Skip to main content

NMEC Hedgehogs and Straw Dogs

By May 13, 2019November 5th, 2021Energy Rant

As the third in a series, we are going to wrap up the normalized metered energy consumption (NMEC) protocol groundwork this week. See the first and second posts to catch up in case you missed those.

  • The first post covered new construction programs for which NMEC doesn’t apply because a baseline of normalized energy use[1] is needed for NMEC.
  • The first post introduced non-routine events (NREs), which are random and not accounted for in any model.
  • The second post explained that NMEC is ideal for residential behavior and weatherization programs.
  • The second post explained that because of several specific NRE examples, NMEC starts to get mucky and difficult, fast.


For many years, Michaels has backed our retro-commissioning projects[2] with proof of impact using NMEC. And you guessed it, we have the bias explained in last week’s post, NMEC Jedi. In other words, if the weather-normalized billing data for commercial buildings demonstrates savings achievement as planned, we declared victory. If the savings were short for some reason, we investigate to determine why and make adjustments as needed.

A big issue with retro-commissioning is the persistence of measures. Will revised schedules, sequences, and strategies be undone in a few weeks or a year?

First, to determine savings, NMEC should be deployed and analyzed as measures are implemented. Many times, customers start optimizing things after intervention (learning about the program and what it is) but before the project begins – i.e., merely describing the sorts of measures that can have big impacts sparks action before the customer is enrolled. Is this a free rider? NO!

Second, documentation of measures, as verified in the field and quantified via NMEC, can be annualized as needed to make the case for sustained, persistent savings.

Third, software can monitor the key strategies deployed to save energy.

In summary, we verify savings with NMEC and demonstrate control strategies are intact by monitoring setpoints and strategy deployment. If consumption changes over time while the setpoints and strategies that accomplished the impacts are still in place, the consumption change is due to something else – somebody else’s mystery to solve. The implementer has proven success, and the roots of the success are demonstrably in place.


Retrofits come in many stripes, including relatively quasi-constant, seasonal, and the uncertain. Let us cover how each of the three should be approached regarding NMEC.

Quasi-Constant Load Retrofits

Quasi-constant load reductions include widget replacements that are not dependent on the weather. Example: commercial and industrial lighting. These measures should be treated like retro-commissioning, only with more rigidity. While lighting projects have been a commodity baseload of our industry, when it is appropriately categorized, impacts of demand and energy savings can be very accurate.

Therefore, like retro-commissioning, lighting measures should be verified as measures are implemented over weeks or months – whatever the case may be. Weather-normalized NMEC can be used to determine impacts over this period. What happens later is someone else’s mystery to solve.

But what about retrofitting vacated commercial office space? This is like new construction. NMEC does not apply.

Seasonal Load Retrofits

Seasonal loads include heating and cooling. Chillers may be replaced on a planned cycle, generally in the off-season. Similarly, large boilers may be scheduled for replacement in the summertime.

How many programs allow energy savings to be determined versus existing heating and cooling equipment? The percentage can be counted on my ten fingers, and if not that, certainly on my ten fingers and ten beat-up toes. So what good is NMEC in situations where existing equipment performance is declared sinful? For that matter, the same issue applies to the quasi-constant scenarios.

Unless we skewer the sacred cow of the code-baseline, NMEC isn’t much good for seasonal loads – or even, quasi-constant loads. Are evaluators, intervenors, and entrenched stakeholders ready? No. For this matter, many folks are dug in like a row of planted hedgehogs.

Uncertain Retrofits

Uncertain retrofits are those that depend on parameters that will produce from minus 100% or more, to plus 200% or more of the predicted impacts, compared to any technical reference manual value. Famous technologies producing random results in the field include variable frequency drives and air compressor control.

The baseline for variable frequency drive applications can be anything. The most opportune case is where a centrifugal fan, blower, or pump operates with no control at all, but this is rarely the case. From there, how is the installed drive controlled? Does the speed vary or is it merely adjustable (set it and forget it)?

Moreover, can such measures even be detected with NMEC?

As for compressed air, if there is no control system to sequence multiple compressors to meet compressed air loads[3], no one knows what is happening in the baseline condition, before metering per IPMVP Options A or B are deployed.


Normalized metered energy consumption offers a mixed bag of opportunity.

  • It is the best for some programs, old (weatherization and building shell measures) and new (engagement and behavior programs).
  • Savings must be substantial; >5% in season.
  • It only applies to as-found baselines. Code baselines must be jettisoned.
  • NMEC approaches need to be baked into Technical Reference Manuals, or the TRMs should reference NMEC best practices, which this author just posted as three straw dogs in recent weeks.

[1] Normalized simply means predicted energy consumption accounts for changes in energy consumption with changes in independent variables such as weather or manufacturing production levels and varying production lines and feedstocks.

[2] Retro-commissioning includes operational and controls-related changes to optimize system performance. The simplest retro-commissioning measures can include optimizing a smart thermostat to match the user’s profile and needs.

[3] Which compressor runs first, second, third, etc., as compressed air demand increases.

Jeff Ihnen

Author Jeff Ihnen

More posts by Jeff Ihnen

Leave a Reply