Short-term load forecasting based on improved fish-swarm algorithm and least squares support vector machine
Song Xuewei1, Liu Tianyu1, Jiang Xiuchen2, Sheng Gehao2, Liu Yuyao3
1. Department of Electrical Engineering, Shanghai Dian Ji University, Shanghai 201306; 2. Department of Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240; 3. State Grid Shandong Province Kenli District Power Supply Company, Dongying, Shandong 257000
Abstract：In the power system, accurate load data is required for both the normal operation and the islanding of the fault repair. Therefore, accurate short-term forecasting of the power load is very important. In this paper, the least squares support vector machine (LS-SVM) is used for prediction. Firstly, the artificial fish-swarm algorithm (AFSA) is improved by the visual field and step size adaptive setting and the introduction of elite reverse learning mechanism, which makes the calculation more advantageous. After that, the improved artificial fish-swarm algorithm (IAFSA) is used to improve the LS-SVM widely used in load forecasting, mainly for its kernel width coefficient and regularization parameters. Finally, the modified LS-SVM is used to the IEEE33 node system. Short-term load forecasting is performed, and the illustration shows the engineering utility of the method.
宋学伟, 刘天羽, 江秀臣, 盛戈皞, 刘玉瑶. 基于改进鱼群算法与最小二乘支持向量机的短期负荷预测[J]. 电气技术, 2019, 20(11): 20-26.
Song Xuewei, Liu Tianyu, Jiang Xiuchen, Sheng Gehao, Liu Yuyao. Short-term load forecasting based on improved fish-swarm algorithm and least squares support vector machine. Electrical Engineering, 2019, 20(11): 20-26.
2019, Vol. 20 
