Renewable chic and the expense of small-scale solar

This is what I came across recently: a luxury brand store window display of “renewable chic,” where a solar panel, as part of the vignette, is being used to partially illuminate the mannequin. Putting aside the irony that the solar panel itself was apparently being illuminated by grid-powered lights, there are broader implications we need to anticipate with the increasing popularization of small-scale solar. Let’s think twice about increasing small-scale rooftop solar, because right now the economics just don’t make sense.

First of all, there is the compromise in quality of service. Small-scale—up to a few kW—rooftop solar systems are set up primarily to inject as much power as can be generated from available sunlight. In small numbers on individual distribution feeders, this is not a problem. But in large numbers, technical challenges such as voltage control and fault protection on distribution feeders become problematic. Until recently, interconnection standards did not allow rooftop solar systems to participate in voltage control. (This may be remedied for new installations with updates to the IEEE 1547 Interconnection Standard.)

This leaves us with at least two salient and interacting concerns.

First, the expense. Small-scale rooftop solar remains much more expensive to install than large-scale solar, because existing household rooftop sites are unlikely to have the best exposure and orientations toward the sun and because the small-scale is almost always going to involve higher costs per unit capacity than large-scale installations. The store display shows this latter issue in microcosm: there was presumably several hours of labor and significant installation cost for this panel of just a few hundred watts! Rooftops of several kW are less expensive per unit capacity than this small panel, but even larger installations are less expensive per unit capacity. In the United States, it is reported that small-scale solar is double or more the cost of large-scale solar.

(An added concern: if large-scale solar is located far from urban centers, the solar may be cheap, but it may also require expensive transmission upgrades. A good compromise will often be medium-scale developments using large commercial and industrial rooftops and community installations. This way, the solar is nearly as cheap and there is no significant extra cost for transmission. Click here to read more about locating distributed generation in urban areas.)

Second, solar production is variable, depending on the available sunlight, which means that the power injection into the grid is variable. But, as I said earlier, small-scale installations are mostly set up to inject as much power as possible from the sunlight. So, as sunlight levels change, the power will vary. Sometimes large amounts of solar power production will coincide with times when the power is not needed.  

In contrast, large-scale installations are easier to set up as being dispatchable. That is, large-scale solar can be controllable to produce less power when that power is not needed. Such circumstances are bound to become more common, as the so-called California duck curve shows. To date, essentially no residential rooftop solar has been dispatchable, so when it is sunny in California but demand is low, there is a need to dispatch down the remaining thermal generation. This might be acceptable by itself, but when the sun goes down in the evening, the demand typically increases in California, and this is threatening to result in situations where the thermal generation has insufficient ramping capacity to cope with the net variation–that is, to cope with the difference between load and solar generation that must be supplied by the rest of the system.

Storage is sometimes put forward as a solution to the duck curve. Moreover, additional power electronics associated with battery storage can help both with the fluctuations in solar production and with the regulation of voltage levels in the distribution system.  Storage can be a cost-effective solution when there are issues such as distribution feeder capacity limits that can be alleviated through storage. So while there are specific situations when investment in battery storage can make sense, it is generally still a relatively expensive solution. There is no doubt in my mind that battery storage will eventually be cost-effective when costs decline significantly from today’s levels. In the meantime, until battery prices become much lower, existing pumped-storage and reservoir hydro facilities are important storage resources to be utilized now. Moreover, dispatchability of solar would mitigate some of the immediate need for storage. In an analogous situation in Texas, dispatchability of large-scale wind has been helpful in ERCOT being able to integrate so much wind power.  

This brings the discussion back to small- versus large-scale. Large-scale installations are likely to be more cost-effective ways to provide large amounts of solar. They also are easier to set up, so that they can be dispatched down when required. However, having gone down the path to installing a lot of expensive, non-dispatchable small-scale rooftop solar installations, jurisdictions such as California, Australia, and Germany have consigned themselves to spending even more money to buy battery storage to balance the solar. It is hard to reconcile this policy with the imperative to decarbonize the electricity system cost-effectively. Other states and countries should take note: until storage becomes cheaper, larger-scale solar installations will reliably bring more cost-effective decarbonization results.