基于遗传单纯形混合算法的热工过程参数辨识

Parameter Identification Based on Genetic-simplex Hybrid Algorithm of Thermal Processes

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。

Thermodynamic process parameter identification is one of the important subjects of thermal automation.Throughout the global search ability of the genetic algorithm and the high speed of the simplex algorithm and the characteristics of the local search ability,genetic-simplex hybrid algorithm for typical thermal process parameter identification is used to analyze the characteristics of the thermal objects.Based on the Matlab simulation program,a typical order jump signal input and other any signal input response curves are created.Compared with the existing literature,the accuracy of the parameter identification is verified.Simulation results show that this method can effectively solve the improper initial value of the simplex algorithm which can easily cause the optimization failure,overcome the shortcomings of the random and scatter of genetic algorithm optimization,improve the breadth and depth of the search space and provide more accurate identification results.