Generating design reference years from the UKCP09 projections and their application to future air-conditioning loads

Abstract

A method is developed to generate future design reference year (DRY) data from the United Kingdom Climate Impact Programme's 2009 (UKCP09) climate change projections for a variety of future time horizons and carbon emission assumptions. The method selects three near-extreme summer months and three near-extreme winter months and weaves them into an existing test reference year (TRY). Risk levels associated with the 85th percentile (broadly equivalent to existing Chartered Institution of Building Services Engineers [CIBSE] design summer years) of the cumulative distribution function of dry-bulb temperature and, for comparison, the 99th percentile are used. A comparison is made with DRYs generated using alternative methods from other research groups. The data are applied to future air-conditioning (cooling) loads analysis for a wide range of non-domestic case study building types. Simulations using a control DRY set applied to these buildings are used to develop a simplified regression-based calculation method for predicting future air-conditioning loads. The simplified model is shown to be applicable to future weather data without loss of accuracy, which makes it possible to carry out large numbers of future cooling loads predictions without the need to perform extensive and complex energy simulations. Practical applications: It is becoming increasingly necessary to design energy and comfort services for buildings with a whole-life perspective. To assist with this, the CIBSE future weather years can be used for building simulations through to the 2080s. In June 2009, the UK’s Department of the Environment, Food and Rural Affairs (Defra) with the support of the United Kingdom Climate Impacts Programme (UKCIP) published updated climate change projections using a probabilistic method. In future, the responsibility will rest with designers to select design data from a large number of probabilistic outcomes. This work develops a technique to select design weather data called a DRY at two alternative risk levels for use in building simulations through to the 2080s. A simplified method is also proposed to allow practitioners to generate large numbers of probabilistic design cooling loads without the need to perform extensive simulations.