【作者】 徐劲力；

【导师】 江征风；

【作者基本信息】 武汉理工大学 ， 机械制造及其自动化， 2009， 博士

【摘要】 随着现代机械设备大型化及自动化程度的不断提高,大型机械在现代工业生产中的作用越来越重要。而随着多层工业厂房的日趋增多,大型旋转机械高层安装已成为发展趋势。由此伴随着两大问题的产生：机械设备的故障诊断问题；大型旋转机械对承载建筑物振动舒适度的问题。因此对大型旋转机械进行故障诊断成为保障生产系统稳定性和可靠性的重要手段,对大型旋转机械引起的承载建筑物舒适度的研究更是为全面进行大型旋转机械对承载建筑物的舒适度评价的建立提供参考依据。本文结合机械学、建筑学、计算机科学。提出了大型旋转机械的故障诊断的智能方法,同时结合结构建筑的舒适度评价方法和标准对原来承载建筑物无定量标准进行定量化设计,为承载建筑物舒适度提供量化依据,同时为采取减振措施提供数据参考。本文以武钢矿业有限公司鄂州球团厂造球机及承载建筑物为研究对象开展研究要想准确模拟诸多复杂环境下大型旋转机械设备的故障诊断过程,需要大量的实验和数据分析。目前,国内外的研究技术多集中在实验、数学模型分析,以及用智能技术做故障诊断等方面。很显然,实验的方法需要花费很大的人力物力,数学模型方法主要是依据实验数据及经验建立相应的数学模型,误差是难以避免的。造球机由承载建筑物高层支撑,造球机运行时的振动直接影响到承载建筑物的振动,建筑物的振动又反作用于造球机。故此造球机-承载建筑物耦合振动造成了故障诊断分析的更复杂性。人工神经网络是一种智能处理技术,力图模拟人类处理问题的方式去理解和利用信息。神经网络控制可通过对网络结构及权值处理的自动调整而实现非生物神经网络系统的部分功能,能处理高维数、强干扰、难建模的工业过程,为大型旋转机械设备的检测和故障诊断提供了另一种可行的办法。但是,神经网络应用过程中存在的主要问题是学习中不具备全局搜索能力,易陷入局部极小,而粒子群算法是基于群体智能理论的优化算法,是一种种群的全局搜索策略,它是通过群体中粒子间的合作与竞争产生的群体智能指导优化搜索。因此,将神经网络与粒子群算法融合可有效克服神经网络学习中的可靠性低的问题,但传统的融合算法只训练网络权值和阈值,存在冗余度高、收敛速度慢的问题,因此,本文采用一种嵌入免疫因子的免疫粒子群与LM(BP神经网络优化算法的一种)融合算法,有效的解决这些问题。造球机运行时由于承载建筑物与造球机的耦合振动,造成楼板产生振动,其振动现象表明：楼板振动造成人对安全性的极大恐慌,通过权威部门的检测承载建筑物及楼板的安全性是没有问题的,这就涉及到舒适度问题,而对于建筑物舒适度的研究,少有涉及到大型旋转机械对承载建筑物的舒适度的问题,本论文已承载建筑物四楼楼板为研究对象,阐述了结构建筑舒适度评价方法及标准,论文采用计权加速度评价方法,根据建筑物舒适度评价指标,结合国际国内标准,对造球机的承载建筑物的楼板舒适度进行系统研究,提出了承载建筑物舒适度的量化标准,为进一步完善舒适度的评价体系提供素材。同时,更是为全面进行大型旋转机械对承载建筑物的舒适度评价方法的建立和量化标准提供参考依据。根据已建承载建筑物的实际情况,本论文对楼板隔振措施进行了系统研究,通过以支撑造球机的立柱的实测振动数据作为激励输入,建立有限元模型,采用浮置楼板作为减振措施进行了仿真分析,其减振效果非常有效,达到了所制定的楼板舒适度标准,符合工作人员的舒适度要求,为武钢二期工程的建设提供了科学依据。更多还原

