【作者】 曲星宇；

【导师】 崔宝侠；

【作者基本信息】 沈阳工业大学 ， 系统工程， 2010， 硕士

【摘要】 双进双出磨煤机是在电厂系统中广泛应用的一种制粉设备,它具有能耗低、生产效率高、研磨煤种范围广等优点。而事实上长期以来绝大部分磨煤机系统保守工况运行,制粉系统耗电率相当大。一个主要原因是其料位精确测量难以实现。此外,双进双出磨煤机制粉系统是一个非线性、大延迟、大惯性、强耦合的被控对象,其优化控制非常困难。为保证磨煤机筒内煤位在适当位置,以实现磨煤机经济工况运行,磨煤机的料位检测及优化控制成为电厂磨煤机使用亟待解决的问题。本文针对以上问题展开研究,首先介绍了双进双出磨煤机的基本结构及工作原理,通过大量实验得到双进双出磨煤机工作特性曲线,根据特性曲线对其运行参数及影响因素进行分析。在对双进双出磨煤机机理分析的基础上,提出基于模糊神经网络多数据融合的智能料位检测方法,将多传感器采集的变量参数按照模糊规则进行模糊化处理,并构造神经网络进行数据融合,融合结果即为检测的料位值。本文为提高料位检测准确度,针对构建的一型模糊神经网络多数据融合系统存在的检测准确度不高的问题进行改进,采用二型模糊神经网络多数据融合系统进行料位检测,通过仿真对比实验表明,应用改进方法在测量准确度方面有明显提高,实现料位更为准确的测量。最后,针对制粉系统存在的控制问题,提出基于二型模糊神经网络控制方法,本方法不仅避免了模糊控制应用中控制规则难以人工提取的限制,简化了多变量系统的模糊控制设计,可以实现双进双出磨煤机自动优化控制,具有较强的应用价值。更多还原

【Abstract】 BBD Ball Mill is a widely used milling device in power plant system. It has the advantages of low energy consumption, high efficiency, widely range of grinding coal, and so on. In fact, most of the mills are working at the conservative conditions for a long time, which makes the milling system a considerable power consumption rate. A major reason for this is the difficult to accurately measure the material level. In addition, BBD Ball Mill milling system is a nonlinear, large delay, large inertia and strong coupling charged object, it is very difficult to implement optimization control. In order to ensure the material in coal mill barrel at the appropriate level and achieve the mill operation at economic conditions, mill’s material level detection and optimal control becomes an urgent problem needed to be solved when using coal mill.According to above problems, firstly, this paper introduces the BBD Ball Mill’s basic structure and working principle. Then, the operator characteristic curve of BBD Ball Mill is achieved, through a large number of experiments. At the same time, based on its operating characteristic curve, analyzes the parameters and influencing factors.Based on the mechanism analysis of BBD Ball Mill, intelligent material measure method by fuzzy neural network multi-data fusion is presented in this paper, fuzzes processing by fuzzy rules with the variable parameters of multi-sensor acquisition, constructs neural networks to fusion the data, and then the integration results are the expected material values. The proposed methods can enhance the accuracy of material measure, improve the problems of the low accuracy of type-1 fuzzy neural network multi-data fusion system, and use two-type fuzzy neural network multi-data fusion system to measure the material. By simulation experiments, the results show that, there is a marked rose in measure accuracy through using this improved method, and then a more accurate measurement of material level is obtained.Finally, based on the control system for the milling issue, the two-fuzzy neural network control method is given, in this paper. It can not only avoid limitation of control rules which is difficult to be extracted by manually in the application of fuzzy control, but also simplify the multi-variable fuzzy control system design. Thus, the proposed method has strong application significance.更多还原