A comparison of control architectures for controlling a CSTR for acrylic acid production

Abstract

In this work, different control architectures are applied and compared for the control of a CSTR reactor producing acrylic acid. The architectures implemented included: a pure PID structure, a dynamic optimization-based approach and an Economic Nonlinear Model Predictive Control (E-NMPC) implementation. In the E-NMPC, the obtained control sequence minimizes an objective function composed by profit, tracking and penalization movement terms, under process and safety constraints. Results have shown the effectiveness of applying the Economic Nonlinear Model Predictive Controller for economic improvement and control performance of the process.