Model predictive control of hybrid thermal energy systems in transport refrigeration

Abstract

A predictive control scheme is designed to control a transport refrigeration system, such as a delivery truck, that includes a vapor compression cycle configured in parallel with a thermal energy storage (TES) unit. A novel approach to TES utilization is introduced and is based on the current and future estimate of the vehicle driving state and load prediction. This assumes vehicle communications are aware of the traffic state along the prescribed delivery route. For the test case under consideration, this paper first shows that a 17% savings in energy use is achieved for charging the TES by simply shifting the charging to the time when vehicle is moving above a threshold speed. Subsequently, a cascade control structure is proposed consisting of (i) an outer loop controller that schedules the TES charging profile using a receding horizon optimization, and (ii) an inner loop model predictive controller (MPC) which regulates the TES state of charge while maximizing a derived efficiency factor. For the test case under consideration, and utilizing a specifically derived performance metric, the cascaded control structure shows a 22% improvement over a baseline logic-based controller that focuses on state of charge regulation for the TES. A detailed nonlinear dynamical simulation tool for thermal system development is employed for control implementations.