Cost Forecast of Intelligent Substation Project Based on Index Evaluation and Artificial Bee Colony Optimized Support Vector Regression

Abstract

With the complexity of the power system, effective control of the cost level of substation engineering projects is of great significance to improve investment efficiency. However, the index screening model in the current cost forecasting model is relatively single and one-sided, which leads to many important factors missing. This paper proposes an index evaluation model based on random forest (RF) and grey relational analysis (GRA), comprehensively evaluates the indexes to achieve reasonable dimension reduction, and then uses the artificial bee colony optimized support vector regression (ABC-SVR) model to predict the static total investment of the substation project. Taking the substation projects in Jiangsu Province in the past three years as a sample, the results show that the model proposed in this paper not only retains important indicators, but also achieves high prediction accuracy, which is of great significance for practical applications.KeywordsIndex evaluationRandom forestGrey relationArtificial bee colony optimized support vector regression