Model reduction and identification for temperature control of the phenol-formaldehyde reaction in batch reactors

This study provides an overall strategy for building an effective model-based control scheme for the highly exothermic phenol-formaldehyde reaction for the production of 1,3,5-methylolphenol, a phenolic resin precursor, carried out in a jacketed batch reactor. A simplified representation of the reactive system is proposed by means of reduced kinetic models, whose unknown parameters are identified with a two-step procedure. The experimental data, used for model identification, have been simulated using a complex reaction network, involving large number of reactions and compounds. The proposed simplified models are adopted to design a model-based feedback control scheme for the temperature control in a batch reactor. The obtained performance is tested in simulation and compared to those obtained via a classical PID controller. According to simulations, it can be stated that the model-based controller always achieves the best performance with respect to the PID controller and the robustness of the control scheme suggests the use of the simplest reduced model without significative performance degradation