基于数据融合的发动机故障诊断系统研究-控制理论与控制工程专业论文.docxVIP

第1I页 第1I页 武汉科技大学硕士学位论文 Abstract With the development of China’S economic and transport， cars have being widely used in many fields．It has become one of the marks of modem society．However,as the complex structures of engine，poor working conditions，making the failure rate hi．gh and maintenance costs much．It is therefore necessary to find simple，practical and effective engine fault diagnosis method． This paper is writen under the premise of in reading large information about fault diagnosis combining the engine fault characteristics．Trying to use the method comparison— multi—SenSOr data fusion which is quite popular in the field of of control in recent years to develop a fault diagnosis system used in the automotive engine．Through theoretical analysis and simulation results show that the data fusion methode of fuzzy logic and neural networks applied in the engine fault diagnosis Can get better results．In order to achieve better results，on the basis this paper design a new method-FNNC，and use it in the training of the system， simulation experiments show that it is all efficient and effective way,have a good futhure in theory and application． This paper includes the content： (1)Introduction of fault diagnosis in particular,the automobile engine fault diagnosis technology and the status； (2)Introduction of multi-sensor data fusion technology development and the status， focused on fuzzy logic and neural network； (3)The combination of engine failures in the characteristics and features of data fusion method and design a fault diagnosis system，give a new method(FNNC)and look forward to improve the Accuracy rate； (4)Simulating the fault diagnosis system for engine fault diagnosis with MATLAB，and give the test results； (5)Summary and Outlook． Key words：engine；fault diagnosis；data fusion；fuzzy；neural networks；Conjugate gradient 武汉科技大学研究生学位论文创新性声明 武汉科技大学 研究生学位论文创新性声明 本人郑重声明：所呈交的学位论文是本人在导师指导下，独立进行研 究所取得的成果。除了文中已经注明引用的内容或属合作研究共同完成的 工作外，本论文不包含任何其他个人或集体已经发表或撰写过的作品成果。 对本文的研究做出重要贡献的个