Dynamic optimization of a continuous Lactide ring-opening polymerization process

An increasing attention has been paid to the production of the polylactic acid (PLA) in recent years, owing to its properties as a biodegradable thermoplastic besides the fact that it is derived from renewable resources. PLA is industrially produced by ring opening polymerization of lactide. This reaction is sensitive to deviations in the operating conditions that highly affect the reaction rate and the polymer properties. Therefore, a process monitoring and control policy is crucial in order to restore the nominal conditions in case of drift. In this paper, a continuous PLA process is considered with three cascade reactors, two tubular and one loop reactor, that is described by partial differential equations. A dynamic optimization strategy is developed to calculate the optimal input flows of the catalyst and co-catalyst in the first reactor to compensate the deviations caused by a sudden change in the contamination level. The optimization objective is to maintain the required properties, mainly the polymer molecular weight and the desired conversion, at the process output. Comparison with classic PI controller showed a better performance of the optimization strategy in terms of off-spec time as it better accounts for coupling of the outputs, non-linearity, delay and process constraints.