Electric quantity demand prediction method and device adapted to new electricity price reform situation, and medium

Abstract

The invention discloses an electric quantity demand prediction method and device adapted to a new electricity price reform situation, and a medium. The method comprises the steps of determining electricity price factor change information and economic factor change information of a target analysis region; acquiring historical electric quantity data and electric quantity demand influence factors of different power consumption subjects in the target analysis region; generating a power consumption curve graph of each power consumption main body, and clustering the power consumption main bodies; performing influence degree analysis on each electric quantity demand influence factor according to historical electric quantity data of each power consumption main body in the power consumption main body class, and screening out key electric quantity demand influence factors; updating the power utilization main body in each power utilization main body class according to the key electric quantity demand influence factor of each power utilization main body in each power utilization main body class; acquiring a power consumption subject class including an electricity price factor and an economic factor as a target power consumption subject class; and carrying out electric quantity demand model training according to the target power consumption main body class, and predicting the electric quantity demand of each power consumption main body in the future prediction time period.

Description

Technical Field

[0001] The present application relates to the technical field of electricity demand forecasting, and in particular to an electricity demand forecasting method, device and medium adapted to the new electricity price reform situation. Background Art

[0002] With socioeconomic development and adjustments to the energy mix, the power industry is facing unprecedented changes. This is particularly true with regard to electricity pricing reform. To promote energy conservation and emission reduction and rationalize resource allocation, governments around the world have introduced new pricing policies, such as tiered pricing and peak-valley pricing. These new pricing policies not only impact the operating costs and revenues of power companies but also directly influence user electricity consumption and demand patterns. Therefore, accurately forecasting electricity demand under these new pricing reforms has become a pressing challenge for power companies and related research institutions.

[0003] Traditional electricity demand forecasting methods primarily rely on historical electricity data, using statistical methods or machine learning algorithms to predict future electricity demand. However, these methods often overlook the impact of factors such as electricity price reforms and economic policies on user electricity usage, resulting in significant deviations in forecast results. For example, after the implementation of a new electricity price policy, some electricity users may reduce their electricity consumption due to price increases, or increase their electricity consumption due to preferential electricity price policies. Traditional forecasting methods struggle to capture these changing trends.

[0004] Furthermore, different types of electricity users (e.g., industrial, commercial, and residential) respond differently to changes in electricity prices and economic policies. Therefore, using a unified forecasting model may not accurately reflect the actual changes in electricity demand across all types of electricity users. This further complicates electricity demand forecasting. Summary of the Invention

[0005] In view of this, the embodiments of the present application provide a method, device and medium for predicting electricity demand that adapt to the new electricity price reform situation.

[0006] According to one aspect of the present application, a method for predicting power demand that adapts to the new electricity price reform situation is provided, the method comprising:

[0007] Based on the new electricity price reform situation information and economic policy impact information of the target analysis area in the future forecast period, determine the electricity price factor change information and economic factor change information of the target analysis area in the future forecast period;

[0008] Obtain historical electricity demand data and electricity demand influencing factors of different electricity users in the target analysis area within a set time period, wherein the electricity demand influencing factors include at least electricity price factors, economic factors, climate factors, and holiday factors;

[0009] Generate the power consumption curves of each electricity-consuming entity based on the historical power consumption data, and cluster each electricity-consuming entity based on the power consumption curves to obtain multiple electricity-consuming entity classes;

[0010] For each electricity-consuming entity class, analyze the influence degree of each power demand influencing factor according to the historical power consumption data of the electricity-consuming entities in the electricity-consuming entity class, and screen out the key power demand influencing factors of each electricity-consuming entity based on the preset influence degree screening conditions according to the influence degree of each power demand influencing factor;

[0011] Update the electricity-consuming entities in each electricity-consuming entity class according to the key power demand influencing factors of the electricity-consuming entities in each electricity-consuming entity class to obtain multiple updated electricity-consuming entity classes;

[0012] Obtain the electricity-consuming entity classes whose key power demand influencing factors include at least one of the electricity price factor and the economic factor as the target electricity-consuming entity classes;

[0013] Train the power demand model according to the historical power consumption data and power demand influencing factors of the electricity-consuming entities in the target electricity-consuming entity class, and predict the power demand of each electricity-consuming entity in the future prediction period based on the electricity price factor change information and the economic factor change information.

[0014] According to another aspect of the present application, there is provided a power demand prediction device adapted to the new electricity price reform situation, and the device includes:

[0015] A change information determination module, configured to determine the electricity price factor change information and the economic factor change information of the target analysis region in the future prediction period based on the new electricity price reform situation information and the economic policy influence information of the target analysis region in the future prediction period;

[0016] A historical data acquisition module, configured to acquire the historical power consumption data and power demand influencing factors of different electricity-consuming entities in the target analysis region within a set time period, and the power demand influencing factors at least include an electricity price factor, an economic factor, a climate factor, and a holiday factor;

[0017] A clustering module, configured to generate the power consumption curves of each electricity-consuming entity based on the historical power consumption data, and cluster each electricity-consuming entity based on the power consumption curves to obtain multiple electricity-consuming entity classes;

[0018] A factor analysis module, configured to, for each electricity-consuming entity class, analyze the influence degree of each power demand influencing factor according to the historical power consumption data of the electricity-consuming entities in the electricity-consuming entity class, and screen out the key power demand influencing factors of each electricity-consuming entity based on the preset influence degree screening conditions according to the influence degree of each power demand influencing factor;

[0019] An update module, configured to update the electricity consumers in each electricity consumer class according to the key electricity demand influencing factors of the electricity consumers in each electricity consumer class, so as to obtain multiple updated electricity consumer classes;

[0020] A model training module, configured to obtain an electricity consumer class whose key electricity demand influencing factors include at least one of an electricity price factor and an economic factor as a target electricity consumer class; perform electricity demand model training according to the historical electricity consumption data and electricity demand influencing factors of the electricity consumers in the target electricity consumer class, and predict the electricity demand of each electricity consumer in a future prediction period based on the electricity price factor change information and the economic factor change information.

[0021] According to another aspect of the present application, there is provided a storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned electricity demand prediction method adapted to the new electricity price reform situation is implemented.

[0022] According to still another aspect of the present application, there is provided a computer device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, and when the processor executes the program, the above-mentioned electricity demand prediction method adapted to the new electricity price reform situation is implemented.

