Paper Conference

Proceedings of uSim Conference 2022: 3rd uSim Conference of IBPSA-Scotland

     

OPTIMISING BUILDING RETROFITS AT THE STOCK LEVEL: THE DEVELOPMENT OF A METHODOLOGY

Shyam Amrith, Paul Ruyssevelt, Ivan Korolija, Pamela Fennell

Abstract: Optimisation has been widely applied to individual building energy models, however there exists a gap in the literature around stock-level optimisation approaches. This paper describes the early development of a novel methodology which aims to allow multi-objective optimisation to be performed over a stock of buildings. This would be valuable to parties such as local authorities who require decision support when mapping retrofit pathways for their stock. An early-stage workflow is described which is facilitated by a custom developed Python module and applied to an EnergyPlus stock model. An optimisation problem can be defined for the area, which influences how the module handles the parametric setup and optimisation. The setup involves specifying restrictions (e.g. listed buildings) and desires for the recommendations (e.g. if a standardised retrofit package should be applied to a certain dwelling type). According to this formulation, the module assists with parameter creation. The EnergyPlus file is automatically altered so that each parameter can be independent. Optimisation objectives and constraints can be derived from the model outputs. The module links EnergyPlus to the NSGA-II algorithm. Results are output in a format that allows easy segmentation, plotting and exporting. The workflow is demonstrated on an area of housing stock to showcase how the methods are informed by an example problem definition. The paper concludes with a discussion around advantages, challenges and future development. The intention is to build on this work and include stages which handle simplification of the stock model, sensitivity analysis and calibration.
Paper:
usim2022_p130