Making sense of falling solar costs
Solar costs have been falling rapidly and the promise of a carbon-free electricity system seems ever closer. An auction for utility-scale solar in Zambia earlier in June resulted in a price of 6 $c/kWh, and a recent auction in Dubai led to a price of 3 $c/kWh. If these prices result in real projects, then they are a great reason to be optimistic about the journey to a decarbonised power system. The implications are especially exciting for Africa where there is both abundant solar resource and where the technology could be displacing diesel generators, which are both expensive and have high emissions.
We’re comparing apples with oranges. However, it is only fair to consumers and taxpayers that we compare these options on a like-for-like basis. A unit of power from a solar panel does not have the same properties as a unit of power from, for example, a gas-fired power station. The solar is not dispatchable and so needs to be complemented by other energy resources. The normalised figure below suggests that we would need to store ~50% of the power generated by a solar panel in order to meet a typical load profile.
Solar is also likely to increase demand for system services such as frequency response and reserve, and it may well increase the need for investment in distribution infrastructure. All of these costs are met by the consumer, and so it is disingenuous to focus on the top line.
But a combined levelised cost approach still tells a positive story. So what do the numbers look like if we take a more holistic view? I’ve put together some very simple analysis below that just considers the first of the issues noted above – but probably the most significant. If we want to maintain a carbon-free approach we can consider a solar/battery combination. Greentech Media reports that lithium-ion battery costs in the 200-250 $/kWh range are increasingly being quoted. If we were to use battery capacity to perform 365 cycles per annum, the levelised cost of each kWh output from the battery would be ~12.5 $c/kWh. So, if we assume that 50% of our solar output needs to be stored, then with 6 $c/kWh solar PV the combined solar/battery system cost comes in at ~12.3 $c/kWh.
In this scenario the combined levelised cost falls to the 6 $c/kWh that we see for solar-only systems in Zambia today. The graph below shows the impact of reduced battery capex, and the final “Future Scenario” indicates the impact of (a) 4 $c/kWh solar PV; (b) 100 $/kWh battery capacity; and (c) improved demand-side actions, such that only 35% of solar PV output needs to be stored. Further cost reductions might seem ambitious, but a recent report from IRENA suggests that solar PV costs can fall a further 59% between 2015 and 2025.
My very simple analysis is of course still incomplete. I haven’t tackled system operation costs, and I haven’t tackled network costs (although those might fall with a solar/battery combination anyway). But it gives a slightly more complete view than just looking at the drop in prices coming out of recent auctions. The good news is that even at these costs solar is competitive with existing grid power in many countries.