Model order reduction, Vapor compression cycle, Dynamic modeling, Piecewise-linear approximations
This paper presents a reduced order modeling approach for dynamic vapor compression cycles (VCC) based on a trajectory piecewiselinear (TPWL) approximation and Proper Orthogonal Decomposition (POD). Unlike other model order reduction techniques that apply to nonlinear dynamic heat exchanger (HX) models, the TPWL approach first represents a finite volume HX model with a weighted combination of linearized models along some state trajectories. Linearization points are selected by the kmeans clustering algorithm, which enables control of the number of linearized pieces. Then, each of the full order linear pieces is reduced by projection onto a POD basis to form a reduced order TPWL model. Reduced order HX models were generated and integrated with other quasistatic component models to complete a reduced order VCC model. Simulation results were compared with experimental data for cycle loadchange transients over a wide range of operating conditions. The proposed method is an alternative to a nonlinear model order reduction framework that is presented in a companion paper (Ma et al., 2020) and may be particularly useful for cases where it is challenging to stabilize nonlinear reduced order models. Comparisons were drawn between these two approaches in terms of generalization, stabilization and computational efficiency.