基于重构相空间FLS-SVM的电力系统混沌预测模型

Power System Chaos Prediction Model Based on FLS-SVM via Reconstructed Phase Space

摘要: 研究电力系统混沌预测以及预测中超参数难以调整的问题，采用基于重构相空间的模糊最小二乘支持向量机（RFLS-SVM）方法进行算法改进．结合Takens嵌入维理论重构数据相空间，并对电力系统混沌状态进行预测，预测精度得到了进一步提高．并研究了RFLS-SVM核参数调整的方法，得到了一般性结论．通过数字仿真实验证明了该方法是有效的．

Abstract: Chaos forecasting of power system and the difficulty in super parameters adjustment are studied.Fuzzy least square support vector machine method based on space reconstruction(RFLS-SVM) is adopted to improve the algorithm. Data space is reconstructed based on Takens embedding theorem,and the chaotic state of power system is forecasted and the forecasting precision is improved.And also RFLS-SVM kernel parameters adjustment method is studied to get a general conclusion.Numerical simulation results show that this method is effective.