The aim of this project is to develop the next generation building energy benchmarking system with following capabilities:
- Fair: to ensure fair comparision of building's energy efficiency.
- Explainable: to make the benchmarking process is understandable to the buliding owners.
ENERGY STAR® is the most widely used energy benchmarking system for buildings and plants in the USA.
We reproduce the ENERGY STAR® models for the following building types:
- Hotels
- K-12 Schools
- Multifamily Houses
- Offices
- Retail Stores and Supermarkets
- Warehouses
- Worship Facilities
In existing benchmarking systems, including ENERGY STAR®, it is difficult for the facility managers to understand how certain score/grade was assigned to their buildings. By leveraging Explainable Artificial Intelligence (XAI) methods, we unbox the complex models to make them exaplaintable to the facility managers. We have created an app using with one can clearly understand the level of influence of each building attribute and how it affects the final benchmarking score/grade.
One of challeneges of benchmarking is clearly define the peer groups. We analyze the feasability of using public datasets for developing a benchmarking system.