CN-121983943-A - Power calculation and power cooperation scale prediction method, system, equipment and medium

CN121983943ACN 121983943 ACN121983943 ACN 121983943ACN-121983943-A

Abstract

The invention discloses a method, a system, equipment and a medium for predicting a computational power collaborative scale, wherein the method comprises the steps of obtaining multi-source data, preprocessing, carrying out data fusion and scene decoupling on the preprocessed multi-source data to obtain scene classification labels, generating a computational power scene decoupling modeling structure by using the scene classification labels, carrying out multidimensional parameter association analysis, respectively carrying out different computational power scale predictions, integrating all computational power scale prediction results to generate a multidimensional computational power scale prediction model, carrying out weight distribution on an area total computational power scale to obtain a preliminary distribution scheme, checking the preliminary distribution scheme to obtain a computational power scale space-time distribution prediction map, constructing a dynamic energy consumption model base, carrying out power demand calculation to obtain a power demand space-time distribution prediction result, carrying out comprehensive evaluation and decision packaging on the multidimensional computational power scale prediction model and the power demand space-time distribution prediction result to generate a standardized evaluation file, and realizing the prediction of the computational power collaborative scale.

Inventors

LIU GUANGSHUO
ZHAO JINGZHI
LIU YAN
ZHAO LIN
YE XIAOCHEN
CHENG MENGZENG

Assignees

国网辽宁省电力有限公司经济技术研究院

Dates

Publication Date: 20260505
Application Date: 20251125

Claims (10)

1. The power calculation and power cooperation scale prediction method is characterized by comprising the following steps of: acquiring multi-source data, preprocessing the multi-source data, performing data fusion and scene decoupling on the preprocessed multi-source data to obtain scene classification labels, and generating a computational field scene decoupling modeling structure by using the scene classification labels; Based on the computational power scene decoupling modeling structure, carrying out multidimensional parameter association analysis, respectively carrying out different computational power scale predictions, integrating all the computational power scale prediction results, and generating a multidimensional computational power scale prediction model; performing weight distribution on the total computational power scale of the region output by the multi-dimensional computational power scale prediction model to obtain a preliminary distribution scheme, and checking the preliminary distribution scheme to obtain a computational power scale space-time distribution prediction map; constructing a dynamic energy consumption model library based on the calculation power scale space-time distribution prediction map, and calculating the power demand to obtain a power demand space-time distribution prediction result; And carrying out comprehensive evaluation and decision packaging on the multi-dimensional calculation power scale prediction model and the power demand space-time distribution prediction result to generate a standardized evaluation file, thereby realizing the prediction of calculation power and power cooperation scale.
2. The method for collaborative power scale prediction according to claim 1, wherein performing data fusion and scene decoupling on the preprocessed multi-source data comprises: carrying out standardization processing on the multi-source data, and carrying out feature extraction on the standardized multi-source data to obtain load feature parameters; and inputting the load characteristic parameters into a pre-trained machine learning classification model to obtain scene classification labels.
3. The method of claim 2, wherein generating a computational power scene decoupling modeling structure using the scene classification tags comprises: Mapping the labeled power calculation facilities as load nodes to corresponding connection points of the power network to construct an integrated coupling diagram model; And according to the scene classification labels, dividing the integrated coupling graph model by using a first community finding algorithm to obtain the computational power scene decoupling modeling structure.
4. A method of collaborative power scale prediction according to claim 3 wherein performing different power scale predictions comprises: acquiring historical data of general calculation force and a planning target, and constructing a macroscopic growth curve to obtain a macroscopic predicted value; Acquiring regional calculation force characteristics, determining a regional duty ratio coefficient, and obtaining a general calculation force predicted value according to the product of the macroscopic predicted value and the regional duty ratio coefficient; Acquiring intelligent power scale data based on regional history, and fitting by using a nonlinear function to obtain a first prediction result; Establishing a multiple linear regression model of the regional intelligent computing power scale and a plurality of key associated industrial scales, and solving regression coefficients; substituting future predicted values of each industry into a multiple linear regression model to obtain a second predicted result; Fusing the first prediction result and the second prediction result to obtain an intelligent calculation power prediction value; And summarizing and packaging the general calculation force predicted value and the intelligent calculation force predicted value to obtain the multi-dimensional calculation force scale prediction model.
5. The method of claim 4, wherein obtaining a power-scale spatiotemporal distribution prediction map comprises: Acquiring an area development index of each geographic unit, setting a weight factor, and carrying out weighted synthesis on the weight factor to obtain comprehensive allocation weight of each geographic unit; the total computational power scale of the area output by the multidimensional computational power scale prediction model is proportionally distributed according to the comprehensive distribution weight of each geographic unit, and a preliminary distribution scheme is obtained; And verifying the preliminary scheme based on a verification rule to obtain a calculation power scale space-time distribution prediction map.
6. The method for collaborative scale prediction according to claim 5, wherein obtaining a power demand space-time distribution prediction result comprises: respectively constructing a general power calculation energy consumption model and an intelligent power calculation energy consumption model; Respectively taking the calculation power scale space-time distribution prediction map as the input of a general calculation power energy consumption model and an intelligent calculation power energy consumption model, and outputting corresponding energy consumption parameters; Calculating the power consumption of each facility according to the energy consumption parameters, and aggregating to generate a basic power demand matrix of each geographic unit; and carrying out load curve synthesis and uncertainty analysis on the basic power demand matrix to generate a power demand space-time distribution prediction result.
7. The method of claim 4 or 6, wherein performing the integrated assessment and decision-making package comprises: Performing association mapping on the multidimensional calculation power scale prediction model and the power demand space-time distribution prediction result, and obtaining a comprehensive evaluation decision matrix through matrix operation; Extracting evaluation parameters of the comprehensive evaluation decision matrix, and simulating different decision preferences by utilizing a multi-objective optimization algorithm to obtain an optimal strategy basis and a priority sequence; And packaging the optimal strategy basis and the priority sequence, and performing encryption coding to obtain a standardized evaluation file.
8. A power-to-power collaborative scale prediction system employing a power-to-power collaborative scale prediction method according to any one of claims 1-7, comprising: the scene decoupling modeling module is used for acquiring multi-source data, preprocessing the multi-source data, carrying out data fusion and scene decoupling on the preprocessed multi-source data to obtain scene classification labels, and generating a computational field scene decoupling modeling structure by utilizing the scene classification labels; the multi-mode prediction engine module is used for carrying out multi-dimensional parameter association analysis based on the computational power scene decoupling modeling structure, respectively carrying out different computational power scale predictions, integrating the prediction results of all computational power scales and generating a multi-dimensional computational power scale prediction model; The space-time distribution module is used for carrying out weight distribution on the total computational power scale of the area output by the multi-dimensional computational power scale prediction model to obtain a preliminary distribution scheme, and checking the preliminary distribution scheme to obtain a computational power scale space-time distribution prediction map; The dynamic energy consumption coupling module is used for constructing a dynamic energy consumption model library based on the calculation power scale space-time distribution prediction map and carrying out power demand calculation to obtain a power demand space-time distribution prediction result; and the evaluation file generation module is used for carrying out comprehensive evaluation and decision packaging on the multi-dimensional calculation power scale prediction model and the power demand space-time distribution prediction result to generate a standardized evaluation file, so as to realize the prediction of calculation power and power cooperation scale.
9. A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps of a method for collaborative power scale prediction according to any one of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium, characterized in that a computer program is stored thereon, which computer program, when being executed by a processor, implements the steps of a method for collaborative power scale prediction according to any of claims 1-7.

