Utilitywise successfully called all three Triad dates and only issued 10 Red alerts.
National Grid, the operator of the UK’s electricity transmission network, recovers its costs through Transmission Network Use of System (TNUoS) charges. While the rates paid vary from consumer to consumer, at the heart of them all is the Triad methodology. The expenses that National Grid recovers reflect system investment and management needed to meet the three half hours of highest national transmission system demand over the winter. They occur between November to February each year and must be split by at least ten clear days. For many businesses billed on a half-hourly basis, TNUoS costs will directly reflect their own peak demand during the three Triads. If a business can avoid using as much energy as possible at these points, they could cut their overall transmission costs. The trick is trying to predict when the Triads will occur and react accordingly.
During the last Triad season, Utilitywise provided daily reports for customers on the potential for a Triad to occur. Over those four months, we indicated there was a possibility for a Triad on 17 days. For ten of these occurrences we warned it was very likely (our strongest warning) that a Triad would occur.
The amount of alerts called by Utilitywise is almost 30% less than last year. This means that businesses would have had to reduce or change production on fewer occasions if they reacted to the alerts to cut their transmissions costs. Overall, those issuing alerts seem to have been more successful in predicting the Triads this year. While the ideal scenario is to call just three alerts through the season, predicting all three Triads accurately at the same time is extremely difficult.
National Grid is the main source of data and information on the operation of the electricity supply system from day-to-day. The information, provided through balancing service administrator Elexon, gives forecasts of demand across the year. The accuracy of these forecasts understandably drops the further in the future the event is. However, even at the day ahead level there are difficulties.
In the graph below, initial demand for the weekdays in the last Triad period is compared to both the National Grid forecast in the morning, and that in the afternoon, closer to the point of peak demand. Even with the later forecast, there are notable differences between forecast and actual demand.
In most cases, actual demand ended up below expectations, which would impact upon any Triad alerts based on National Grid forecasts.
There are many factors which affect the accuracy of demand forecasts. A key element is weather conditions, such as temperature, which may vary from expectations and dramatically alter demand. The weather could also change the level of wind generation, affecting peak electricity system demand, as more businesses use their onsite generation for energy. Additionally, there is the very act of Triad avoidance itself. Businesses reacting to a potential Triad will cut overall system demand. This reduction in peak system demand may mean that the day turns out not to be a Triad after all. These factors are all changeable, volatile and affect demand in a relatively short time.
The scale of error in National Grid demand forecasts is having an increasing impact on Triad alerts, particularly as UK demand is generally falling, helped by the push for energy efficiency.
Demand levels from day-to-day are also varying less over the Triad period. At points when demand has reached levels where a Triad is likely, the daily variation in demand has sometimes been as little as 300MW. With potential levels of error of 450MW, it becomes difficult to easily identify which of the days will be a Triad.
This is particularly a concern when you are seeking the third Triad. In recent years, it has been quite clear what the highest or even second highest demand levels will be due to the scale of demand at these points. However, chasing that third Triad, becomes more problematic if there is very little to distinguish demand levels from day-to-day.
What you see is not what you get
Matters are complicated as demand used to define a Triad is provided over the winter. The calculation of system demand goes through different stages of ‘settlement’ before it is considered final. This is due to more accurate data becoming available over time. With demand variations being smaller, the changes caused by settlement become more important. This was highlighted in 2015/16 when a day which many thought was a Triad – and one National Grid itself was forecasting as one of the three peaks – did not turn out to be a Triad. This meant that some actually missed one of the three Triads – though Utilitywise did not. In the end, there was less than 1MW difference between the two potential Triad points under debate, so we’re proud to have identified this correctly in such close circumstances.
This year saw something similar occurring. The National Grid operational demand data, which equates to initial demand estimates, suggested different Triad points to the data from current settlement data. The data ultimately ended up being more accurate. As Triads need to be ten clear days apart, it does not take much of a change in demand to affect when the Triads could ultimately be called.
The next evolution
The final results of this year’s Triad period will not only confirm the scale of Transmission costs consumers will face, but will help inform and improve the forecasting of these potentially very costly demand points.