[0023] With the above technical solution, an electricity demand forecasting method, device and medium provided by an embodiment of the present application are based on the new electricity price reform situation information and economic policy impact information of the target analysis area in the future forecast period to determine the electricity price factor change information and economic factor change information of the target analysis area in the future forecast period; obtain the historical electricity consumption data and electricity demand influencing factors of different electricity consumers in the target analysis area within a set time period, and the electricity demand influencing factors at least include electricity price factors, economic factors, climate factors, and holiday factors; generate an electricity consumption curve graph for each electricity consumer based on the historical electricity consumption data, and cluster each electricity consumer based on the electricity consumption curve graph to obtain multiple electricity consumer classes; for each electricity consumer class, analyze the influence degree of each electricity demand influencing factor based on the historical electricity consumption data of the electricity consumers in the electricity consumer class, and screen out the key electricity demand influencing factors of each electricity consumer based on the preset influence degree screening conditions according to the influence degree of each electricity demand influencing factor; update the electricity consumers in each electricity consumer class according to the key electricity demand influencing factors of the electricity consumers in each electricity consumer class to obtain multiple updated electricity consumer classes; obtain the electricity consumer class whose key electricity demand influencing factors include at least one of electricity price factors and economic factors as the target electricity consumer class; perform electricity demand model training based on the historical electricity consumption data and electricity demand influencing factors of the electricity consumers in the target electricity consumer class, and predict the electricity demand of each electricity consumer in the future forecast period based on the electricity price factor change information and the economic factor change information. By carefully analyzing the electricity demand influencing factors of different electricity consumers, especially electricity price and economic factors, and clustering analysis based on these factors, the embodiment of the present application can more accurately predict future electricity demand and reduce prediction errors. This method can adapt to the new electricity price reform situation and economic policy changes, and maintain the timeliness and accuracy of the prediction model by continuously updating the electricity consumer classes and key electricity demand influencing factors. Further, it helps power system operators, energy suppliers and policymakers better plan and allocate power resources, improve energy utilization efficiency, reduce waste, and can better balance power supply and demand, reduce power shortages or surpluses, and promote the sustainable development of the power industry.

[0024] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the specific embodiments of the present application are given below. Brief Description of the Drawings

[0025] The accompanying drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0026] Figure 1 A schematic flow chart of a method for predicting electricity demand that adapts to the new electricity price reform situation provided by an embodiment of the present application is shown;

[0027] Figure 2 A schematic flow chart of another method for predicting electricity demand that adapts to the new electricity price reform situation provided by an embodiment of the present application is shown;

[0028] Figure 3 A schematic structural diagram of a device for predicting electricity demand that adapts to the new electricity price reform situation provided by an embodiment of the present application is shown. Detailed implementation manners

[0029] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0030] In this embodiment, a method for predicting electricity demand that adapts to the new electricity price reform situation is provided. As Figure 1 shown, the method includes:

[0031] Step 101: Based on the new electricity price reform situation information and economic policy impact information of the target analysis area in the future prediction period, determine the electricity price factor change information and economic factor change information of the target analysis area in the future prediction period.

[0032] Step 102: Obtain the historical electricity consumption data and electricity demand influencing factors of different electricity consumers in the target analysis area within a set time period. The electricity demand influencing factors at least include electricity price factors, economic factors, climate factors, and holiday factors.

[0033] Step 103: Generate an electricity consumption curve graph for each electricity consumer based on the historical electricity consumption data, and cluster each electricity consumer based on the electricity consumption curve graph to obtain multiple electricity consumer classes.

[0034] Step 104: For each electricity consumer class, analyze the influence degree of each electricity demand influencing factor according to the historical electricity consumption data of the electricity consumers in the electricity consumer class, and screen out the key electricity demand influencing factors of each electricity consumer based on the preset influence degree screening conditions according to the influence degree of each electricity demand influencing factor.

[0035] Step 105: Update the electricity consumers in each electricity consumer category according to the key electricity demand influencing factors of the electricity consumers in each electricity consumer category, so as to obtain multiple updated electricity consumer categories.

[0036] Step 106: Obtain the electricity consumer categories whose key electricity demand influencing factors include at least one of the electricity price factor and the economic factor as the target electricity consumer categories.

[0037] Step 107: Train the electricity demand model according to the historical electricity consumption data and the electricity demand influencing factors of the electricity consumers in the target electricity consumer category, and predict the electricity demand of each electricity consumer in the future prediction period based on the electricity price factor change information and the economic factor change information.

[0038] In the embodiments of the present application, information on new electricity price reform situations (such as changes in time-of-use electricity prices, tiered electricity prices, etc.) and economic policy impact information (such as economic growth rate, industrial structure adjustment, etc.) that the target area may experience during the future prediction period is collected and analyzed. Historical electricity consumption data of different electricity consumers (such as residents, industrial users, commercial users, etc.) in the target area within a set time period (such as the past few years) is collected, as well as various factors affecting these electricity demands, including electricity prices, economy, climate (such as temperature, humidity), holidays, etc. These data are the basis for subsequent analysis and modeling. By generating electricity consumption curves for each electricity consumer, the electricity consumption trends of different electricity consumers can be visually seen. Based on these curves, cluster analysis is performed on the electricity consumers, and electricity consumers with similar electricity consumption trends are grouped into one category for more accurate analysis of the influencing factors of the electricity demand of each category of electricity consumers in the future. For each category of electricity consumers, an impact degree analysis is carried out according to its historical electricity consumption data and various electricity demand influencing factors. Specifically, statistical analysis and machine learning techniques can be used to determine which factors have a significant impact on the electricity demand. Then, according to the preset impact degree screening conditions (such as the impact coefficient being greater than a certain threshold), the key electricity demand influencing factors of each electricity consumer are screened out. Based on the screened key electricity demand influencing factors, the category of electricity consumers is updated. The purpose of this step is to further refine the category of electricity consumers to more accurately reflect the electricity demand characteristics of different electricity consumers. From the updated category of electricity consumers, those categories of electricity consumers whose key electricity demand influencing factors include at least one of the electricity price factor and the economic factor are selected as the target categories of electricity consumers. These categories of electricity consumers are more likely to be affected by the new electricity price reform and economic policy changes, so they are the focus of the prediction. The historical electricity consumption data and electricity demand influencing factors of each electricity consumer in the target category of electricity consumers are used to train the electricity demand model. This model should be able to capture the impact of changes in electricity price and economic factors on the electricity demand. Among them, the electricity demand model can adopt a time series model. After training is completed, based on the collected information on changes in electricity price factors and economic factors, the electricity demand for the future prediction period is predicted.

