Location: JHE 342
In the first part of this talk, the advanced process control and state estimation for the distributed parameter system, modeled by the partial differential equation (PDE) will be addressed by using the spectral method with Galerkin approximation. The major theoretical contribution of this work is the integration of residual bounds into the controller and observer synthesis such that the closed-loop stability and hard constraint satisfaction are guaranteed. A successful industrial application will also be presented to show the superiority of the proposed method.
In the second part of this talk, three studies for the large scale process optimization under uncertainties will be presented. The first study is the blending problem solved by the chance constrained programming method, which is enhanced by the data analysis technique over the large scale process network. The second study is the crude oil procurement under uncertainties, modeled by the stochastic programming and solved by the nonconvex generalized Benders decomposition (NGBD) approach. The third study is the refinery optimization integrated with a nonlinear crude distillation unit (CDU) model. Through the numerical simulations, we demonstrate that the proposed optimization methods can handle highly nonlinear, nonconvex process models and find a non-conservative solution or the global optimum more efficiently than the conventional approaches and even the state-of-the-art software.
Return to list of all seminars