National Grid is set to publish the three confirmed Triads for winter 2015/16. Triads form a vital part of calculating the costs for electricity transmission charges. For those who have been seeking to reduce their exposure to transmission costs via triad response, the publication of the dates will also confirm how successful they have been. However, the task of avoidance is getting increasingly complex.
The Triads are the three half hours with the highest system demand between November and February each year. Crucially, though they have to be a clear 10 days apart. The scale of demand recorded during these periods is used in National Grid’s calculation of its transmission charges. For consumers, with all-inclusive contracts Triads have little impact. However, for business on energy-only contracts, the Triad dates have a direct effect on their transmission charges.
In an ideal situation, a business would be able to predict the Triads and cut demand accordingly. However, there are many factors which make this difficult and complex. At the heart of many of the barriers towards a ‘perfect’ warning system is demand itself.
Trusting the forecast
National Grid demand forecasts have been above the actual final demand figure for most of the last winter. Forecasts on the day, were on average 1.4% above the actual demand outturn for the day. This rose to 5% for longer-term forecasts. This is in stark contrast to the level of error the previous year, which was around 1%. This has ramifications on estimating the accuracy of the demand data used in any assessment of a possible Triad.
The accuracy of forecasts is a vital part of any Triad avoidance scheme. The level of error is important, not just for estimating the validity of demand on the day, or on the next, but also for the coming weeks. Triads need to be at least 10 days apart, so peak demand on any day must be compared to that in the following and preceding weeks. If future demand is expected to exceed demand on the day, this reduces the likelihood of a Triad occurring in the present.
National Grid’s forecasting error is the result of many uncertain factors in its forecasts, such as weather and seasonal demand changes. This has been complicated in recent years by the growth in small-scale renewable generation. These sites offset the use of electricity from the main supply system. National Grid does not have full details of their actual output, making it very difficult for the system operator to estimate the scale of demand reduction they might cause. Embedded wind is just one contributor to reducing the accuracy of system demand from National Grid.
Triad avoidance – a victim of its own success?
Further issues surround the act of trying to avoid the Triad itself. As people become more aware of the Triad, they have also taken more steps to deal with it. This has created a situation where businesses are cutting demand on a particular day due to the possibility of a Triad and in so doing reducing the possibility for that day being a Triad.
National Grid has estimated that demand side response to a possible Triad has increased by over 40% in recent years. It has also found that the scale of demand reduction has also risen. Given the other factors influencing National Grid forecasts, it is difficult to focus on the full impact of a response to a Triad alert. However, National Grid has estimated that, at times, demand response cut forecast demand by between 1.6 and 2.0GW. During the 2014/15 winter, the level of response struggled to reach 1.5GW.
Such a swing in demand would easily lower peak demand to a point at which it would no longer be deemed a Triad. In such a situation, businesses would have reacted, cut demand, potentially incurring a cost for doing so, but with no benefit from avoiding a Triad.
Duos and Tuos – an unhappy overlap
Businesses may be making it harder to predict Triads by seeking to cut their distribution charges. So-called ‘Red-Band Avoidance’ is helping businesses cut their energy bills, but it may also be affecting overall energy demand.
Distribution networks divide their charges between colour coded time bands to reflect the different demand pressures they face over the day. The highest charges are the ‘Red-Band’ and are paid for usage in the late afternoon. However, the Red-Band time period is designed to overlap with the period of highest system demand. This means it also overlaps with when a Triad called be called.
The general push for energy efficiency has contributed to a general trend in demand reduction over the last few years, while schemes such as the Red—Band avoidance has helped to flatten demand from day-to-day. In the last winter, the mild weather conditions for much of the season enhanced this effect, keeping demand both subdued and virtually flat from day to day.
For identifying potential Triads, this provides a significant challenge. If demand on any particular day is very close to that on the previous day or the day after, how certain are you that it will end up being a Triad? Once any possible forecasting error or demand side response to a Triad alert is factored in, the margin of demand between the days could fully disappear. Every day starts to look like a possible Triad and you return to the key difficulty of identifying the peak days with the least amount of false positives.
It is clear, with issues over demand side avoidance adding to the standard concerns over uncertain weather conditions and forecast error, the task of identifying a Triad is getting much more complex.