Two degrees of freedom control of a multistage power-to-methanol reactor

Power-to-methanol processes use green hydrogen, which is generated by electrolysis using regenerative energy, e.g. wind or solar energy. In this paper a novel control concept is proposed to handle fluctuations in the hydrogen feed due to unavoidable fluctuations in the energy supply. Focus is on a robust multistage reactor, with variable feed distribution as additional degrees of freedom. The controller uses dynamic optimization with a hybrid model for feedforward control of the feed distribution and simple PI control of the total carbon feed to compensate plant model mismatch and unforeseen disturbances. The hybrid model combines modeling from first principles with a neural network to capture the influence of catalyst dynamics on the reaction rates. The concept is validated with a simulation study using a detailed reference model. • A novel control strategy for a multistage power-to-methanol reactor is proposed. • Feedforward control is done by optimization of a hybrid physics and data based model. • Additional feedback control is applied to compensate plant model mismatch. • Systematic in silico testing shows that the method is robust and efficient.