【作者】 时光；

【导师】 吴朋；

【作者基本信息】 重庆大学 ， 控制理论与控制工程， 2011， 博士

【摘要】 调节阀在石油、化工、纺织、冶金、制药等大型过程控制领域具有广泛的应用,调节阀按照驱动方式可以分为电动阀门、气动阀门和液动阀门。气动调节阀由于其结构简单、操作方便、使用可靠、易于维护和防火防爆等优点,在工业中更是得到了广泛的应用。智能电-气阀门定位器是气动调节阀的核心控制附件,可以改善阀门特性、提高控制的精度、速度和增加控制的灵活性。在工业应用中,对流量控制方面的要求越来越高,相应的亦对智能电-气阀门定位器的性能提出了更高的要求。气动执行器作为智能电-气阀门定位器的控制对象具有非线性、时变性和滞后性等特性,而且多应用于工业的多干扰环境中,这使得电-气阀门定位器采用传统的控制算法很难达到更高的性能要求。为了提高智能电-气阀门定位器的控制性能,本文以智能控制为理论基础,以863项目(2006AA040303)“智能化通用仪器仪表的开发及产业化”为应用背景,针对气动执行器的非线性、时变性和滞后性等特性,研究了智能电-气阀门定位器的智能控制策略,本文主要做了以下工作:①论文以气动执行机构为研究对象,分析了气动执行机构存在非线性、时变性与滞后性的原因。针对气动执行机构的特性分析了智能电-气阀门定位器采用模糊PD控制策略与仿人智能控制策略的可行性以及模糊控制策略需要解决的参数优化与控制精度的问题。②针对模糊控制器参数与规则的优化设计问题,结合模糊推理、遗传算法和单纯形法,提出了一种改进的遗传算法——模糊自适应混合遗传算法。该遗传算法以模糊推理为基础,完成对遗传算法中的交叉、变异概率以及反馈式灾变个体数量的自适应调整,并通过单纯形法加强局部寻优能力。仿真结果表明,模糊自适应混合遗传算法较之于标准遗传算法与自适应交叉、变异算子的自适应遗传算法具有更快的收敛速度和更高的寻优精度。模糊自适应混合遗传算法用于对模糊控制器的量化因子与比例因子的寻优,以及对模糊控制器的隶属函数的优化。③针对模糊控制器精度不高的问题,提出了一种CMAC网络与模糊并行控制器的设计方法以提高控制精度。该方法以模糊控制实现反馈控制,CMAC网络实现前馈控制,并采用模糊自适应混合遗传算法完成对模糊控制器中的量化因子与比例因子的寻优。仿真与实验测试结果表明了该并行控制器的有效性。④研究了智能电-气阀门定位器的仿人智能控制策略。针对智能电-气阀门定位器采用的积分分离PID控制算法存在的问题,提出了一种仿人智能控制算法。该算法依据智能电-气阀门定位器的系统自检得到气动调节阀的动态信息和辨识得到的模型信息,查询基于专家经验的参数知识库,完成仿人智能控制器运行控制级参数的整定;依据运行过程中的误差相轨迹和滞后时间等特征信息,结合专家规则,完成仿人智能控制器参数校正级的设计。并对气体泄漏量采取补偿措施,以减少由气体泄漏造成的不利影响。应用测试结果表明,参数自整定的仿人智能控制算法较之原智能电-气阀门定位器采用的积分分离的PID控制算法在调节时间与超调量等性能方面,均有显著提高。通过上述工作,本文研究了CMAC网络与模糊并行控制和仿人智能控制作为基于喷嘴-挡板原理的压电线圈式智能电-气阀门定位器控制策略各自的特点以及实际用中需要解决的问题。实际的应用测试证明了本文所提出的智能控制策略的有效性。更多还原

