Reliability Evaluation Based On Fuzzy Neural Network

With the development of reliability design theory, machine design technique, the work ability of large agricultral machine is getting stronger and stronger. But, with the growth of structral complexity of agricultral machine, the safety is becoming more and more important. How to evaluate the reliability of machine system uder the permissibility of economic and technique condition and offer references for improvement of reliability and maintainability of tractor engines, and for establishment of maintenance strategy is an important and urgent task.Current agricultral engines has more and more reliability problem and failure problem when working on the farm has become a universal problem. This dissertation uses the east-red 1002 track-tractor that put into production recent years as research object, and combines reliability engineering theory,artificial neural network,fuzzy mathematics to give a theory analysis and applied study on tractor's use reliability. This dissertation gives a deep study on tractor's reliability evaluation problem, The main achievement of this dissertation are briefly mentioned as follows:1. This dissertation introduce the meaning of study on reliability and the importance of reliability evaluation to this work. This dissertation expatiate the existing methods of evaluate method based on fuzzy theory ,and probe into its connotation and developing direction.2. Fuzzy theory and Artificial Neural Network technique are systemically studied, especially how to combine with these two thories are throughout studied. Give a five layers onwards-feedback Fuzzy Neural Network model according to the practice of reliability evaluation. The evaluation model includes five parts: input layer, subject function layer, rule layer, unitary layer, output layer. When the evaluate swatch that created by insert arithmetic is fuzzy-disposed, it can be imported into BP Neural Network to train the Network, and we can get the real tractor's reliability evaluation model when this train is finished.3. This dissertation gives compositive evaluation and analysis to tactor's reliability using observed value of reliability,maintaining and validity that get by data of failure track test.4. The method of reliability evaluation model based on Fuzzy Neural Network is applied on the tractor's reliability evaluation, Select 6 units'reliability index data in practice work as evaluate object, and then get its reliability grade of this type by writing Matlab program, and this result is considered to be external.The reliability evaluation of engine system is studying field which is a more synthesis subject and correlates with many subjects, and those applied prospect inengineering are vast,the theoretical worth are significant. Along with through research and an urgent need for engineering application, this studying field will be developed in depth.