Prediction of power grid investment demand based on GRA and IPSO-BP neural network

Abstract With the continuous increase in global energy demand and the increasing complexity of the electricity market, accurately predicting the investment demand for the power grid has become an important issue that governments and power companies around the world urgently need to pay attention to. In this paper, grey correlation analysis (GRA), improved particle swarm optimization (IPSO), and BP neural network (BPNN) methods were described. On this basis, the GRA-IPSO-BPNN model for power grid investment demand forecast was proposed. In this model, the GRA method was first used to screen the key influencing factors of investment demand, and these key factors were used as input variables for the BPNN, reducing the dimensionality of the network’s input data. Subsequently, the IPSO algorithm was employed to optimize the initial weights and thresholds of the BPNN, solving the problems of the BPNN easily getting stuck in local minima and slow training speed, overcoming the shortcomings of the BPNN, and improving the prediction accuracy of the BPNN. The GRA-IPSO-BPNN model was applied to the investment demand forecasting of FJ Provincial Power Grid Company, and the research results showed that the GRA-IPSO-BPNN model has good forecasting performance. Compared with the traditional BPNN model and IPSO-BPNN model, the proposed method effectively improves the effectiveness and accuracy of investment demand forecasting. This offers innovative approaches and techniques for predicting the demand for power grid investment within the framework of precise investment strategies.