Study on the Detection Method of Refrigerant Leakage Amount in Air Heat Pump System

Abstract

In this study, the two detection methods of refrigerant leakage amount in air heat pump systems are suggested and verified by modeling and experiment. This is important because one of the most frequently occurred and the most costly fault in heat pumps is refrigerant leakage. When a leak occurs, the leak point needs to be found and it is relatively easy once we have proper leak detector. After that, charge amount leaked from the system needs to be determined by the system behavior and to be recovered. The modeling process of the heat pump system requires lubricant consideration because the optimal charge amount in the modeling without lubricant is largely underpredicted compare with experimental results. This difference is mainly due to the refrigerant amount dissolved in the lubricant. The effect of heat transfer and pressure drop from the lubricant is relatively small but the refrigerant dissolves in the oil as a form of the liquid, which makes its amount large significantly. When the lubricant is considered in the modeling, the results show good matches with experiment. With the modeling and the experiment from two different heat pump systems, which are the residential and commercial type, the refrigerant detection method is suggested. The residential system only has a small number of sensors and within this limitation, the current charge amount needs to be found. The method utilizes the temperature sensor at evaporator side, which are located at the indoor air inlet and evaporator midpoint. The temperature difference between these two is a detection judgment index. It is larger than the 5K at normal charge condition because evaporator midpoint stays at the two-phase condition. At leaked condition, the evaporator midpoint state becomes superheated so the temperature difference at evaporator becomes smaller than 5K. With this principle, charge level around 55 to 65% is judged. Additionally with the EEV control method, which intentionally reduces the opening value of EEV, the detection range of charge level increases to 80%. For commercial type, the detection method with log mean temperature difference (LMTD) at condenser is suggested. This is because the relationship between the charge level and condensing pressure showed linearity. The data divided into groups according to the compressor rotational speed. With the compensated LMTD at condenser charge amount in the system is predicted with RMS error of 5.9%. This study is useful for any air heat pump system and the proposed methods can be applied immediately to existing systems.