【作者】 邓胜祥；

【导师】 周孑民； 文敦伟；

【作者基本信息】 中南大学 ， 热能工程， 2004， 博士

【摘要】 石灰的烧成是一门古老的技术,近代以前石灰多用于建筑工程。随着近代化工和冶金工业的兴起,石灰的用途急剧增大,石灰生产技术和设备的开发研究工作才得到重视。石灰生产设备以前多用土窑,近代开始采用机械化立窑,其生产能力和劳动生产率成倍增长而能耗成倍降低,但仍难以满足现代工业快速增长的高效益、低成本、高自动化的要求,这种状况制约着化工、冶金及造纸等相关行业的发展。 石灰的生产是一个连续的过程,石灰炉本身是一个连续的反应器,原料和燃料不断输入,产品和二氧化碳等尾气不断输出。以往产品质量的判断主要靠产品产出后样品的化验,化验的结果固然可靠,但存在严重滞后,不能作为在线决策控制的依据。本文以Φ4×21m石灰竖炉为研究对象,在原有检测系统的基础上,通过增加必要的在线检测手段,充分收集石灰炉生产过程的在线信息,实现了石灰炉的“全息监测”。 通过综合运用炉窑热工及反应动力学原理,开发了石灰炉在线仿真优化模型及软件。在线仿真系统能实时计算石灰石煅烧分解率、出口RO₂（指二氧化碳、二氧化硫等）浓度、“三带”（预热带、煅烧带、冷却带）高度及炉内温度分布,并由计算机在线显示、记录运行历史曲线,为现场的考核与管理提供依据。在“全息监测”与在线仿真的基础上,结合石灰炉工艺的特点与工程技术人员、熟练操作工人的经验,首次建立了石灰炉炉况诊断专家系统,为生产运行和操作提供在线指导。炉况诊断专家系统具有自动报警、故障解释及对策分析等多重功能。 基于神经网络理论,本文开发了具有自组织与自学习功能的生石灰质量预报模型,该系统以历史数据作为训练样本,对神经网络进行离线训练,训练后的网络用在线信息检验发现,生石灰质量的实测值与预报值有很好的拟合程度,平均预报误差为3.2%,为控制操作提供了准确依据。 石灰生产过程中原料成分、含水量、进料速度、焦炭成分等生产条件经常变化,根据传热、传质过程建立的模型中有很多参数必须通过机理分析、假设或大量实验来确定,此外还需要检测难以测量更多还原

【Abstract】 Lime calcining is an old technology. About two century ago lime was only used in building construction. With latter-day’s development of chemical engineering and metallurgy, the application and consumption of lime are largely increased, and people find that it is very important to study lime calcining process and equipment. In old times, lime was mostly produced by medieval kiln, and mechanization limekiln wasn’t used until recent times. With the application of mechanization kiln lime throughput and productivity were doubly increased, and energy expenditure was decreased quickly. But this improvement can’t satisfy the increasing requirement of high benefit, low cost and high automatization for chemical engineering, metallurgy, paper making and correlative industries.Lime production is a sequential processes, limekiln itself is a sequential reactor. Raw material and fuel are steadily input, calcium oxide and carbon dioxide are continuously produced. The estimation of production quality formerly depends on the assay of production sample. As there is a serious lag, so the assay result, which is reliable, can’t be regarded as presumptive evidence of on-line decision-making control. A limekiln with diameter of four meters and height of twenty one meters is studied systematically in this paper. Based on existing detection system, by equipping necessary on-line detect instrument and collecting enough on-line information in the process of lime production, limekiln hologram monitor is achieved.By applying synthetically the principle of kiln thermal engineering and reaction kinetics, on-line simulation and optimization model and software were worked out. The concentration of carbon dioxide in the exit, "three zone" (warm-up zone, calcinations zone, cooling zone) heights and temperature distribution in kiln could be calculated by on-line simulation system, meanwhile the running historical curves were recorded by the computer which are important evidences for company assessment and management. On the basis of hologram monitor andon-line simulation, combining technical characteristic of limekiln with experiences of engineer and specialized worker, Lime Kiln Status Diagnostic Expert System (LKSDES) was established for the first time to offer on-line guidance for operation. LKSDES can provide multiple functions involving auto alarm, fault interpretation and countermeasure analysis.Based on Neural Network(NN) theory, the author devised a prediction model of quicklime quality with self-organization and self-learning functions. In the system training samples were history data, NN were off-line trained and tested by on-line information. The author found that the average forecasting error is 3.2 percent, thus exact information was offered for manipulation.In the process of lime production, some technical conditions always fluctuate, such as the compositions of raw material, the content of water, the feedrate of charge and the compositions of coke. Many parameters of the model, forming in the process of heat and matter transfer, need be made certain by means of mechanism analysis, hypothesis or experiments. Furthermore, all the state variables should be measured for the operation and management. Based on chaos fuzzy theory, the author adopted system identification method, and studied the interrelation between main control variables and output variables by considering limekiln as a "black-box". By referencing operating knowledge, the author worked out a furnace roof temperature fuzzy controller and an intelligent control system of multi-kiln lime discharging. The former dynamically controlled the roof temperature in the suitable range which was confirmed by simulation and optimization in different working conditions. The latter considered synthetically the complexity of a lime discharging system in the plant, the remarkable nonlinearity, the intense coupling between kilns, the difficulty in detecting important parameters, as well as the fact that only one lime belt system was used by multi-kilns. Then an intelligent control system of material height of limekiln was designed. Based on the chaos theory and the artificial intelligence technology, the intelligent control for the process of calces discharging was realized.The creativeness of the article lies in that the mathematical modelingand on-line simulation are combined to realize the multi-mode kiln status diagnosis and intelligent control of limekiln for the first time.更多还原