Is there a limit to solar photovoltaic energy in the generation mix?

  ⁽²⁾Faculty of Sciences, University of Lisbon & IDL - Instituto Dom Luiz

In many sunny countries, such as Portugal, solar photovoltaic (PV) energy, with reference costs of 0.03 to 0.04€/kWh, is the cheapest alternative: ̴1/3 of that from new nuclear or biomass generation plants. This is one reason why solar energy will be one of the main instruments for the decarbonisation of the electric system – and of the economy.

The obvious question is: what are the limits to the penetration of solar PV energy in an electric system? What fraction of the generation mix may be solar?

Let us first consider the impact on territory: if it were to satisfy, for instance, 50% of future generation needs, wouldn’t it occupy an unacceptably large area of the territory?

If we assume a future total need of 90TWh/year in Portugal 2050, and a PV generation of 1.5GWh/ha.year (resulting from present technology), in order to satisfy 50% of the total demand, an area of 283km2 would be necessary. This is about the same area as that of the Alqueva lake, and only 0.3% of the territory. If the area is carefully chosen, with involvement of local population, its implantation is not a problem (contrary to biomass, which, for the same generation of electric energy, would require an area of dedicated forest >40% that of the country).

Let us now consider the impact of massive penetration of solar PV in the balance of the electric system. Here we do have a problem, resulting from the time variability of PV generation, and its incapacity to adjust to demand, just like wind. Presently, solar+wind generation is less than 40%, hydropower is ̴20%; with ̴3GW of cross-border connections with Spain, and 4.6GW of natural-gas powered generation capacity, it has been possible to balance the system. However, in the future decarbonised system, solar+wind will be >80%, hydro only ̴10%, and natural-gas is undesirable: the central issue of the future electric system will be imbalance. The central issue will cease to be the economic and environmental costs of generation, it will be the cost of balancing the system.

The problem is illustrated in Figure 1A, where we represent, for a winter week in Portugal 2050, the time-series of demand and of the various generation sources (their relative proportions inspired by the data in the Roadmap for Decarbonisation Portugal 20501), directly reflecting the resources (sun, wind, water, and biomass), with no attempt yet to balance the system. In Figure 1B, we represent demand and the system generation deficit function (at times of generation excess, the function is negative), after all classical means to balance the system have been put to work at their maximum possibilities: water management in the dammed reservoirs, including pumped storage in periods of excess generation, and power management of the thermal biomass plants (with maximum power in times of system deficit, and minimum in times of excess), leaving out only import/export. The imbalance is still brutal, with deficits reaching 9GW (when the sun sets and evening demand is high), and excess generation reaching >20GW (when the sun shines and the wind blows strong).

Does this mean that, in this decarbonised model system, the limit of penetration for solar PV and wind has been largely exceeded, because the system is irremediably imbalanced?

No, it is always possible to balance it, through a combination of (1) additional storage systems (beyond the foreseen reservoirs and pumped hydro plants), (2) generation overcapacity (with curtailment in times of excess) and (3) import/export. The problem is the costs associated with said solutions, which may easily surpass generation itself if badly designed, resulting in excessively expensive electric energy.

There is, however, another possibility, which may be the key for low-cost solutions: to turn a substantial fraction of demand into a flexible load. In other words, instead of trying to adjust generation to demand, find substantial sectors of demand that may adapt to generation availability.

First, we observe that domestic, commercial, and industrial general demand has some capacity to adjust to availability (e.g., management of heating/cooling systems), but very limited, quite insufficient for a relevant system balancing effect. Which relevant sectors may then adjust consumption to availability?

One of them is green hydrogen production via electrolysis. Electrolysers may work at maximum power in times of availability in the system, and reduce or even cancel consumption, in times of lesser abundance or deficit. In addition, stored hydrogen might be used in fuel cells to inject power back into the system in times of need, acting as a firm power reserve. However, the degree of penetration of hydrogen technologies (influenced by high costs and low efficiency) in the economy is still not clear. If green hydrogen production remains limited to present values, the weight of the sector will be insignificant; on the contrary, if hydrogen becomes an energy vector with an impact like that of natural gas, its relevance for system balance becomes substantial.

Another sector, the impact of which is much safer to estimate, is road transportation, which presently resorts to an energy equal to 135% of electrical consumption in the form of fossil fuels and must be urgently decarbonised. The electrification of vehicles with batteries brings about an efficiency gain, relative to internal combustion technology, of a factor of ̴2.5, so the additional electric energy demand by road transportation would be ̴54%, a substantial fraction indeed. Moreover, the mechanical power on wheels of the present ̴7 million vehicles is ̴600GW, 100 times the average power in the grid. The impact of electrification of road transportation on the electric system will be huge. Hence, the relevant question: is it possible to turn this massive demand into a flexible load, which adjusts to availability?

If we carry on with present technology, with battery charging being carried out by plugging the vehicle to a power socket, the answer is yes, but in a very limited way. The vehicles which consume most are the heavy, long-range trucks, and they need immediate fast charging, they cannot afford to wait for times of excess generation in the grid. On the contrary, private light vehicles may be parked 99% of the time, have all the time available, but no consumption, except during long-range trips, when their owners again want fast, immediate charging.

Let us imagine, instead, that vehicles exchange their discharged batteries for charged ones, an operation that takes a couple of minutes, in service stations, where batteries remain coupled to the grid for an average time of 24h, during which period they act as a totally flexible load while charging. The nominal power of these resident batteries, which amount to ̴10% of total on-board batteries, will be 60GW, much more than necessary

• to absorb all power generation excess1.Efficiency in energy use;


• to support the system, as a firm power reserve, in case of deficit.

In Figure 1C we demonstrate the impact of these models of decarbonisation of road transportation, by displaying

• the generation deficit function resulting from a rigid total demand, already shown in 1B, as a reference;


• the deficit function resulting from partial flexibilisation of road transportation demand, in a model of battery charging by plugging the vehicle to a power socket (including a vehicle-to-grid component of grid support with use of vehicle batteries);


• the deficit function resulting from the total flexibilisation of road transport demand, in a model of battery swapping, BS (also including a battery-to-grid component, which interestingly is much less relevant than the dominant effect of demand flexibility).

This last function demonstrates that system balance is easily achievable, even with a penetration of 44% of solar PV generation (and 85% of solar+wind), if we adopt an intelligent model of powering road transportation, turning its consumption into a flexible load.

Finally, the answer to the original question: there is no foreseen limit to the penetration of solar PV generation. The optimum fraction for Portugal will be around 40-50%, as already pointed out by the National Roadmap, and we demonstrated that there will be no problem with its implantation nor with the electric system stability – provided we embrace the correct options.

References

1. Ministry of Environment and Energy Transition (2019), Roadmap for Decarbonisation Portugal 2050 (RNC2050).

2. A.M. Vallera, P.M. Nunes, M.C. Brito (2021), Why we need battery swapping technology, Energy Policy, 157:112481. doi.org:10.1016/j.enpol.2021.112481.