Ingress of warm moist air during cold-store door opening is a major issue in refrigeration plants, as it increases energy consumption and can also cause an undesirable rise in product temperature. Moisture infiltration itself can produce food safety problems in chill stores and can jeopardize personnel safety in frozen stores as the moisture turns to ice.

The British Department for Environment, Food and Rural Affairs (DEFRA) has commissioned the University of Bristol (FRPERC) to look at this problem in partnership with Anglo Dutch Meats (UK) Ltd; ACS&T Wolverhampton Ltd., Thermoscreens Ltd., Northern Foods Plc and Ballymoney Foods Ltd. The aim of the project, which is part of the LINK programme, is to reduce infiltration through doorways and thereby reduce the effects on temperature control, safety and energy consumption.

We created a numerical model of the experimental test room using CFX-5. The geometry of the model covered the room itself and the structure immediately outside the opening. To model the effect of buoyancy we used the Boussinesq approximation; this gave better convergence than a full buoyancy model, while results for the two models showed negligible differences. With initial conditions of stationary air at -20°C inside the room and +20°C outside, we ran the model for a period of 30 seconds from the moment the door was opened.

We used the CFD model to investigate different methods for reducing air infiltration. Air curtains were predicted to be the most promising method without restricting access, and measurements on an air curtain fitted to the test room later confirmed this to be the case.

One of the greatest benefits of the CFD models was their ability to show the process by which infiltration occurs. It can be difficult to convey the problems associated with cold-store door openings to operators of these stores by conventional verbal or written means; CFD however allows the process to be viewed in more easily understood pictorial and animated forms. The greater understanding can also stimulate further investigation and directions for study. Indeed, it was the stimulation, investigation and validation of these 'what-if' scenarios that was the prime reason for using CFD in this project.

Flow through the entrance after opening the door. Cold air spills out of the door and warm air enters.

LDA-measured and CFX-predicted air velocities entering (negative) and exiting (positive) the room. The data shows that the gradient of the transition point from inflow to outflow is well predicted.

Photograph of the Thermoscreens air curtain fitted to the test room.