Research On Technical Condition Integrated Evaluation For Power Plant Of Ships Based On Information Fusion

This paper is supported by the National Natural Science Foundation of P.R. China"Research on Fusion Information Entropy of Quantitative Evaluation of Vibration Condition of Rotating Machinery"and Naval Youth Science of P.R. China"Research on Fusion Information Entropy of Vibration Condition of Rotating Machinery". Taking plant power of ships as the object, technical condition evaluation of plant power of ships based on information fusion is constructed, which is studied on the theory and application.The general situations on the technique of device condition monitoring and fault diagnosis, the technique of technical condition evaluation for ships, and the technique of information fusion are summarized. The question in technical condition evaluation for ships is pointed out. It is brought out that the technical condition is evaluated by applying the information fusion.The concept and development of technical condition management are expatiated. The signification of technical condition evaluation is discussed. The model on the technical state evaluation for ships based on the information fusion is designed. Some top index systems are founded to be used in the technical state evaluation for ships.It is put forward that the character indexes for technical condition evaluation are optimized by the reduction function of rough set theory. Taking the QGF-1 steamer-fan as object, the character indexes and decision-making rules for technical condition evaluation of steamer-fan are made by rough set theory.In view of the uncertainty in the integrated evaluation, it is brought forwarded that various evidences are combinated by Dempster-Shafter (D-S) evidence theory, to reduce the uncertainty and enhance the nicety. Therefore, the development of D-S evidence theory is described. Secondly, the basic means, lack, and some mend measure of D-S evidence theory are studied and the decision-making means based on evidence theory is set forth. Lastly, the technical condition evaluation based on D-S evidence theory is made.Because many devices of plant power are rotating machinery, the fault of the rotating machinery based on information entropy is studied. Firstly, the various information entropy features are set up, which show the vibration energy of the fault of the rotating machinery such as singular spectrum entropy, power spectrum entropy, wavelet space state feature entropy, wavelet power spectrum entropy. By a lot of experiment on the rotor, the expectation of the value of the various entropy based on the different typical faults of the rotating machinery is set up, which can be the standard feature vector for fault diagnosis based on the information entropy. According to the thought that the more short of the distance, the more similar among the models, a method based on close degree to information entropy is set up. Considering the function of D-S evidence theory, the method of rotating machinery fault diagnosis based on close degree of information entropy and evidence theory is set up. The applicability of the various measures is depicted by the instances given.A kind of integrated technical condition evaluation method based on mission sections for complex systems is presented to deal with the engineering application, in which some theories are applied, such as rough set theory, D-S evidence theory, device cluster and fuzzy theory. Devices and device clusters needed are confirmed by the mission sections and the concept of device cluster in advance. Indexes of simple device are optimized by rough set theory. The technical condition can be gained by the evaluation method based on index classification or by the evaluation method based on D-S evidence theory. The technical condition of device cluster can be gained by the evaluation method based the technical condition and the weigh of simple devices. According to the technical condition of device clusters, the technical condition of the complex system is gained by the evalution method base D-S evidence theory or the weigh of device clusters.The weigh model based on expert fuzzy mark assures that the evaluation result accords with the actual instance.