Calibrated whole building energy simulation: An evidence-based methodology
MetadataShow full item record
This item's downloads: 898 (view details)
Climate change, driven by global energy consumption, is one of the major challenges to humanity today. The buildings sector consumes a significant portion of global energy resources and much of this is due to inefficient design and operation. Whole building energy simulation provides a means of assessing building performance at the design stage of the building life cycle. Calibration of these models allows for performance assessment and efficiency improvements at the operational stage. Also, the information output from the calibration process can be used to identify mistaken assumptions made in design stage models, to improve best practice modelling techniques and to drive the development of simulation tools. However, there are issues with current approaches to calibrated simulation. Many existing methodologies are informal, ad-hoc, and not firmly based on clearly referenced evidence. In addition, many calibrated simulation case studies use simplified models and limited measured data. This thesis presents a novel, evidence-based methodology that uses version control techniques to track the entire calibration process. This improves the reproducibility and credibility of the calibrated model as future users can review the model and the evidence on which it is based at any stage of the calibration process. The methodology also includes a new zoning-strategy that more closely represents the actual building than current strategies, and is particularly applicable to deep floor plan buildings. This methodology was applied to a 4-storey, 30,000m2 industrial office building - the Intel IR6 building in Leixlip, Co. Kildare, Ireland. The final calibrated model uses measured hourly internal load data in the simulation instead of scheduled approximations. The calibrated model has a MBE of 4.16% MBE and a CVRMSE(hourly) of 7.80% when compared to measured hourly HVAC electricity consumption for 2007. The data from this case study was also used to present a new visualisation technique that combines 'bin' analysis with 3-dimensional surface plots.