【作者】 张立岩；

【导师】 柴天佑；

【作者基本信息】 东北大学 ， 控制理论与控制工程， 2010， 博士

【摘要】 中国虽然有丰富的铝土矿,但铝土矿的铝硅比普遍比较低,广泛采用烧结工艺来生产氧化铝。在烧结法氧化铝生产过程中,氧化铝回转窑制粉系统的作用是使用回转窑熟料烧结过程中产生的余热来制备细度和水分合格的煤粉,然后将煤粉供给回转窑熟料烧结使用。在制粉过程中,保证煤粉的细度和水分合格,可以提高煤粉燃烧热效率、降低烧结煤耗；提高制粉出力可以提高制粉效率、降低制粉电耗,因此,在保证系统安全、稳定运行的条件下,将煤粉细度和水分控制在工艺设定范围内,并且使磨机具有尽可能大的出力,对于提高氧化铝的产品质量和生产效率具有重要意义。煤粉细度工艺指标难以在线检测,它与磨机负荷之间具有强非线性,受热风温度、原煤性质、钢球负载等因素影响,难以建立精确数学模型；磨机负荷和磨机出口温度与给煤机转速之间具有强非线性,且受热风温度的影响；同时用于制粉系统的热风温度低且变化频繁,因此难以采用已有的控制方法实现磨机负荷的自动控制和煤粉细度的控制。目前氧化铝回转窑制粉过程处于人工控制状态,当原煤性质和热风温度频繁变化时,操作员难以及时准确地判断运行工况,并对制粉过程进行调整,常常造成煤粉包裹钢球和衬板以及煤粉堵塞磨机的故障发生,导致停磨；煤粉细度指标不合格,且波动较大；制粉能耗高等问题。本文依托《复杂工业过程能耗指标的智能优化控制技术及应用》的国家863计划课题,以实现在保证系统安全、稳定运行的条件下,将煤粉细度、水分稳定控制在工艺要求的范围内,同时尽可能提高磨机出力为控制目标,开展了氧化铝回转窑制粉过程智能控制系统的研究,取得了如下研究成果：提出了由磨机负荷智能设定层和过程控制层两层结构组成的氧化铝回转窑制粉过程智能控制方法；设计和开发了实现上述控制方法的由硬件平台、软件平台和智能控制软件组成的回转窑制粉过程智能控制系统；并在某铝厂制粉车间进行安装、调试、运行,取得了显著的应用效果。本文的主要研究工作如下：1.首先分析了煤粉细度与磨机负荷之间的动态特性以及磨机出口温度和磨机负荷与给煤机转速之间的动态特性；并针对该过程的控制难点和问题,提出了由磨机负荷智能设定层和过程控制层组成的氧化铝回转窑制粉过程智能控制方法。负荷设定层根据煤粉细度控制目标,给出磨机负荷设定值；过程控制层通过磨机负荷控制,调节给煤机转速使磨机负荷跟踪设定层给出的负荷设定值,从而实现煤粉细度工艺指标的控制。(1)提出了由磨机负荷智能预设定模型,基于最小二乘-支持向量机的煤粉细度软测量模型和基于模糊推理的反馈补偿模型组成磨机负荷智能设定方法。负荷预设定模型由预设定主模型和钢球磨损补偿模型组成,根据当前原煤水分、原煤细度、热风温度、以及取样煤粉的细度测定分析值和取样时刻的负荷设定值、原煤水分、原煤细度和热风温度,得到磨机负荷的预设定值；煤粉细度软测量模型根据给煤机转速、热风温度、热风温度、磨机进出口负压和布袋收尘器进出口负压,估计煤粉细度；反馈补偿模型根据煤粉细度软测量估计和煤粉细度目标的偏差,采用模糊推理技术调整负荷设定值。(2)提出了由切换机制、磨机负荷和磨机出口温度协调控制器、磨机负荷PI控制器和过负荷控制器组成的磨机负荷智能切换控制方法。切换机制采用基于规则推理的方法识别当前系统工况。当磨机处于低温工况时,切换到负荷与温度协调控制器,将磨机负荷和磨机出口温度控制在工艺规定的范围内；否则识别磨机负荷工况,在过负荷时采用基于规则推理的过负荷控制器,避免出现煤粉堵塞磨机的故障,在正常负荷采用负荷PI控制器,将磨机负荷控制在由设定层给出的负荷设定值上。2.研制了实现上述智能控制方法的制粉过程智能控制系统,由硬件平台、软件平台和智能控制软件组成。其中硬件平台由PLC控制系统、监控计算机、检测装置、执行机构以及相关网络组成；软件平台采用过程控制组态软件；智能控制软件包括磨机负荷智能设定软件、负荷控制切换软件、过程控制软件和监控软件。3.将上述研制的制粉过程智能控制系统应用到某铝厂制粉车间。长期应用结果表明,采用负荷切换控制方法,在煤粉低温工况时,能够将磨机负荷和出口温度控制在工艺规定的范围内,在正常工况下可以将磨机负荷控制在设定层给出的负荷设定值上,在运行期间内避免了由于煤粉低温、过负荷造成故障停磨。同时采用磨机负荷智能设定方法,在边界条件频繁变化的条件下,可以将煤粉细度控制在工艺设定的范围内,且尽可能提高磨机出力。制粉系统能耗统计结果表明与人工控制相比,制粉单耗降低了4.56%。更多还原

