节点文献
复杂机电系统机电耦合分析与解耦控制技术
Analysis of Electromechanical Coupling Facts and Decoupling Control Strategies for Complex Electromechanical Systems
【作者】 贺建军;
【导师】 喻寿益;
【作者基本信息】 中南大学 , 机械电子工程, 2004, 博士
【摘要】 论文以“复杂机电系统机电耦合分析与解耦控制技术”为主题,研究了复杂机电系统非线性电磁谐波分量与主机构运动的动态耦合关系;分析了机电耦合的类型、特点和作用规律;提出了机电耦合模型的建立方法;探求了系统参数最优匹配方法;开发了复杂机电系统解耦控制技术;实现了冷轧带钢平整机张力-速度的解耦控制。主要内容如下: 对典型复杂机电系统——CM04平整机的机电传动部分作了全面地调研。提出了系统的总体测试方案,设计了高保真度的信号隔离和缓冲器,解决了从大功率、大电流复杂机电系统主回路中提取真实谐波分量、而又能不影响系统正常运行的难题;发现了谐波电流对平整机工作状态的影响规律;确认谐波电流是可能引起平整机工况奇异、产生机电耦合振动、导致带材产品表面出现明暗振纹的一个因素;提出了一种多相整流电路谐波电流的分析计算方法,用该方法对平整机三相桥式整流供电电路的谐波分量进行了理论分析和计算,计算结果与实测结果能得到相互验证,说明该方法是可行的,实测结果是可靠的。 提出了复杂机电系统机电耦合模型的建立方法。复杂机电系统都是由机电元件组成的,再复杂的机电系统都能被分解成简单的机电元件或等效机电元件,将机电元件按照系统的组成原理进行正确的连接,写出各机、电回路方程;对于四辊或多辊冷轧机、连轧机列等一类的复杂机电系统,先用机理建模方法得到各个机电传动子系统的数学模型;再通过传动对象之间的耦合参数和物理量将各个子系统耦合起来,建立整个机电传动系统的数学模型;最后根据时间尺度原则,在保留原系统模型的主导特征值和重要的状态变量的前提下,对模型进行降阶处理,得到系统可实用的数学模型;通过对CM04平整机机电传动系统耦合模型的建立和仿真实验,验证了该建模方法的有效性和可行性。 分析了复杂机电系统中存在的机电耦合行为和作用规律,归纳总结出四种形式的机电耦合,即电磁转矩直接耦合,谐波转矩直接耦合,微变量控制回路耦合,多传动加工件耦合。其中,电磁转矩直接耦合是通过电动机定子一转子之间电磁场相互作用产生机电能量转换而实现的,它是机电系统中实现系统设计功能的基本形式,也是机电系统普遍存在的主要物理过程;谐波转矩直接耦合会对系统的主运动产生扰动,影响系统的主运动形态;控制变量对于主体运动是另一种扰动,与主体运动的力学参数构成参数弱变量耦合形式;电枢回路中的谐波电流经过反馈回路进入控制系统中,通过控制回路的“放大”后,再作用于机械主体运动,它对机械主体运动的影响可能比谐波电流直接耦合对主传动产生的影响还要大;通过对多传动复杂机电系统耦合模型的建立发现了机电参数耦合的实质和途径,得到了机电参数耦合矩阵的统一结构形式。 分析了遗传算法和模拟退火算法寻优性能和各自的不足与局限性,将两种算法巧妙地结合起来加以修正、取长补短构成新的退火遗传算法,提高了算法的整体寻优性能:并以典型的复杂机电系统一平整机为研究对象,以其主传动速度控制系统为切入点,应用退火遗传算法,解决了复杂机电系统中多参数、相互祸合、难于整定和组态的难题,使控制系统的性能指标在参数可调范围内达到最优。 开发了复杂机电系统智能解祸控制技术。针对复杂机电系统非线性、模型未知、参数时变、多重网络式祸合的特点,提出了基于退火遗传算法前馈解祸控制方法、基于退火遗传算法的输出反馈解藕控制方法和基于神经网络的快速自适应解祸控制方法。详细讨论了算法的原理和实现;证明了一类复杂机电系统实现输入输出完全解祸的充分必要条件和基于神经网络的快速自适应解祸控制方法能够实现复杂机电系统输入输出的完全解祸;仿真实验也验证了复杂机电系统的智能解祸控制方法是简单的、可行的和鲁棒的,能以任意精度满足控制要求。 实现了典型复杂机电系统—CM04平整机张力一速度的解祸控制。针对平整机张力、速度控制的特点,提出了基于极点配置的解祸控制策略,将相互作用的前张力、后张力和工作辊速度解祸成三个几乎独立的控制参数,并将闭环控制系统的极点配置在所期望的位置,使系统具有良好的动、静态性能。从理论分析、计算机仿真到工业实验都表明该解祸控制策略能实现对平整机工作辊速度、前张力和后张力的解祸控制,明显地降低了三个参数的祸合程度,改善了系统的控制效果,有利于提高带材产品的质量。
【Abstract】 Taking Analysis of electromechanical coupling facts and decoupling control strategies for complex electromechanical systems as a main topic,this doctoral dissertation researches the dynamic coupling relation between nonlinear harmonic current and main motion of complex electromechanical system;analyzes electromechanical coupling types,characteristics and their acting laws;explores coupling modeling method,optimal parameter tuning method and decoupling control methods of complex electromechanical systemjand realizes decoupling control among the forward tension,the backward tension and the main drive speed of a typicalcomplex electromechanical system--CM04 temper mill.The main ideas areexpressed as following:The study investigates thoroughly the electromechanical drive subsystems of thetypical complex electromechanical system--CM04 temper mill.It proposes a totaltesting scheme of the system,designs signal isolating buffers with high fidelity,solves the difficult problem that the real signals of harmonic currents are picked up from the main armature circuit with high voltage and strong current,meantime,the normal work of the system is not affected at all. It finds the law that the harmonic currents act