Research On The Evaluation Of Overhead Line Engineering Based On Attribute Measure Theory

As a essential part of the transmission and distribution project,the evaluation of 220 k V overhead line engineering design scheme has an important impact on the whole project engineering quality and engineering cost,and has a great effection on the operation and maintenance of the power grid after the project is completed;220k V overhead line is one of the vital parts in the power grid,and its complex operation environment always causes the line breakdown and seriously threatens the safe operation of the power grid.It will help to reduce the system failure and improve the reliability of the whole power grid when the 220 k V overhead line engineering design implements comprehensive evaluation scientifically and reasonably.Firstly,this paper researches the basic theory of evaluation in detail from five aspects,including evaluation object,evaluation index,index weight coefficient,comprehensive evaluation model and evaluation results;This paper focuses on the principles and methods which are used to establish index system,and combines Delphi with Analytic Hierarchy Process to select indicators in constructing evaluation index system,and establishes a scientific and objective evaluation index system of 220 k V overhead line engineering design scheme.Secondly,the paper introduces the attribute recognition theory in attribute measure theory,and analyzes how to determine the weight in the traditional attribute recognition model,and attains the conclusion that the weight is greatly influenced by the subjectivity of experts,but sometimes it cannot truly reflect the intrinsic rules of the evaluation object and the preference of the evaluation;So this paper proposes a supervised learning method which is based on Neural Network with Random Weights to determine the weights,this method weakens the evaluation results which depend s on the subjectivity of individual experts to some extent,and this method is more accurate and profound in characterizing the inherent rules and assessment preferences through the intelligent learning of the historical data.Finally,this paper researches the precision estimation of the adaptive evaluation in the overhead line engineering design scheme,and proposes a confidence dispersion to describe the influences of confidence parameters in attribute recognition model,then researches the sensitivity of confidence parameters and simultaneously confirms the validity of the confidence dispersion by detailed theoretical derivation and analysis of simulation examples,thus will provide a theoretical basis for the selection of confidence parameters.This paper proposes the CD-SAPL algorithm which is based on confidence dispersion,and adopts adaptive method to select the confidence parameter values scientifically and reasonably.According to example analysis on the 220 k V overhead line engineering projects des ign scheme,it has been verified that the proposed algorithm is scientific and reasonable.The effectiveness of CD-SAPL algorithm has been verified though simulation results that compared with the support Vector Machine,Extreme Learning Machine and Analyt ic Hierarchy Process on the evaluation database of 220 k V overhead line engineering projects design scheme.CD-SAPL algorithm is applied in the evaluation of 220 k V overhead line engineering design scheme,which can ensure the quality of evaluation,improve the efficiency of the evaluation,and promote the development of evaluation from the traditional manual evaluation to the intelligent evaluation.It has significant meaning for the improvement of engineering quality,the control of engineering cost,the implementation of the project and the later stage operating of the project.