A nonlinear reduced-order modeling method for dynamic two-phase flow heat exchanger simulations

Abstract

This article presents a nonlinear reduced-order modeling method for evaporators in vapor compression systems. This method utilizes linearized reduction techniques to replace the fast dynamic modes of a system with nonlinear functions calculated from the full-order equations of slow dynamic modes. This process creates a nonlinear reduced-order model that is capable of matching full-order nonlinear model results without suffering from linearization error and simulates over 20 times faster than the full-order model. Simulation results are presented for the application of this reduction method to both moving-boundary and finite-control-volume evaporator modeling paradigms. A wide array of case studies are performed for various model input perturbations to illustrate this method's performance advantages. Simulation time step results are presented that show the reduced-order nonlinear model's capability of numerical time steps two orders of magnitude larger than the full-order nonlinear model, resulting in drastically reduced computational times. This reduced computational time allows the use of more heavily discretized models, increasing the feasibility of using more detailed and accurate models for real-time simulation and control.