Nonlinear State Estimation of Moving Boundary Heat Exchanger Models for Organic Rankine Cycle Waste Heat Recovery

This paper presents results for nonlinear state estimation of a nonlinear, control-oriented Moving Boundary heat exchanger model derived from energy and mass conservation principles. The estimator design assumes pressure and temperature measurements typically available in waste heat recovery (WHR) applications. An Extended Kalman Filter (EKF) and a Fixed-Gain state estimator are developed for an open Organic Rankine Cycle (ORC). The ORC model assumes a nonlinear evaporator dynamic model connected to static expander and throttle valve models. Simulations show that the Fixed-Gain state estimator diverges when initial estimation error is present, and thus is not applicable for the nonlinear model. The EKF provides state estimates regardless of initial estimation error for both the Approximated and Full Jacobians used in the linearization update equations. The estimation error is slightly higher for the Approximated case only at the onset of mass flow rate changes, but shortly converge to zero in both cases. The results suggest the Approximated and Full Jacobians are valid for estimation of a nonlinear ORC in the presence of the examined transient inputs. Furthermore, the results are useful for state feedback control design and heat exchanger performance monitoring.