论文部分内容阅读
化工过程中,由于外部环境变化等因素引起的模型失配及不可测扰动会对预测控制效果带来一定的影响。对化工过程施加预测控制时,预测控制中通常采用反馈校正的方式来解决此类问题,即直接将预测输出与实际输出的差值作为校正量补偿到预测控制算法中。其不足之处是当模型失配程度增大时,无法实现良好的控制。鉴于状态估计能够利用实际测量值来估计系统状态的优点,本文提出一种卡尔曼滤波校正的预测控制方法,通过分析模型内在机制来补偿上述问题对控制效果造成的影响。该方法首先将底层回路和被控对象当作广义对象模型,预测控制器直接控制广义对象模型;然后将模型参数的变化等效成扰动,通过分析得到扰动的统计特性后,用卡尔曼滤波方法估计模型参数发生变化后系统的真实状态;最后将估计状态代入预测控制算法的优化求解中,实现系统的优化控制。仿真实例验证了该方法的有效性。
In the process of chemical industry, the model mismatch and unmeasurable disturbance due to the change of external environment will bring a certain influence on the predictive control effect. When predictive control is applied to chemical processes, feedback correction is usually used to solve such problems in predictive control. That is, the difference between predictive output and actual output is directly compensated for as predictive control algorithm. The disadvantage is that good control can not be achieved when the degree of model mismatch increases. In view of the fact that state estimation can use the actual measured values to estimate the state of the system, this paper presents a method of predictive control based on Kalman filter. By analyzing the inherent mechanism of the model, the effect of the above problems on the control effect is compensated. In this method, the bottom loop and the controlled object are regarded as the generalized object model. The predictive controller directly controls the generalized object model. Then the change of the model parameters is equivalent to perturbation. By analyzing the statistical properties of the disturbance, the Kalman filter method Estimating the true state of the system after the parameters of the model change. Finally, the estimated state is substituted into the optimal solution of the predictive control algorithm to realize the optimal control of the system. Simulation results show the effectiveness of the method.