on working states of temper mill,confirms that the harmonic currents are one of the causes which results in singular operating mode,coupling chatter,and leads to shade and light streaks on the surface of strip steel.It puts forward a analyzing and computing method of harmonic current of multiphase rectifying circuit.By this method,we compute theoretically the harmonic currents in the 3-phase-bridge circuit of temper mill.The theoretical results are testified by the measuring values and vice versa.Therefore,the computing method is practical and the measuring values are trustworthy.The study presents a coupling modeling approach of complex electromechanical system.Complex electromechanical system consists of electromechanical components.No matter how complex it is,it can be generally decomposed electromechanical components or equivalent electromechanical components. Connecting correctly these component symbols according to the structural principle of electromechanical system,we can get all dynamic equations of electromechanical loops.Aiming at a class of complex electromechanical systems,such as 4-high or multi-high cold mills and continual mills,we firstly construct mathematical models of every subsystems by means of structural principle modeling method;nextly,get theoverall mathematical model of electromechanical drive system by joining together these subsystems’ models according to coupling variables and parameters; we finally simplify the overall model according to time-scale principle which can retain main eigenvalues and significant state variables, and obtain a practical mathematical model of complex electromechanical system. By simulating the CM04 temper mill’s model constructed by above-mentioned method,we test and verify that the modeling method is effective and practical.The study analyzes and researches the electromechanical coupling facts and acting laws existing in complex electromechanical system.lt summarizes four modes of electromechanical couplings.They are electromagnetic torque direct coupling,harmonic torque direct coupling,microvariable control loop coupling and multi-drive system coupling by workpiece.Among them,the electromagnetic torque direct coupling occurs in electromechanical energy conversion through electromagnetic field between stator and rotor of electric motor,it is basic mode implementing design functions,and it is also a main physic process existing universally in electromechanical system;the harmonic torque direct coupling disturbs main motion states of system;the microvariables are also a disturbance to main motion,they and dynamic parameters form parameter weak coupling;the harmonic currents get into current-adjusting system via current feedback units, and is amplified by current regulator, then act on mechanical main motion again,this coupling action affects more sev
【Key words】 complex electromechanical system; temper mill; electromechanical coupling; harmonic torque; coupling model; time-scale principle; coupling chatter; annealing-genetic algorithm; parameter tuning; neural network; intelligent decoupling; pole assignment; forward tension; backward tension; optimal control;