Negative prices
Fri 4 Sep · Sat 5 Sep · Sun 6 Sep
Research notes
Figures are for London unless they say otherwise.
What happened. Fri: −2.7p at 12:00, below zero 4 hours. Sat: −4.4p at 15:00, below zero 7 hours from 09:00, average 12.6p. Sun: −4.5p at 12:30, below zero 7.5 hours. Evening peaks stayed high (35p, 43p, 47p).
Likely causes
- Strong wind arrived on 3 Sep. Transmission wind averaged 12.0 GW on 3 Sep and 13.0 GW on 4 Sep (14.4 GW at Friday's low point, with Britain exporting). Over the weekend wind was about 9 GW with solar peaks of 12.4 GW and 11.5 GW (confirmed by data) Elexon FUELHH 3 to 6 Sep, PV_Live 4 to 6 Sep
- Day-ahead baseload fell from £114.14 to £94.88 (3 Sep, for Fri) and to £72.76 (4 Sep, for the weekend), with the peak contract £28 to £29 below baseload two sessions running: "solar and wind are covering the daytime block", "while the expensive plant is only called after dark" (confirmed by source) Catalyst 4 Sep, Catalyst 7 Sep
- The balancing market went further: −£16.10/MWh at 12:18 on 3 Sep and −£84.88/MWh at 13:14 on Sat 5 Sep (confirmed by source, same two reports).
- Weekend demand on 5 and 6 Sep (inference).
Wider context that week. Gas kept climbing regardless (175.50p on 3 Sep, 179.00p on 4 Sep), which is why the evenings were expensive. Heysham 2 unit 8 (660 MW) went off for 80 days on 4 Sep. Russia struck Ukrainian energy infrastructure over the weekend. Carbon Brief published its record-solar-summer analysis on 4 Sep. On 5 Sep the US and Iran had a standoff in the strait (US attacked three Iranian ships; Iran fired at a US carrier), per the Wikipedia search summary; I did not open that page section.
Confidence. High.
Gaps. No NESO or press item on these specific days. Energy Live News says a daily balancing-cost record of £32m was set on "Friday" before the £35m record on Tue 8 Sep; that Friday is probably 4 Sep but the article doesn't give the date.