[0039] By applying the technical solution of this embodiment, by carefully analyzing the influencing factors of the electricity demand of different electricity consumers, especially the electricity price and economic factors, and the cluster analysis based on these factors, the future electricity demand can be predicted more accurately, and the prediction error can be reduced. This method can adapt to the new electricity price reform situation and economic policy changes, and maintain the timeliness and accuracy of the prediction model by continuously updating the category of electricity consumers and the key electricity demand influencing factors. It further helps power system operators, energy suppliers, and policymakers better plan and allocate power resources, improve energy utilization efficiency, reduce waste, and can better balance power supply and demand, reduce power shortages or surpluses, and promote the sustainable development of the power industry.

[0040] In an embodiment of the present application, optionally, step 105 includes:

[0041] Step 105-1: For each type of electricity-consuming entity, based on the key electricity demand influencing factors and their influencing degrees of each electricity-consuming entity in the type of electricity-consuming entity, determine the category key electricity demand influencing factors and their influencing degrees corresponding to the type of electricity-consuming entity. Based on the key electricity demand influencing factors and their influencing degrees of each electricity-consuming entity and the category key electricity demand influencing factors and their influencing degrees of the type of electricity-consuming entity, determine the electricity-consuming entities among each electricity-consuming entity whose key electricity demand influencing factors are different from the category key electricity demand influencing factors, and the electricity-consuming entities whose deviation between the influencing degree of the key electricity demand influencing factor and the influencing degree of the category key electricity demand influencing factor is greater than a preset first deviation, and jointly use them as the electricity-consuming entities to be classified;

[0042] Step 105-2: For each electricity-consuming entity to be classified, determine each candidate electricity-consuming entity type in which the category key electricity demand influencing factor is the same as the key electricity demand influencing factor of the electricity-consuming entity to be classified among each type of electricity-consuming entity, and calculate the influencing degree deviation between the electricity-consuming entity to be classified and each candidate electricity-consuming entity type based on the influencing degree of the key electricity demand influencing factor of the electricity-consuming entity to be classified and the influencing degree of the category key electricity demand influencing factor of each candidate electricity-consuming entity type, and classify the electricity-consuming entity to be classified into the type of electricity-consuming entity based on the influencing degree deviation.

[0043] In this embodiment, for each type of electricity-consuming entity, first, based on the key electricity consumption demand influencing factors and their influencing degrees of each electricity-consuming entity within it, the key electricity consumption demand influencing factors and their influencing degrees of the entire type of electricity-consuming entity are determined. This is an aggregation process aimed at refining the common characteristics of this type of electricity-consuming entity. Next, the differences between the key electricity consumption demand influencing factors and their influencing degrees of each electricity-consuming entity and those of the key electricity consumption demand influencing factors and their influencing degrees of the type are compared. Specifically, electricity-consuming entities are sought whose key electricity consumption demand influencing factors are different from those of the key electricity consumption demand influencing factors of the type, or whose deviation between the influencing degrees of the key electricity consumption demand influencing factors and those of the key electricity consumption demand influencing factors of the type is greater than a preset first deviation. These electricity-consuming entities are regarded as electricity-consuming entities to be classified due to their uniqueness, and they may need to be reclassified into a more appropriate type of electricity-consuming entity. Further, for each electricity-consuming entity to be classified, candidate types of electricity-consuming entities whose key electricity consumption demand influencing factors are the same as those of the electricity-consuming entity to be classified are searched for among all types of electricity-consuming entities. This is a matching process aimed at finding the type of electricity-consuming entity that is closest to the electricity-consuming entity to be classified in terms of key electricity consumption demand influencing factors. Then, the deviation between the influencing degree of the key electricity consumption demand influencing factors of the electricity-consuming entity to be classified and the influencing degrees of the key electricity consumption demand influencing factors of each candidate type of electricity-consuming entity is calculated, that is, the influencing degree deviation. This deviation value reflects the similarity degree between the electricity-consuming entity to be classified and the candidate type of electricity-consuming entity. Finally, the electricity-consuming entity to be classified is classified based on the influencing degree deviation. Specifically, the electricity-consuming entity to be classified can be classified into the candidate type of electricity-consuming entity with the smallest influencing degree deviation. In this way, the update of the types of electricity-consuming entities is completed, making each type of electricity-consuming entity more accurately reflect the electricity consumption demand characteristics of the electricity-consuming entities within it. By comparing the differences between the key electricity consumption demand influencing factors and their influencing degrees of each electricity-consuming entity and those of the key electricity consumption demand influencing factors and their influencing degrees of the type, the embodiments of the present application can identify those electricity-consuming entities that do not conform to the type characteristics and reclassify them, which helps to more accurately divide the types of electricity-consuming entities and improve the accuracy of subsequent electricity consumption demand prediction. Through continuous update and optimization of the types of electricity-consuming entities, it can be ensured that the electricity consumption demand prediction model can handle more diverse electricity-consuming entity characteristics, so as to enhance the generalization ability of the model and enable it to better adapt to electricity consumption demand prediction tasks in different scenarios.

[0044] In an embodiment of the present application, optionally, the classification of the electricity-consuming entity to be classified into an electricity-consuming entity class based on the impact degree deviation in step 105-2 includes: if the minimum deviation in the impact degree deviation is less than a preset second deviation, then classify the electricity-consuming entity to be classified into the candidate electricity-consuming entity class corresponding to the minimum deviation; wherein, the preset second deviation is less than the preset first deviation; if the minimum deviation in the impact degree deviation is greater than or equal to the preset second deviation, then return the electricity-consuming entity to be classified to the electricity-consuming entity class where it originally belonged.

