Gnarum | Empowering your energy
The new QH market in Spain allows for more accurate grid management and deviation handling. Market participants will need to update forecasts and intraday plans as settlements move to 15-minute intervals.
The implementation of the quarter-hourly (QH) market in the Spanish electricity sector is a significant change that allows for more precise deviation allocation and a more optimized operation of the power system. However, it requires market participants to adapt their forecasting, operational, and settlement systems to handle these higher-granularity intervals.
Starting on May 1, 2026, deviation settlements will be based on 15-minute measurements, according to the Resolution of March 28, 2025, issued by the Secretary of State for Energy.
This marks the end of the transitional period in which REE settled deviations using a single measurement applied to all four quarter-hour intervals. From May onward, settlements will be carried out on a quarter-hourly basis, either using actual meter data or interpolations, for all production facilities and consumption points in the power system.
This update to the electricity system also requires forecasting systems to perform hourly-to-quarter-hourly interpolation, enabling market participants to adjust their positions in the intraday market, define 15-minute programs, and reduce the impact of deviations.
The availability of historical data in 15-minute intervals always represents a significant competitive advantage over the traditional hourly approach, as it allows intrahour patterns to be captured more accurately, leading to more robust models adapted to the requirements of the new 15-minute settlement scheme.
However, because this quarter-hourly approach entails higher technological costs for market participants, or due to insufficient historical data, market agents typically work with hourly-based forecasts, which are then interpolated to estimate values at 15-minute intervals.
The choice of interpolation method will depend on the type of measurement point and the specific requirements of the electricity market:
Measurement Points Type 4 and 5: In these cases, linear interpolation will continue to be used, following the method established by REE. This approach is sufficient to meet the settlement requirements for these types of points.
Measurement Points Type 1, 2, and 3: For these points, where settlements will be based on actual quarter-hourly (QH) measurements, it is necessary to implement an advanced interpolation process with specific properties:
– Continuity of the function and its derivatives (up to the second order): This ensures smooth transitions between adjacent segments, avoiding abrupt changes in slopes or curvature.
– Flexibility to model nonlinear behavior: The interpolation function must have adjustable parameters to capture complex patterns.
– Overall consistency: The interpolation must ensure that the sum of the 15-minute interpolated values exactly reproduces the original hourly measurement. This condition is essential to guarantee coherence between daily hourly market operations and 15-minute intraday operations.
In the particular case of photovoltaic (PV) energy, studies and analyses conducted by Gnarum show that once sufficient historical quarter-hourly data is available, results obtained from training and calibration at this same granularity will outperform those derived from hourly training followed by a subsequent interpolation process.
In the case of electricity consumption forecasting, it is not necessary to accumulate an extensive 15-minute historical dataset, as future behavior depends more on recent trends and short-term patterns. This allows for immediate results without the need for large volumes of historical data.
Another key factor for improving forecast accuracy is having meteorological data with the same resolution as the measurement intervals.
Sometimes, a lack of temporal consistency between meteorological data and QH measurement intervals can introduce noise and distort predictive models. Therefore, it is essential to align the temporal resolution of both datasets.
This is especially critical for technologies such as wind power, as rapid fluctuations in wind speed make it essential to have meteorological data at QH intervals to capture these variations and improve forecast accuracy. According to predictive models developed by Gnarum, significant improvements are only achieved when meteorological data have the same granularity as the generation data.
The future of the electricity market, as we can see, points toward the adoption of very short-term forecasting systems that combine real-time data with advanced predictive models.
With the entry into force of the Electricity Market Design Reform (EMDR), scheduled for January 1, 2026, trading in the Continuous Intraday Market (MIC) will play a central role, as the 30-minute trading frequency ensures that Nowcasting predictions generated every 15 minutes can be utilized in the market, maximizing their accuracy and usefulness.
For plants participating in Balancing Services (SSAA), the model would be complemented by the integration of producible energy provided by SCADA systems.
In this context of transformation and increasing demand for accurate 15-minute forecasts, Gnarum stands out as a key player, actively working to deliver the best predictive solutions on the market.
Their efforts focus on several core pillars:
Training algorithms directly from 15-minute historical measurements: We use historical data with 15-minute temporal resolution to train predictive algorithms. As mentioned, this approach results in more robust models. It is especially useful for high-consumption CUPS, as it directly improves deviation management without the need to accumulate extensive historical data.
Higher-resolution meteorological models: Gnarum employs meteorological sources with high temporal resolution to improve renewable energy generation forecasts, particularly for wind farms.
Nowcasting models integrating real-time high-frequency measurements: These models developed by Gnarum allow anticipating critical events, such as sudden demand changes or generation interruptions, enabling rapid and efficient responses.
Automated trading systems for high-frequency generation operations: Gnarum has implemented automated trading systems that operate with high-frequency generation of offers calculated based on energy imbalances, 45 minutes before the offer submission. These systems allow market operators to execute optimized strategies in real time and unattended.
Thanks to these advances, Gnarum positions itself as a leader in providing predictive solutions for the electricity market, helping its clients successfully navigate the transition to 15-minute settlements.
Need guidance? Don’t hesitate to get in touch with us.
