Commercial Solar Parks

-do they make financial or engineering sense?

Recently, plans have been put forward to create several massive UK solar parks of multi-megawatt peak output. The economics of such projects differ considerably from domestic installations. On the positive side, economies of scale and bulk purchase will bring equipment costs down somewhat. The key issue faced by such projects, however, is that they must sell their electricity into the Grid at wholesale rates rather than the retail price paid by consumers. The wholesale price of electricity varies considerably due to supply and demand factors. Typically it is in the region of £45 to £60 per megawatt-hour, for energy from coal or gas plant.  Nuclear energy is somewhat dearer at around £80, whilst onshore wind energy is typically bought at £100 per MWh.

The financial arithmetic is much the same as for a domestic install, and the obvious issue is that a solar park faced with only £60 of return per thousand units generated, would take several decades to recoup its equipment costs. To provide the investor with any sensible level of payback will require a massive subsidy from the public purse, far larger percentage-wise than the domestic FIT. Typically the present levels of subsidy to such installations take the price paid per unit up to around the same as that of retail power.

There are other issues, beside the astronomic cost. Whilst the relatively small output from a domestic install can be readily absorbed by the Grid, a multi-megawatt solar park is a different proposition. We are already familiar with the situation where windfarm output variation demands the use of CCGT backup rather than coal or nuclear, gas turbines being the only source able to track wind output variations over periods of typically tens of minutes or a few hours.

With small-scale solar, very rapid variations are a problem due to clouds obscuring the sun, etc. The variations in  large scale solar output are more gradual, however, as the effects of partial cloud cover tend to average out over a large array. What we do see from German output figures though, is that useful output tends to occupy a relatively narrow time-slot centered on mid-day. Thus, the issue here is more one of dispatchability - Can the energy be put to use during the time slot in which it is available? If not, and in the absence of storage, it goes to waste.

Germany, with well over 20GW of installed solar, is already suffering load-balancing issues due to a glut of solar electricity around mid-day on sunny days, coupled with little solar output at other times or on overcast days. These load-balancing issues give rise to wastage at fossil-fueled stations, and thus reduce the benefit of solar.

Aside from the limited time-slot the solar electricity is available over, there is the sheer amount of land required, should solar be required to supply any significant proportion of UK demand. UK demand is typically 40GW. (Forty million one-bar fires, for illustration purposes) 

Suppose we wished to supply 8-10% of UK demand from solar -around the same level as present wind capacity. If you recall our ballpark figure, a solar install typically gives the equivalent of 2h of full output per day. So, to equate to 24h per day of 4GW output, our solar farm would require  12x4GW peak capacity, or 48GW. Costwise, at an optimistic £0.50 per watt,  the panels alone would be £24 billion. To put this in perspective it's about the cost of thirty conventional gasfired power stations, nearly enough to supply the entire UK electricity demand. For your £24 billion, you will get a system whose output probably exceeds the current-carrying capacity of the local Grid cables when the sun is strong and nearly overhead, but which delivers very little on an overcast day, and nothing at all at night.

We know that a square metre of panel gives 150W peak output, so our 48GW array would have to measure 48/150 billion, or 320,000,000 square metres at absolute minimum. That is pure collecting area, and of course any realistic installation would need to allow space between rows for servicing access. In practice, 130MW per square mile is a more typical density, which would give us a groundspace requirement of 369 square miles for our 48GW array. This is about three fifths of the area of Greater London. Or, a three-mile wide swath of panels reaching almost from Edinburgh to Aberdeen. And, on average, that will supply only a tenth of the UK energy demand.

Where solar farms are placed on agricultural land, they represent a loss of land which could be productive in other ways. Thus, the cost of losing this production has to be factored-in. At least, the city rooftop installations don't have this issue.

German solar output, june/july 2015. Source: Fraunhofer ISE.

The main issue, though, is that the best part of the energy is only available over a relatively limited time of day. In March, this may be seven hours of at least 50% of peak output. In summer, possibly ten hours. The promoters of solar technology are always talking about energy storage becoming available to solve this issue, but in reality no such technology currently exists. If our solar farm happens to generate its expected maximum watt-hour rating on a favourable day, but everyone is out sunbathing so it isn't immediately needed, then that would require us to find somewhere to store 48x12 gigawatt-hours, or 576,000,000 units of energy. That is one colossal amount of energy to store. For comparison purposes, typical car battery stores one or two units of energy, and costs maybe £40 at trade prices. Other more exotic battery types can store larger amounts, but are proportionately more costly per unit stored.

How about hydro storage, then? The UK presently has a few pumped storage stations which are useful for meeting surges in demand, but their capacity is simply nowhere near the amount required to back-up a solar farm network. 

There are other suggestions, such as turning the energy into hydrogen gas by electrolysis, but all are unproven ideas, and all would reduce the efficiency and increase the cost of the installation still further. Basically, it has yet to be demonstrated that any such ideas are feasible, and the nature of the situation makes them somewhat implausible as solutions.

Overall, it should become apparent that the economics of large-scale solar farms are quite different from those of domestic solar installations, especially as they must compete with the much lower pre-distribution wholesale energy prices. If the government were to heavily promote such projects, it would most likely lead to a huge price hike for domestic electricity users. In combination with wind infrastructure, it has had precisely this effect in the countries with the greatest deployment.