminal device 200 shown.

[0048] See also Figure 3 , Figure 3 The present invention provides a schematic block diagram of the structure of a terminal device 200 according to an embodiment of the present application. The terminal device 200 includes a processor 201 and a memory 202. The processor 201 and the memory 202 are connected via a device bus 203, wherein the memory 202 may include a non-volatile storage medium and an internal memory.

[0049] The non-volatile storage medium may store a computer program. The computer program includes program instructions, and when the program instructions are executed by the processor 201, the processor 201 may execute any of the above-mentioned air source heat pump load aggregation modeling methods.

[0050] The processor 201 is used to provide computing and control capabilities to support the operation of the entire terminal device 200.

[0051] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 201, the processor 201 can execute any of the above-mentioned air source heat pump load aggregation modeling methods.

[0052] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the terminal device 200 involved in the scheme of the present application. The specific terminal device 200 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0053] It should be understood that the processor 201 may be a central processing unit (CPU), and the processor 201 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0054] In some embodiments, the processor 201 is used to run a computer program stored in the memory to implement the following steps: Acquire air source heat pump operation data information of several regions; the air source heat pump operation data information includes climate characteristic information, and a mapping relationship table between air source heat pump load and environmental parameters; Based on a preset classification algorithm, the multiple sets of air source heat pump operation data information obtained are classified to obtain multiple data sets; wherein each of the data sets includes multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data set is greater than a preset similarity; The preset neural network model is trained based on each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is the environmental parameter, and the output information is the air source heat pump load; The model parameters of each of the local load forecasting models are integrated to obtain integrated model parameters, and the model parameters of any of the local load forecasting models are replaced with the integrated model parameters to obtain a global load forecasting model; The prediction loss function of the global load prediction model is determined based on the operation data information of each of the air source heat pumps, and the global load prediction model is optimized based on the prediction loss function to obtain a target load prediction model.

[0055] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the terminal device 200 described above can refer to the corresponding process of the aforementioned air source heat pump load aggregation modeling method, and will not be repeated here.

[0056] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by one or more processors, the one or more processors implement the air source heat pump load aggregation modeling method provided in the embodiment of the present application.

[0057] The computer-readable storage medium may be an internal storage unit of the terminal device 200 in the aforementioned embodiment, such as a hard disk or memory of the terminal device 200. The computer-readable storage medium may also be an external storage device of the terminal device 200, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped with the terminal device 200.

[0058] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for modeling load aggregation of air source heat pumps, characterized in that: include: Acquire air source heat pump operation data information of several regions; the air source heat pump operation data information includes climate characteristic information, and a mapping relationship table between air source heat pump load and environmental parameters; Based on a preset classification algorithm, the multiple sets of air source heat pump operation data information obtained are classified to obtain multiple data sets; wherein each of the data sets includes multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data set is greater than a preset similarity; The preset neural network model is trained based on each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is the environmental parameter, and the output information is the air source heat pump load; The model parameters of each of the local load forecasting models are integrated to obtain integrated model parameters, and the model parameters of any of the local load forecasting models are replaced with the integrated model parameters to obtain a global load forecasting model; The prediction loss function of the global load prediction model is determined based on the operation data information of each of the air source heat pumps, and the global load prediction model is optimized based on the prediction loss function to obtain a target load prediction model.

2. The air source heat pump load aggregation modeling method according to claim 1 is characterized in that: The method classifies the acquired multiple groups of air source heat pump operation data information based on a preset classification algorithm to obtain multiple data sets, including: Determining a plurality of initial cluster numbers based on the amount of the air source heat pump operation data information; Respectively obtain high-dimensional feature coordinate points of climate feature information corresponding to each of the air source heat pump operation data information; Calculate the distance between each high-dimensional feature coordinate point and the Euclidean origin respectively; Arranging each of the high-dimensional feature coordinate points in sequence based on the distance values ​​corresponding to each of the high-dimensional feature coordinate points to obtain a high-dimensional feature coordinate point sequence; wherein the distance values ​​corresponding to the high-dimensional feature coordinate point sequence increase in sequence; For each of the initial clustering numbers, determining a clustering error corresponding to the initial clustering number based on the high-dimensional feature coordinate point sequence; Determining a target number of clusters based on each of the clustering errors; Based on the target cluster quantity and each of the high-dimensional feature coordinate points, the acquired multiple groups of air source heat pump operation data information are classified to obtain multiple data sets.

