基于梯度信息的实时优化与控制集成策略

针对工业过程的最优操作问题，分析过程系统的层次模型，提出实时优化与控制集成的级联结构. 对于优化层采用基于梯度信息的稳态实时优化方法，通过在线采集过程测量值，估计过程的梯度信息，更新设定值. 不需要使用显式的过程模型，可以有效抑制模型失配对优化目标的影响. 利用最小二乘的思想求解梯度向量，降低计算成本，可应用于大规模工业过程的稳态实时优化. 提出非线性过程中被控变量的选取方法，利用非线性模型计算平均损失，优化效果具有全局性. 为了快速求解非线性规划问题，对某些条件进行合理假设，从而获得次优解，给出求解被控变量的解析方法，提高计算效率，同时将优化层与控制层联系起来. 通过对数值算例、蒸发过程和放热反应过程的研究，验证所提出方法的有效性.

Gradient information-based strategy for real time optimization and control integration

The hierarchy model of process system was analyzed, and the cascade structure of real time optimization and control integration was proposed, aiming at the optimal operation problem for industrial process. Gradient information-based steady state real time optimization approach was installed in the optimization layer. The setpoint was updated by collecting measurements online and estimating the gradient information of process. The proposed approach can effectively suppress the impact of plant model mismatch on optimization objective, since it avoided using an explicit process model. A least square thought was introduced to compute the gradient vector. The proposed algorithm not only had a low computational burden, but also can be applied to the steady state real time optimization of large-scale industrial processes. A method for selecting controlled variables of nonlinear process was discussed. The proposed method minimized the global average loss based on the nonlinear model. Reasonable assumptions were made for some conditions, so that the suboptimal solution was obtained, in order to solve the nonlinear programming problem efficiently. The analytical solution of controlled variables was given and the calculation efficiency was improved as well as the optimization layer was connected with the control layer. A numerical example, evaporation process and exothermic reaction process were studied to illustrate the effectiveness of the proposed method.