[Sorry for the sometimes technical nature of the language – we’ve tried to make this as accessible as possible, but some boffin-ish terminology is inevitable!]
Less is More is an innovative community-based approach to electricity demand management. It aims to reduce demand among householders connected to selected electricity substations, and rewards the local community with cash prizes when they are successful.
Up to £5,000 can be won by each community, to be spent on improvements or events for their local area.
So, how do we allocate the funds?
The rewards are for i) reducing the ‘peak’ demand (i.e. the point at which the community is using the most electricity, which usually happens in the evening) and ii) reducing the total overall consumption.
(Energy providers are particularly interested in peak electricity demand because this defines the capacity of the substation; a higher peak may mean having to upgrade the substation even though total consumption may stay the same, or even go down.)
In order to know whether peak and/or overall consumption are being reduced we need values to compare to, and because this project wasn’t running last year, we have to estimate them. This is how we did it:
Step 1: Collect baseline data
First of all, we collected consumption data from each electricity substation from July to November 2013. (Data for the whole year would have been better, but wasn’t possible in the timescales available.)
This became our baseline data and is used as the benchmark against which to compare future consumption.
Step 2: Estimate and add PV generation
Some households use electricity which they generate themselves with solar (PV) panels. This electricity doesn’t come through the substation but still needs to be factored in to the baseline data, otherwise it might appear that a community is using less energy, when in fact it’s simply been a sunny day!
To do this, we estimated the electricity generated by PV installations during the data collection period (July to November 2013) using this online tool and added this generation to the baseline figures. This gave us a total electricity usage estimate.
Of course, solar electricity isn’t generated evenly, so has been assigned proportionally (and minute by minute) to appropriate times of the day.
Using the same online tool we forecast future electricity generation over the length of the project and will incorporate this into the overall results.
Step 3: Collect weather data at substations
Another variable to factor in to our calculations is the weather. Weather, and particularly temperature, has a significant effect on how much electricity people consume. In general more is used during winter and when it is darker, colder or wetter than normal.
So we needed to know both what the weather was like at the various substations when we collected our baseline data and also whether this was warmer or colder than average. This way, if the baseline data was collected during a particularly warm period, the group wouldn’t be penalised if during the ‘competition’ phase the weather is much colder.
We collected weather data in the neighbourhood of each substation during the baseline time period, including the temperature at noon on each day. This was then compared to the average for every individual day during the past ten years and corrected accordingly.
Estimates for future electricity consumption are based on an assumption that the weather will be similar to the average weather over the last ten years – knowing that this assumption may well prove incorrect!
Step 4: Estimate annual consumption for each community
Using a proprietary model for energy consumption we have used the weather-corrected baseline data to estimate the total annual consumption of electricity for each community going forward. This approach is similar to how energy companies estimate household electricity consumption for bills from past meter readings.
The estimated annual electricity consumption has been allocated minute by minute over the length of the project and weighted according to the day of the week.
We now have a very good forecast of how much electricity each community is likely to consume based on previous usage and adjusted for weather and power generated by solar panels.
Step 5: Estimate peak consumption
Estimating the peak consumption requires a similar approach to that of general consumption. First we need to isolate the daily peaks, before we apply our model to adjust for PV generation, correct for weather effects and finally estimate the future peak consumption for the project.
How we calculate rewards
As noted above, Less is More has two goals; to reduce electricity use at peak times and to reduce total consumption.
A peak-reduction target may be something like ‘bringing daily electricity consumption peak down by 20%.
A consumption target might be reducing the community’s daily electricity consumption by 30%. Note that if the community hits this target, it will not only earn a reward but also save money on its bills.
Specific targets have been set for every community in the project and are set daily. Over time the targets may be adjusted to take into account additional or improved data, or differences between the communities. For example, communities where the local people were already being very frugal with their electricity use before the project started may find it harder to make continual reductions in use as the project continues. We may therefore adjust their targets later in the project. Any rewards that have already been earnedup to the point of changing a target will, of course, be paid.
Each community can earn up £5,000 if they meet their targets. The split is not equal, there is £3,000 allocated across the peak target and £2,000 across the consumption target.
The targets set by the project are quite ambitious. But the rewards are on a sliding scale of “reward bands”, so even if a group misses the target on a particular day, by getting close they will still earn some of the available reward.
For example, if a target was to cut total consumption for that day by 30%, the group would get all the reward if that target was hit, but only half of it if they could only cut their consumption by 10%.
Here’s a table showing the portion of reward given for varying percentages of target achieved:
|Percentage of estimated consumption||Percentage of target achieved||Percentage of reward earned|
This table shows that if, at the end of the day, a community has reduced its total electricity consumption by 70%, it would receive 100% of the reward. A kind of sliding scale is at work, which means that a reduction of between 80% and 90% would earn half to three-quarters of the reward, depending on if they were closer to 80% than 90%.
It also shows that some of the reward (in this case 25%) will be paid even when no apparent saving has been made. This is because we have built in a degree of tolerance to reflect the fact that the baseline data is based on estimates, which may be out by an average of as much as 10%. In other words, we’re giving the communities the benefit of the doubt.
The reward bands vary between communities. This is because hitting targets may be easier for some than others. Larger communities may find it more difficult to hit targets because the impact of a few individual households won’t make as much difference.
The same mechanism applies to the peak reduction target as well: with the reward earned reflecting the percentage reduction of peak consumption compared to the estimate.