Commercial Building Modeling with AMI Data

Despite the massive opportunity for energy reduction, operational savings in commercial buildings have long been low on utility priority list because of the challenges of big data: acquisition, analysis, and action. There are energy savings that are easy to prove, and energy savings that are harder to prove. Proving energy savings from capital measures, like changing out lightbulbs, is relatively simple. Incandescent light bulbs use a certain amount of power to operate, and LEDs use less power. Measuring the energy savings from operational savings, like adjusting air conditioning schedules to only operate during occupied hours, is more complicated. Power TakeOff calculates energy savings by measuring energy consumption over time and demonstrating what the customer would have done without the help of the program (the counterfactual) and comparing that to the actual measured consumption. The energy saved is the difference between the measured consumption and the counterfactual.

In order to prove the counterfactual, Power TakeOff needs to know all of the different drivers of consumption for the building in question. Commercial buildings have a huge range of operations that make them very difficult to model using a one-size-fits-all model. Schools and stores are both considered to be commercial buildings, but are not at all similar in how they are used, who they serve, and when they are occupied. We collect data on energy use at each hour of the day, when the buildings are in use and how, any deviations from normal schedules, building locations and weather data, and any changes the occupants are making to try to save energy. Weather at each hour of the day is pretty easy to collect, but hourly energy use and hourly building operations are not. Without advanced technology, this information is simply impossible to collect.

Utilities are rapidly solving the problem of data acquisition with Advanced Metering Infrastructure (AMI) deployment. Power TakeOff handles the analysis and provides the customers with exactly what actions they need to take to save energy. Luckily, the complexities of commercial modeling are greatly simplified with AMI data. Now, we can algorithmically detect when a building is occupied and build that information into the model without needing to visit individual school websites and painstakingly input information from a fuzzy pdf calendar into an Excel spreadsheet. And for retail, Black Friday sales and holiday schedules can be detected and automatically accounted for, so the yearly sales events no longer eliminate any savings detected from previous attempts to save energy.