Optimizing the Open-and Closed-Loop Operations of a Batch Reactor

The purpose of this paper is to optimize the function of a batch reactor operating in open and closed loops. A simultaneous optimization method is employed. This optimization method provides optimal solution at a finite number of points in time. However, the actual batch reactor operation may deteriorate if the computed optimal input is improperly applied to the process. In the open-loop operation, an appropriate wave-shaping method is employed to modify the computed optimal discrete input profile to achieve an improved batch operation while simultaneously guaranteeing the enforcements of the path and end-point constraints. In the closed-loop operation, to improve the disturbance rejection capability of the reactor, two optimizer structures based on the nonlinear model predictive control (NMPC) are developed. The first structure is founded on a modified version of a conventional approach which utilizes a variable number of finite elements (VNFE). The second optimizer introduced in this paper is a newly developed scheme which is compared against the first approach. This approach makes use of a variable number of collocation points (VNCP). Simulation studies, considering both of the fixed batch-time problem (FBTP) and minimum batch-time problem (MBTP), are presented in this paper. tch reactor, dynamic optimization, path and end-point constraints, nonlinear model predictive control