航空发动机性能参数约简方法研究

Research on Reduction of Aeroengine Parameters

对航空发动机进行性能参数约简以提取关键参数,是提高故障识别准确率和可靠度的必要条件,为此提出两种性能参数约简算法。基于灰色聚类算法的参数约简,首先计算参数间的灰色关联度,然后选取聚类中心进行聚类以提取特征参数。基于模糊粗糙集的参数约简,首先利用模糊聚类算法对连续数据进行离散化处理,然后用粗糙集理论进行参数约简。用某型航空发动机实际监测数据进行验证,结果表明:两种方法约简结果一致,以此约简结果进行故障诊断可提高诊断率,缩减运算时间。

Reducing aeroengine parameters to extract key attributes is the necessary condition to improve the accuracy and reliability of diagnosis recognition. So, two methods of parameters reduction is proposed. For the method based on grey clustering algorithm, the grey correlations between parameters are calculated firstly and then the clustering center is selected to cluster and extract the key parameters. For the method based on fuzzy rough set, the. continuous data is discrete using by fuzzy clustering algorithm firstly and then the rough set is used to reduce the parameters. The validation result of the monitoring data of an aeroengine show that two methods get the same characteristic parameters and using it to diagnose can improve accuracy and shorten computation time.