北票风电场发电机组的齿轮箱故障诊断研究

发布时间：2018-01-04 20:25

本文关键词：北票风电场发电机组的齿轮箱故障诊断研究　出处：《辽宁工程技术大学》2011年硕士论文　论文类型：学位论文

【摘要】：齿轮箱是风力发电机组的重要组成部分,如何及早发现并诊断齿轮的故障,对维护系统正常运行,经济合理地安排维修设备时间,减少设备故障发生,避免重大人身伤亡事故有着十分重要的意义。故障诊断方法很多,诸如:传统故障诊断、数学故障诊断、智能故障诊断方法(模糊逻辑、神经网络、专家系统)等。通过综合比较,本文提出了基于遗传算法(GA)的模糊神经网络模型(FNN),并通过在神经网络框架下引入模糊规则,从而使网络权值有明显意义,并且保留神经网络的学习机制。使用遗传算法在搜索解的过程中,能够达到最佳收敛,优化全局。在神经网络训练之前,引入GA对染色体的交叉、变异运算寻找BP网络的最优初始权值,训练网络时再次引入GA优化网络参数,可以有效避免网络收敛过早。本文针对权值的学习采用进化算法,避免了原有BP算法容易陷入局部最优的缺点。本文首先使用了模糊规则专家系统,对齿轮箱进行故障诊断,得出诊断结果,本文分析了这种方法在故障诊断中具有实用性,同时也存在很大的局限性.由于神经网络强大的学习能力,被广泛应用与故障诊断。在文章中采用了BP网络建立故障诊断模型和基于遗传算法、神经网络、模糊逻辑结合建立模糊神经网络模型,对齿轮箱故障进行诊断,均可以得到正确的故障诊断结果。通过对于两种方法训练时间,相对误差值等方面的比较,显示了GA-FNN的优越性,表明了该方法的有效性、可行性,达到了预期效果。
[Abstract]:Gearbox is an important part of wind turbine. How to detect and diagnose the fault of gear as early as possible, to the normal operation of the maintenance system, to arrange the maintenance equipment time economically and reasonably, and to reduce the fault of the equipment. It is of great significance to avoid serious personal injury and injury. There are many fault diagnosis methods, such as: traditional fault diagnosis, mathematical fault diagnosis, intelligent fault diagnosis (fuzzy logic, neural network, expert system). In this paper, a fuzzy neural network model based on genetic algorithm (GA) is proposed, and the fuzzy rules are introduced under the framework of neural network to make the weight value of the network have obvious significance. The genetic algorithm can achieve the best convergence and optimize the whole situation in the process of searching the solution. Before the neural network training, the genetic algorithm is introduced to the crossover of chromosomes. Mutation operation can find the optimal initial weight of BP network, and introduce GA to optimize the network parameters again when training the network, which can effectively avoid premature convergence of the network. In this paper, evolutionary algorithm is used to study the weights. It avoids the disadvantage that the original BP algorithm is easy to fall into local optimum. In this paper, a fuzzy rule expert system is first used to diagnose the gearbox, and the result is obtained. This paper analyzes the practicability of this method in fault diagnosis. At the same time, there are also great limitations. Because of the powerful learning ability of neural network, it is widely used and fault diagnosis. In this paper, BP neural network is used to establish fault diagnosis model and based on genetic algorithm, neural network. Fuzzy logic combined with fuzzy neural network model can be used to diagnose the gearbox fault, and the correct fault diagnosis results can be obtained. The training time and the relative error of the two methods are compared. The superiority of GA-FNN is demonstrated, and the effectiveness and feasibility of this method are demonstrated.
【学位授予单位】：辽宁工程技术大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3

【引证文献】