基于改进支持向量机的微网负荷预测方法及其仿真分析

Microgrid Load Forecasting Based on Improved SVM and Simulation Analysis

为满足微电网建设和能源互联网技术的发展要求,根据微电网负荷难以统计、负荷波动性大、随机性强的特点,以基于用户侧微电网历史数据为基础,提出了一种基于多种群遗传算法优化支持向量回归机的微电网负荷预测模型,根据对遗传算法的改进,应用在支持向量机的参数寻优中对微电网负荷进行预测。仿真实验表明,与传统的微网负荷预测方法进行对比,其对微电网的负荷预测具有更高的精度,满足实际要求。

In order to meet the construction of microgrid and the development of the energy intcrnet technology requirements, according to the characteristics of the microgrid load, such as the difficulty of statistics, large load fluctuation, random, based on the user side historical data of microgram, a kind of microgrid load forecasting model based on multi population genetic algorithm optimized support vector machine is proposed is predicted. According to the improvement of genetic algorithm in the parameter optimization of support vector machine, the microgird load is predicted. Simulation experiments show that compared with the traditional method of load forecasting, the improved algorithm has higher accuracy and meet the actual requirements.