[0045] In this embodiment, a preset second deviation is set in advance, and this deviation value is less than the preset first deviation. The setting of the preset second deviation is to provide a more stringent classification standard to ensure that the similarity between the electricity-consuming entity to be classified and the candidate electricity-consuming entity class reaches a certain level. In the classification decision, first check whether the minimum deviation in the impact degree deviation is less than the preset second deviation. If the minimum deviation is less than the preset second deviation, it indicates that the similarity between the electricity-consuming entity to be classified and the candidate electricity-consuming entity class corresponding to the minimum deviation is very high. Therefore, the electricity-consuming entity to be classified can be classified into this candidate electricity-consuming entity class. However, if the minimum deviation is greater than or equal to the preset second deviation, it means that the similarity between the electricity-consuming entity to be classified and all candidate electricity-consuming entity classes is not sufficient to meet the classification standard. In this case, to avoid classification errors, the electricity-consuming entity to be classified is selected to be returned to the electricity-consuming entity class where it originally belonged. This can maintain the stability and consistency of the electricity-consuming entity class, and at the same time reduce the error of electricity demand prediction caused by inaccurate classification. By setting the preset second deviation as the classification standard in the embodiment of the present application, the similarity degree between the electricity-consuming entity to be classified and the candidate electricity-consuming entity class can be more strictly controlled, which helps to improve the classification accuracy and ensure that the electricity-consuming entity to be classified is correctly classified into the most suitable electricity-consuming entity class. When the similarity between the electricity-consuming entity to be classified and all candidate electricity-consuming entity classes is not sufficient to meet the classification standard, the strategy of returning it to the original class can reduce the error of electricity demand prediction caused by classification errors, maintain the stability and consistency of the electricity-consuming entity class, contribute to the training and prediction of the electricity demand prediction model in subsequent steps, and improve the generalization ability and prediction accuracy of the model.

[0046] In an embodiment of the present application, optionally, as Figure 2 shown, the training of the electricity demand model according to the historical electricity consumption data and electricity demand influencing factors of the target electricity-consuming entity class in step 107 includes:

[0047] Step 107-1, obtain the electricity-consuming entity classes other than the target electricity-consuming entity class in the electricity-consuming entity class as the reference electricity-consuming entity classes;

[0048] Step 107-2: Construct an initial electricity demand model based on the influencing factors of electricity demand, and copy the initial electricity demand model based on the total number of the target electricity-consuming entity classes and the reference electricity-consuming entity classes to obtain multiple initial electricity demand models that match the total number, and respectively determine the initial electricity demand models that match each target electricity-consuming entity class and each reference electricity-consuming entity class;

[0049] Step 107-3: Train the initial electricity demand model corresponding to each target electricity-consuming entity class respectively according to the historical electricity data and the influencing factors of electricity demand of each electricity-consuming entity in each target electricity-consuming entity class, and train the initial electricity demand model corresponding to each reference electricity-consuming entity class respectively according to the historical electricity data and the influencing factors of electricity demand of each electricity-consuming entity in each reference electricity-consuming entity class;

[0050] Step 107-4: Optimize the model parameters of the initial electricity demand model corresponding to the target electricity-consuming entity class based on the model parameters of the initial electricity demand model corresponding to the reference electricity-consuming entity class to obtain the electricity demand models corresponding to each target electricity-consuming entity class.

[0051] In this embodiment, an innovative method for training an electricity demand model is proposed. This method combines the data of target electricity-consuming entity classes and reference electricity-consuming entity classes. By replicating, training, and optimizing the initial electricity demand model, accurate electricity demand models corresponding to each target electricity-consuming entity class are finally obtained. First, electricity-consuming entity classes other than the target electricity-consuming entity classes are selected from all electricity-consuming entity classes as reference electricity-consuming entity classes. These reference electricity-consuming entity classes provide electricity demand characteristics and influencing factors different from those of the target electricity-consuming entity classes, which helps introduce more diversity and generalization ability in the model training process. Next, an initial electricity demand model is constructed according to electricity demand influencing factors (such as electricity price, economy, climate, holidays, etc.). To process the data of multiple electricity-consuming entity classes, the initial electricity demand model is replicated according to the total number of target electricity-consuming entity classes and reference electricity-consuming entity classes to obtain multiple initial electricity demand models that match the total number. Each electricity-consuming entity class (whether target or reference) is assigned a matching initial electricity demand model. Then, the historical electricity data and electricity demand influencing factors in each electricity-consuming entity class are used to train the corresponding initial electricity demand model respectively. This step aims to enable the model to learn the relationship between the specific electricity demand characteristics and influencing factors of each electricity-consuming entity class and the electricity demand. Finally, based on the model parameters of the initial electricity demand model corresponding to the reference electricity-consuming entity class, the model parameters of the initial electricity demand model corresponding to the target electricity-consuming entity class are optimized. This step utilizes the idea of transfer learning to improve the model of the target electricity-consuming entity class by learning the knowledge from the reference electricity-consuming entity class. The optimized model can better capture the electricity demand characteristics of the target electricity-consuming entity class and improve the prediction accuracy. By introducing reference electricity-consuming entity classes in the embodiment of this application, more diversity and generalization ability can be introduced in the model training process, which helps the model perform better when dealing with unseen data. By replicating, training, and optimizing multiple initial electricity demand models, we can obtain an accurate electricity demand model for each electricity-consuming entity class (including the target electricity-consuming entity class). These models can better capture the specific electricity demand characteristics and influencing factors of each electricity-consuming entity class, thereby improving the prediction accuracy.

[0052] In an embodiment of the present application, optionally, step 107-3 includes: when training the initial power demand model corresponding to each target electricity-consuming entity class, determining the non-critical power demand influencing factors corresponding to the target electricity-consuming entity class according to the power demand influencing factors and the key power demand influencing factors corresponding to the target electricity-consuming entity class, freezing the model parameters corresponding to the non-critical power demand influencing factors, and training the initial power demand model based on the historical power data and power demand influencing factors of each electricity-consuming entity in the target electricity-consuming entity class to optimize the model parameters corresponding to the key power demand influencing factors; when training the initial power demand model corresponding to each reference electricity-consuming entity class, not freezing any model parameters, and training the initial power demand model according to the historical power data and power demand influencing factors of each electricity-consuming entity in the reference electricity-consuming entity class to optimize all model parameters of the initial power demand model, so as to obtain the power demand model corresponding to the reference electricity-consuming entity class.

