A Comparison And Decision System Of Power Grid Planning Based On SEC

Electric power industry is an important basic industry,and power grid planning is the specific implementation of power system planning for the electric power industry,so the research of power grid planning has important practical significance.With the rapid development of computer technology,technologies such as machine learning and others has been applied to the cost control of power system,load prediction of power grid and fault prevention and maintenance of power system.Some methods such as the projection of ideal vector space are also applied to the comparison of power grid planning.A power grid planning comprehensive evaluation model based on life-time safetyefficiency-cost is designed and implemented.Applying the vector space model and K-means clustering algorithm and other methods in the analysis of the power grid planning decisionmaking is put forward and implemented.In terms of the data processing of historical power grid planning schemes,the normalization model is used to eliminate the difference of feature parameters on the premise of maintaining the proportional characteristics,which more reasonably reflects the characteristic differences among schemes.Moreover,in order to integrated the different weight vectors given by experts,a combination weight vector optimization model is used to find the combination weight vector with the smallest deviation from the original weight vectors and weight the feature vector of the scheme,so as to reflect the importance of different feature parameters in the scheme.The test results of the comparison and decision system show that the software platform of the comprehensive evaluation model is completed,and the deviation between the actual similarity and the solution similarity in vector space model is small.The recommendations by the vector space model for the current power grid construction projects show 83.3% accuracy and recall rate,and 97.6% accuracy,and the recommendations by the K-means clustering algorithm show 91.7% accuracy and recall rate,and 98.8% accuracy,which are both feasible in application.