基于局部线性嵌入算法的柴油机测点优化  

Measuring Points Optimization of Diesel Engine Based on Locally Linear Embedding Algorithm

作　　者：董安[1] 潘宏侠[1] 龚明[1] 

机构地区：[1]中北大学机械工程与自动化学院,太原030051

出　　处：《车辆与动力技术》2013年第4期53-57,62,共6页Vehicle & Power Technology

基　　金：国家自然科学基金资助项目(50875247)

摘　　要：为了能够高效地进行柴油机故障诊断,需要找出敏感性和稳定性好的振动测点,提出了一种改进的局部线性嵌入算法对柴油机振动测点进行优化处理.该算法实质是在保持原始数据性质的情况下,将各测点的高维特征数据映射到低维空间中,并以2-范数为各测点的阈值作为测点优化的标准.以某12缸柴油机为例,应用小波包能量谱分析方法来提取各振动测点信号的特征值;应用局部线性嵌入算法进行测点优化,通过对比实验结果可知,局部线性嵌入算法能够有效地区分不同故障测点敏感程度,从而可找出各故障敏感的测点位置,在一定程度上可提高故障识别率与效率.In order to promote efficiency of the fault diagnosis in a diesel engine, it is needed to seek out some measurement points with sensitivity and stability to the vibration. An improved locally linear embedding algorithm is proposed to optimize the measurement points of vibration in a diesel engine. The essence of the algorithm is to map those high-dimensional characteristic data into a low dimensional space under the case of maintaining features of the original data, and to act the threshold of each measurement point with a 2-norm as the criteria to optimization of the measurement points. Taking a 12-cylinder diesel engine as an example, the features of the vibration signal of the measurement points are extracted by using wavelet packet energy spectrum analysis method, and the measurement points are optimized by means of the locally linear embedding algorithm. The comparison with the experimental results show that, with the help of the locally linear embedding algorithm, the sensitivity of different fault measurement points can be effectively distinguish, the positions of the sensitive points can be located, and the accuracy and efficiency of the fault recognition can be improved.

关 键 词：局部线性嵌入算法 测点优化 小波包变换 

分 类 号：TH113[机械工程—机械设计及理论] TK421[动力工程及工程热物理—动力机械及工程]