基于ISSA-ELM的短期电力负荷预测研究

Research on short-term load forecasting in smart grid based on ISSA-ELM

负荷预测的准确性与电力系统的运行安全关系密切,寻求更为精准的预测方法是当前面临的重要问题。提出一种基于改进麻雀算法优化极限学习机的负荷预测模型。在标准麻雀算法中引入Sobol序列、自适应收敛因子以及柯西变异,以增强初始种群历遍性、平衡算法重点区域搜索能力与全局的搜索能力。相较于标准麻雀算法,采用改进麻雀算法对极限学习机的关键参数寻优,可实现对负荷更为精准的预测。仿真结果表明,该模型拥有更为良好的回归精度和泛化能力,在短期电力负荷预测方面能够得到更精确的预测结果。

Whether the load of power network can be predicted accurately has an important influence on the security of power system.In this context,more accurate load forecasting methods are needed.In this paper,the load prediction model is based on the standard sparrow algorithm,adding Sobol sequence,adaptive convergence factor and Cauchy variation simultaneously.This approach can expand the global search scope,improve the search efficiency,and achieve more accurate prediction function.The simulation results confirm the above improvement ideas,especially in the short-term prediction,the regression goodness of fit is higher,the results are closer to the real situation.