李宝月,余永华,曹炳鑫,等.基于多源信息融合的柴油机典型故障诊断方法[J].内燃机工程,2025,46(1):73-79.
基于多源信息融合的柴油机典型故障诊断方法
Diagnosis Method for Typical Faults of Diesel Engines Based on Multi-Source Information Fusion
DOI：
关键词:柴油机  多源信息融合  t分布–随机邻近嵌入  故障诊断
Key Words:diesel engine  multi-source information fusion  t-distributed stochasticneighbor embedding（t-SNE）  fault diagnosis
基金项目:国家自然科学基金重点项目（52271328）
作者单位E-mail
李宝月 武汉理工大学 船海与能源动力工程学院，武汉 430063 li.by@whut.edu.cn 
余永华* 武汉理工大学 船海与能源动力工程学院，武汉 430063 yyhua@whut.edu.cn 
曹炳鑫 武汉理工大学 船海与能源动力工程学院，武汉 430063 bxcao@whut.edu.cn 
叶剑平 武汉理工大学 船海与能源动力工程学院，武汉 430063 2325808287@qq.com 
马炳杰 船舶动力工程技术交通运输行业重点实验室，武汉 430063  
尧阳烽 武汉理工大学 船海与能源动力工程学院，武汉 430063 yaoyangfeng@whut.edu.cn 
赵国旭 武汉理工大学 船海与能源动力工程学院，武汉 430063 zgx1049722103919@whut.edu.cn 
摘要点击次数: 315
全文下载次数: 156
摘要:针对基于单一振动信号难以区分柴油机不同部件故障，提出一种基于t分布–随机邻近嵌入（t-distributed stochastic neighbor embedding， t-SNE）多源信息融合的故障诊断方法。首先，通过试验对柴油机故障仿真模型进行标定，基于仿真模型获取不同故障状态下的热工参数与缸盖振动，选取相关性低的热工参数，提取振动信号的时域和频域特征参数，并利用t-SNE将振动特征参数与热工参数进行融合降维，基于支持向量机（support vector machine， SVM）方法对降维后的数据进行分类识别，构建柴油机故障诊断模型，最终取得了95.7%的故障识别准确率。与基于振动单一信号的故障诊断方法相比，多源信息融合能有效区分不同故障类别，提高柴油机故障识别准确率。
Abstract:A fault diagnosis method based on t-distributed stochastic neighbor embedding（t-SNE） and multi-source information fusion was proposed because it is difficult to distinguish the faults of different components of diesel engines based on a single vibration signal. The fault simulation model of the diesel engine was calibrated through experiments. Based on the simulation model， the thermal parameters and cylinder head vibration under different fault conditions were obtained. Thermal parameters with low correlation were selected， and the time domain and frequency domain characteristic parameters of the vibration signal were extracted. The vibration characteristic parameters and thermal parameters were fused and dimensionally reduced using t-SNE. The data after dimensional reduction was classified and recognized based on the support vector machine（SVM） method to construct a fault diagnosis model for the diesel engine， and a fault recognition accuracy of 95.7% was finally achieved. Compared with the fault diagnosis method based on a single vibration signal， multi-source information fusion can effectively distinguish different fault categories and improve the fault recognition accuracies of the diesel engine.
查看全文  HTML   查看/发表评论