基于BDPCA聚类算法的航空发动机故障数据标记

航空发动机作为飞行器的动力核心对飞行器的安全飞行有着举足轻重的作用,保证航空发动机的平稳运行对飞行安全有着重大意义。在基于有监督学习的航空发动机故障诊断技术不断取得进展的同时,如何将平时获取的大量未标记数据转换为能够用来训练故障诊断模型的带标记数据,成为了制约行业发展的一大瓶颈。针对这一问题引入了基于无监督学习的DPCA算法,用以实现对未标记数据集的准确分类与标记,并针对DPCA算法在局部密度计算与簇类别数选择方面的缺陷进行了优化:针对原始DPCA算法应用标准高斯核计算局部密度易造成误识别的状况,引入共享邻域算法对局部密度的计算方法进行优化;针对原始DPCA算法需要人工研判确定簇类别数易造成的误识别状况,引入BIC选择准则对簇类别数的选择方法进行优化;提出了原始DPCA算法与共享邻域算法以及BIC选择准则相结合的BDPCA算法。最后应用航空发动机转子故障数据对BDPCA算法进行了性能验证并取得了良好的结果,证实了BDPCA算法在航空发动机气路故障诊断领域有较高的实用价值。

Aeroengine fault data tag using on improved BDPCA algorithm

As the power core of an aircraft,the aeroengine plays an important role in the safe flight of the aircraft.It is of great significance to ensure the smooth operation of aeroengine for flight safety.While the technology of aeroengine fault diagnosis based on supervised learning has been progressing continuously, how to convert a large number of unlabeled data obtained at ordinary time into labeled data that can be used to train the fault diagnosis model has become a bottleneck restricting the development of the industry.In the paper, the DPCA algorithm based on unsupervised learning was introduced to realize the accurate classification and marking of unmarked data sets.Aiming at some defects of the DPCA algorithm, the algorithm was optimized by using the shared neighborhood algorithm and BIC selection criteria.Finally, the performance of the improved algorithm was verified by applying gas path fault data of some aeroengine, and good results were obtained.