Last week in Behavior Freak, we explored the ratio of energy savings to beer consumption at the Delta house (answer = χ/∞ = 0 ). Specifically, that post explained why college students’ only motivation to save energy is currency, and absent currency, the needle does not move. This week we explore more interesting behavioral success with commercial building actors, also operating under no specific personal monetary drivers. The secret ingredient: timely, useful information, and maybe a little human nature – energy management success or failure for large office building facility operators. This post is based on a paper presented by Buildings Alive at the 2018 ACEEE Summer Study for Buildings. Nice paper, guys!
I’m Different. You’re Different.
Let me start with another flaw of the home energy report, as gathered by informal, unscientific studies over the years with consultants in our field of interest.
Home energy reports compare me against my neighbors. My neighbors include an elderly retired widow, a husband, wife, and one child, a widow and adult child, a husband, wife and two children. The homes are all 80+ years old with disparate efficiency upgrades for sure. I don’t care how I compare with neighbors. My home’s operation, use, and occupancy pattern is nothing like theirs.
It’s like running a road race. The only people who care about the neighbors are the ten elites at the front with a chance to win. Everyone else is gunning for their PR (personal record), maybe hoping for a silent companion with whom to share a few miles of the journey. The key to improvement is gaging against oneself past performance – am I improving or not?
This is demonstrated in the paper. Energy users respond positively to feedback on how they are doing compared to their own past performance, considering weather, day of week, time of day and so forth. Accounting for these variables is called normalization. Observe once again the plot of natural gas consumption from Tooling Pay for Performance. The chart is reprinted below. The savings accumulate as the sum (total) of differences between the predicted base use and actual consumption. From the start of the measure through peak savings up to the point where one of the measures is inadvertently undone, you can see savings total about 3,000 therms on the right axis.
The chart to the left is normalized by weather and day type, but this is not directly observable. To make it observable, I took the data and plotted predicted and actual consumption by weather conditions, or in this case, heating degree days. To clarify, the higher the heating degree days, the colder the weather. For example, 75 degree days translates to an average -10 degrees F outside.
We can see from the chart below that the predicted and model data clearly correlate with weather conditions. Due to the nature of the controls measures, which is beyond the scope of this post, the gas savings are greatest in milder, warmer weather. When I sum the values of the gray dots, I get the same answer as shown above – about 3,000 therms.
Note, consumption is a function of the weather. Savings is not! And in case you are a dot counter, the data below cover the period from 12/01/2016 to 05/31/2017.
Information is Power
The test specimens in the case of the Buildings Alive paper include the operators of 140 large office buildings. They were conditioned to receiving daily feedback on their buildings’ actual energy consumption versus their predicted consumption. As shown in our example above, some days are better than others. The green dots (days) above the orange ones represent more natural gas use than the predicted status quo. Green dots below represent less consumption and savings. The gray savings dots show daily savings or overage. As noted, they sum to about 3,000 therms savings.
Happy Talk Helps
Initial daily reporting also included relevant information to incite curiosity about fluctuations in energy consumption, and to motivate internal dialogue with experimental activities to reduce energy consumption.
Results are shown in the following figure, which is a bit confusing to me, but I think I get it. The green dots shown represent combined savings of all buildings in the sample. Some were participating for the entire period of 1,360 days while others joined along the way, but apparently, the latecomers were added to the numerator (current, or actual as it is referenced in our data above, for all buildings) and denominator (base, or predicted as it is referenced above, for all buildings). If your brain is tuned to math, the chart below seems hard to believe, unless buildings were all added in the first couple hundred days.