摘要
双进双出磨煤机是火力电厂中广泛采用的一种制粉设备。它具有能耗低、生产效率高、研磨煤种范围广和不受异物影响等优殿,但是磨煤机运行时也存在一定的问题,有时出力不均匀,功耗变大,甚至发生堵煤和乏煤现象。这是由煤位检测设备精度低造成的,而国外的煤位检测设备虽然检测精度高,但进口价格昂贵。因此自主进行检测设备的研发对于降低整体成本,具有重要意义。
论文首先介绍了双进双出磨煤机的工作原理、特点及传统的煤位检测方法,分析了传统的煤位检测方法的不足,即煤位检测精度较低的问题。为了解决此问题,论文首次提出了压差法与噪声法相结合的分段煤位检测方法。当磨煤机工作在乏煤和堵煤的情况下,噪声法很难准确测量煤位,此时采用压差法进行煤位检测可以取得良好的效果;而当磨煤机工作在正常煤位情况下时,压差法精度降低,不再适合此状态,而此段是噪声法测量精度最高的煤位段,因此采用噪声法。为了提高噪声法正常工作时的检测精度,在原有方法的基础上,对其进行算法的改进。将现场收集的磨煤机的噪声,利用小波变换的方法进行频率分解,并找到反映磨煤机简体内煤位的特征频率段,之后对分解后的特征向量采用智能分析的方法,利用神经网络建立磨煤机煤位的统计模型,最终实现煤位检测。
在优化控制方面,为了使双进双出磨煤机制粉系统在低功耗的前提下,研磨效果理想,出力均匀,需要对磨煤机的煤位进行优化控制,使磨煤机工作在最佳的煤位段。煤位控制方法采用内模控制,有效的抑制了系统滞后时间长、惯性大的问题。内模控制模型由神经网络训练生成,在数据量大的情况下此方法建立系统模型要优于传统的机理建模。最后采用MATLAB进行仿真分析,验证了算法的有效性。
BBD Ball Mill is well used in power plant, and it has advantages of low energy consumption, high efficiency, wide application in grinding mill and stable running. However, when the equipment is running, problems such as occasionally asymmetric power, energy consumption becoming larger, even the phenomena that the equipment is jammed or lacked, exist in the Mill because of low precision of the material level detecting equipment. While foreign equipments have higher precision, but the price is high simultaneously. Therefore, it is very meaningful to develop detecting equipment independently to decrease entire cost. The thesis will study on this question.
Firstly, the thesis introduces working principle, characteristics of BBD Ball Mill and traditional coal level measuring method, then analyses shortage of low precision of traditional coal level measurement. Therefore, the thesis put forward the subsection measuring method through pressure difference method mixed with noise measuring method. When the Mill works under condition that coal is lacked or jammed, it's difficult for noise method to measure the coal level exactly. And pressure difference measuring method can get well result under this condition. But when the Mill works under the condition that coal level is normal, pressure difference method precision decreases and isn't appropriate for the condition. As in normal level, the noise method gets the highest precision. Therefore, the noise method is adopted. In order to acquire better control effect, the noise measure method needs to be improved. The wavelet packet transformation, which is used to decompose the noise frequency, is adopted to process the Mill noise from the field to find characteristic frequency segment inside the Mill. After that, the decomposed eigenvector is processed by intelligent analysis method, and statistics model of the coal level is built using nerve network. Finally, coal level measurement is achieved.
Secondly, in optimum control part, for the consideration of low power consumption, the mill pulverizing optimum control system of BBD Ball Mill is presented in the thesis to guarantee the asymmetric power and the machine to work under best coal level. The coal level control method effectively restrains the shortages of long lagging time and large inertia of the system. Because internal model is obtained by nerve network, and if data are plenty, the system model from above method is better than traditional mechanism modeling. And simulation results demonstrate effectiveness of the method.
引文
[1] 吕权息,汪思源,张翔等.振动信号在球磨机料位监测系统的运用研究.控制系统及其应用,2002,32(3):32-34.
[2] 陶文华,柴天佑,岳恒.钢球磨煤机的动态参数模型与仿真研究.系统仿真学报,2004,16(4):778-780.
[3] 陶文华,岳恒,柴天佑.钢球磨中储式制粉系统得建模与控制.控制工程,2003,10(3):146-150.
[4] 翟廉飞,柴天佑,高忠江等.制粉系统智能解耦控制的分布式仿真实验平台,系统仿真学报,2006,18(7):118-122.
[5] 王东风,宋之平,李遵基.评价电站制粉系统效率的模糊综合评判方法.热能动力工程,2001,16(93):140-145.
[6] 李遵基,王俊蕊,唐歆熙.一种智能型球磨机载煤量测试系统的研究.中国电力,2001,34(3):119—121.
[7] 李遵基,姜萍,梁伟平.球磨机模糊控制在分散控制系统中的应用.华北电力大学学报,2000,27(2):42-46.
[8] 刘长良,梁伟平,李长青.火电厂球磨机制粉系统的自调整模糊控制.中国电机工程学报,2001,21(12):140-143.
[9] 刘长良,梁伟平,董泽.钢球磨煤机制粉系统的递阶模糊控制.动力工程,2002,22(5):150-154.
[10] 王东风.制粉系统球磨机的神经网络预测控制.华北电力大学学报,2001,28(3):124-127.
[11] 王东风,于希宁,宋之平.制粉系统球磨机的动态数学模型及分布式神经网络逆系统控制.中国电机工程学报,2002,22(1):140-144.
[12] 王东风.钢球磨煤机制粉系统的优化控制.动力工程,2002,22(3):150-154.
