Challenges and Lessons Learned in Applying Sensitivity Analysis to Building Stock Energy Models

Abstract

Uncertainty Analysis (UA) and Sensitivity Analysis (SA) offer essential tools to determine the limits of inference of a model and explore the factors which have the most effect on the model outputs. However, despite a wellestablished body of work applying UA and SA to models of individual buildings, a review of the literature relating to energy models for larger groups of buildings undertaken by Fennell et al. ( 2019) highlighted very limited application at larger scales. This contribution describes the efforts undertaken by a group of research teams in the context of IEA-EBC Annex 70 working with a diverse set of Building Stock Models (BSMs) to apply global sensitivity analysis methods and compare their results. Since BSMs are a class of model defined by their output and coverage rather than their structure and inputs, they represent a diverse set of modelling approaches. Key challenges for the application of SA are identified and explored, including the influence of model form, input data types and model outputs. This study combines results from 7 different modelling teams, each using different models across a range of urban areas to explore these challenges and begin the process of developing standardised workflows for SA of BSMs.

Uncertainty and sensitivity analysis of building-stock energy models: sampling procedure, stock size and Sobol’ convergence

“…Accordingly, the following research gaps were identified: (i) as a result of (a) and (d), a vast amount of UA-SA research concerning BSEMs is based on inadequate methods (i.e. LSA instead of GSA) (Fennell et al, 2023;Saltelli et al, 2019;Saltelli & Annoni, 2010); (ii) due to (b) and (d), UA-SA practitioners find it challenging to define uncertainty ranges and distributions for input parameters (Langevin et al, 2020;Fennell et al, 2021); and (iii) because of (a), (b), (c) and (d), UA-SA practitioners often limit the UA-SA scope and/or computation time of their study by reducing the number of uncertain parameters and/or building stock segments, the building stock segment's size (i.e. the number of buildings in the stock segment used for the analysis) and/or the number of model evaluations (i.e.…”

“…This limitation originated once more (i) from the research's main focus on UA-SA methodology and (ii) from the authors' BSEM structure since not all parameters in the model were easily accessible at the top level (N.B. a finding echoed by other participants within the BSEM UA-SA common exercise (Langevin et al, 2020;Fennell et al, 2021)): a considerable number of parameters were buried deep within inaccessible sublevels of the BSEM's structure. For example, (i) the ventilation air changes per hour between the zones as well as the ventilation airflow paths in the model were currently predefined in inaccessible sublevels of the model's structure.…”

“…The guiding objective of IEA-EBC Annex 70 has been to improve the use of data and models of building energy use to facilitate dramatic reductions in building energy use and greenhouse gas emissions. In support of this objective, a common exercise was drafted in which the researchers aimed to undertake a co-ordinated effort to apply existing Uncertainty Analysis (UA) and Sensitivity Analysis (SA) methods for model quality assurance to a diverse set of BSEMs (Langevin et al, 2020;Fennell et al, 2021). The modelling teams in the BSEM UA-SA common exercise aimed to explore (i) the challenges of defining input parameter uncertainty ranges for large-scale BSEMs and collecting the appropriate data; (ii) the applicability of different global SA techniques in terms of robustness of results, quality assurance and computational burden; (iii) key drivers of uncertainty in the models.…”

Quality assurance for building-stock energy models: a performance comparison of eleven uncertainty and sensitivity analysis methods

“…This limitation originated once more from the authors' BSEM structure since not all parameters in the model were easily accessible at the top level (N.B. a finding echoed by other participants within the BSEM UA-SA common exercise (Langevin et al, 2020;Fennell et al, 2021)): a considerable number of parameters were buried deep within inaccessible sublevels of the BSEM's structure. For example, (i) the ventilation air changes per hour between the zones as well as the ventilation airflow paths in the model were currently predefined in inaccessible sublevels of the model's structure.…”

“…The guiding objective of IEA-EBC Annex 70 has been to improve the use of data and models of building energy use to facilitate dramatic reductions in building energy use and greenhouse gas emissions. In support of this objective, a common exercise was drafted in which the researchers aimed to undertake a co-ordinated effort to apply existing Uncertainty Analysis (UA) and Sensitivity Analysis (SA) methods to a diverse collection of BSEMs (Langevin et al, 2020;Fennell et al, 2021). The modelling teams in the BSEM UA and SA common exercise aimed to explore (1) the challenges of defining input parameter uncertainty ranges for large-scale BSEMs and collecting the appropriate data; (2) the applicability of different global SA techniques in terms of robustness of results, quality assurance and computational burden and (3) key drivers of uncertainty in the models.…”