Rough Set Information Analysis In The Application Of Fault Diagnosis And Self-Repairing Flight Control System Effectiveness Evaluation

In the domain of rough sets, an algorithm for finding reduct based on approximation accuracy of rough set is proposed in this dissertation. This algorithm takes the approximation accuracy of rough set as the iterative criterion to assure that the classification ability of the resulted reduct do not decline. It avoids the sophisticated process of finding out discernibility matrix and discernibility function, and is simple and practical. This algorithm can find out reduct with little time complexity.For noise pollution problem of feature extraction, we analyze the reason of eliminating noises using upper approximation. This dissertation proves that upper approximation quality worsens as the feature set is pruned down, and puts forward an algorithm for feature extraction using upper approximation. This algorithm takes the upper approximation quality as the iterative criterion and finds the 0-reduct (which is the feature which we need) of the system. This algorithm can deal with pollution problem of feature extraction.A fault diagnosis method based on rough set modal is put forward. The information of system's faults is denoted as a kind of information system based on rough set, furthermore is denoted as a decision table. Rough set based algorithms are used to attributes reduction, rules extraction, and fulfilling the diagnosis of the faults. This method directly deals with measurable signal of the system, mathematical modal and relevant transcendental knowledge isn't needed, so it is more practical than other methods.As the complexity of the practical diagnosis problem and the limitation of the diagnosis methods, it is almost impossible to solve the diagnosis problem of the practical completely with a single diagnosis method. So the dissertation puts forward a diagnosis method for fighter faults based on rough neural networks. Based on this diagnosis scheme, the system's diagnosis performance is improved greatly. This method can diagnose the composed faults, can identify the grade of the fault, and retains good robustness.This dissertation introduces some problems such as basic definition, fault classification, the task and the framework of fault diagnosis. The performance index system of the fault diagnosis is proposed. This dissertation overviews and analyze the fault diagnosis methods, a qualitative evaluation conclusion is given as a result. The steps of evaluation the diagnosis method is given. Then several fault diagnosis methods is evaluated, and some valuable conclusion is get.The selection principals of the effectiveness evaluation index system and the evaluation index system itself of the self-repairing flight control system is proposed. Some basic concepts is given. A synthetical effectiveness evaluation method based on analytic hierarchy process and rough set theory is proposed. The general process and steps of effectiveness evaluation of the self-repairing flight control system is given. It can provide help and standard for setting up effectiveness evaluation criterion of the self-repairing flight control system.