Island grid: how much of a 20 MW solar array is actually usable
An island nation of 84,000 people generates 62 GWh of electricity a year from six diesel generators, with peak demand of 45 MW and a daytime minimum of 24 MW. Diesel fuel costs US$0.28 per kWh and the utility emits 0.72 t CO2 per MWh generated. The government holds a US$34 million concessional loan and has a 2030 target of 40% renewable electricity. Two options are on the table. Option A is 20 MW of solar PV with a 15 MWh lithium battery, built on 32 ha of abandoned sugarcane land bordering a RAMSAR wetland. Option B is 6 MW of rooftop solar on commercial buildings in the capital, with no battery. Grid studies show that instantaneous solar output above 25% of demand on any feeder must be curtailed unless it is stored. Solar capacity factor is 17%, battery round-trip efficiency 88%, and usable depth of discharge 80%.
1. Calculate the annual generation and CO2 saving of the 20 MW array if none of it were curtailed, then explain why the real figure will be lower.
Read the worked answer
Gross generation: 20 MW × 8,760 h × 0.17 = 29,784 MWh, or 29.8 GWh.
Share of demand: 29,784 / 62,000 = 48%. On paper that alone nearly meets the 40% target.
CO2 avoided: 29,784 × 0.72 = 21,444 t CO2 a year, against 62,000 × 0.72 = 44,640 t for the diesel-only system.
Why the real figure is lower - the instantaneous problem: the 25% limit is on output as a share of demand at that moment, not on annual energy. Daytime minimum demand is 24 MW, so the allowable solar output is 0.25 × 24 = 6 MW. The array produces up to about 17 MW at midday (near 85% of capacity). Roughly 11 MW has no permitted home at that instant.
What the battery can absorb: 15 MWh × 0.80 usable = 12 MWh, and at 88% round trip about 10.6 MWh is delivered back. That is about one hour of midday output, shifting perhaps 3,800–4,400 MWh a year - around 13–15% of gross solar generation.
Realistic capture: perhaps 55–65% of 29.8 GWh, so roughly 12,000–15,000 t CO2 avoided, not 21,444. Quoting the gross figure would overstate both the climate benefit and the fuel saving by nearly half. The constraint is grid flexibility, not sunshine.
2. Evaluate the choice of site for Option A against Option B, considering land use, the protected area, and the curtailment constraint.
Read the worked answer
Option A, siting: the sugarcane land is already cleared, so direct habitat loss is low compared with clearing intact vegetation - that is a genuine advantage, and it is the argument the developer will lead with. But it borders a RAMSAR wetland: construction sediment, runoff, panel glare and bird strike are plausible impacts, and an environmental impact assessment is likely to be required, with schedule and cost risk. 32 ha is also a permanent land-use commitment in a small island economy.
Option A, curtailment: the array is sized against annual energy, not against the 6 MW instantaneous limit. Building 20 MW of generation behind a 6 MW interconnection is a design error; the loan would buy a great deal of capacity that is regularly switched off.
Option B, siting: 6 MW × 8,760 × 0.17 = 8,935 MWh, about 14% of demand, no land take, no wetland conflict, and generation is at the point of use, so it is spread across feeders and does not concentrate penetration on one. Higher cost per kW and harder installation logistics are the trade-offs.
Judgement: B first. It clears the wetland objection and the curtailment constraint, and it delivers roughly 6,400 t CO2 a year with low ecological risk. Reserving the sugarcane site until storage, demand response or an interconnector lifts the 25% ceiling is the defensible sequence - the constraint is the grid, not the land.
3. The minister wants the full US$34 million spent before 2030. Propose a spending sequence and the criteria you would use to justify each step.
Read the worked answer
Criteria (agreed before options are ranked): tonnes CO2 avoided per dollar of loan; cost of displaced diesel (US$0.28/kWh fuel plus logistics and price volatility); ecological risk to the RAMSAR site; and whether the investment relieves or worsens the 25% penetration limit.
Sequence:
- Rooftop solar, up to 6 MW (~US$8–10 million). Clears both constraints, immediate fuel saving, no land use. Highest CO2 per dollar once curtailment is properly costed.
- Grid and storage enablers - feeder automation, demand response, and additional battery capacity beyond 15 MWh. These have no generation of their own but they raise the ceiling that makes any large array bankable. Spending here is what converts option A from a 60%-captured asset into an 85%-captured one.
- Ground-mounted solar at the sugarcane site, sized to the interconnection limit that now exists, with an EIA, bird monitoring and a construction runoff plan for the wetland. Scale it after step 2, not before.
- Wind, only after a year of site data, because its output profile is complementary to solar but its ecological risk to birds on the island is unknown.
On the deadline: spending US$34 million by 2030 is not a target, it is an input. A 20 MW array built into a grid that curtails half its output meets the deadline and fails the objective. The defensible answer to the minister is a smaller first phase with a firm option on the second, and a published capacity factor and curtailment figure each year so the 2030 target is judged on delivered renewable electricity, not on installed capacity.