基于模糊聚类的刮板输送机故障分析与诊断

发布时间：2018-05-17 13:04

本文选题：模糊聚类算法 + 遗传算法　；参考：《西安科技大学》2017年硕士论文

【摘要】：目前,随着科技的发展,煤矿自动化水平的提升,刮板输送机也朝着自动化、智能化方向发展。刮板输送机的主要功能是转载和运送,同时还与破碎机、采煤机等设备组成一个串联系统。但是刮板机故障频发,常常使整个采煤系统陷入困境,导致煤矿蒙受巨大的经济损失同时给工作人员带来安全隐患。为使刮板机安全可靠运行,人们投入了更大的精力和财力用于对设备的故障进行诊断研究。本文针对SGZ1250/2565型煤矿刮板机进行研究,分析了故障诊断的意义。在广泛了解刮板输送机研究的国内外现状、以后的发展方向和目前常用故障诊断分析方法上,提出了基于模糊聚类分析的故障诊断方法。首先,从机械原理方面出发分析了刮板输送机的整体构造和故障频发的传动部系统的故障机理,进而阐释了故障的原因和故障导致的结果。然后,对模糊聚类算法的概念进行了介绍,提出了在刮板输送机故障诊断中应用模糊C-均值聚类(FCM)。通过实例发现,在样本数据较少、样本之间差别比较大或者没有交错的数据存在的话,FCM算法具有良好的分类能力且收敛速度快。当样本数据量很大、彼此之间交错重叠的话,FCM算法错误率比较高,分类结果不理想。最后针对FCM算法存在的问题,提出了以优化分类为目的的遗传模拟退火算法并将其应用在刮板机故障诊断中,实验证明优化后的模糊聚类算法能够有效快速的解决聚类问题。
[Abstract]:At present, with the development of science and technology and the improvement of coal mine automation level, scraper conveyors are also developing towards automation and intelligence. The main function of scraper conveyor is to transfer and transport, but also with crusher, shearer and other equipment to form a series system. However, the frequent faults of scraper often make the whole coal mining system in a difficult situation, resulting in huge economic losses to the coal mine and bring safety risks to the staff. In order to make the scraper run safely and reliably, people put more energy and financial resources into the fault diagnosis of the equipment. In this paper, SGZ1250/2565 type coal mine scraper is studied, and the significance of fault diagnosis is analyzed. A fault diagnosis method based on fuzzy clustering analysis is put forward on the basis of a wide understanding of the present situation of scraper conveyor research at home and abroad, the future development direction and the common fault diagnosis and analysis methods. Firstly, the whole structure of scraper conveyor and the fault mechanism of transmission system with frequent faults are analyzed from the aspect of mechanical principle, and the causes and results of failure are explained. Then, the concept of fuzzy clustering algorithm is introduced, and the application of fuzzy C-means clustering algorithm in fault diagnosis of scraper conveyor is put forward. It is found that the FCM algorithm has good classification ability and fast convergence rate if the sample data is less and the difference between samples is large or there is no interlaced data. When the sample data is very large, the error rate of FCM algorithm is high and the classification result is not ideal. Finally, aiming at the problems of FCM algorithm, a genetic simulated annealing algorithm for the purpose of optimizing classification is proposed and applied to the fault diagnosis of scraper. The experiment shows that the optimized fuzzy clustering algorithm can solve the clustering problem effectively and quickly.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD528.3

【参考文献】