This week the IEA (International Energy Agency) published its 2019 Renewables report. The report foresees strong growth in renewables – driven by decentral solar power.

We are seeing an increase in summerly imbalance costs driven by residential solar, which if the IEA is right, will continue to increase. What effect will this have on imbalance?

Solar growth

Renewable power capacity is set to expand by 50% between 2019 and 2024, led by solar PV, driven by a cost reduction of another 15% – 35%, concludes the IEA report. It is expected that there will be 100 million residential rooftop PV systems in 2024.

This growth is happening in Europe: The Netherlands, Belgium and Austria, which are Dexter’s primary markets, are in the top five most attractive markets for residential PV installations per capita. This brings major challenges (and opportunities) to energy companies and grid systems. 

Synthetic load profiles and summerly imbalances

One of the challenges is that most small consumers are allocated using synthetic load profiles instead of quarterly data that is collected in smart measurement devices (smart meters). This means that imbalance costs are calculated and rolled over to energy suppliers using a so-called measurement correction factor (MCF).

The MCF is the difference between the measurements on grid-area level deducted by the expected load, using the pre-defined synthetic profiles and scaled by the expected yearly consumption. 

The problem is that many households, due to a solar roof, EV charging or e-heat pumps will have a very different real load than the allocated synthetic load profile. In certain net area’s, a MCF < 0.5 is not the exception anymore. In other words, the actual consumption was only half as was ‘planned’. With the lack of knowing the load per consumer per quarterly data, the imbalance is socialised pro-rata to all suppliers. The costs are subsidised

In the last summers we have seen a strong increase in imbalance related to forecasting errors caused by solar effects. Given the expected growth rates also expected by the IEA, we may expect that these summerly imbalances are likely to increase. 

The plot depicts the variability of allocated loads of all small consumers plotted against the hour of the day. The subplot rows are years 2015 – 2019, the columns are week of day. It can be observed that the variability during midday hours throughout the year is increasing. As the synthetic load profiles have remained rather constant over the years, this increase in variability in allocated load shows the variability of the MCF.  

What to do about it?

Solar and wind power is inherently challenging to forecast (even if you have 15min data). But with advanced forecasting techniques and by combining physical models with statistical techniques (machine learning), these errors (and imbalance costs) can be managed and reduced with increasing amounts of data. 

When looking at synthetic profile clusters implementing good forecasts is more challenging. The root cause of this challenge is not having quarterly data on which forecasting models are trained on. That is what the IEA report means when saying: 

“Major policy and tariff reforms are required to make distributed PV growth sustainable.”

In the Netherlands, a way of handling this is to transfer residential PV customers out of the synthetic profiles and into smart meter allocation, the so-called rollout of the new allocation proces “Allocatie 2.0”. With this change the small consumers will (for the most part) be allocated using quarterly data.

This allocation technique is called smart meter allocation (SMA). This causes the residual unallocated load to be much smaller. The advantage is that there are less costs on a system level as well as on a BRP level (balancing responsible party). Secondly, imbalance costs are driven down as forecasting algorithms have historical data to work with, thereby increasing the accuracy.  Allocation 2.0 will be implemented in phases and how long the full change will take is still a bit unclear. 

Another opportunity with smart meter allocation will be the marketing of new energy products. With a more exact allocation, energy companies can offer smarter products tailor-made to consumers with PV, EV or e-pumps. 

Last but not least, when registering negative loads – feeding into the system produced by solar – one could market and label this electricity as green. Currently, green energy is sold by buying certificates of origin. However, these certificates are only obtained from large consumer categories (Grootverbruik, 3x80A connections). An argument could be made that residential consumer with PV feeding into the grid in the portfolio of a supplier reduces the net positive load and is therefore still selling energy in a green way – even though there are no certificates of origin.

An executive summary of the IEA renewables 2019 report can be found here. If you want to know more about load forecasting on small consumers with PV (smart meter forecasting), the Smart Meter Allocation 2.0 or MCF forecasting, please reach out: hubert@dexterenergy.ai

Play Video