内燃机振动的数据可视化诊断方法研究  

Data visualization diagnosis method of internal combustion engine vibration

作　　者：张世雄 蔡艳平 石林锁 张恒 

机构地区：[1]火箭军工程大学,西安710025

出　　处：《空军预警学院学报》2017年第5期380-383,共4页Journal of Air Force Early Warning Academy

摘　　要：为提高内燃机故障诊断的智能化程度和诊断精度,提出将内燃机振动数据进行图像表征的可视化诊断方法.根据振动数据时频表征和多元图表征方法,分析了内燃机故障在图像域的表现形式及分类识别.在6135型柴油机上模拟气阀机构的8种常见工况,利用可视化诊断技术对振动信号进行图像表征、特征提取及分类识别,最高识别率达93.75%.实例结果验证了可视化诊断技术的有效性.In order to improve the intelligentization degree and diagnostic precision of internal combustion engine fault diagnosis, this paper proposes a visualization diagnosis method for image representation of the vibration data of the internal combustion engine. According to the time-frequency characterization and multivariate representation method of vibration data, the paper analyzes the performance and classification recognition of internal combustion engine failure in image domain. The paper also simulates eight different operating conditions of valve mechanism on Model 6135 diesel engine, and utilizes the visual diagnostic technology to realize the image representation, feature extraction and classification recognition of the vibration signal with the highest recognition rate of 93.75%. The examples test verifies the effectiveness of the proposed visual diagnostic techniques.

关 键 词：内燃机 图像表征 特征提取 模式识别 

分 类 号：TK428[动力工程及工程热物理—动力机械及工程] TP391.4[自动化与计算机技术—计算机应用技术]