3. The air source heat pump load aggregation modeling method according to claim 2 is characterized in that: The determining the clustering error corresponding to the initial number of clusters based on the high-dimensional feature coordinate point sequence includes: For each of the initial cluster numbers, determining whether the initial cluster number is 1; If so, determine a center point in the high-dimensional feature coordinate point sequence; when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is an odd number, the center point is a high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point sequence; when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point sequence is an even number, the center point is a center point between two high-dimensional feature coordinate points located in the middle of the high-dimensional feature coordinate point sequence; Calculating the square of the Euclidean distance between each high-dimensional feature coordinate point in the high-dimensional feature coordinate point sequence and the center point respectively, and taking the sum of each of the square values ​​as the clustering error when the initial cluster number is 1; If not, segmenting the high-dimensional feature coordinate point sequence based on the initial cluster number to obtain a plurality of high-dimensional feature coordinate point subsequences; the number of the high-dimensional feature coordinate point subsequences is consistent with the initial cluster number; For each of the high-dimensional feature coordinate point subsequences, a center point is determined in the high-dimensional feature coordinate point subsequence, and the square value of the Euclidean distance between each high-dimensional feature coordinate point in the high-dimensional feature coordinate point subsequence and the center point is calculated respectively, and the sum of each of the square values ​​is used as the intermediate clustering error corresponding to the high-dimensional feature coordinate point subsequence; wherein, when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point subsequence is an odd number, the center point is a high-dimensional feature coordinate point located in the middle of the high-dimensional feature coordinate point subsequence, and when the number of high-dimensional feature coordinate points in the high-dimensional feature coordinate point subsequence is an even number, the center point is a center point between two high-dimensional feature coordinate points located in the middle of the high-dimensional feature coordinate point subsequence; The sum of the intermediate clustering errors is taken as the clustering error corresponding to the initial number of clusters.

4. The air source heat pump load aggregation modeling method according to claim 2 is characterized in that: The determining the target number of clusters based on each of the clustering errors comprises: With the initial number of clusters as the horizontal axis and the clustering error as the vertical axis, a mapping curve between the initial number of clusters and the clustering error is drawn; The elbow rule is used to determine the target cluster quantity on the mapping relationship curve.

5. The air source heat pump load aggregation modeling method according to claim 2 is characterized in that: The multiple sets of air source heat pump operation data information obtained are classified based on the target cluster quantity and each of the high-dimensional feature coordinate points to obtain multiple data sets, including: Based on the target cluster number, a plurality of initial cluster center points are randomly selected from each of the high-dimensional feature coordinate points; wherein the number of the initial cluster center points is consistent with the target cluster number; For each of the high-dimensional feature coordinate points, the Euclidean distance between the high-dimensional feature coordinate point and each of the initial cluster center points is calculated respectively, and the cluster corresponding to the initial cluster center point corresponding to the shortest Euclidean distance is determined as the initial cluster corresponding to the high-dimensional feature coordinate point; wherein each of the initial clusters includes a plurality of high-dimensional feature coordinate points; For each of the initial clusters, a first intermediate cluster center point is determined among the high-dimensional feature coordinate points in the initial cluster; the first intermediate cluster center point is the cluster center of the initial cluster; For each of the high-dimensional feature coordinate points, the Euclidean distance between the high-dimensional feature coordinate point and each of the first intermediate cluster center points is calculated respectively, and the cluster corresponding to the first intermediate cluster center point corresponding to the shortest Euclidean distance is determined to be the first intermediate cluster corresponding to the high-dimensional feature coordinate point; wherein each of the first intermediate clusters includes multiple high-dimensional feature coordinate points; For each of the first intermediate clusters, determining a second intermediate cluster center point in each high-dimensional feature coordinate point in the first intermediate cluster; the second intermediate cluster center point is the cluster center of the first intermediate cluster; Iterate the steps after determining the first intermediate cluster center point in each high-dimensional feature coordinate point in each initial cluster for each initial cluster until a plurality of target clusters are obtained; wherein each target cluster converges; For each of the target clusters, the air source heat pump operation data information corresponding to each high-dimensional feature coordinate point in the target cluster is combined to obtain a data set corresponding to the target cluster.

6. The air source heat pump load aggregation modeling method according to claim 1 is characterized in that: The method of determining the prediction loss function of the global load prediction model based on the operation data information of each of the air source heat pumps includes: Inputting each environmental parameter in all the air source data information into the global load prediction model respectively to obtain the air source heat pump predicted load corresponding to each environmental parameter; For each of the environmental parameters, calculating the difference between the predicted load of the air source heat pump corresponding to the environmental parameter and the load of the air source heat pump; Constructing an environmental parameter-difference mapping relationship table based on the difference values ​​corresponding to each of the environmental parameters; The prediction loss function is generated based on the environmental parameter-difference mapping relationship table.

7. An air source heat pump load aggregation modeling system, characterized in that: include: An acquisition module is used to obtain the operating data information of air source heat pumps in several areas; The air source heat pump operation data information includes climate characteristic information, a mapping relationship table between the air source heat pump load and environmental parameters; A classification module, for classifying the acquired multiple groups of air source heat pump operation data information based on a preset classification algorithm to obtain multiple data sets; wherein each of the data sets includes multiple air source heat pump operation data information, and for each of the data sets, the similarity between the climate characteristic information corresponding to any two air source heat pump operation data information in the data set is greater than a preset similarity; A training module, used to train a preset neural network model based on each of the data sets to obtain a local load prediction model corresponding to each of the data sets; wherein the input information of the local load prediction model is environmental parameters, and the output information is the air source heat pump load; A parameter fusion module, used to fuse the model parameters of each of the local load forecasting models to obtain a fusion model parameter, and use the fusion model parameter to replace the model parameter of any of the local load forecasting models to obtain a global load forecasting model; The model optimization module determines the prediction loss function of the global load prediction model based on the operation data information of each of the air source heat pumps, and optimizes the global load prediction model based on the prediction loss function to obtain a target load prediction model.

8. A terminal device, characterized in that: It comprises a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the air source heat pump load aggregation modeling method as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the air source heat pump load aggregation modeling method according to any one of claims 1 to 6 is implemented.

Images