Model-Based Evaluation of a Data-Driven Control Strategy: Application to Ibuprofen Crystallization

This work presents a methodology that relies on the application of radial basis functions network (RBF)-based feedback control algorithms to pharmaceutical crystallization process. Within scope model-based evaluation proposed strategy, firstly strategies for data treatment, structure and training methods reflecting possible scenarios in industry (Moving Window, Growing Window Golden Batch strategies) were introduced. was followed by incorporation such RBF within soft sensor nonlinear predictive data-driven application. The performance tested undisturbed cases as well presence disturbances promising results from both demonstrated great potential these techniques In particular, Moving performed best an sensor, outperformed remaining methodologies control.