Description

Power calculation and power cooperation scale prediction method, system, equipment and medium Technical Field The invention relates to the technical field of intersection of digital infrastructure and power system planning, in particular to a method, a system, equipment and a medium for predicting a coordinated power and electricity scale. Background Currently, collaborative planning of computing infrastructure and power systems has now become a key challenge. In the aspect of computational load prediction, the existing technical scheme has the common problems of high scene coupling, single prediction dimension, coarse space-time granularity, insufficient consideration of technical evolution factors and the like. The traditional power load prediction method is difficult to adapt to the new characteristics of high increase, high density and strong technical driving performance of the power load, and the pure power industry planning lacks accurate mapping to power requirements, so that the source-network-load planning is disjointed. The existing method is mainly simple in converting by adopting static unit power consumption indexes, cannot construct a dynamic prediction model which is from the essence of the power calculation service and penetrates through a policy-technology-service-energy consumption full chain, and cannot meet the prospective planning requirement of the collaborative development of the power calculation and the power. Therefore, the invention provides a method, a system, equipment and a medium for collaborative scale prediction of computational power to solve the problems, and the method realizes accurate, fine and prospective prediction of future computational power scale and corresponding power requirements thereof by decoupling heterogeneous computational power scenes, fusing multi-source data, constructing a collaborative prediction model and introducing a dynamic energy consumption coupling mechanism. Disclosure of Invention In view of the above existing problems, the present invention provides a method, a system, a device and a medium for collaborative scale prediction of power and computing. The invention provides a computational power and electric power collaborative scale prediction method, a system, equipment and a medium, which solve the problems of high coupling, single dimension, coarse space-time granularity, lack of dynamic evolution mechanism and the like of the traditional computational power load prediction method. In order to solve the technical problems, the invention provides the following technical scheme: In a first aspect, the present invention provides a method for collaborative scale prediction of power of computing, comprising: acquiring multi-source data, preprocessing the multi-source data, performing data fusion and scene decoupling on the preprocessed multi-source data to obtain scene classification labels, and generating a computational field scene decoupling modeling structure by using the scene classification labels; Based on the computational power scene decoupling modeling structure, carrying out multidimensional parameter association analysis, respectively carrying out different computational power scale predictions, integrating all the computational power scale prediction results, and generating a multidimensional computational power scale prediction model; performing weight distribution on the total computational power scale of the region output by the multi-dimensional computational power scale prediction model to obtain a preliminary distribution scheme, and checking the preliminary distribution scheme to obtain a computational power scale space-time distribution prediction map; constructing a dynamic energy consumption model library based on the calculation power scale space-time distribution prediction map, and calculating the power demand to obtain a power demand space-time distribution prediction result; And carrying out comprehensive evaluation and decision packaging on the multi-dimensional calculation power scale prediction model and the power demand space-time distribution prediction result to generate a standardized evaluation file, thereby realizing the prediction of calculation power and power cooperation scale. As a preferable scheme of the power and computing cooperation scale prediction method, the method for performing data fusion and scene decoupling on the preprocessed multi-source data comprises the following steps: carrying out standardization processing on the multi-source data, and carrying out feature extraction on the standardized multi-source data to obtain load feature parameters; and inputting the load characteristic parameters into a pre-trained machine learning classification model to obtain scene classification labels. As a preferable scheme of the power calculation and power cooperation scale prediction method, the method for generating the power calculation scene decoupling modeling structure by using the scene classif