基于PCA-KFCM的船舶柴油机故障诊断

为提高船舶柴油机故障诊断的准确率和深刻反映船舶柴油机的运行状况,结合主元分析(PCA)的特征提取优势和模糊核聚类(KFCM)具有较好聚类效果的特点,提出了一种新的船舶柴油机故障诊断方法。该方法首先利用主元分析对船舶柴油机故障的训练和测试数据集进行特征提取,消除了故障征兆之间的相关性;然后对经特征提取后的训练样本进行模糊核聚类,并用网格法确定其中的参数,得到聚类中心。通过计算测试样本集中各样本与聚类中心在高维特征空间中的欧氏距离,得出最终的故障诊断结果。对MAN B&W 10L90MC型船用柴油机的故障诊断结果验证了该方法的有效性。因此,应用提出的方法对船舶柴油机进行故障诊断具有重要的实际意义。

Fault Diagnosis of Marine Diesel Engine Based on PCA - KFCM

In order to improve the accuracy of fault diagnosis in marine diesel engine and reflect the operation conditions of marine diesel engine profoundly,combining with the advantage of principal component analysis for extracting feature and better clustering effect of kernel fuzzy c-means clustering,a new fault diagnosis method of marine diesel engine was proposed.Firstly,this method used principal component analysis to extract the feature from the training and test fault data set of marine diesel engine,by which the correlation among fault symptoms was eliminated.Secondly,kernel fuzzy c-means clustering was applied to classify the faults of training samples whose features had been extracted,the parameters in which was determined by grid method,and the clustering centers were obtained.Finally,by calculating the Euclidean distance between each sample in testing sample set and the clustering centers in high dimensional feature space,the final fault diagnosis results were obtained.The effect of fault diagnosis in MAN B&W 10L90MC marine diesel engine verified the effectiveness of this method.Therefore,the method proposed has important practical significance towards the fault diagnosis of marine diesel engine.