ShelTherm: An aid-centric thermal model for shelter design

Abstract

The growing displacement crisis leaves many housed in simple shelters, which are bespoke to each setting, often sited in harsh climates and normally unconditioned. Measurements have shown that the thermal environment within the shelters can be life-threatening. Hence aid agencies are interested in improving these conditions. Unfortunately, given the skill set within agencies, there is little possibility to use complex thermal modelling to predict the impact of different materials or constructions. Hence the need to provide simple methods. In this work, we introduce a new physics-based thermal model (ShelTherm) for simple structures and encapsulate it within a tool designed to be used by humanitarian staff. Unlike other reduced models, the method is capable of dealing with high ventilation/infiltration rates, thin materials (such as tarpaulin) and high U-values. The model was validated over a wide variety of architectures, materials and climates. Mean error against observed data from a variety of real shelters was +0.94 °C (s.d. 0.79 °C), comparing favourably against the well-known reduced model CIBSE Admittance method (+4.95 °C, s.d. 0.89 °C) and the industry standard Energy+ (+0.60 °C, s.d. 0.32 °C). ISO 13792:2012(E) classified it as a Class 1 model, the highest possible. The differences between the new method and Energy+ across block, tarpaulin and stone shelters in Ethiopia, Bangladesh and Nepal were always less than 1.3 °C, and +0.24 °C for shelters in a mild climate. Humanitarian workers classed the tool as “easy” or “very easy” to use. The typical time to model a shelter by a first-time user was 34 minutes. ShelTherm can hence be seen as highly fit-for-purpose due to its overall accuracy and ease of use.