IGSA优化LSSVM的短期风电功率预测研究

发布时间：2018-04-18 14:54

本文选题：短期风电功率预测 + 引力搜索算法　；参考：《可再生能源》2017年11期

【摘要】：提出一种基于改进引力搜索算法(Improved Gravitational Search Algorithm,IGSA),优化LSSVM的短期风电功率预测方法。对引力搜索算法采用混沌映射学习策略初始化种群位置,引入全局记忆策略来改进速度公式,提高最优解质量,利用高斯变异算子及贪婪策略来更新最优解位置。为对比不同核函数对LSSVM预测模型性能的影响,选取了4种常用的核函数(RBF,Sigmoid,Poly及Linear)构建LSSVM预测模型,并用IGSA优化构建的模型。以安徽某一风电场实测数据为例,仿真结果表明,选择RBF核函数的IGSA-LSSVM模型的风电预测性能优于其它核函数;同时,与反向传播神经网络(Back Propagation Neural Network,BPNN)相比,以及与GA,PSO和GSA优化LSSVM相比,IGSA优化LSSVM方法对短期风电功率预测具有更好的稳定性和更高的准确性。
[Abstract]:A short-term wind power prediction method based on improved Gravitational Search algorithm and optimized LSSVM is proposed.The chaotic mapping learning strategy is used to initialize the population position, the global memory strategy is introduced to improve the speed formula and the quality of the optimal solution is improved. Gao Si mutation operator and greedy strategy are used to update the optimal solution position.In order to compare the effect of different kernel functions on the performance of LSSVM prediction model, four kinds of common kernel functions, LSSVM Sigmoid Poly and Linear. were selected to construct LSSVM prediction model, and the model was optimized by IGSA.Taking the measured data of a wind farm in Anhui Province as an example, the simulation results show that the wind power prediction performance of IGSA-LSSVM model with RBF kernel function is superior to that of other kernel functions, and compared with backpropagation neural network back Propagation Neural network.Compared with Ga-PSO and GSA optimized LSSVM, the LSSVM method has better stability and higher accuracy in short-term wind power prediction.
【作者单位】：安徽工程大学安徽省电气传动与控制重点实验室;
【基金】：安徽省自然基金(1508085ME74) 安徽省教育厅自然科学研究重点项目(KJ2014A282)
【分类号】：TM614;TP18

【相似文献】