水泥熟料回转窑燃烧效率软测量研究

发布时间：2018-11-26 16:37

【摘要】：如今,节能与减排成为水泥行业发展的趋势,而水泥生产中能耗最大的设备之一就是回转窑。生产过程中,生料由窑尾送入,在窑体不断的旋转中生料自窑尾向窑头(自上而下)逐渐运动;经磨煤机磨碎的煤粉自窑头向窑内喷入与窑内空气混合,在窑内燃烧形成热力场,提供由生料煅烧为熟料的热量;窑内温度场分布均匀,自窑尾到窑头依次为过渡带、烧成带及冷却带,热量经三种传热方式辐射、对流、传导进行热交换;回转窑能够满足熟料形成过程中不同阶段化学和物理反应的需求;熟料煅烧过程产生的热力场和温度场能够降解有毒废弃物、固化重金属,减少环境污染。目前,回转窑燃烧效率没有现成的在线测量方法,离线测量手段也仅有使用GB/T 26281-2010《水泥回转窑热平衡、热效率、综合能耗计算方法》进行物料统计。以统计手段获得回转窑燃烧效率,既浪费人力物力又费工费时,其大滞后性极大减小了对回转窑优化控制的参考意义,无法及时反应回转窑的运行工况。因此实现回转窑燃烧效率软测量,改变水泥生产过程中人工看火的现状对回转窑后期优化控制和水泥行业节能减排具有重大的意义。本文以吉林亚泰水泥有限公司双阳建材公司2000T/D水泥熟料生产线为研究对象,在深入分析回转窑结构和水泥生产工艺的基础上,结合厂家DCS生产线,初步选出一次风速、二次风温、窑尾压力、生料喂料量、喂煤量、窑尾氧气浓度、窑尾一氧化碳浓度、窑头压力、窑尾烟气温度、分解炉窑尾分解率为软测量辅助变量,回转窑燃烧效率作为主导变量的水泥熟料回转窑燃烧效率软测量原始数据集合。然后使用格罗布斯准则对原始数据集合中存在的粗大误差进行处理,最后对原始数据进行灰色关联度分析,最终确定了辅助变量为一次风速、喂煤量、窑尾氧气浓度、窑尾一氧化碳浓度与窑尾烟气温度。由于回转窑的特点,机理建模不适用于回转窑燃烧效率测量,为此本文分别使用人工神经网络中的BPNN和基于统计学习理论的SVR建立软测量模型,并分别对其进行改进,仿真结果验证了本文回转窑燃烧效率软测量的有效性和可行性。最后选取软测量效果最佳的遗传算法改进SVR的回转窑燃烧效率软测量,使用VC++6.0平台的MFC图形化编程系统、ADO技术和OPC技术开发了回转窑燃烧效率软测量软件,软件现场运行状况良好,符合回转窑燃烧效率软测量需要,具有一定的实用价值。
[Abstract]:Nowadays, energy saving and emission reduction have become the development trend of cement industry, and one of the most energy-consuming equipment in cement production is rotary kiln. In the process of production, raw material is sent from kiln tail to kiln head (from top to bottom) in the continuous rotation of kiln body; The pulverized coal pulverized by the pulverizer is injected into the kiln from the kiln head and mixed with the air in the kiln to form a thermal field in the kiln to provide the heat from raw meal calcination to clinker. The distribution of temperature field in kiln is uniform, from kiln tail to kiln head is transition zone, firing zone and cooling zone, heat is radiated by three heat transfer modes, convection and conduction are carried out heat exchange. The rotary kiln can meet the needs of chemical and physical reactions in different stages of clinker formation, and the thermal field and temperature field of clinker calcination can degrade toxic wastes, solidify heavy metals and reduce environmental pollution. At present, there is no ready-made on-line measurement method for rotary kiln combustion efficiency, and only GB/T 26281-2010 is used to calculate material statistics for cement rotary kiln heat balance, thermal efficiency and comprehensive energy consumption. To obtain the combustion efficiency of rotary kiln by statistical means, it not only wastes manpower and material resources, but also takes time and labor. Its large lag greatly reduces the reference meaning for the optimization control of rotary kiln, and it can not respond to the operating condition of rotary kiln in time. Therefore, it is of great significance to realize the soft measurement of combustion efficiency of rotary kiln and to change the present situation of manual fire watching in cement production process for the optimization control of rotary kiln and the energy saving and emission reduction in cement industry. This paper takes the 2000T/D cement clinker production line of Shuangyang Building Materials Company of Jilin Yatai cement Co., Ltd. as the research object. On the basis of in-depth analysis of the rotary kiln structure and cement production technology, combined with the manufacturer DCS production line, the primary wind speed is preliminarily selected. Secondary air temperature, kiln tail pressure, raw material feed, coal feed, oxygen concentration in kiln tail, carbon monoxide concentration in kiln tail, kiln head pressure, flue gas temperature at kiln tail, decomposition rate of kiln tail are soft sensing auxiliary variables, Cement clinker rotary kiln combustion efficiency as the dominant variable of the original data set of soft measurement of rotary kiln combustion efficiency. Then using the Gro Booth criterion to deal with the gross errors in the original data set, and finally to analyze the grey correlation degree of the original data, and finally determine the auxiliary variables as primary wind speed, coal feed amount, oxygen concentration in the kiln tail, and then use the Gro Booth criterion to deal with the gross errors in the original data set. Carbon monoxide concentration and flue gas temperature in kiln tail. Because of the characteristics of rotary kiln, mechanism modeling is not suitable for measuring combustion efficiency of rotary kiln. In this paper, BPNN in artificial neural network and SVR based on statistical learning theory are used to establish soft sensor model and improve it respectively. The simulation results verify the effectiveness and feasibility of soft measurement of combustion efficiency in rotary kiln. Finally, genetic algorithm, which has the best effect of soft sensing, is selected to improve the combustion efficiency of rotary kiln in SVR. The soft sensing software of rotary kiln combustion efficiency is developed by using MFC graphical programming system based on VC 6.0, ADO technology and OPC technology. The software is running well and meets the need of soft measurement of combustion efficiency in rotary kiln, which has certain practical value.
【学位授予单位】：长春工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TQ172.622

【相似文献】