摘要
针对传统电力电子电路在软故障诊断领域的特征区分度低、诊断效率低等一系列问题,提出一种变分模态分解(VMD)结合改进的黑翅鸢搜索算法(IBKA)优化极限学习机(ELM)的故障诊断方法。首先,利用VMD技术将采集到的故障信号进行分解重构,并得到故障诊断的特征向量。其次,用改进后的黑翅鸢搜索算法对ELM的参数进行优化,得到IBKA-ELM分类模型;IBKA采用Sine映射初始化种群,随机选择3个不同的个体进行差分变异操作,更新领导者位置,在领导者位置更新处引入自适应惯性权重因子,可有效提高算法的寻优能力和收敛速度。最后,通过150W的Boost电路对本文方法进行实验验证。实验结果显示,VMD结合IBKA-ELM的故障诊断方法在实际诊断中的精度均达到99%以上。
To address the problems of low feature differentiation and low diagnostic efficiency of traditional power elec-tronic circuits in the field of soft fault diagnosis,a fault diagnosis method based on variational mode decomposition(VMD)combined with improved black-winged kite search algorithm(IBKA)to optimize extreme learning machine(ELM)is proposed in this article.Firstly,the collected fault signal is decomposed and reconstructed by using VMD technology,and the fault diagnosis feature vector is obtained.Secondly,the improved BKA is used to optimize the ELM parameters,and the IBKA-ELM classification model is obtained;IBKA initializes the population by Sine mapping,randomly selects three different individuals for differential mutation operation,updates the leader position,and introduces adaptive inertial weight factor in the leader position update,which can effectively improve the optimization ability and convergence speed of the algorithm.Finally,the proposed method was verified by a 150W Boost circuit.The experimental results showed that the accuracy of VMD combined with IBKA-ELM in the actual diagnosis reached more than 99%.
作者
陈苗
姜媛媛
CHEN Miao;JIANG Yuanyuan(College of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan 232001,China;Institute of Environment-Friendly Materials and Occupational Health,Anhui University of Science and Technology,Wuhu 241003,China)
出处
《天津科技大学学报》
CAS
2024年第6期57-65,共9页
Journal of Tianjin University of Science & Technology
基金
安徽省重点研究与开发计划项目(202104g01020012)
安徽理工大学环境友好材料与职业健康研究院研发专项基金资助项目(ALW2020YF18)。
关键词
软故障诊断
变分模态分解
黑翅鸢搜索算法
极限学习机
DC–DC电路
soft fault diagnosis
variational mode decomposition
black-winged kite search algorithm
extreme learning machine
DC-DC circuit
