Loughborough University doctoral researcher, Ahmed I Ahmed, who is part of the LoLo CDT EPSRC program has published his research in Applied Energy, a highly reputable scientific journal in the field of energy research.
The title of the published paper: Forecasting underheating in dwellings to detect excess winter mortality risks using time series models.
Research background
Excess avoidable deaths occur each winter in temperate climate countries. According to the Office for National Statistics, 50,100 excess deaths were recorded in 2017/2018 winter in England and Wales. Excess winter mortality is strongly linked with cold homes which is associated with inadequate heating, poorly insulated homes and fuel poverty. The aim of this research is to explore statistical predictive modelling tools in general and in particular focus on the selection and implementation of the most appropriate time series forecasting model for the purpose of achieving multi-step ahead predictions of low internal temperatures of dwellings in the winter with acceptable accuracy. This study develops a time series forecasting model using AutoRegressive with eXogenous inputs (ARX) to provide reliable multi-step ahead predictions of internal temperatures of homes in the winter. The model was validated using a case study home located in Loughborough, UK. The ARX model proved capable of achieving reliable forecasts for shorter prediction horizons (< 9 h). This study demonstrated the potential for using time series forecasting as an early warning system to detect low winter temperatures; reduce EWM, promote energy efficiency and highlight fuel poverty.