Research And Application Of Asset Management Level Evaluation Method For Power Grid

Power grid assets are an important part of state-owned assets.With the deepening of power system reform,the formulation of transmission and distribution price under the new situation is directly linked to the power grid assets,which undoubtedly puts forward higher requirements for the power grid assets management level.How to scientifically evaluate the power grid assets management level of power grid enterprises,find out the shortcomings of asset management of various departments and achieve accurate asset management and control is an important research topic for power grid enterprises under the new situation of power reform.Firstly,the index system of power grid assets management level evaluation is constructed.Based on the analysis of the current situation and characteristics of power grid assets management,the core index of power grid assets management level is the coincidence rate of power grid assets accounting which reflects the loss degree of power grid assets.The main influencing factors are summarized and quantified,and the evaluation index system of power grid assets management level is established.This paper presents an evaluation method of power grid asset management level based on multiple linear regression algorithm.This paper establishes a multivariate linear regression model with the account-fact consistency rate as an independent variable and four main impact indicators as dependent variables,and explores the linear relationship between the account-fact consistency rate and the main impact indicators.The model carries out goodness of fit and significance test,and will input the actual data of assets management of a power grid enterprise to realize the prediction and verification of the reconciliation rate.This paper also presents an evaluation method of power grid assets management level based on BP neural network algorithm.Using the ability of BP neural network to approximate any non-linear function,this paper designs a three-layer BP neural network model,determines the number of hidden neurons by empirical formula and trial method,divides the sample of assets management data of a power grid enterprise into training data and testing data,establishes BP neural network and sets up multiple iteration training.Test the accuracy of thetest results.This paper compares the multiple linear regression method and BP neural network method for the evaluation of power grid assets management level through practical cases,and designs and develops a power grid assets management level evaluation system based on JAVA language.The system is divided into historical data input module,evaluation model selection module,evaluation results module and guidance and suggestion module.According to the evaluation results,the system intelligently provides users with asset management guidance and suggestions.Through the above analysis,this paper draws the following conclusions:(1)Compared with the multiple linear regression model,the asset management level evaluation model based on BP neural network has better prediction results.(2)The evaluation method of power grid assets management level based on BP neural network and multiple linear regression algorithm realizes the quantitative evaluation of power grid assets management level,and provides effective management means for power grid assets management.(3)The power grid asset management level evaluation method proposed in this paper can predict the power grid enterprise’s asset management level for a period of time in the future,and provide effective guidance and methods for the allocation of power grid asset management resources.