Van Dronkelaar, Chris; (2018) Quantifying and mitigating differences between predicted and measured energy use in buildings. Doctoral thesis (Eng.D), UCL (University College London).

Preview

Text EngD Thesis Chris van Dronkelaar (1).pdf - Submitted Version Download (16MB) | Preview

Abstract

Simulation is commonly utilised as a best practice approach to assess building performance in the building industry, and can help facility managers and engineers identify energy saving potentials, forecast future scenarios and evaluate the energy and cost performance of energy saving measures. However, the built environment is complex and influenced by a large number of independent and interdependent variables, making it difficult to achieve an accurate representation of real-world building energy in-use. This gives rise to significant discrepancies between simulation results and actual measured energy consumption of real buildings, termed ‘the performance gap’. This is partly fueled by a lack of understanding of the procedural differences between national calculation methodologies and energy certificates commonly employed in presenting energy use. As such, a classification was adhered to, which distinguishes between three different performance gaps; the regulatory gap (predictions from compliance modelling), static gap (predictions based on performance modelling), and dynamic gap (calibrated predictions taking a longitudinal perspective). This research added to knowledge by making three separate contributions. The first contribution was the exploration of industry practices and stakeholders, which identified common barriers to delivering high building performance, and made suggestions on how to overcome such barriers. Through semi-structured interviews and round-table discussions with industry experts, five key factors were suggested for delivering better building performance. The second and third contributions emerged from case research, for which an overarching methodology was developed, aiming to quantify and mitigate differences between predicted and measured energy use. Fundamental tasks within the methodology were based upon previous research efforts, while new techniques were introduced to include the uncertainty of typically static input parameters to improve the calibration process. In particular, the second contribution was the quantification of the underlying causes of the performance gap and mitigation of differences between predicted and measured energy use in four case study buildings, through the application of sensitivity, and uncertainty analysis and manual calibration. Subsequently, the third contribution investigated the effect of data granularity on model calibration accuracy through meta-model based optimisation.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item