【Abstract】 Control valve is used extensively in large-scale process control such as petroleum, chemical, textile, metallurgy and pharmaceutical. Control valve can be divided into electric valves, pneumatic valves and hydraulic valves according to its drive mode. Pneumatic valve has been widely used because of its advantages of simple structure, convenient manipulation, reliable operation, easy maintenance, fireproof and explosion-proof. Intelligent valve positioner is the most important accessory of control valve. It can improve the character of valve, enhance the precision and speed of control and increase the flexibly of control.In industrial applications, the requirement of control process is more and more strict; the request of performance towards intelligent electro-pneumatic valve positioner is higher and higher accordingly. The pneumatic actuator as the controlled object has the characteristics of nonlinear, time-variation and time delay and it is widely used in the industrial environment with multi-disturbance. So the electro-pneumatic valve positioner with the traditional control algorithm cannot achieve the higher request of performance.In order to improve the performance of the intelligent electro-pneumatic valve positioner, the dissertation based on the theory of intelligent control has study the intelligent control tactic of the intelligent electro-pneumatic valve positioner, which takes the national“863”project (2006AA040303)“Intelligent general instrumentation development and industrialization”as application background, aiming at the characteristics of nonlinear, time-variation and time delay of the pneumatic actuator.①The dissertation takes the pneumatic actuator as the target, analyses the reason why the pneumatic actuator has the characteristics of nonlinear, time-variation and time delay. Aim at the characteristics of pneumatic actuator, analyses the feasibility of fuzzy control and human-simulated intelligent control tactic which are used in the intelligent electro-pneumatic valve positioner, and the problem of parameters optimization and control precision which should be solved in fuzzy control tactic.②Aim at the problem of control parameters and rule optimization in fuzzy controller, the dissertation presents an improved genetic algorithm. The improved genetic algorithm based on fuzzy inference of genetic algorithm adjusts the crossover and mutation probability and the number of reckoning adaptively, and strengthens local search ability through the simplex method. Convergence performance and convergence rate of the improved genetic algorithm are analyzed and estimated. The simulate results shows that the fuzzy adaptive hybrid genetic algorithm has the better performance of convergence rate and preventing pre-maturity than the standard genetic algorithm and the adaptive genetic algorithm with the adaptive crossover operator and mutation operator. The fuzzy adaptive hybrid genetic algorithm finished the optimization of quantitative factors and proportion factors of fuzzy controller.③Aim at the problem that the fuzzy controller has the low precision, the dissertation proposes CMAC network and fuzzy parallel controller in order to improve the control precision. The algorithm uses fuzzy control to realize the feedback control and CMAC network to realize the feedforword control and determines the quantizing factors of fuzzy control based on fuzzy adaptive hybrid genetic algorithm. The simulation reasults show that CMAC and fuzzy combined control is available.④The human-simulated intelligent control tactic of the intelligent electro-pneumatic valve positioner is studied. Aim at the problem of the integral separation PID control algorithm used in intelligent electro-pneumatic valve positioner, a kind of human-simulated intelligent control algorithm is proposed. This algorithm based on self-test of intelligent electro-pneumatic valve positioner gets the dynamic information and identification model of pneumatic valve and inquires about the parameters knowledge based on expert experience to determine the parameters of operation control level. Human-simulated intelligent parameter-adjusting level is designed according to the running process error phase track, the time lag etc feature information, and the expert rules. Compensation measure for gas leakage is adopted to reduce the harmful effect caused by gas leakage. Application results show that, the parameter self-setting human-simulated intelligent control algorithm compared with the original integral separation PID control algorithm which is used in intelligent electric-pneumatic valve positioner has better performance in regulating time and overshoots.This dissertation studies respective characteristics of the CMAC network and fuzzy parallel control and human-simulated intelligent control as control strategy of piezoelectric-coil intelligent electric-pneumatic valve positioner based on the principle of nozzle–baffle, and the actual use of problems needed to be solved. The actual application test proved intelligent control strategy proposed by the dissertation is feasible.更多还原