粗糙集神经网络在注水泵机组故障诊断中的研究

发布时间：2018-07-17 08:40

【摘要】：目前,计算机智能已经在故障诊断中得到了广泛应用。粗糙集理论是波兰教授Z.Pawlak提出的,是一种研究不精确知识及不完整数据的表达、学习、归纳的数学工具。人工神经网络是模拟人类思维的非线性动力学系统,具有并行协同处理和可学习能力,可以实现识别与分类、优化计算、联想记忆、知识处理等功能。注水泵机组的振动数据含有注水泵机组工作状态的大量信息,是注水泵机组故障诊断的重要资料。本文利用粗糙集的属性约简、知识处理和神经网络的可学习、分类、并行处理的能力,建立粗糙集神经网络完成注水泵机组的故障诊断。本文融合粗糙集与神经网络的研究成果,先利用粗糙集理论有效降低样本特征的维数,再利用约简后样本构造网络,减少神经网络学习和运行的时间。最后利用MATLAB软件建注水泵机组故障诊断系统。网络诊断结果表明对于已学习过的样本知识,仿真结果与实际相吻合,表明该网络能够正确地进行故障诊断,粗糙集神经网络在注水泵机组故障诊断中的的研究具有一定的理论意义和实践价值。
[Abstract]:At present, computer intelligence has been widely used in fault diagnosis. Rough set theory is a mathematical tool for studying imprecise knowledge and incomplete data representation, learning and induction, which is proposed by Z.Pawlak, a Polish professor. Artificial neural network (Ann) is a nonlinear dynamic system that simulates human thinking. It has the ability of parallel cooperative processing and learning. It can realize the functions of recognition and classification, optimization calculation, associative memory, knowledge processing and so on. The vibration data of the water injection pump unit contain a lot of information about the working state of the water injection pump unit, and it is an important data for the fault diagnosis of the water injection pump unit. In this paper, the ability of attribute reduction, knowledge processing and neural network for learning, classification and parallel processing of rough set is used to establish rough set neural network to complete fault diagnosis of water injection pump unit. In this paper, the research results of rough set and neural network are combined. Firstly, rough set theory is used to effectively reduce the dimension of the sample feature, then the network is constructed by using the reduced sample to reduce the learning and running time of the neural network. Finally, the fault diagnosis system of water injection pump unit is built by MATLAB software. The results of network diagnosis show that the simulation results are in agreement with the actual situation for the sample knowledge that has been learned, which shows that the network can correctly diagnose the faults. The research of rough set neural network in fault diagnosis of water injection pump unit has certain theoretical significance and practical value.
【学位授予单位】：西安石油大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3;TP183

【参考文献】