基于水质的污水处理过程事件触发模型预测控制

Event-Triggered Model Predictive Control in Wastewater Treatment Process Based on Effluent Quality

研究了将模型预测控制(MPC)应用于污水处理(WWTP)这类大型非线性系统时的求解耗时性和计算复杂性问题,提出了一种新的事件触发控制策略加以解决,模型预测控制器只有在满足相应的触发条件时才进行信息的更新和优化问题的求解。为了保障系统性能,该方法将出水水质融入事件触发逻辑条件的设计中,并引入出水水质系数,用于建立输出变量与出水水质之间的关系,通过这种关系间接反映了WWTP的性能,并根据该性能指标调整控制动作以降低计算成本,同时可以采用不同的出水水质系数增强算法的灵活性。实验结果表明,在保证出水水质和总成本指数基本不受影响的情况下,计算时间可以减少27.1%~42.7%。

The time-consuming and computational complexity of the application of model predictive control(MPC)to large-scale nonlinear systems such as wastewater treatment plants(WWTP)were studied,and a new event-triggered control strategy was proposed to solve the problem,in which the model predictive controller updates the information and optimizes the solution of the problem only when the corresponding trigger conditions are met.In order to ensure the performance of the system,the effluent quality was integrated into the design of the event triggering logic condition,and the effluent water quality coefficient was introduced to establish the relationship between the output variable and the effluent water quality,which indirectly reflected the performance of the WWTP,and the control action was adjusted according to the performance index to reduce the calculation cost,and different effluent water quality coefficients could be used to enhance the flexibility of the algorithm.The experimental results show that the calculation time could be reduced by 27.1%to 42.7%while the effluent quality and total cost index are basically unaffected.