Charge-sensitive modelling of organic Rankine cycle power systems for off-design performance simulation

This paper focuses on a charge-sensitive model to characterize the off-design performance of low-capacity organic Rankine cycle (ORC) power systems. The goal is to develop a reliable steady-state model that only uses the system boundary conditions (i.e. the supply heat source/heat sink conditions, the mechanical components rotational speeds, the ambient temperature and the total charge of working fluid) in order to predict the ORC performance. To this end, sub-models are developed to simulate each component and they are assembled to model the entire closed-loop system. A dedicated solver architecture is proposed to ensure high-robustness for charge-sensitive simulations. This work emphasizes the complexity of the heat exchangers modelling. It demonstrates how state-of-the-art correlations may be used to identify the convective heat transfer coefficients and how the modelling of the charge helps to assess their reliability. In order to compute the fluid density in two-phase conditions, five different void fraction models are investigated. A 2 kWe unit is used as case study and the charge-sensitive ORC model is validated by comparison to experimental measurements. Using this ORC model, the mean percent errors related to the thermal power predictions in the heat exchangers are lower than 2%. Regarding the mechanical powers in the pump/expander and the net thermal efficiency of the system, these errors are lower than 11.5% and 11.6%, respectively