In the world of Measurement and Verification (M&V), proving the true ROI of an energy-saving project requires more than basic math. Relying on a simple “before and after” subtraction leaves your data vulnerable to real-world changes, like unexpected weather shifts or sudden drops in building occupancy. That is where multivariate regression modeling comes in. A multivariate regression model that accounts for multiple independent variables creates a highly accurate energy baseline equation, so your M&V reports reflect precise, verifiable energy consumption and savings.

How Multivariate Regression Calculates Energy Consumption

In Measurement and Verification (M&V), a multivariate regression model estimates energy consumption (kWh, therms, kW, etc.) based on influential factors such as weather, occupancy, and production.

Understanding Dependent vs. Independent Variables

In this type of model, your consumption is the dependent variable and your influential factor is the independent variable. You can include multiple independent variables in a model to capture the true drivers of your energy use.

Building an Accurate Energy Baseline Model Equation

The regression model takes these factors into account when creating its baseline model equation. A baseline model equation simply states that there is a baseline level of consumption for each included independent variable.

A Real-World Example: School Energy Consumption

For example, a school’s energy consumption is influenced by both outdoor temperature and building occupancy. In the school’s example, this baseline equation would mean there is a baseline kWh consumption per day, and that consumption increases or decreases depending on changes in temperature and building occupancy.

Facility-Level vs. Isolated Equipment Modeling

Multivariate regression modeling is highly adaptable. It can be done at the whole facility level or for an isolated piece of equipment, depending on the scope of your project.

Why Multivariate Models Outperform Simple Subtraction

Ultimately, this modeling approach helps to reflect real-world conditions in your energy-saving project. This provides a rigorous and accurate standard of proof, as opposed to simply subtracting the consumption data before and after the measures are installed.