[0053] In this embodiment, a detailed explanation of the refined implementation of step 107-3 and the derivation of beneficial effects

[0054] Step 107-3 Refinement Implementation: First, based on the electricity demand influencing factors and the key electricity demand influencing factors corresponding to the target electricity-consuming entity class, determine the non-critical electricity demand influencing factors corresponding to the target electricity-consuming entity class. These non-critical factors may have less or unstable influence on electricity demand, so they do not need to be overly concerned in model training. When training the initial electricity demand model corresponding to the target electricity-consuming entity class, freeze the model parameters corresponding to the non-critical electricity demand influencing factors. This means that these parameters will not be updated during the training process, thereby reducing the complexity and computational amount of model training. Then, based on the historical electricity data and electricity demand influencing factors (especially the key electricity demand influencing factors) of each electricity-consuming entity in the target electricity-consuming entity class, train the initial electricity demand model. During the training process, only the model parameters corresponding to the key electricity demand influencing factors will be optimized to improve the prediction accuracy of the model for the electricity demand of the target electricity-consuming entity class. For the initial electricity demand model corresponding to the reference electricity-consuming entity class, a more comprehensive training method is adopted. Specifically, when training the initial electricity demand model corresponding to the reference electricity-consuming entity class, do not freeze any model parameters, and all parameters will be updated during the training process to fully utilize the diversity and generalization ability provided by the reference electricity-consuming entity class. Then, based on the historical electricity data and electricity demand influencing factors of each electricity-consuming entity in the reference electricity-consuming entity class, train the initial electricity demand model. During the training process, all model parameters will be optimized to obtain a more accurate electricity demand model. In the embodiment of the present application, by freezing the model parameters of the non-critical electricity demand influencing factors corresponding to the target electricity-consuming entity class, the complexity and computational amount of model training are reduced, thereby improving the model training efficiency. By focusing on optimizing the model parameters corresponding to the key electricity demand influencing factors, the electricity demand characteristics of the target electricity-consuming entity class can be more accurately captured, thereby improving the model prediction accuracy. By comprehensively training the initial electricity demand model corresponding to the reference electricity-consuming entity class, the diversity and generalization ability provided by the reference electricity-consuming entity class are fully utilized, enhancing the generalization ability of the model.

[0055] In an embodiment of the present application, optionally, optimizing the model parameters of the initial power demand model corresponding to the target electricity-consuming entity class based on the model parameters of the initial power demand model corresponding to the reference electricity-consuming entity class to obtain the power demand models corresponding to each target electricity-consuming entity class includes: for each target electricity-consuming entity class, extracting the model parameters corresponding to the non-critical power demand influencing factors from the power demand models corresponding to each reference electricity-consuming entity class, and determining the fusion weights of the model parameters corresponding to each non-critical power demand influencing factor according to the influence degree of the non-critical power demand influencing factors in the target electricity-consuming entity class and the influence degree of each non-critical power demand influencing factor in the reference electricity-consuming entity class, fusing the model parameters corresponding to each non-critical power demand influencing factor based on the fusion weights to obtain the model parameters corresponding to each non-critical power demand influencing factor, and setting the model parameters corresponding to each non-critical power demand influencing factor in the initial power demand model corresponding to the target electricity-consuming entity class; performing a second round of training on the set initial power demand model based on the historical power data and power demand influencing factors corresponding to each electricity-consuming entity in the target electricity-consuming entity class to obtain the power demand model corresponding to the target electricity-consuming entity class.

[0056] In this embodiment, for each target electricity-consuming entity class, first, model parameters corresponding to non-critical electricity demand influencing factors are extracted from the electricity demand models corresponding to each reference electricity-consuming entity class. These parameters reflect the influence of non-critical factors on electricity demand in the reference electricity-consuming entity class. Next, according to the influence degree of non-critical electricity demand influencing factors in the target electricity-consuming entity class and their influence degree in the reference electricity-consuming entity class, the fusion weights corresponding to the model parameters of each non-critical electricity demand influencing factor are determined. This step aims to ensure that in the fusion process, non-critical factors with a greater impact on the target electricity-consuming entity class can obtain greater weights. Based on the determined fusion weights, the model parameters of each non-critical electricity demand influencing factor are fused to obtain the optimized model parameters corresponding to each non-critical electricity demand influencing factor in the target electricity-consuming entity class. Then, the model parameters corresponding to the fused non-critical electricity demand influencing factors are set in the initial electricity demand model corresponding to the target electricity-consuming entity class. Finally, based on the historical electricity data and electricity demand influencing factors corresponding to each electricity-consuming entity in the target electricity-consuming entity class, the set initial electricity demand model is trained in a second round, and all parameters of the model are iterated, so as to further optimize the model parameters and improve the prediction accuracy of the model for the electricity demand of the target electricity-consuming entity class. In the embodiment of the present application, by fusing the model parameters of non-critical electricity demand influencing factors in the reference electricity-consuming entity class and determining the fusion weights in combination with the influence degree, more accurate model parameters can be provided for the target electricity-consuming entity class, thereby improving the prediction accuracy of the model. Using the data of the reference electricity-consuming entity class to optimize the model parameters of the target electricity-consuming entity class helps to introduce more diversity and generalization ability, enabling the model to perform better when dealing with unseen data. By guiding the model parameter optimization process through the fusion weights, the optimal combination of model parameters can be found more quickly, thus accelerating the process of model optimization.

[0057] In the embodiment of the present application, optionally, the predicting the electricity demand of each electricity-consuming entity in the future prediction period based on the electricity price factor change information and the economic factor change information in step 107 includes: determining the electricity demand model corresponding to each electricity-consuming entity, and obtaining the prediction information of each electricity demand influencing factor in the target analysis area in the future prediction period, where the electricity price factor change information is used as the prediction information of the electricity price factor, and the economic factor change information is used as the prediction information of the economic factor; predicting the electricity demand of each electricity-consuming entity in the future prediction period through the corresponding electricity demand model based on the historical electricity data of each electricity-consuming entity and the prediction information of each electricity demand influencing factor.

