Retain useful gas capacity. Give coal changing roles.

What the tests establish

Five-year gas retention reduces transition shortages, but the retained capacity later retires. It therefore does not translate one-for-one into permanently avoided new plants. The central candidate with30 new combined-cycle stations matches the tested pass-year pattern of the34-station candidate. Cutting to20 stations leaves gaps for longer. Ten-year retention and additional coal are separate sensitivities, not engineering-approved commitments.

Several constrained cases still fail normal or synthetic supply screens. Their costs are not equal-service comparisons and do not price unserved electricity. The original higher-throughput presets remain available as controls; these alternatives do not prove the original schedule deliverable or the constrained cases impossible.

Choose a completed test

Gas capacity through the transition

Blue: total gas nameplate capacity. Orange: additional inherited capacity retained beyond its original estimated retirement. Availability and operating constraints are applied separately.

New equipment through2050

CCGT:920MWstation; engine site:100MW comprising5×20MWengines; OCGT:300MWsite; coal:2GWstation. Unit counts are not comparable without their capacities. Deferred/unplaced CCGTs are excluded from commissioned counts.

How retention is represented

The five-year case retains75% of eligible named CCGT capacity after an assumed inspection programme. It uses85% availability,54% full-load efficiency, a25% refurbishment-year outage allowance,£150/kW refurbishment expenditure per five-year tranche and£10/kW/year extra fixed maintenance. These are conditional analyst estimates, anchored to published refurbishment evidence.

The ten-year sensitivity uses80% availability,52% efficiency,£250/kW per tranche and£15/kW/year extra O&M. A second overhaul and inspection are paid for after five years. Unknown forecast residuals, small peakers and ambiguous pre1990site dates receive no extension. No operator-approved station list is implied.

Original inherited-gas capital recovery has ended before the estimated40-year retirement boundary. Retention pays for refurbishment, maintenance and fuel; it does not incur a hypothetical new-plant capital annuity. Refurbishment borrowing uses the same4.5% real rate as the plan, with five-year recovery. Construction cash is reported separately and is not added again to annual capital recovery.

Coal can have mixed objectives

Up to4 GW nameplate—two station equivalents—is regularly committed through2044. Remaining coal follows an economic schedule. From2045 all coal moves to reserve duty: gas, scheduled imports and storage cover demand first, while coal starts in anticipation of residual needs under the existing ramp/start model.

The late six-station case adds units in2039/2040 under the original one-station-per-year limit. The early six-station case adds units in2035/2036 and explicitly raises the coal envelope to two stations/year. The latter tests a stronger construction assumption; it does not secure suppliers, sites, finance, domestic mining or logistics. The existing domestic coal supply assumption remains unchanged.

Procurement, fuel and interpretation

Slow cases allow at most2 new CCGTs/year with a five-year order-to-online lead,1.5 GW engines/year and0.9 GW open-cycle turbines/year, with a shared4.24 GW annual ceiling. No discretionary orders precede2029. Fast controls retain the earlier4 CCGT/4GWengine/7.36 GW shared envelope. These are procurement sensitivities, not measured global manufacturing limits.

The engine/OCGT ladder and CCGT reductions are bounded heuristics. Every final dated candidate is tested against7 historical-weather profiles and2 synthetic28-daycold/low-wind/nuclear-outage stresses for each year2030–2050. No stress probability, dynamic frequency certification or global least-cost optimum is claimed.

Costs cover generation, storage and bounded trade. Separate new planning/grid connections, wider networks, retail, complete fiscal settlements and fuel-security infrastructure remain outside the headline subtotal. Carbon and support payments are separate transfers.

Sources and reproducibility

• CEPA/GHD sections 2.3–2.4; five-year extension £81/kW in 2024 prices for a modern H-class archetype; inspections are required. Ten-to-25-year repowering is distinct and replaces major equipment.
• SSE Thermal existing-fleet life-extension programmes; does not substantiate age-50 retention of every modelled station.
• Wärtsilä production expansion and lead-time evidence
• GEVernova turbine production and backlog evidence

The complete package contains per-year capacity searches, all connected replay results, dated plans, assumptions and source register. MethodsM50–M52 describe retention, procurement and staged coal. Raw results are in retention_study/annual_results.csv and summary.csv.