Short-term load forecasting method based on cuckoo search algorithm and support vector machine considering demand price elasticity
1.College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China; 2. State Grid Lanzhou Power Supply Company, Lanzhou, Gansu 730070, China； 3. State Grid Energy Research Institute, Beijing 102209, China
Abstract： To solve the problems that electricity demand was affected by multiple disturbance factors in the electricity market, and the price elasticity of demand was increased, for obtaining more accurate and comprehensive power load forecast value, the cuckoo search algorithm and support vector machine (CSSVM) shortterm load forecasting method was proposed with considering demand price elasticity. The Pearson correlation coefficient method was used to analyze the load autocorrelation and the correlation among load and historical load temperature, humidity, electricity price, load difference and electricity price difference. Based on the definition of demand price elasticity, a demand price elasticity model was established to reflect the impact of electricity market transactions on load. The cuckoo search algorithm was used to optimize the parameters of support vector machine, and the CSSVM shortterm load forecasting model was established with considering the price elasticity of demand. The actual data of a certain state in the PJM power market in the United States was taken as example to conduct the prediction, and the prediction error was compared with that of comparison model. The results show that the average relative percentage error of the proposed model is 13.43%, and the prediction accuracy is improved by 5.31% compared with that of the model without introduction of demand price elasticity.