[0058] In this embodiment, the electricity demand models corresponding to each electricity-consuming entity are determined, and the prediction information of each electricity demand influencing factor in the target analysis area during the future prediction period is obtained. This prediction information can be obtained through methods such as time series analysis, regression analysis, and expert prediction. Among them, the prediction information of the electricity price factor uses the electricity price factor change information in the above text, and the prediction information of the economic factor uses the economic factor change information in the above text. Finally, based on the historical electricity consumption data of each electricity-consuming entity and the prediction information of each electricity demand influencing factor, the electricity demand of each electricity-consuming entity in the future prediction period is predicted through the corresponding electricity demand model. The prediction information is input into the electricity demand model, and the model is run to obtain the prediction result. The electricity price and economic factors are important factors affecting electricity demand, and their changes can often be directly reflected in electricity demand. By combining the prediction information of the electricity price factor and the economic factor in this embodiment of the application, various factors affecting electricity demand can be considered more comprehensively, thereby improving the accuracy of the prediction.

[0059] By applying the technical solution of this embodiment, through the refined construction and training of the electricity demand model, combined with the prediction of the electricity price factor change information and the economic factor change information, and the differential processing of the target electricity-consuming entity class and the reference electricity-consuming entity class, including the distinction between key and non-key electricity demand influencing factors, the freezing and optimization of model parameters, and the parameter fusion strategy based on the fusion weight, this technical solution significantly improves the accuracy of electricity demand prediction, enhances the generalization ability and adaptability of the model, provides more accurate decision-making support for power system operators, promotes the optimal allocation and efficient utilization of power resources, and at the same time promotes the sustainable development of energy, reduces energy waste, and improves energy utilization efficiency.

[0060] Furthermore, as Figure 1 a specific implementation of the method, this embodiment of the application provides an electricity demand prediction device adapted to the new electricity price reform situation, as Figure 3 shown. The device includes:

[0061] A change information determination module, configured to determine the electricity price factor change information and the economic factor change information in the target analysis area during the future prediction period based on the new electricity price reform situation information and the economic policy impact information in the target analysis area during the future prediction period;

[0062] A historical data acquisition module, configured to acquire the historical electricity consumption data and the electricity demand influencing factors of different electricity-consuming entities in the target analysis area during a set time period, where the electricity demand influencing factors at least include an electricity price factor, an economic factor, a climate factor, and a holiday factor;

[0063] A clustering module, configured to generate power consumption curves for each power consumption entity based on the historical power consumption data, and cluster each power consumption entity based on the power consumption curves to obtain multiple power consumption entity classes;

[0064] A factor analysis module, configured to, for each power consumption entity class, analyze the influence degree of each power demand influencing factor according to the historical power consumption data of each power consumption entity in the power consumption entity class, and screen out the key power demand influencing factors of each power consumption entity based on the preset influence degree screening condition according to the influence degree of each power demand influencing factor;

[0065] An update module, configured to update the power consumption entities in each power consumption entity class according to the key power demand influencing factors of each power consumption entity in each power consumption entity class, so as to obtain multiple updated power consumption entity classes;

[0066] A model training module, configured to obtain a power consumption entity class whose key power demand influencing factors include at least one of a power price factor and an economic factor as a target power consumption entity class; train a power demand model according to the historical power consumption data and power demand influencing factors of each power consumption entity in the target power consumption entity class, and predict the power demand of each power consumption entity in a future prediction period based on the power price factor change information and the economic factor change information.

[0067] In an embodiment of the present application, optionally, the update module is further configured to:

[0068] For each power consumption entity class, determine the class key power demand influencing factors and influence degree corresponding to the power consumption entity class according to the key power demand influencing factors and influence degree of each power consumption entity in the power consumption entity class, and determine, according to the key power demand influencing factors and influence degree of each power consumption entity and the class key power demand influencing factors and influence degree of the power consumption entity class, the power consumption entities in each power consumption entity whose key power demand influencing factors are different from the class key power demand influencing factors, and the power consumption entities whose deviation between the influence degree of the key power demand influencing factors and the influence degree of the class key power demand influencing factors is greater than a preset first deviation, and jointly use them as power consumption entities to be classified;

[0069] For each power consumption entity to be classified, determine each candidate power consumption entity class in which the class key power demand influencing factors are the same as the key power demand influencing factors of the power consumption entity to be classified in each power consumption entity class, and calculate the influence degree deviation between the power consumption entity to be classified and each candidate power consumption entity class according to the influence degree of the key power demand influencing factors of the power consumption entity to be classified and the influence degree of the class key power demand influencing factors of each candidate power consumption entity class, and classify the power consumption entity to be classified into a power consumption entity class based on the influence degree deviation.

[0070] In an embodiment of the present application, optionally, the update module is further configured to:

[0071] If the minimum deviation among the influence degree deviations is less than a preset second deviation, classify the power consumption entity to be classified into the candidate power consumption entity class corresponding to the minimum deviation; wherein, the preset second deviation is less than the preset first deviation;

[0072] If the minimum deviation among the influence degree deviations is greater than or equal to the preset second deviation, replay the power consumption entity to be classified to the power consumption entity class where the power consumption entity to be classified originally belonged.

[0073] In an embodiment of the present application, optionally, the model training module is further configured to:

[0074] Obtain the power consumption entity classes other than the target power consumption entity class in the power consumption entity class as the reference power consumption entity classes;

[0075] Construct an initial power demand model according to the power demand influencing factors, and based on the total number of the target power consumption entity classes and the reference power consumption entity classes, copy the initial power demand model to obtain a plurality of initial power demand models matching the total number, and respectively determine the initial power demand models matching each target power consumption entity class and each reference power consumption entity class;

[0076] Train the initial power demand model corresponding to each target power consumption entity class according to the historical power data and power demand influencing factors of each power consumption entity in each target power consumption entity class, and train the initial power demand model corresponding to each reference power consumption entity class according to the historical power data and power demand influencing factors of each power consumption entity in each reference power consumption entity class;

[0077] Optimize the model parameters of the initial power demand model corresponding to the target power consumption entity class based on the model parameters of the initial power demand model corresponding to the reference power consumption entity class to obtain the power demand models corresponding to each target power consumption entity class.

