The energy consumption of building systems can be estimated from nameplate data and the operating schedule. However, this is a very rough estimate that does not consider the local climate and the interaction between different construction systems. When deciding between energy efficiency measures for a building project, energy modeling is a powerful tool.
A study by the American Institute of Architects (AIA) compared energy savings achieved in buildings with and without energy modeling. Building owners who used energy efficiency measures without modeling achieved savings of 37%, but this figure increased to 52% in projects where energy modeling was used. In other words, energy efficiency projects in buildings achieved 15% more savings when energy modeling was used.
Energy models can be classified into three main types, depending on how they process information: white box, black box and gray box. This classification is generally used by data scientists, and is not exclusive to energy models.
|
Modeling Approach |
Description |
|
White box |
This model is based on the physics and properties of the system. |
|
Black box |
This model is based on historical data and statistical analysis. |
|
Gray Box |
This model uses a combination of simplified physics and historical data. |
White-Box: Energy Modeling with Physics
Physics-based energy models are the most accurate, and this is the approach used by software like DOE-2 and EnergyPlus. However, creating white box models is demanding as they must include all necessary equations and data. Due to their complexity, white-box models are also the most demanding in terms of computational power, which makes their simulation slow.
In addition to accuracy, white box models have the advantage of not requiring any historical data. This means that they can simulate a building that does not yet exist, as long as its physical properties are known. If engineering expertise and computing power are available, white box energy models can provide valuable information to real estate developers.
Black-Box: Energy Modeling with Data
Black box modeling is a completely different approach: while white box models attempt to predict behavior based on physics, black box models are reverse engineered with existing data. These models can be easily calibrated with available data and can be processed much faster than white box models. Some examples of data-driven modeling methods are artificial neural networks (ANN), support vector machines (SVM), and statistical regression.
The main limitation of black box models is their dependence on historical data. In energy efficiency applications, these models are only valid for the building that produced the data, or other buildings with very similar properties. Creating a black box model for a building that doesn't exist is impossible since there is no data to calibrate the model.
Black box models are very useful in the management of existing buildings, as they can be used to simulate the impact of energy efficiency measures before implementing them. Once the model is created, it can be used to analyze construction problems and identify their causes.
Gray Box: Hybrid Energy Modeling
Gray box models have elements of both white box and black box models. They use physical equations to represent building behavior, but they are simpler equations than those used in white box models. As a result, a gray box model can be simulated more quickly after it is calibrated.
However, simplified physics leads to a loss of accuracy. To compensate for this, gray box models are calibrated with historical data, just like black box models. A gray-box power model offers a balance between the accuracy of a white-box model and the speed of a black-box model.
The process of calibrating black-box and gray-box models is often called model “training,” as simulation parameters are adjusted until model results match the behavior of the system being modeled.
Using Energy Modeling Effectively
None of the energy modeling approaches can be considered better than the others, as they all have applications in the construction sector. White box models are the only option when no data is available. However, they can also be used to compare the actual behavior of a building with its ideal behavior according to the principles of physics. On the other hand, black box and gray box models can be used to represent an existing building and predict its energy consumption.
Energy modeling can be very useful for New York building owners, as Local Law 97 of 2019 imposes strict emissions limits starting in 2024. With so many different options for updating buildings, owners should make sure to that find the best combination to reduce emissions. Ideally, a building retrofit should maximize energy savings and avoid emissions for every dollar invested upfront.
