Abstract: In order to solve the turning problem of PID controller parameters, an automatic tuning method based on Internal Model Control (IMC) algorithm and system identification was proposed. In this approach, an identification method based on the open-loop unit step response was employed. The input/output data during the transient process were used to obtain a First Order Plus Dead Time (FOPDT) or Second Order Plus Dead Time (SOPDT) model. Then the parameters of PID controller were determined by IMC algorithm. As to the determination of the IMC filter parameter λ, two parameters, γ and σ were introduced in this method. Then the parameter λ was determined by the relationship between the square of output error and the two parameters above. In the simulation experiments, compared with the traditional IMC based PID controller, the Integral Absolute Error (IAE) index can be improved by about 20% without the input disturbance, while the index can be improved by about 10% with disturbance. The simulation results show that in the premise of ensuring the system robustness, the proposed algorithm not only improves the speed of the transient response, but also effectively restrains the overshoot of the system output.

Key words: Internal Model Control (IMC), system identification, Proportional-Integral-Derivative (PID) control, automatic tuning

摘要： 为解决传统工业控制中比例-积分-微分(PID)控制器参数整定的问题,提出了一种基于内模法(IMC)以及系统辨识的控制器参数确定算法。该方法首先利用被控过程在开环阶跃信号激励下,输入与暂态输出的对应关系,将被控对象辨识为一阶加滞后(FOPDT)或二阶加时滞(SOPDT)的模型;再利用IMC算法确定控制器的参数。对于在内模法中引入的滤波器参数λ的确定问题,提出通过引入γ和σ两个参数,并与输出误差的平方建立关系来确定λ 的方法。仿真显示,对于输出误差绝对值之和(IAE)这个指标,该种算法与传统基于IMC的PID控制算法相比,在无输入扰动时可提高20%左右,在有输入扰动时可提高10%左右。仿真结果表明:在用单位阶跃信号激励系统时,提出的整定方法在保证了系统鲁棒性的前提下,提高了系统的瞬态响应速度,并有效抑制了系统输出的超调。

关键词: 内模控制, 系统辨识, 比例-积分-微分控制, 自整定