[0078] In an embodiment of the present application, optionally, the model training module is further configured to:

[0079] When training the initial power demand model corresponding to each target power consumption entity class, determine the non-critical power demand influencing factors corresponding to the target power consumption entity class according to the power demand influencing factors and the key power demand influencing factors corresponding to the target power consumption entity class, freeze the model parameters corresponding to the non-critical power demand influencing factors, and train the initial power demand model based on the historical power data and power demand influencing factors of each power consumption entity in the target power consumption entity class to optimize the model parameters corresponding to the key power demand influencing factors;

[0080] When training the initial power demand model corresponding to each reference electricity-consuming entity class, no model parameters are frozen. According to the historical power consumption data and power demand influencing factors of each electricity-consuming entity in the reference electricity-consuming entity class, the initial power demand model is trained to optimize all model parameters of the initial power demand model, and the power demand model corresponding to the reference electricity-consuming entity class is obtained.

[0081] In an embodiment of the present application, optionally, the model training module is further configured to:

[0082] For each target electricity-consuming entity class, extract the model parameters corresponding to the non-critical power demand influencing factors from the power demand models corresponding to each reference electricity-consuming entity class, and determine the fusion weights of the model parameters corresponding to each non-critical power demand influencing factor according to the influence degree of the non-critical power demand influencing factors in the target electricity-consuming entity class and the influence degree of each non-critical power demand influencing factor in the reference electricity-consuming entity class. Based on the fusion weights, fuse the model parameters of each non-critical power demand influencing factor to obtain the model parameters corresponding to each non-critical power demand influencing factor, and set the model parameters corresponding to each non-critical power demand influencing factor in the initial power demand model corresponding to the target electricity-consuming entity class;

[0083] Based on the historical power consumption data and power demand influencing factors corresponding to each electricity-consuming entity in the target electricity-consuming entity class, perform a second round of training on the set initial power demand model to obtain the power demand model corresponding to the target electricity-consuming entity class.

[0084] In an embodiment of the present application, optionally, the model training module is further configured to:

[0085] Determine the power demand models corresponding to each electricity-consuming entity, and obtain the prediction information of each power demand influencing factor in the future prediction period of the target analysis area, where the change information of the electricity price factor is used as the prediction information of the electricity price factor, and the change information of the economic factor is used as the prediction information of the economic factor;

[0086] Based on the historical power consumption data of each electricity-consuming entity and the prediction information of each power demand influencing factor, predict the power demand of each electricity-consuming entity in the future prediction period through the corresponding power demand model.

[0087] It should be noted that for other corresponding descriptions of each functional unit involved in the power demand prediction device adapted to the new electricity price reform situation provided in the embodiment of the present application, reference can be made to Figures 1 to 2 the corresponding description in the method, which will not be elaborated here.

[0088] The embodiments of the present application further provide a computer device, which may specifically be a personal computer, a server, a network device, etc. The computer device includes a bus, a processor, a memory, and a communication interface, and may further include an input / output interface and a display device. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store location information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in the method embodiments are implemented.

[0089] Those skilled in the art can understand that the structure of the above computer device is only a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components, or combine some components, or have different component arrangements.

[0090] In one embodiment, a computer-readable storage medium is provided. The computer-readable storage medium may be non-volatile or volatile, and has a computer program stored thereon. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0091] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0092] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties.

[0093] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.

[0094] 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 as within the scope described in this specification.

[0095] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for predicting electricity demand that adapts to the new electricity price reform situation, characterized in that, The method includes: Based on the new electricity price reform situation information and economic policy impact information in the future prediction period of the target analysis area, determining the electricity price factor change information and economic factor change information in the future prediction period of the target analysis area; Obtaining the historical electricity consumption data and electricity demand influencing factors of different electricity users in the target analysis area within a set time period, where the electricity demand influencing factors at least include electricity price factors, economic factors, climate factors, and holiday factors; Generating an electricity consumption curve graph for each electricity user based on the historical electricity consumption data, and clustering each electricity user based on the electricity consumption curve graph to obtain multiple electricity user classes; For each electricity user class, analyzing the influence degree of each electricity demand influencing factor based on the historical electricity consumption data of the electricity users in the electricity user class, and screening out the key electricity demand influencing factors of each electricity user based on the preset influence degree screening conditions based on the influence degree of each electricity demand influencing factor; Updating the electricity users in each electricity user class according to the key electricity demand influencing factors of the electricity users in each electricity user class to obtain multiple updated electricity user classes; Obtaining the electricity user classes whose key electricity demand influencing factors include at least one of the electricity price factors and economic factors as the target electricity user classes; Training an electricity demand model based on the historical electricity consumption data and electricity demand influencing factors of the electricity users in the target electricity user classes, and predicting the electricity demand of each electricity user in the future prediction period based on the electricity price factor change information and the economic factor change information.

2. The method according to claim 1, characterized in that The step of updating the electricity users in each electricity user class according to the key electricity demand influencing factors of the electricity users in each electricity user class to obtain multiple updated electricity user classes includes: For each electricity user class, determining the category key electricity demand influencing factors and influence degree corresponding to the electricity user class according to the key electricity demand influencing factors and influence degree of the electricity users in the electricity user class, and determining the electricity users whose key electricity demand influencing factors are different from the category key electricity demand influencing factors among the electricity users, and the electricity users whose deviation between the influence degree of the key electricity demand influencing factors and the influence degree of the category key electricity demand influencing factors is greater than a preset first deviation, as the electricity users to be classified together; For each electricity user to be classified, determining the candidate electricity user classes with the same category key electricity demand influencing factors as the key electricity demand influencing factors of the electricity user to be classified among the electricity user classes, and calculating the influence degree deviation between the electricity user to be classified and each candidate electricity user class according to the influence degree of the key electricity demand influencing factors of the electricity user to be classified and the influence degree of the category key electricity demand influencing factors of each candidate electricity user class, and classifying the electricity user to be classified into an electricity user class based on the influence degree deviation.

3. The method according to claim 2, wherein Classifying the electricity consumption entity to be classified into electricity consumption entity classes based on the deviation of the influence degree includes: If the minimum deviation in the deviation of the influence degree is less than a preset second deviation, classify the electricity consumption entity to be classified into the candidate electricity consumption entity class corresponding to the minimum deviation; wherein, the preset second deviation is less than the preset first deviation; If the minimum deviation in the deviation of the influence degree is greater than or equal to the preset second deviation, return the electricity consumption entity to be classified to the electricity consumption entity class where the electricity consumption entity to be classified originally belonged.

