Distributed Economic Nonlinear Model Predictive Control for Flexible Electrified Biodiesel Production─Part I: Sequential Architectures

Abstract

We present distributed economic nonlinear model predictive control (DeNMPC), employing sequential communication protocols, to optimize the flexible operation of electrified biodiesel production under fluctuating electricity prices. By incorporating buffer tanks for intermediate and final products within the production process, we decouple process dynamics, segmenting it into three distinct subprocesses. The DeNMPC strategy yields significant energy cost savings of 20% compared to steady-state operation while being real-time tractable and ensuring operational feasibility. This stands in contrast to conventional scheduling employing quasi-stationary models, which results in infeasible outcomes. Furthermore, the DeNMPC strategy handles unexpected disturbances in production demand and feed composition, mitigated by the buffer tanks. These tanks prove essential not only for enhancing operational flexibility but also for enabling realizable DeNMPC applications through system decomposition. Additionally, we extend the DeNMPC with two stability formulations from the literature, assessing their suitability and implications within the specific context of the biodiesel production application.