Review behaviour of electricity storage forecasting algorithm

[mabijkerk]

We have been notified by some of our clients of several potential shortcomings of the forecasting algorithm. These warrant further investigation.

• 1. After some changes with batteries in a 2050 II3050 scenario, the sum of elements in the chart "total costs" was not equal to the total presented in the table (not able to reproduce yet)
• 2. When the installed capacity of batteries in a 2050 II3050 scenario is reduced, the full load hours decrease in some cases
• 3. Local forecasting with large amounts of solar PV in households and demand in households, despite large installed capacity of household batteries, the batteries only charge with a limited capacity
• 4. Forecasting has unexpected consequences for price-setting behaviour in this scenario: on January 1st curtailment is active but prices jump from 0 tot ~5 euros/MWh

[mabijkerk]

The following scenario reproduces issue 1: https://energytransitionmodel.com/saved_scenarios/16595.

The error seems to be caused by a rounding error in the Energy carriers import/export group. For 2 GW li-ion batteries the decimals reappear and the sum total is equal to the total row.

[kaskranenburgQ]

When I open this scenario, the correct numbers seem to appear:

[mabijkerk]

The issue for 4. is likely the same that we encountered for hybrid offshore wind. As @kaskranenburgQ puts it:

Merit sets the price equal to the lowest flexible consumer that does not completely fulfills it's demand. Since in these situations the electrolyser does not completely meet it's demand, it will set the price, while it does not reflect the actual market price.

When battery forecasting is on, the price is likely set to the max_consumption_price and marginal_costs as specified on the battery node. We can quickly check to see whether this is indeed the case.

For hybrid offshore wind we solved it by exclude the must-run part of the electrolyser from pricesetting. This worked because the electricity would otherwise have been curtailed anyway. For batteries this is not the case. We therefore need to consider a suitable solution for pricesetting for batteries with forecasting.

[github-actions]

This issue has had no activity for 60 days and will be closed in 7 days. Removing the "Stale" label or posting a comment will prevent it from being closed automatically. You can also add the "Pinned" label to ensure it isn't marked as stale in the future.