基于改进GWO-GM(1,1)模型的直流充电桩在线计量误差预测方法研究  

Research on DC charging pile online measurement error prediction method based on improved GWO⁃GM(1,1)model

作　　者：陈平 周娟 吴名功 CHEN Ping;ZHOU Juan;WU Minggong(College of Quality and Safety Engineering,China Jiliang University,Hangzhou 310018,China;Anhui Institute of Metrology,Hefei 230051,China)

机构地区：[1]中国计量大学质量与安全工程学院,浙江杭州310018 [2]安徽省计量科学研究院,安徽合肥230051

出　　处：《现代电子技术》2024年第5期112-117,共6页Modern Electronics Technique

基　　金：国家重点研发计划项目基金(2021YFF0601102)。

摘　　要：针对传统灰色理论预测精度不高和基本的灰狼算法容易陷入局部最优的情况,提出改进的灰狼算法与灰色理论融合的直流充电桩在线计量误差预测模型。首先,通过差分变异策略进行向量合成,引入非线性变异概率k,增强前期全局搜索能力,平衡灰狼算法的全局和局部搜索能力,避免陷入局部最优的问题;然后,将改进的算法应用于GM(1,1)模型,通过多次迭代寻找适应度值最好的一组灰狼位置,寻找到最优背景值对灰色模型进行优化,进一步提高模型的预测精度;最后,将改进前与改进后的验证模型进行对比,改进的灰色预测模型相较于基础的灰色模型均方误差与平均绝对误差分别降低了70.7%和27.2%,验证了所提方法的有效性。Since the prediction accuracy of the traditional grey theory is not high and the basic grey wolf optimization(GWO)algorithm is prone to falling into the local optimum,a DC charging pile online measurement error prediction model based on improved GWO algorithm and grey theory is proposed.Vector synthesis is carried out by differential mutation strategy,and nonlinear mutation probability k is introduced to enhance the global search ability in the early stage,balance the global and local search ability of GWO algorithm,and avoid falling into the local optimal.Then,the improved algorithm is applied to the GM(1,1)model,and a group of grey wolf positions with the best fitness value are found by several iterations,and the optimal background value is found to optimize the grey model and further improve the prediction accuracy of the model.In comparison with the verification model before and after improvement,the mean square error and mean absolute error of the improved grey prediction model are reduced by 70.7%and 27.2%,respectively,which verifies the effectiveness of the proposed method.

关 键 词：充电桩 工作误差 灰色预测 在线计量 模型融合 灰狼算法 

分 类 号：TN06-34[电子电信—物理电子学]