This article describes the development of a new three-dimensional model of the British building stock, called '3DStock'. The model differs from other 3D urban and stock models, in that it represents explicitly and in detail the spatial relationships between 'premises' and 'buildings'. It also represents the pattern of activities on different floors within buildings. The geometrical/ geographical structure of the model is assembled automatically from two existing national data sets. Additional data from other sources including figures for electricity and gas consumption are then attached. Some sample results are given for energy use intensities. The first purpose of the model is in the analysis of energy use in the building stock. With actual energy data for very large numbers of premises, it is possible to take a completely new type of statistical approach, in which consumption can be related to a range of characteristics including activity, built form, construction and materials. Models have been built to date of the London Borough of Camden and the cities of Leicester, Tamworth and Swindon. Work is in progress to extend the modelling to other parts of Britain. Because of the coverage of the data, this will be limited however to England and Wales.

A brief history is provided of models of energy use in the UK building stock, with the focus on the non-domestic sector. This history leads to an account of the development, since 2009, of the 3DStock method for modelling complete building stocks, both domestic and non-domestic. The paper explains how 3DStock models are built and the data sources used. Special emphasis is placed on the relationship of premises (the floorspace occupied by organisations) to buildings. Energy use may be metered at the level of premises, buildings or groups of buildings. Representing the patterns in which premises relate to buildings is therefore crucial to the modelling process, and in particular to the precise measurement of energy intensities. Applications of 3DStock models in building science and policy tools are reviewed, including the London Building Stock Model (LBSM), delivered to the Greater London Authority (GLA) in 2020. This 'digital twin' can be used for the monitoring, simulation and analysis of the building stock. Implications for research and policy are discussed, particularly for energy epidemiology, density, high-rise buildings, retrofit potential, energy-use intensity and benchmarking. Data are in place to extend 3DStock modelling to the whole of England and Wales.