基于GA-VMD-BiLSTM算法的风电功率预测

Research on wind power predition based on GA-VMD-BiLSTM algorithm

针对传统风电功率预测精度低、效果差的问题,设计一种风电功率预测模型.首先,采用密度峰值聚类对实测数据去噪,并结合遗传算法(genetic algorithm,GA)优化变分模态分解(variational mode decomposition,VMD)获取最优分解个数,完成初始信号分解;其次,以双向长短期记忆(bi-directional long short-term memory,BiLSTM)网络为基础建立预测模型,预测分解后的各组数据;最后,对比其他模型验证参数正确性和模型精度.结果表明,遗传算法可获取变分模态分解中的最优分解个数,且BiLSTM网络模型的精度和适应性优于其他模型.

Aiming at the problems of low precision and poor effect of traditional wind power prediction,a wind power prediction model is designed.Firstly,density peak clustering is used to denoise the measured data,and genetic algorithm(GA)is used to optimize the variational mode decomposition(VMD)to obtain the value range of the optimal number of decomposition to complete the initial signal decomposition.Secondly,a prediction model is established based on bi-directional long short-term memory(BiLSTM)network to predict the decomposed data of each group.Finally,the parameter correctness and model accuracy are verified by comparing with other models.The results show that the genetic algorithm can obtain the value range of the optimal number of decomposition in variational mode decomposition,and the accuracy and adaptability of BiLSTM network model are better than that of other models.