首页|期刊导航|电力系统保护与控制|基于SHAP的异构双分支KAN-LightGBM变压器故障诊断模型

基于SHAP的异构双分支KAN-LightGBM变压器故障诊断模型OA

Transformer fault diagnosis model based on SHAP and a heterogeneous dual-branch KAN-LightGBM architecture

中文摘要英文摘要

针对现有变压器故障诊断模型对易混淆样本的特征信息利用率不足等问题,提出一种基于沙普利加性解释(Shapley additive explanations,SHAP)的主次特征选择的异构双分支 B 样条-小波 KAN(basis spline-wavelet Kolmogorov-Arnold network,B Spline-WavKAN)的变压器故障诊断模型.首先,采用SHAP算法量化各特征对不同故障类型的贡献度,完成主次特征有效划分.然后,提出异构双分支B Spline-WavKAN特征提取网络及自适应融合架构,主特征输入 B SplineKAN 以学习全局特征,次特征输入 WavKAN 以深度挖掘易混淆样本边界信息.并通过置信度引导的自适应融合机制实现主次特征动态加权互补.最后,将融合后的增强特征输入轻量级梯度提升机(light gradient boosting machine,LightGBM)进行故障诊断.实验结果表明,所提方法能够高效地挖掘故障特征信息,增强对易混淆样本的识别能力,诊断准确率高达98.7%,具有较高的工程应用价值.

To address the insufficient utilization of feature information from easily confusable samples in existing transformer fault diagnosis models,this paper presents a transformer fault diagnosis model based on Shapley additive explanations(SHAP)primary-secondary feature selection and heterogeneous dual-branch basis spline-wavelet Kolmogorov-Arnold network(B Spline-WavKAN).First,the SHAP algorithm is used to quantify the contribution of each feature to different fault types,achieving the effective differentiation between primary and secondary features.Second,a B Spline-WavKAN heterogeneous dual-branch feature extraction and adaptive fusion framework is introduced.Primary features are fed into the B SplineKAN to learn global features,while secondary features are fed into the WavKAN to deeply extract boundary information of easily confusable samples.The primary and secondary features are dynamically weighted and complemented through a confidence-guided adaptive fusion mechanism.Finally,the fused enhanced features are fed into a lightweight gradient boosting machine(LightGBM)for fault diagnosis.Experimental results demonstrate that the proposed method can efficiently capture fault feature information and improve the recognition of easily confusable samples,achieving a diagnostic accuracy of up to 98.7%and demonstrating strong potential for engineering application.

李文豪;王艳;陈翰林;司永震;赵洪山

华北电力大学电气与电子工程学院,河北 保定 071003华北电力大学电气与电子工程学院,河北 保定 071003华北电力大学电气与电子工程学院,河北 保定 071003华北电力大学电气与电子工程学院,河北 保定 071003华北电力大学电气与电子工程学院,河北 保定 071003

变压器故障诊断特征选择SHAP算法主次要特征KAN网络

transformer fault diagnosisfeature selectionSHAP algorithmprimary and secondary featuresKAN network

《电力系统保护与控制》 2026 (20)

28-39,12

This work is supported by the National Natural Science Foundation of China(No.51807063). 国家自然科学基金项目资助(51807063)

10.19783/j.cnki.pspc.260391

评论