Experimental Validation of Nonlinear Model Predictive Control for a Heavy-Duty Diesel Engine Waste Heat Recovery System

This paper discusses an experimental validation of a real-time augmented control scheme for an organic Rankine cycle (ORC) waste heat recovery (WHR) system. A nonlinear model predictive control (NMPC) is designed to regulate the working fluid vapor temperature after the evaporator. The NMPC utilizes a six-state reduced order moving boundary (MB) evaporator model. The state estimator is constructed using an extended Kalman filter (EKF) given the working fluid outlet vapor temperature and exhaust gas outlet temperature as measurements. Working fluid evaporation pressure is controlled by an external proportional-integral-derivative (PID) control loop. The experimental validation first compares the performance of the augmented control scheme with that of a traditional multiple loop PID control with a feedforward term over an engine transient. The experimental study shows that the augmented control scheme outperforms the baseline multiloop PID control in both terms tracking error and settling time during transient engine operation. The performance of the augmented control scheme is further validated over three additional transient conditions with alterations to both the engine transient and the working fluid reference temperature. The NMPC validation shows that the working fluid vapor temperature can be controlled within 1% error margin relative to the targeted reference.