Dylan Brock
CLEAResult Consulting | Statistical Analyst
Speaker
Track B: Data Analytics & Energy Services
Session B2: Measuring, Monitoring, and Analyzing Energy Data
June 26, 2025 | 10:00 am - 10:30 am
Whole Facility Modeling Viability Based on Annual Usage
The Strategic Energy Management (SEM) program regularly employs whole-facility energy models for monitoring and tracking of participant performance within the program. Due to strict model validation criteria, not all models built can be used for savings estimations. Leveraging over 10,000 successfully created whole-facility energy models and failed model attempts, we investigate how annual energy usage and other facility characteristics impact the ability to create a viable SEM energy model. The results show a positive influence of annual usage on model viability, as well as a mix of influences based on other facility characteristics. The trends found in this study will offer insights to inform future program design, emphasizing the importance of exploring alternative measurement and verification methods that extend beyond traditional whole-facility modeling. This research is expected to transform the historical approach to modeling, moving toward a more forward-thinking and flexible framework for addressing the evolving needs of the industry.
Speaker Bio
Dylan Brock is a Statistical Analyst for the Strategic Energy Management (SEM) division at CLEAResult Consulting. The team’s primary focuses include avoided-use analysis, conditional and dynamic forecasting, and causal inference. Dylan is responsible for building statistical energy models and providing statistical validation to the energy modeling process. Model specification relies heavily on exploratory data analysis, regression analysis, as well as a knowledge of engineering concepts in relation to whole-building operations.
Originally from Austin, Dylan studied Quantitative Economics at the University of Texas. While at SEM, he has built countless energy models, established dynamic forecasting methodologies, and developed recursive coefficient estimation algorithms. He holds certifications in several areas of statistics including Causal Inference, Applied Econometrics, Time-Series Analysis and Forecasting, among others.