两种差别矩阵约简算法在故障诊断中应用分析

Analysis of Fault Diagnosis Based on The Two Discernibility Matrix Reduction Algorithms

军用飞机故障诊断知识库的信息存在不完备和知识的不确定性,严重增加了实时诊断的难度。粗糙集针对不完整、不精确信息处理具有显著优势,首先介绍了粗糙集理论中的知识表达系统和差别矩阵,分析了基于差别矩阵的两种决策表属性约简算法,并应用于某型机载设备进行故障诊断,结论表明两种算法均能对数据进行必要约简,并准确地做出故障诊断。通过两种约简算法的复杂度比较,得出基于差别矩阵的属性约简算法实时性更高,实际故障诊断中更可取。

There are incompleterness and uncertained infromation in fault diagnosis of military aircraft,which increase the difficulty of real-time diagnosis.Rough sets theory has much more superiority of processing uncomplete and uncorrect information.The paper introduces the knowledge expression system and discernibility matrix,and then analies the two attributive reduction algorithms of decisiontable based on discernibility matrix.The results show that the two algorithms can reduce the data and the fault diagnosis are correct through applying these two algorithms to some aero equipments.By comparing with the complex degree of the two algorithms,it can be induced that the attribute reduction algorithm has the higher real-time and is more advisable in practical fault diagnosis.