How To Validate Energy Efficiency Improvements With Regression Analysis

Energy efficiency improvements offer businesses opportunities to reduce electricity use, reduce fuel demand, and save on operational expenses. Standardized best practices are available to effectively implement Energy Conservation Measures (ECMs) and provide stakeholders with confidence that the ECM is delivering verifiable energy and cost savings.

Having a sound strategy to verify ECMEnergy conservation measure More results is paramount to the overall success of the project. The International Performance Measurement and Verification Protocol (IPMVPInternational Measurement and Verification Protocol More) and other related frameworks stress the use of regression analysis as the preferred option.  In fact, this approach is essential in order to decouple the impacts of changes at the site, including seasonable or production changes, from the energy savingsElectricity or gas savings or both. More from resulting from the ECM.  Without regression analysis there is no way to independently validate the claimed savings and to ensure the business outcomes were achieved.

Regression analysis is a powerful and flexible tool to gauge the impact of various operational and environmental factors (independent variables) on a facility’s energy consumption. It is a valuable exercise to reliably communicate the effectiveness of an ECM. Doing so can help build momentum and support for future energy efficiency projects.

As with any useful analysis, it is critical to start with reliable data and to understand the level of accuracy required. The process for regression analysis can be laid out in six steps.

The Steps To Developing Regression Model

First, all independent variables must be identified. With regard to verifying ECMs, these are parameters that are understood to influence the facility’s energy consumption. All the variables can be included into a single model, so that the pre- and post-performance can be compared. Common independent variables included in energy regression models include:

• ambient temperature
• plant production
• facility occupants
• heating degree days
• cooling degree days

The next step is to collect data. It is impossible to determine the extent to which something is absent without having a solid reference to how much was present before. For energy conservation, this means establishing a baseline period for a complete operating cycle, sometimes a full calendar year, immediately before ECM implementation. As the reference point to gauge ECM effectiveness, the baseline period must be representative. Anomalies and apparent outliers in the data should be addressed to ensure only valid data points are used.

Graphing the data to get a visual representation of the independent variable impact on energy consumption can be useful if a single variable is responsible for the majority of the variation in energy use, however, if there are multiple variables a graph is less useful.

Performing a regression analysis establishes a model, or algorithm for the relationship between energy consumption and the variables. In some cases, a simple regression will suffice. In many cases, however, there are multiple independent variables that influence energy consumption. In these situations, more complex multiple regression models will need to be developed to accurately depict the dynamic effect that the variables have on energy usage.

The final step is to ensure the effectiveness of the regression model. Several statistical tests can help determine if the model is a fair and accurate representation of the system.  While not an exhaustive list, the fundamental tests to validate a regression model are R-squared t-statistics and residuals.

Planning Out Your Next Efficiency Project

Energy consumption is often impacted by a complex interaction of multiple factors. Therefore, solely examining energy consumption over time will paint a very narrow picture. It is important that the effect of an ECM is validated in a scientific and thorough manner.

Northmore Gordon uses a regression analysis approach based on the IPMVP to provide a statistically valid measure of the savings from an energy efficiency project or group of projects. Using this approach, we can also give the statistically verifiable likelihood of the savings results being in between upper and lower bounds. Additionally, we use regression analysis to create energy efficiency and emission reduction certificates that can generate additional revenue for businesses.

From inception to execution, Northmore Gordon supports commercial and industrial businesses to implement ECMs and realize savings through proven decarbonisation strategies. Energy efficiency improvements are cornerstone measures in accelerating toward net zeroNet-zero is a term meaning to get to zero emissions, there is no current standard on what getting to net-zero means. For governments a net-zero target is one that covers all scope 1 emissions in their jurisdiction. For business it usually covers scope 1 and 2 emissions, although some companies also include some scope 3 emissions. The Science Based Targets Initiative is currently working on a standard for companies. More emissions — and every ECM project should include a plan to validate the project’s effectiveness.