【Abstract】 Although bauxite ore is very rich in China, the alumina-silica ratio of the ore is generally low. Therefore, the alumina sintering process technology has widely been used to produce alumina. The pulverizing system of alumina sintering rotary kiln is utilized to pulverize raw coal into coal powder and make the particle size and moisture content within their required ranges, with the remaining heat produced by cooling clinker. In pulverizing process, qualifications of the particle size of coal powder and moisture content can increase the heat efficiency of combustion and decrease the assumption of coal powder. Moreover, increase of the mill output is very important in terms of improvement in pulverizing efficiency and reduction of electrical energy assumption. Thus, it is of importance and significance for improving the alumina quality and the production efficience to control the particle size and moisture content of the coal powder within their required ranges and in the meanwhile to maximize the mill output.In pulverizing process, the index of particle size of coal powder can not be directly measured online, and has the complicated relations with the mill load, the temperature of hot air, the moisture content of raw coal, the particle size of raw coal and ball loads in mill; relationship among mill loads, the temperature of mill outlet and the speed of coal feeder is of a strong nonlinear nature, which is also affected by the temperature of remaining heat. Moreover, the remaining heat of pulverizing process is low and varies frequently, thus the exsiting control methods are difficult to control the index of particle size and mill load.In industrial process, manual operation on pulverizing process is mainly used. However, on-site operators cannot accurately recognize the operatioanl conditions, and can not dertermin the coal feeding in time. Such a manual control can lead to possible operational faults such as the "mill blockage" and the "powder-enwrapped balls". Moreover, for the manual control, the particle size of the coal powder fluctuates violently as the boundary conditions vary, such that the particle size of coal powder is unqualified and the electric energy assumption is large.The work conducted in this thesis was supported by the National 863 High Technology Program project of "the intelligent optimal control technology and its application of energy assumption index for complex industry process". The objective of this thesis is to control the particle size and moisture content within their required ranges and in the meantime to maximize the mill output. This thesis focuses on the research on the intelligent control system for the pulverizing process of alumina sintering rotary kiln. The achievements are as follows: based on the dynamic character analysis of this system, an intelligent operational control method which consists of an intelligent setting layer of mill load and a process control layer is proposed; an intelligent control system which includes hardware platform, software platform and intelligent control software has been successfully applied in a plant in China. The main research of this paper concludes the following contributions:1. The dynamic characteristics between the particle size of coal powder and mill load, and the dynamic characteristics between the mill load, temperature of mill outlet and the coal feeding are first analyzed separately. Then, an intelligent control method of pulverizing process of alumina sintering rotary kiln has been proposed, which includes a mill load intelligent setting layer and a process control layer. The intelligent setting control method for mill load produces a setpoint of mill load, according to the target of the particle size of coal powder. The process control layer using the mill load switching control method makes the mill load to track its setpoint by adjusting the speed of coal feeder, thus this index can be controlled.(1) An intelligent setting control method for mill load has been presented to make the particle size of coal powder within its required range and to maximize the mill output. This control method includes an intelligent setting model for mill load, a soft measurement model based on the least square-support vector machines (LS-SVMs), and a compensation model based on fuzzy rules. The intelligent setting model consists of a pre-setting model and a compensation model for the variation of iron ball load. This model realizes the setting of the mill load by combining the multi-model technique, the rule based reasoning and the compensation method. The soft measurement model estimates the particle size of the coal powder according to the auxiliary variable using the LS-SVMs method. The feedforward compensation model modifies the setpoint of mill load using the error between the output of soft measurement model and the target of the particle size.(2) An intelligent switching control method of mill load has been proposed to control the mill load and the temperature of mill outlet within their target ranges, and to make the mill operational noise represent the mill load at the optimal setpoint. It consists of a switching mechanism, a coordinate controller for the mill load and the temperature of mill outlet, a PI controller for the mill load, and a rule based overload controller. First, a rule based reasoning method is used to recognize the system conditions. The three following controllers are designed to cope with the three different conditions in a switching manner, respectively. Detailedly, the coordinate controller for mill load and the temperature of mill outlet is implemented under low the temperature condition. The rule based reasoning overload controller is used under the overload condition, and the PI controller is used under the normal condition.2. The intelligent control system has been designed and developed for the pulverizing system of alumina sintering rotary kiln. It includes hardware platform, software platform, and intelligent control softwares. The hardware platform includes the measurement instruments, the transmitters, the actuators, the supervision computer and touch screen, etc. The software platform is based on the configuration software of process control. The softwares including the intelligent setting control software for mill load, the process control software, the switching control softeware for mill load, and monitoring software are designed and developed to implement the above mentioned control methods.3. The intelligent control system has been applied in a pulverizing process of alumina sintering rotary kiln. This application in practice industry shows that the switching control method copes with the automatic control of mill load, and makes the mill load and the temperature of mill outlet within their required ranges, and thus avoids the faults such as "mill blockage" and "powder enwrapped ball" caused by low temperature of coal powder or overload conditions. And the intelligent setting control for mill load makes the particle size within its required range, and the electric energy assumption reduced by 4.56% compared with the manual operation.The successful application has shown that the proposed intelligent operational control method and the designed control system have important and extensive applications for other pulverizing processes. Also, it can be as a reference for the operational control of complex industrial process which can hardly be described using any accurate mathematical models.更多还原