Normalized Metered Energy Consumption (NMEC) Open-Source R Code: Tools for Meter-Based, Whole-Building M&V

We are very excited to announce that kW Engineering released open source R code today for analyzing site-level, meter-based energy savings. Our will make it much easier for energy efficiency program implementers to analyze energy efficiency project impacts using the normalized metered energy consumption (NMEC) approach.

Background

For a long time, like twenty years or so since Monitoring-Based Commissioning became an efficiency program in California, we’ve been using energy use data from time-of-use meters to make better and more accurate regression-based energy models. The data and improved modeling have improved our ability to reliably quantify savings in ratepayer-funded efficiency programs. These developments have led to other significant advancements:

  • Development of cloud-based energy management platforms
  • Development of tools, guidelines, and protocols
  • Further research into automated M&V and savings analysis
  • Legislation to adopt the use of meter-based data for energy efficiency programs

This is our contribution to scaling up meter-based methods and efficiency programs. I work at an energy engineering company and we get into a lot of buildings helping owners with energy efficiency, renewable, and demand management opportunities. There are hundreds of firms in the energy efficiency industry across the nation. And there are multiple meter-based efficiency programs in action or in development in many states. In other words, a lot of interest and activity in meter-based M&V.

About a year or more ago, I wrote about why I think meter-based approaches are superior to our traditional method of quantifying savings – which was based mainly on calculations for individual measures. Meter-based approaches align the savings estimations with how customers consider them – whether their utility bills go down. There are several other advantages of the methodology – many identified on our previous blogs and elsewhere throughout the industry. The approach:

  • Encourages investment into deeper savings and management of energy performance of buildings
  • Reduces transaction costs – savings calculations are made based on the number of meters, not the number of measures
  • Reduces the technical review and evaluation costs – because there are fewer savings methods (based on meters, not measures), and the method is well known and practically standardized per IPMVP
  • Incentivizes owners and program implementers to assure savings are achieved and persist over time
  • Provides tools to help owners maintain savings and energy performance
  • Is based on analysis methods our industry is already familiar with (regressions)

So What?

Since 2015, when California passed Assembly Bill 802, which enabled efficiency program administrators to offer incentives based on the overall reduction in ‘normalized metered energy consumption’ or NMEC, utility program administrators, and now soon-to-be selected third-party program administrators, have begun offering incentives under their ‘NMEC’ offerings. This has increased attention on the details of the method. In the same year, California passed Senate Bill 350 which dramatically increased the state’s efficiency goals.

Achieving these goals will be a huge challenge and it will help to leverage the advantages of meter-based M&V. If we can streamline the transaction costs of efficiency projects, we can devote more resources to finding savings in more buildings. There are many underserved building sectors such as small and medium sized commercial buildings, that do not participate in efficiency programs because it takes time, can be a hassle, and rewards may be modest. This makes efficiency programs less cost effective and therefore more challenging to administer. Large building owners can also benefit by devoting more resources to finding more efficiency opportunities, and quickly see the impact of their investments. Industry also benefits – currently strategic energy management (SEM) approaches are learning to use meter-based methods to quantify savings from their own internal efforts to reduce energy use and incorporating the method into their business practices. Soon, demand side resources (efficiency, demand reduction, renewables, storage, and grid responsiveness) will become more integrated, and determining their benefits will be based on meter data.

There is much to do, and time is running short. To speed progress, kW Engineering has developed site-level M&V analysis software that we’ve used to help California efficiency program managers qualify participants and determine savings in NMEC programs. , released today as open source on our GitHub public repository of the same name, is R code to help develop advanced regression models from baseline period data and use them to quantify savings in the reporting period after measures have been installed. R was chosen because it is also open source, in the public domain, and is language familiar to many professionals. It has robust statistical analysis packages. For projects, we have packaged the code and submitted it along with project M&V Plans and Savings Reports so that technical reviewers and evaluators can see directly the ‘live calcs’ on how savings was calculated from raw data.

NMEC R Library Features

nmecr code features include:

  • Inclusion of multiple modeling algorithms to enable the user to develop the best fitting model for their application, including:
    • Lawrence Berkeley National Laboratory’s time-of-week and temperature model
    • Change-point models based on ASHRAE’s inverse modeling toolkit (RP1050)
    • Simple linear regression
    • Heating Degree-Day and Cooling Degree Day algorithms
  • Capability for users to develop energy models based on hourly, daily, or monthly time intervals.
  • Assessment of temperature coverage, model goodness of fit metrics, savings uncertainty, model coefficient statistics, and distributions of residuals, to assure the models are statistically valid and appropriate for quantifying the savings for their application.
  • Addition of independent variables to improve model accuracy, such as occupancy or production rates, or indicator variables to address different operation modes in buildings, such as a school’s vacation and summer schedules.
  • Capability to pre-screen building energy use patterns to assure the method is appropriate for the application.
  • Normalization of baseline energy use to reporting period conditions for calculating ‘avoided energy use’ (per IPMVP nomenclature).
  • Normalization of both baseline and reporting period energy use to a common set of conditions for calculating “normalized savings.”

Future Plans

kW Engineering will continue to improve the code. Among our immediate plans are to:

  • Test the tool’s algorithms using EVO’s Tool Testing Portal
  • Test the tool’s algorithms in the ASHRAE Great Energy Predictor Shootout III
  • Add non-routine event (NRE) identification and impact quantification capability
  • Add vignettes on various applications of meter-based M&V
  • Identify and develop ways the code may be implemented in owner’s energy management and information systems (EMIS).

Why? What’s the Vision?

Scalability. We want more people to test and use the code and do meter-based projects. Try it; use the tool to quantify savings and promote its persistence. Become familiar with R and nmecr. Gain insight. Contribute to the development of the code. Advance the energy efficiency industry.

We’d love to get your feedback. As you get into it, you’ll have more questions about different aspects of its application, and desire to add features and capabilities to it. You can do this privately ☹ or share your advancements ????. But most importantly, as the tool becomes more widely used in projects, the methods become better known, and improvements may be made that benefit the industry.

Please feel free to download and use nmecr. It is available here: https://github.com/kW-Labs/nmecr

We look forward to hearing from you!

Scroll to Top