Paper Conference

Proceedings of BSO Conference 2020: Fifth Conference of IBPSA-England

     

Towards reproducible dwelling retrofit evaluation: Data, tools & opportunity to address uncertainty

Kate Simpson

Abstract: Reproducible research allows scientific work to be trusted. It enables data to be shared and methods developed. Retrofit evaluation relies on mixed method data collection, processing and analysis techniques. This includes building performance simulation and indoor environment evaluation. However, few studies to date have successfully combined and shared these data. Reproducible retrofit evaluation is an emerging topic. In building simulation, there is a recognised gap between modelled and measured energy use. In previous research, the most sensitive parameters influencing domestic space heating energy demand were found to be temperature set points and heating duration. These are aleatory uncertainties and almost impossible to eliminate, but more data could increase knowledge. Government, organisations and researchers do share data openly, where ethical and possible, or required. However, not all data in academic research is designed to be open. Where ethical principles are followed, opportunity exists to progress this space. This paper aims to highlight the opportunity for space heating to be better characterised with more open data, particularly on indoor air temperature. A synthesis of recent work on open and reproducible research in relation to retrofit evaluation is presented. A reproducible guide is highlighted. This is followed by a summary of open datasets that inform retrofit decisionmaking. A gap is identified for more measured indoor air temperature. A selection of open-source tools and suitability for retrofit evaluation are discussed. The novel contribution is the identification of opportunities for reproducible retrofit research, using open data to address aleatory uncertainty in terms of heating behaviour, derived from air temperature data. In addition, ideas towards data repositories are made.
Pages: 104 - 112
Paper:
bso2020_Simpson