Optimization of a fluid bed dryer by the implemention of a model predictive controller

Optimization of the drying operation requires a model containing the operating variables for fluid bed dryers (FBD). Modeling of FBD is still uncertain since the drying process is unpredictable and relies on empirical parameters. The objective was to develop a first principle model that would describe the moisture content of a pharmaceutical powder as function of operating variables and use it for optimization purposes. Granulations consisting of lactose monohydrate, pregelatinized starch, and distilled water were dried in a FBD to generate the data to develop the models. Basic energy and material balances led to the description of the drying curve and it was used to determine the parameters necessary to design a customized Model Predictive Controller (MPC).; The simulation shows that by implementing an MPC the desired set point is achieved in half the time and consumes less energy as compared to working at the FBD's maximum settings. Compared to traditional MPC, this customized controller provided a more logical response of the drying process since the traditional strategy is more accurate when it is applied to stable processes. Compared to a Proportional-Integral-Derivative (PID) strategy, the MPC proved to be a better tool for optimizing the drying process.