×
SERVICES
ABOUT US

Helping you make sense of the Irish energy sector

Publications

SEM Chart of the Week

2019

Prev
Next

Wind of Change: Analysing SEM wind forecast accuracy

Joe Camish
Joe Camish

Following on from the publication of the Tomorrow’s Energy Scenarios (TES) 2019 last month, we can expect to see greater penetration of wind generation in the coming 20 years, with installed capacity forecast between 6.0GW to 8.2GW by 2040.

As the growth in wind generation continues, forecasting wind accurately will become increasingly important in order to balance demand and supply in the SEM. In this week’s SEM Chart of the Week, we analyse EirGrid’s wind forecast against actual generation over the past two years.

Observing forecasts throughout 2019, we can see that actual generation has predominantly fallen in the ‘5% higher than forecast’ bracket. Wind generation, more broadly, can be seen to have been under forecast across 2019, with 228 days experiencing higher than forecast wind output, with more than half of these days 5-10% higher than expected.

It should be noted that this trend has eased year-on-year, with 2018 experiencing 242 days of higher than projected wind output. Improved forecasting accuracy this year is represented in the chart with bars falling closer to the middle of the distribution. The previous year experienced more extreme episodes of generation, being >25% lower or higher than forecast.

As we get closer to the year’s end, we can see that while forecasting remains conservative, fewer extreme events have occurred throughout the year compared to 2018. This indicates that the system operators’ forecasts have improved year-on-year, an important development for balancing the system and for market participants actively trading and operating in the SEM.

Figure 1 shows the distribution of the percentage difference between actual wind generation against EirGrid’s wind forecast year-on-year. The y-axis represents the number of days, with the x-axis representing percentage ranges that each day fell into (e.g. 0-5% higher than forecast).

Can we help?