论文部分内容阅读
热工过程参数辨识是热工自动化领域的重要课题之一。针对热工对象的特点,利用遗传算法全局搜索能力强,单纯形算法收敛速度快、局部搜索能力强的特点,引入遗传-单纯形混合算法对典型热工过程数学模型进行参数辨识。编写Matlab程序进行仿真实验,通过描绘典型阶跃信号输入和其他任意信号输入下的响应曲线,并与已有文献中的辨识结果进行比较来验证参数辨识的精度。仿真实验表明,这种方法有效地解决了单纯形初值设置不当容易导致寻优失败的问题,克服了遗传算法寻优结果随机、分散的缺点,提高了搜索空间的广度和深度,能够得到精确的辨识结果。
Thermal process parameter identification is one of the important topics in the field of thermal automation. According to the characteristics of thermal object, genetic algorithms are used to identify the parameters of a typical thermal process mathematical model by using the advantages of global search ability of genetic algorithm, quick convergence of simplex algorithm and strong local search ability. The Matlab program is programmed to simulate the experiment, and the response curve of typical step signal input and any other signal input is plotted, and the accuracy of parameter identification is verified by comparison with the identification results of the existing literature. Simulation results show that this method effectively solves the problem that the initial value of simplex is not set easily, which leads to the failure of optimization. It overcomes the shortcomings of random and scattered search results and improves the breadth and depth of search space, Identification results.