汽轮机故障诊断神经网络的研究.docVIP

PAGE 2 PAGE 2 PAGE PAGE 1 摘要 火力发电是当今中国社会用电的主要来源，所以火力发电机组的故障诊断非常重要，尤其是火力发电机组中的汽轮机的故障诊断。 本文首先对汽轮机故障诊断的意义进行概述，并简要介绍了汽轮机故障诊断研究现状以及神经网络在汽轮机故障诊断中的应用。然后，本文详细分析了汽轮机振动故障的机理和信息处理方法，并根据BP神经网络、Elman神经网络和D-S证据理论的应用方法和步骤，在MATLAB平台上进行了对汽轮机故障诊断的实验。最后根据实验结果对BP学习算法、Elman神经网络和D-S证据理论的优缺点进行了分析。 关键词:神经网络??BP神经网络??Elman神经网络 D-S证据理论??汽轮机故障诊断 PAGE PAGE 2 PAGE PAGE 1 Abstract Thermal power generation is main source of electric power used by Chinese society, so the fault diagnosis of thermal generator set, especially the steam turbines that is one of three main bodies of thermal generator set, is extremely crucial. Firstly, this paper outline significance of fault diagnosis of the steam turbines, and it briefly introduces the research status of fault diagnosis of the steam turbines and the application of neural network in fault diagnosis of the steam turbines. Then this paper analyzes the mechanism of fault of vibration of steam turbine and processing oriented of information in detail, and according to the method of application of BP (Back Propagation) neural network ,Elman neural network and D-S (Dempster-Shafer) evidence theory and the procedure, the paper has conducted the experiment of fault diagnosis of steam turbine on MATLAB (Matrix Laboratory). Finally, based on the experimental results, this paper analyze the advantages and disadvantages of BP learning algorithm and D-S evidence theory, and that introduce the improved method of BP learning algorithm. Key words: neural network; BP neural network; Elman neural network; D-S evidence theory; fault diagnosis of the steam turbines 目录 TOC \o 1-3 \h \u 7035 摘要 1 9029 Abstract 2 2475 第一章：绪论 1 12570 1.1 课题意义和目的 1 28436 1.2?汽轮机故障诊断研究现状 1 3012 1.2.1 国内外概况 1 6027 1.2.2 国内研究现状 1 23970 1.2.3 国外研究现状 2 25162 1.3 神经网络信息融合在汽轮机故障诊断中的应用 2 22007 1.4?本论文研究内容 2 8885 1.5本章小结 3 32539 第二章:汽轮机组故障机理及信息处理 4 13373 2.1 引言 4 16204 2.2 汽轮机组的典型振动故障机理 4 17006 2.2.1转子质量不平衡 4 8455 2.2.2轴系不对中 7 3697 2.2.3动静碰