4. The method according to claim 1, characterized in that, Training the electricity demand model according to the historical electricity consumption data and electricity demand influencing factors of the target electricity consumption entity class includes: Obtain the electricity consumption entity classes other than the target electricity consumption entity class in the electricity consumption entity class as the reference electricity consumption entity classes; Construct an initial electricity demand model according to the electricity demand influencing factors, and based on the total number of the target electricity consumption entity class and the reference electricity consumption entity classes, copy the initial electricity demand model to obtain multiple initial electricity demand models matching the total number, and respectively determine the initial electricity demand models matching each target electricity consumption entity class and each reference electricity consumption entity class; Train the initial electricity demand model corresponding to each target electricity consumption entity class according to the historical electricity consumption data and electricity demand influencing factors of each electricity consumption entity in each target electricity consumption entity class, and train the initial electricity demand model corresponding to each reference electricity consumption entity class according to the historical electricity consumption data and electricity demand influencing factors of each electricity consumption entity in each reference electricity consumption entity class; Optimize the model parameters of the initial electricity demand model corresponding to the target electricity consumption entity class based on the model parameters of the initial electricity demand model corresponding to the reference electricity consumption entity class to obtain the electricity demand models corresponding to each target electricity consumption entity class.

5. The method according to claim 4, wherein Training the initial electricity demand model corresponding to each target electricity consumption entity class according to the historical electricity consumption data and electricity demand influencing factors of each electricity consumption entity in each target electricity consumption entity class, and training the initial electricity demand model corresponding to each reference electricity consumption entity class according to the historical electricity consumption data and electricity demand influencing factors of each electricity consumption entity in each reference electricity consumption entity class includes: When training the initial electricity demand model corresponding to each target electricity consumption entity class, determine the non-critical electricity demand influencing factors corresponding to the target electricity consumption entity class according to the electricity demand influencing factors and the key electricity demand influencing factors corresponding to the target electricity consumption entity class, freeze the model parameters corresponding to the non-critical electricity demand influencing factors, and train the initial electricity demand model based on the historical electricity consumption data and electricity demand influencing factors of each electricity consumption entity in the target electricity consumption entity class to optimize the model parameters corresponding to the key electricity demand influencing factors; When training the initial electricity demand model corresponding to each reference electricity-consuming entity class, no model parameters are frozen. Based on the historical electricity consumption data and electricity demand influencing factors of each electricity-consuming entity in the reference electricity-consuming entity class, the initial electricity demand model is trained to optimize all model parameters of the initial electricity demand model, and the electricity demand model corresponding to the reference electricity-consuming entity class is obtained.

6. The method according to claim 5, characterized in that, Optimizing the model parameters of the initial electricity demand model corresponding to the target electricity-consuming entity class based on the model parameters of the initial electricity demand model corresponding to the reference electricity-consuming entity class to obtain the electricity demand model corresponding to each target electricity-consuming entity class includes: For each target electricity-consuming entity class, extract the model parameters corresponding to the non-critical electricity demand influencing factors from the electricity demand models corresponding to each reference electricity-consuming entity class, and determine the fusion weights of the model parameters corresponding to each non-critical electricity demand influencing factor according to the influence degree of the non-critical electricity demand influencing factors in the target electricity-consuming entity class and the influence degree of each non-critical electricity demand influencing factor in the reference electricity-consuming entity class. Based on the fusion weights, fuse the model parameters of each non-critical electricity demand influencing factor to obtain the model parameters corresponding to each non-critical electricity demand influencing factor, and set the model parameters corresponding to each non-critical electricity demand influencing factor in the initial electricity demand model corresponding to the target electricity-consuming entity class; Based on the historical electricity consumption data and electricity demand influencing factors of each electricity-consuming entity in the target electricity-consuming entity class, conduct a second round of training on the set initial electricity demand model to obtain the electricity demand model corresponding to the target electricity-consuming entity class.

7. The method according to any one of claims 4 to 6, characterized in that, Predicting the electricity demand of each electricity-consuming entity in the future prediction period based on the electricity price factor change information and the economic factor change information includes: Determine the electricity demand model corresponding to each electricity-consuming entity, and obtain the prediction information of each electricity demand influencing factor in the target analysis area in the future prediction period. Among them, use the electricity price factor change information as the prediction information of the electricity price factor, and use the economic factor change information as the prediction information of the economic factor; Based on the historical electricity consumption data of each electricity-consuming entity and the prediction information of each electricity demand influencing factor, predict the electricity demand of each electricity-consuming entity in the future prediction period through the corresponding electricity demand model.

8. An electricity demand forecasting device adapted to the new electricity price reform situation, characterized in that, The device includes: A change information determination module for determining the electricity price factor change information and economic factor change information of the target analysis area in the future prediction period based on the new electricity price reform situation information and economic policy impact information of the target analysis area in the future prediction period; A historical data acquisition module for acquiring the historical electricity consumption data and electricity demand influencing factors of different electricity-consuming entities in the target analysis area within a set time period, where the electricity demand influencing factors at least include electricity price factors, economic factors, climate factors, and holiday factors; A clustering module for generating an electricity consumption curve graph of each electricity-consuming entity based on the historical electricity consumption data, and clustering each electricity-consuming entity based on the electricity consumption curve graph to obtain multiple electricity-consuming entity classes; A factor analysis module is configured to, for each type of electricity-consuming entity, analyze the influence degree of each electricity demand influencing factor based on the historical electricity consumption data of the electricity-consuming entities in the type of electricity-consuming entity, and screen out the key electricity demand influencing factors of each electricity-consuming entity based on the preset influence degree screening conditions for the influence degrees of the electricity demand influencing factors; An update module is configured to update the electricity-consuming entities in each type of electricity-consuming entity according to the key electricity demand influencing factors of the electricity-consuming entities in each type of electricity-consuming entity, so as to obtain multiple updated types of electricity-consuming entities; A model training module is configured to obtain a type of electricity-consuming entity whose key electricity demand influencing factors include at least one of a electricity price factor and an economic factor as a target type of electricity-consuming entity; perform electricity demand model training according to the historical electricity consumption data and electricity demand influencing factors of the electricity-consuming entities in the target type of electricity-consuming entity, and predict the electricity demand of each electricity-consuming entity in a future prediction period based on the electricity price factor change information and the economic factor change information.

9. A storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by a processor, the method described in any one of claims 1 to 7 is implemented.

10. A computer device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, the method described in any one of claims 1 to 7 is implemented.

Images