[13] 尚雪莲,王东风,韩璞.基于微利群优化的球磨机单神经元自适应解耦控制系统.电力科学与工程,2006,3(4):30-34.
[14] 冯健,杨久蓉,周克毅等.中间储仓式制粉系统流体网络数学模型及数值方法.中国电机工程学报,2000,20(1):140-144.
[15] 崔宝侠,苏桂华.采用智能方法对双进双出钢球磨煤机进行优化控制:(硕士论文).沈阳:沈阳工业大学,2006.
[16] 崔宝侠,苏桂华.噪声法在球磨机煤位监测系统中的应用.中南大学学报,2005,36(1):622-625.
[17] 易继锴,侯媛彬.智能控制技术.北京:北京工业大学出版社,1999.
[18] Horikawa Set al. On fuzzy modeling using fuzzy neural networks with BP algorithm. IEEE Trans. Neural Networks, 1992, No. 2.
[19] Jang J S. ANFIS: adaptive-network-based fuzzy inference system. IEEE Trans. System, Man and Cybernet, 1993, No. 3.
[20] Kong S G, Kosko B. Adaptive fuzzy system for backing up a truck-and-trailer. IEEE Trans. Neural networks, 1992, No. 2.
[21] Sahin M. Adaptive robust neural controller for robots. Robotics and Autonomous Systems, 2004, 46 (3): 175-184.
[22] 王孝红,江海鹰,袁铸钢.磨煤机的预测控制,山东建材学院学报,1994,8(2):155-158.
[23] 徐立鸿,冯纯伯.适用于降阶模型的新型多部预测控制算法.控制理论应用,1995,12(4):151-157.
[24] Andre P, Daniel H. A survey of grinding circuit control methods: from decentralized PID controllers to multivariable predictive controllers. Powder Technology, 2000,108:103 - 115.
[25] Allison B J, Ball J B. Constrained model predictive control of blow tank consistency. Control Engineering Practice, 2004, 12:837-845.
[26] Peukert W. Control of aggregation in production and handling of nanparticles. Chemical Engineering and Processing, 2005, 44:245-252.
[27] Ramasamy M. Narayanan, Control of ball mill grinding circuit using model predictive control scheme. Journal of Process Control, 2005, 15:273-283.
[28] Radhakrishnan V.R. Model based supervisory control of a ball mill grinding circuit. Journal of Process Control, 1999, 9:195-211.
[29] 梁庾,白焰,李文.神经元解耦控制方案在火电厂球磨FCS中的设计.计算机测量与控制,2004,12(1):150-154.
[30] 周洪,钟明慧.球磨机智能解耦控制系统的一种实现方法.华东电力,2003,6(4):7-11.
[31] 瞿瞾.神经元解耦模糊控制器在球磨机控制系统中的应用.武汉大学学报,2004,37(1):125-128.
[32] Yao X. Evolving Artificial Neural Network. Proceedings of the IEEE, 1999, 9(9):1423-1439.
[33] Topalov h V, Kaynak O. Neural network modeling and control of cement mills using a variable structure systems theory based on-line learning mechanism. Journal of Process Control, 2004, 14:581-589.
[34] Talebi H h, Khorasani k. Neural network based control schemes for flexible-link manipulators: simulations and experiments. Neural Networks, 1998, 11 (7): 1357-1377.
[35] Callant S I. Connectionist expert systems. Communication of ACM, 1998, 31:152-196.
[36] Wu C J. Huang C H. h neural network controller with PID compensation for trajectory tracking of robotic manipulators. Journal of The Franklin Institute, 1996, 333 (4): 523-537.
[37] 林新田,吴志雄.直吹式双进双出球磨机自动控制系统分析和改进优化.热力发电,2006,6(40):40-43.
[38] 李兆吉.双进双出钢球磨煤机半直吹制粉系统运行分析.山西电力技术,1994,14(2):24-25.
[39] 曲守平,赵登峰,宋协春等.双进双出钢球磨煤机的煤位监测技术.矿山机械,2000,28(286):6-9.
[40] 刘建华.BBD4060型双进双出磨煤机调试及运行特性分析.电力学报,2006,21(2):238-240.
[41] 陈钢.压差及电耳在钢球磨中的应用.电站辅机,2005,3(1):46-48。
[42] 张怡强.功率法对磨煤机煤位的控制与应用.华东电力,2001,7(2):48-49.
[43] 陈荐.钢球磨煤机噪声控制技术.北京:中国电力出版社,2002.
[44] 李辉,宋智勇,孙丰瑞.基于小波包包络分析的故障特征提取方法.振动、测量与诊断,2003,23(4):291-295.
[45] 李萌,陆爽,陈岱民.基于小波神经网络的滚动轴承智能故障诊断系统.仪器仪表学报,2005,26(8):609-610.
[46] 吕琛,王桂增,叶昊等.基于噪声小波包络谱的主轴承磨损故障诊断.中南工业大学学报,2003,34(4):459—462.
[47] 周海平,周建新,王志明.钢球磨中间仓储式制粉系统的计算机优化控制技术和应用.中国电力,1994(9):27-31.
[48] 李春香,钟碧良,毛宗源.基于神经网络实现的PID控制器。石油化工高等学校学报,1999,12(2):78-81.
[49] Yicken K, Celal B. Model reference based neural network adaptive controller. ISA Transactions, 1998, 37 (1):21-39.
[50] 刘荣,吕震中.基于内模PID控制的球磨机负荷控制系统的设计.电力设备,2005,6(1):30-33.
[51] 王东风,王剑东.一种多变量系统得内模解耦控制设计方法.控制工程,2003,10(5):463—465.