Announcing New Demand Flexibility Methodologies
Whole building metered Demand Response and Load Shifting
Since the launch of the OpenEAC Alliance, we’ve been working on methods suitable for measuring the impacts of all types of distributed energy resources. For demand response and daily load shifting, the challenges are well-known. Dozens of existing protocols for measuring demand flexibility exist across the different global energy markets. We wanted to establish a foundational approach here that could serve the vast majority of the market most of the time, but also be flexible enough to accommodate edge cases that would be spelled out in individual M&V plans.
This work has been amplified by the fact that the stakes have gone up considerably over the past six months. Demand flexibility is now the most valuable energy resource on the planet, as it uniquely solves the problem of load growth from new data center construction, EV charging, and other large electrical loads.
Today, we are releasing draft methods for calculating “whole-building metered demand response” as well as “whole-building metered load shifting”. These are the two most prevalent forms of demand flexibility, outside of direct device telemetry.
Demand Response
For demand response, we built on LBNL’s Time-of-Week-and-Temperature (TOWT) framework, which creates a weather-normalized regression of a rolling 28 day baseline. Each hour of the week is given a dummy variable, which allows the model to endogenously account for occupancy patterns without overfitting on small samples. This is an event-driven methodology, so it also requires event data in order to calculate proper hourly load reductions.
See full methodology here.
Load Shifting
For load shifting, we altered the TOWT model in two important ways. First, the baseline is fixed to pre-installation conditions, with weather-sensitive loads requiring a full year to capture seasonal impacts. Second, for a given day, we identify a set of “measurement windows”. We refer to these as a “peak window” and an “off-peak window”, but the way to think about it is that you have a window in which energy consumption is expected to be higher than the baseline (off-peak, where you are adding load), and a second window in which consumption is expected to be lower than the baseline (peak, where you are removing load). In order to get credit for peak window reductions, you must also show off-peak increases. That is, load shifting requires an actual measurable shift, rather than just a reduction during the peak window. Credit for peak reductions is capped at the amount of increase relative to baseline in the off-peak window.
See full methodology here.
Comment Period and Presentation
The next OpenEAC Alliance working group meeting is scheduled for Thursday, November 20th at 8am PT. Expect to see a more detailed presentation of these methods at the working group meeting. Between now and then is the open comment period. Copies of each of the methodologies can be found at https://methods.openeac.org/. We will collect comments on the methods between now and November 10th and then post any updates to the methodologies prior to the working group meeting.