粗糙集神经网络在注水泵机组故障诊断中的研究-流体机械及工程专业论文.docxVIP

英文摘要 英文摘要 目 目 录 PAGE PAGE III Subject： Rough Set Neural Network in the Injection Riciprocating Pump in Research on the Application of Fault Diagnosis Specialty： Mechanism Engineering Name： Li Kunjie （Signature） Instructor: Wu Wei （Signature） Wei Hangxin （Signature） ABSTRACT At present, the computer intelligence has been in fault diagnosis are widely application. Rough set theory is proposed Z.P awlak professor Poland, is a kind of research are not accurate knowledge and incomplete data expression, learning, inductive mathematical tools. Artificial neural network is a simulation of the human mind nonlinear dynamics system, with parallel coordination treatment and learning ability, can realize the identification and classification, optimization calculation, associative memory, the processing of knowledge, and other functions. Note the vibration of the pump unit data contain note pump unit of work state of information, is note pump unit of the fault diagnosis of important information. By using the attributes reduction of rough set, the processing of knowledge and neural network can learn, classification, parallel processing ability, establish rough set neural network complete note pump unit of fault diagnosis. This paper fusion rough sets with neural network of research results, first effectively by using rough sets theory, and then after reduction sample network structure, reduce the neural network learning and running of the time. Finally, using the MATLAB software built note pump unit fault diagnosis system. Network diagnosis results show that it has studied the knowledge of the sample, and the simulation results and the practical match, show that the network can correctly to fault diagnosis, based on rough sets and neural network combined with the note pump fault diagnosis system of research has certain theoretical meaning and practical value. Keywords: Fault diagnosis，Neural network，Rough set，Injection pump Thesis：Application Study 目 录 第一章 绪论 1