改进的模型参考自适应在磨矿过程中的应用

Improved Model Reference Adaptive and Its Application in Grinding Process

针对磨矿过程中一段球磨机与分级机之间存在的多变量、强耦合、参数时变等特性,将其数学模型解耦后,提出一种以差分进化算法在线调整自适应增益的模型参考自适应控制方法,因返砂和给水等因素给被控对象模型带来不确定性,通过差分进化不断在线调整自适应增益,以克服磨矿运行过程参数变化和各种干扰对系统的影响。仿真结果表明,与以梯度法为自适应律的模型参考控制效果相比,该控制方法具有超调小、响应时间快、鲁棒性能好等特点,实现分级机溢流浓度的稳定控制,证明基于差分进化算法的模型参考自适应控制方法的有效性和实用性。

Aiming at the characteristics between a ball mill and a grader during the grinding process such as multi variable, strong-coupling and parameter time variation, after decoupling the mathematical model, a model reference adaptive control （MRAC） method for online automatic gain adjustment by differential evolution algorithm is proposed. Due to return sand and water supply and other factors bringing uncertainty to the controlled object model, it needs to adjust constantly adaptation gain online through the differential evolution so as to overcome the effects of parameter variations in mine operation process and various kinds of interference to system. Simulation results show that the control method performs a smaller overshoot, quicker responsivity and better robustness than the model reference adaptive control with gradient method; the stable control of the grader overflow concentration is achieved; and the model reference adaptive control method with differential evolution algorithm is effective and practical.