【作者】 王介生；

【导师】 王伟；

【作者基本信息】 大连理工大学 ， 控制理论与控制工程， 2006， 博士

【摘要】 链篦机-回转窑氧化球团烧结过程是一个涉及到传质、传热和复杂化学反应的工业过程。从控制角度来看，球团烧结过程具有非线性、分布参数、慢时变和大时滞等特性，属于典型的复杂被控对象。大量不确定信息、多样化海量数据使得传统控制方法难以对其进行有效控制，而智能控制为解决这类复杂被控对象的控制问题提供了有效途径。 本文在查阅大量文献资料、广泛收集链篦机-回转窑氧化球团烧结生产过程历史数据、专家经验和操作规程和对现场工艺进行充分调研的基础上，综合运用球团烧结工艺、计算机控制技术、现代控制理论和人工智能技术等多学科知识，提出了基于球磨机制粉系统多变量解耦控制模型、回转窑温度控制模型和球团矿化学成分软测量模型的球团烧结过程智能控制策略。论文的主要工作如下： (1)针对煤基球团厂球磨机制粉系统的多变量强耦合特性，以球磨机的数学模型为依据，提出一种新型的多变量PID解耦控制策略。它由比例因子α模糊自整定单元、PID控制器和基于对角矩阵法的解耦补偿器组成，PID控制器参数由粒子群算法进行优化。仿真研究表明所建模型和所提控制方法的有效性。 (2)基于粗糙集理论的知识约简方法和T-S模糊神经网络的非线性映射理论，建立了回转窑自动喷煤系统。基于一种新的聚类有效性准则函数的模糊C均值聚类算法对连续属性进行最优离散化；采用粗糙集理论对球团烧结生产过程中一些代表性工艺参数所构成的决策表进行知识约简分析；最后根据最小规则约简建立权值不完全连接的T-S模糊神经网络模型。 (3)结合球团烧结过程工艺机理，采用T-S模糊神经网络建立了球团矿化学成分(FeO含量和碱度R)软测量模型。模型采用灰色关联分析方法进行辅助变量选取，以对软测量模型输入进行降维；粒子群优化算法和最小二乘法相结合的混合算法来调整T-S模型的前提参数和结论参数，该混合学习算法提高了网络参数辨识的收敛速度。然后基于球团矿化学成分软测量模型建立了球团烧结过程状态粗糙集专家控制系统，实现链篦机过程状态的优化控制。 (4)将上述智能建模和优化控制方法在球团烧结过程进行了工业应用实验，结果表明使用智能建模和优化控制方法提高了产品的产量和成品球团矿的质量，对提高球团企业的整体经济效益具有推动作用，同时为复杂工业过程的优化控制提供了实用的方法。更多还原

【Abstract】 Grate-rotary kiln oxide pellets sintering process is a dynamic process with matter transmission, heat exchange and complicated chemical reaction. From the control view, the sintering process is a kind of typical complex object due to its nonlinearity, distributed parameters, time-varying and model uncertainty. A great deal of uncertainty information and multiplex magnanimity datum make it difficult to perform the control task of total sintering process effectively by using conventional control methods. The intelligent control theory provides an efficient approach to realize the control of this kind of complex systems.Based on the review of the related research references, the detailed investigation on the mechanism and technique and the comprehensive collection of production history datum, expert experience and manipulation regulations of pellets sintering process, an intelligent control strategy is proposed by comprehensive utilization of pellets sintering theory, computer control technology, modern control theory and artificial intelligent technique. The strategy includes the multivariable PID decoupling control model of ball milling pulverizing systems, the temperature controller of the rotary kiln process and the soft sensing model of chemical composition of the finished sinter mineral. This dissertation has carried on the following research.(1) A ball mill coal pulverizing system of pelletizing plant is a complex nonlinear multivariable process with strongly coupling and time-delay. A new multivariable PID decoupling controller is proposed based on the mathematics model of pulverizing system, which consists of diagonal matrix method-based decomposition compensatory unit, PID controller and fuzzy self-tuning components unit with scaling factor α . Particle swarm optimization (PSO) algorithm is adopted to optimize parameters of the PID controller. Simulation results show the validity of the obtained model and the proposed control method.(2) Based on the idea of the knowledge reduction of the rough sets (RS) theory and the nonlinearity mapping of Takagi-Sugeno fuzzy neural network (FNN), a kind of RS-FNN control method is presented and applied to the rotary kiln sintering process. The fuzzy c-means (FCM) clustering method based on a new cluster validity index function is used to obtain the optimal discrete values of the continuous attributes. RS theory is adopted to obtain the reductive rules using industrial history datum and corresponding T-S fuzzy model has better topology configuration reflecting system characteristics.更多还原