Comparison of MPC Algorithms for Continuous Stirred Tank Reactor

Reaction processes are widely used in the process industry. Continuous Stirred Tank Reactor (CSTR) is an example of such processes. Due to the fact that CSTR temperature control plays a vital role in system performance, there are lots of control strategies proposed to address this need. There are lots of control strategies. Model Predictive Control (MPC) is a popular strategy in industry and academic research. Some popular MPC algorithms are MAC, DMC, GPC, Adaptive GPC, State Space MPC, Linear MPC, and Explicit MPC. This paper provides a performance comparison between these algorithms using performance characteristics and performance indices. The algorithms are evaluated by conducting several series of simulations which illustrate a better performance achieved by State Space MPC along with disturbance and noise robustness, so State Space MPC is the best choice to implement for CSTR.