CN-121980487-A - Electric load adjustment potential evaluation method, device, medium and equipment based on multi-model fusion

CN121980487ACN 121980487 ACN121980487 ACN 121980487ACN-121980487-A

Abstract

The application discloses an electric load adjustment potential evaluation method, device, medium and equipment based on multi-model fusion, wherein the method comprises the steps of acquiring electric load data to be predicted of owners aiming at each of a plurality of owners, and constructing electric load characteristics to be predicted based on the electric load data to be predicted, wherein the electric load characteristics to be predicted comprise electric load characteristics and time period characteristics; the method comprises the steps of inputting electric load characteristics to be predicted into a multi-model fusion model, respectively carrying out electric load prediction on the electric load to be predicted through each electric load prediction model in the multi-model fusion model, fusing the predicted results of the electric load prediction models to obtain electric load prediction fusion values, respectively determining the maximum interruptible power, the transferable electric quantity and the power adjustment range of a proprietor based on the electric load prediction fusion values, and determining the comprehensive electric load adjustment potential of the proprietor based on the maximum interruptible power, the transferable electric quantity and the power adjustment range.

Inventors

LV RAN
QIAN YANYUAN
GUO MINGXING
WANG SU
LAN LI
FU CHEN
ZHANG YONG

Assignees

国网上海市电力公司

Dates

Publication Date: 20260505
Application Date: 20251203

Claims (11)

1. An electrical load regulation potential assessment method based on multi-model fusion, which is characterized by comprising the following steps: Aiming at each of a plurality of owners, acquiring electric load data to be predicted of the owners, and constructing electric load characteristics to be predicted based on the electric load data to be predicted, wherein the electric load characteristics to be predicted comprise electric load characteristics and time period characteristics; inputting the electric load characteristics to be predicted into a multi-model fusion model, respectively carrying out electric load prediction on the electric load characteristics to be predicted through each electric load prediction model in the multi-model fusion model, and fusing the predicted results of the electric load prediction models to obtain an electric load prediction fusion value, wherein the electric load prediction model comprises a XGBoost model, a random forest model, a support vector regression model and a generalized additive model; And respectively determining the maximum interruptible power, the transferable electric quantity and the power adjustment range of the owners based on the electric load prediction fusion value, and determining the comprehensive electric load adjustment potential of the owners.
2. The method of claim 1, wherein prior to inputting the electrical load characteristic to be predicted into a multi-model fusion model, the method further comprises: Acquiring an electric load training sample, and respectively training each electric load prediction model by utilizing the electric load training sample; And optimizing the fusion weights corresponding to the electric load prediction models through a Bayesian optimization method to obtain the fusion weights of the electric load prediction models.
3. The method of claim 2, wherein fusing the predicted results of each electrical load prediction model to obtain an electrical load prediction fusion value comprises: and fusing the predicted results of the electric load prediction models based on the fusion weights of the electric load prediction models to obtain electric load prediction fusion values.
4. A method according to claim 3, wherein the optimizing the fusion weights corresponding to the electric load prediction models by using a bayesian optimization method to obtain the fusion weights of the electric load prediction models includes: and optimizing the fusion weights corresponding to the electric load prediction models by using an electric load verification sample through a Bayesian optimization method to obtain the fusion weights of the electric load prediction models by taking the prediction mean square error of the multi-model fusion models as a target.
5. The method of claim 1, wherein after obtaining the electrical load prediction fusion value, the method further comprises: Grouping the electric load prediction fusion values based on a preset threshold value, and compressing the electric load prediction fusion values based on the average value of the electric load prediction fusion values in the group so as to update the electric load prediction fusion values based on a compression result.
6. The method of claim 1, wherein determining the maximum interruptible power, the transferable amount of power, the power adjustment range, and the integrated electrical load adjustment potential of the owner based on the maximum interruptible power, the transferable amount of power, and the power adjustment range, respectively, based on the electrical load prediction fusion values, comprises: Determining a maximum interruptible power for the owner based on a product of a preset interrupt scaling factor and a user base load power corresponding to the electrical load prediction fusion value; determining the transferable electric quantity of the owner based on the product of a preset electric quantity transfer proportion coefficient and the daily electric quantity of the user corresponding to the electric load prediction fusion value; The method comprises the steps of determining an interruptible potential index of a proprietor according to maximum interruptible power, preset maximum interruptible power optimal value and preset maximum interruptible power worst value of the proprietor, determining a transferable potential index of the proprietor according to transferable electric quantity of the proprietor, preset transferable electric quantity optimal value and preset transferable electric quantity worst value of the proprietor, determining an adjustable potential index of the proprietor according to power adjusting range, preset power adjusting range optimal value and preset power adjusting range worst value of the proprietor, carrying out fusion calculation on the interruptible potential index, the transferable potential index and the adjustable potential index of the proprietor, and determining comprehensive electric load adjusting potential of the proprietor.
7. The method of claim 6, wherein the method further comprises: Acquiring a corresponding owner range of the virtual power plant; and determining the owner adjustable potential of each owner according to the sum of the interruptible potential index, the transferable potential index and the adjustable potential index of each owner in the owner range, and determining the total adjustable potential corresponding to the virtual power plant based on the sum of the owner adjustable potentials of all owners in the owner range.
8. The method of claim 7, wherein the method further comprises: carrying out power consumption typical mode clustering on each owner according to the power load prediction fusion value of each owner; Clustering the interruptible potential indexes, the transferable potential indexes and the adjustable potential indexes of owners in the electricity consumption typical mode class respectively aiming at each electricity consumption typical mode class to obtain a plurality of interruptible potential index classes, a plurality of transferable potential index classes and a plurality of adjustable potential index classes, and classifying the owners in the electricity consumption typical mode class into different levels of interruptible potential owners, different levels of transferable potential owners and different levels of adjustable potential owners based on each cluster; And carrying out power utilization scheduling on each owner according to the real-time load characteristics corresponding to the virtual power plant and the interruptible potential grade, the transferable potential grade and the adjustable potential grade of each owner in the range of the owners.
9. An electrical load regulation potential assessment device based on multimodal fusion, the device comprising: The characteristic construction module is used for acquiring electric load data to be predicted of each of a plurality of owners, and constructing electric load characteristics to be predicted based on the electric load data to be predicted, wherein the electric load characteristics to be predicted comprise electric load characteristics and time period characteristics; The load prediction module is used for inputting the electric load characteristics to be predicted into a multi-model fusion model, respectively carrying out electric load prediction on the electric load characteristics to be predicted through each electric load prediction model in the multi-model fusion model, and fusing the predicted results of the electric load prediction models to obtain an electric load prediction fusion value, wherein the electric load prediction model comprises a XGBoost model, a random forest model, a support vector regression model and a generalized additive model; And the potential evaluation module is used for respectively determining the maximum interruptible power, the transferable electric quantity and the power adjustment range of the owner based on the electric load prediction fusion value, and determining the comprehensive electric load adjustment potential of the owner based on the maximum interruptible power, the transferable electric quantity and the power adjustment range.
10. A storage medium having stored thereon a computer program, which when executed by a processor, implements the method of any of claims 1 to 8.
11. 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 the processor implements the method of any one of claims 1 to 8 when executing the computer program.

