An experimentally validated, energy focused, optimal control strategy for an Organic Rankine Cycle waste heat recovery system

This paper presents a Nonlinear Model Predictive Controller (NMPC) designed to provide optimal control input for maximum turbine power generation in an Organic Rankine Cycle (ORC) Waste Heat Recovery (WHR) system. While the literature is rich in ORC-WHR system modeling and control approaches in simulation environments, the fundamental dynamic analysis, system aging, thermal inertia, and experimental implementation of power optimization based optimal ORC-WHR control are still lacking. These factors are key to fully understanding and controlling the dynamic behavior of the system and are the main focus of this study. In contrast to prior literature, this work experimentally evaluates the nonlinear dynamics of the ORC system to comprehensively understand the controller design requirements. A power optimization-based Nonlinear Model Predictive Controller (NMPC) is derived utilizing an Extended Kalman Filter (EKF) as a state estimator. Simulation results indicate that optimal turbine power generation is obtained with minimal working fluid superheat for the system under study. Consequently, a superheat-tracking controller is designed, and the performance of the controller is simulated over step inputs. The designed controller is then experimentally validated on an ORC test rig with a 13L Heavy Duty Diesel Engine (HDDE). During experimental evaluation of the controller, it was discovered that the control-oriented model is susceptible to system aging effects and therefore, the model was calibrated online to match the behavior of the aged system. Moreover, evaporator thermal inertia was found to play a vital role attenuating the fluctuating frequency components of the exhaust conditions. The tuned controller provided satisfactory control response for transient engine conditions and maintained the working fluid temperature within acceptable limits.