Building Energy Models: What Are Black-Box, White-Box, and Gray-Box Models?

Building Energy Models (BEM) are crucial tools in the energy-efficient design of buildings. The two fundamental modeling techniques are data-driven (black-box models) and simulation-based (white-box models). Each of these models has unique characteristics and capabilities.

The black-box modeling approach encompasses statistical, machine learning, and data-driven approaches to predicting building energy consumption. This approach leverages data analysis and mining tools, including machine learning algorithms, linear regression, and statistical regression methods. In this approach, programs use energy consumption data and statistical analysis to generate predictive models. The problem with this approach is insufficient accuracy limited by the observed data. 

The white-box modeling approach depends on fundamental laws of physics, such as the fundamentals of heat transfer and thermodynamics. This approach encompasses building energy modeling tools that produce a detailed analysis of the building energy systems based on the defined properties of the building systems. The problem with this approach is the inability to incorporate available data into the model.

The gray-box approach combines white-box and black-box modeling techniques. It is a hybrid of data-driven approaches and white-box modeling techniques. This approach combines physical principles with prior knowledge of building performance.

ENERlite Consulting’s automated BEM calibration tool, ENERtune, is a hybrid model that provides a starting point for connecting building data to BEMs. ENERtune uses gray-box modeling techniques, combining energy modeling with historical data analysis and machine learning. ENERtune creates a simpler and more efficient surrogate of BEM and creates a calibrated model that is more realistic and accurate than the original model. 

Please sign up to get more information about ENERtune at our website.