Parameter Solving of Hydraulic Pressure Control System and Simulation Analysis Based on Improved GA
Abstract:The model of GA had been studied in this paper, to point against the feature of GA, the self-adapting plot had been draw into mapped GA, and the math model had been provided, at the same time, the arithmetic flew had been provided too. Through the self-adapting plot, the probability of cross and variation can automatic adjust by fit degree; the convergence speed and the quality of result had been increased. The parameters of system transfer function had been achieved by this method, and the open loop model had been developed. By comparing the simulation results of the open loop and the actual testing curves of the system, the possibility of using GA to gain the system parameters of hydraulic servo system was proven. While, by analyzing the simulation results, the differences of identification methods are found on material testing machine's proportional and servo pressure control systems.
Keywords:generic algorithm;self-adapting plot;Simulink;control system; system identification
1 引言
准确的数学模型对于有效地控制液压系统来说是至关重要的。由于实际系统复杂的工况因素,往往只能进行闭环辨识,但辨识的模型并不是标准开环或标准闭环的,需要通过求解非线性方程组获得系统参数,这就涉及到非线性方程的求解问题。非线性方程组求解是实际工程领域的一个重要问题[1]。长期以来,人们在此方面进行了大量研究。一些学者对采用遗传算法求解非线性方程组已经有一定的研究,而且证明是有效的方法[2,3]。
本文在隔代映射遗传算法基础上引入自适应策略,使IP_GA中的交叉、变异概率根据适应度大小自动调节,提高了收敛速度及解的质量。将该算法用于对液压系统参数的求解,并利用 Matlab/Simulink工具对其准确性进行了仿真验证。通过比较表明,该方法在辨识比例压力控制系统和伺服压力控制系统中是行之有效的。
2 控制系统模型描述[4]
典型材料试验机控制系统是液压压力控制系统,控制系统如图1所示。
参考文献
[1]罗亚中,周黎妮,唐国金.遗传算法求解非线性方程组的应用研究[J].华南理工大学学报(自然科学版),2003,31(S):140-142.
[2]罗亚中,袁端才,唐国金.求解非线性方程组的混合遗传算法[J].计算力学学报,2005,22(1):109-114.
[3]曾毅.改进的遗传算法在非线性方程组求解中的应用[J].华东交通大学学报,2004,21(4):132-134.
[4]王春行.液压控制系统[M].北京:机械工业出版社,1999.
[5] de Jong K A. An Analysis of the Behavior of a Class of Genetic Adaptive Systems[D]. Ann Arbor, MI: University of Michigan, 1975.
[6]胡小兵,吴树范,江驹.一种基于遗传算法的求代数方程组数值解的新方法[J].控制理论与应用,2002,19(4):507-580.
[7]王小平,曹立明.遗传算—理论、应用与软件实现[M].西安:西安交通大学出版社,2002.
[8]KarrCL,WeckB,FreemanLM. Solutions to Systems of Nonlinear Equations via a Genetic Algorithm[J].Engineering Applications of a Artificial Intelligence,1998,11(3):369-375.