Description

Electric load adjustment potential evaluation method, device, medium and equipment based on multi-model fusion Technical Field The application relates to the technical field of electric power, in particular to an electric load adjustment potential evaluation method, an electric load adjustment potential evaluation device, a storage medium and computer equipment based on multi-model fusion. Background In the field of electricity utilization in China, load management is developed in large users, and is an important development direction for long-term support and promotion. The load management has a key effect of not neglecting balance of power supply and demand, improvement of power grid operation efficiency and guarantee of power safety and stable supply. Through reasonable regulation and control of the power consumption load of large users, the power shortage in the power consumption peak period can be effectively avoided, the occurrence of the condition of switching off and limiting electricity is reduced, the utilization efficiency of power resources is improved, and the energy waste is reduced. In the operation management of an electric power system, accurate assessment of electric load regulation potential is important, whether the electric power system can stably and efficiently operate or not is related, and reasonable distribution and effective utilization of energy sources are affected. With the continuous development of the electric power market and the wide application of the emerging technologies such as distributed energy sources, smart grids and the like, the complexity and uncertainty of the electric load are greatly improved, and a plurality of difficulties are brought to the evaluation work of the electric load adjustment potential. How to meet the requirements of modern power systems on the evaluation of the potential of electrical load regulation is an important topic in the art. Disclosure of Invention In view of the above, the embodiment of the application provides an electrical load adjustment potential evaluation method, an electrical load adjustment potential evaluation device, a storage medium and computer equipment based on multi-model fusion. According to one aspect of the present application, there is provided a method for evaluating electrical load regulation potential based on multi-model fusion, the method comprising: aiming at each of a plurality of owners, acquiring electric load data to be predicted of the owners, and constructing electric load characteristics to be predicted based on the electric load data to be predicted, wherein the electric load characteristics to be predicted comprise electric load characteristics and time period characteristics; inputting the electric load characteristics to be predicted into a multi-model fusion model, respectively carrying out electric load prediction on the electric load characteristics to be predicted through each electric load prediction model in the multi-model fusion model, and fusing the predicted results of the electric load prediction models to obtain an electric load prediction fusion value, wherein the electric load prediction model comprises a XGBoost model, a random forest model, a support vector regression model and a generalized additive model; And respectively determining the maximum interruptible power, the transferable electric quantity and the power adjustment range of the owner based on the electric load prediction fusion value, and determining the comprehensive electric load adjustment potential of the owner based on the maximum interruptible power, the transferable electric quantity and the power adjustment range. Optionally, before inputting the electrical load characteristic to be predicted into the multi-model fusion model, the method further comprises: Acquiring an electric load training sample, and respectively training each electric load prediction model by utilizing the electric load training sample; optimizing the fusion weights corresponding to the electric load prediction models through a Bayesian optimization method to obtain the fusion weights of the electric load prediction models; correspondingly, fusion is carried out on the predicted result of each electric load prediction model to obtain an electric load prediction fusion value, which comprises the following steps: and fusing the predicted results of the electric load prediction models based on the fusion weights of the electric load prediction models to obtain electric load prediction fusion values. Optionally, the optimizing the fusion weight corresponding to each electrical load prediction model by using a bayesian optimization method to obtain the fusion weight of each electrical load prediction model includes: and optimizing the fusion weights corresponding to the electric load prediction models by using an electric load verification sample through a Bayesian optimization method to obtain the fusion weights of the electric load prediction mo