【Abstract】 With the improvement of the modern mechanical equipments which become more large and more automatic, the large-scale machinery is getting more and more important in modern industry production. As a result, the number of multi-layered industry workshops are increasing day-by-day; the high-level installment of large-scale rotating machinery has become the trend of development. There are two major issues emerge:Equipment fault diagnosis and the degree of large rotating machinery’s vibration comfort evaluation of load building. Therefore, large-scale rotating machinery fault diagnosis is considered as a important means to guarantee the stability and reliability of an production system; Analysis on comfort of load buildings which caused by large-scale rotating machinery can provide a reference on its establishment of evaluation system. This article unifies mechanics, architecture, and computer science. It proposes a intelligent method of large-scale rotating machinery fault diagnosis, meanwhile combines with the comfort of structures and criteria for evaluation of the design of the original buildings in a quantitative way, as result provides reference data to evaluation of comfort of Load building and damping measures.This paper researches on pelletizer and the pelletizer’s load buildings of Ezhou Pelletizing Plant of Wuhan Iron and Steel Mining Co. Ltd.It needs an analysis with large number of experiments and data to accurately simulate the complex environment of many large-scale rotating machinery fault diagnosis process. At present, research on this field no matter at home and abroad, is concentrating on experiment, analysis of mathematical model, as well as intelligent fault diagnosis technology and so on. It is clear that the methods of the experiment will take a lot of human and material resources, while the way of mathematical models is based mainly on experimental data and experience, so errors are difficult to avoid.Pelletizer is built on load building by carrying high-level support. When the pelletizer is running, there is a direct impact on the vibration of the load building; meanwhile the vibration of the load building has counteraction on the pelletizer. As a result the coupled vibration of pelletizer and load building makes the analysis of fault diagnosis more complex.Artificial Neural Network (ANN) is considered as a kind of intelligent technology, which manages to deal with problem by way of human beings to understand and use information. ANN can achieve part of functions of non-living neural network system through automatically adjust of the network structure and right value, and it make a good deal on industrial process which is high dimensionality, strong interference and hard to model. So it provides a possible solution on testing and fault diagnosis for large-scale rotating machinery equipment. However, the main issues in the application of ANN have not global search capabilities, and are easy to fall into local minima, but the PSO (Particle Swarm Optimization) is an optimization algorithm based on the theory of swarm intelligence, it’s a search strategy of population global. And it can optimize the search by guidance arising from the of swarm intelligence which are inter-group cooperation and competition. Therefore, the combination of ANN and PSO can effectively overcome the problem of low reliability in the study process of ANN. But the traditional way of integration only focus on weights and thresholds, which cause problems of high redundancy and slow convergence. Therefore, this article uses particle swarm optimization which is immune, to give an effective solution to these problems.The vibration caused by the coupled vibration of pelletizer is serious harmful for people’s health when the equipment is running, even causes extreme discomfort and gives people the feeling of panic. Detection of the authority department proves the safety of the load buildings are no problem. So it involves the issue of comfort, but there rarely are issues about comfort of the large-scale rotating machinery running on the load building. This paper studys on the fourth floor of load building has described the structure of the building comfort evaluation methods and standards. It uses method of weighted acceleration thesis, in accordance with building comfort evaluation index and combined with international and domestic standards to comprehensively investigate the comfort of load buildings of pelletizer, propose quantitative criteria of load building, and provide some elements for further improvement of the comfort evaluation system. At the same time, it established methods of evaluation of large-scale rotating machinery on the comfort of load building and provided a reference basis for quantitative criteria.According to the actual situation of the load building, a systematic study is made on the measures of the floor’s vibration isolation. Through taking the vibration of the columns which support the pelletizer as excitory input, building finite element model, and using a floating floor as damping measures to simulate, this paper present an effective method of vibration reduction. This method has reached the standards for comfort of the load building’s floor, satisfied the comfort requirements of staff and provided scientific basis for the second phase of the construction of Wuhan Iron and Steel Mining Co., Ltd.更多还原