IBPSA-India hosted a technical talk with live webcast on “Analysis of Uncertainty in Building Design Simulation” by Professor Godfried Augenbroe at Malaviya National Institute of Technology, Jaipur.Professor Godfried Augenbroe, Georgia Tech, USA has a 35-year track record of teaching and research in the building science. Professor Jyotirmay Mathur (MNIT Jaipur) moderated the talk and 40+ participants (including 22 on the webinar) participated by a wide range of academicians, students,and industrialists.
Speaker highlighted the uncertainties related to building performance simulation. While it is widely used in both academic and industrial applications but there is still need to understand its capability of predicting actual performance and informing associated decision-making. These uncertainties may lead to its under-utilization and lost opportunities in practice due to lack of confidence in BPS tools. He addressed two intriguing questions:
- What order should a predictive energy model be?
- How close can an energy audit model be to actual?
To address the questions and other lingering concerns about the performance gap between predicted and actual energy performance, Professor suggested to deal with firstly parameter uncertainties and secondly with scenario uncertainties. He advised to focus on parameters exposed in theinput file or through application program interface (API) and execute through standard Monte Carlo algorithms.
Scenario uncertainties arise from external (weather, occupancy) and internal (HVAC control and operation) sources. Occupancy scenarios cannot typically be parameterized and exposed.Since Weather is external to simulation code, thus weather research and forecasting (WRF) models can be of great help to reduce these uncertainties. Operation scenarios are most significant unknown uncertainties, and “Operation Uncertainty factor” is used as a post multiplier in efficiency. All the uncertainties constitute the model for uncertainty (MFU) of building performance simulation. Massive measurements or a higher fidelity model is needed to quantify MFU which requires substantial effort and enough quality data.
Model fidelity and software benchmarking are trending topics in building simulation industry these days. Practical insights are required for straightforward implementation of uncertainty quantification in risk-conscious decision making processes and protect the interests of stakeholders. Research contributions for reliability and resiliency decisions, right-sizing practices, and litigation protection can demonstrate remarkable impact on building industry.