by Energize staffwriter –
As we enter the summer rainfall season, we can expect some decrease in the amount of electricity generated by rooftop PV systems.
This year’s summer rainfall season might be higher than average. This has been attributed to the La Nina phenomenon, a weather pattern which results in the abnormal cooling of sea surface temperatures in the Pacific, which enhances the probabilities for summer rains in Southern Africa.
In addition to the normal Highveld afternoon storms, we have periods of overcast weather, sometimes lasting for several days. Overcast weather can extend over the whole country or significant parts. Figure 1 shows an example of cloud cover recorded by the SAT24 system.
Figure 1: Satellite view of cloud cover over South Africa.
It is fairly obvious that PV is affected by overcast weather, but the extent is not always realised. A South African study  found that generation can be reduced by more than 70% under overcast conditions, with 75% being considered accurate, with a high figure of more than 90%.
Overcast weather is normally also accompanied by low winds, and this means low production from both solar and wind power plants. Long periods of overcast weather are not uncommon in South Africa, and need to be taken into account when planning renewable energy installations. The record for consecutive periods of overcast weather in South Africa is nine days .
In December 2019, a period of seven days of consecutive overcast weather was recorded in Gauteng. It may be pure coincidence that this was accompanied by a period of extensive load shedding.
The phenomenon of extended periods of bad weather is not unique to South Africa, but occurs in other countries as well. Germany even has a name for it, the “Dunkelflaute”. during this period Germany has to import power from other countries, a luxury which this country does not have.
Utility solar and wind are generally not affected severely by inclement weather as they are planned in areas generally not subject to bad weather and the contribution of utility PV is not significantly affected by countrywide weather conditions.
The same does not apply to rooftop solar systems which are co-located with the consumer and are mostly installed urban centres and industrial areas, which are subject to both wide area and long duration overcast weather conditions.
Figure 2: The German “Dunkelflaute” phenomenon.
Current rooftop or own generation is estimated to be 700 MW , and IRP 2019 makes the assumption of 2600 MW of embedded generation by 2030, which would be mainly solar PV. The recent lifting of the 1 MW limit on own generation is likely to encourage larger own generation plant in the future, and the actual figure could be much higher than assumed.
The impact of reduced generation from rooftop PV on the system can be significant. The consumer load does not go away and, the power is drawn from the grid, increasing the total demand during overcast days by an amount equal to the rooftop PV decrease. 2600 MW of rooftop solar could lose as much as 2000 MW of generation capacity under heavy and widespread overcast conditions. This would place an additional load of 2000 MW on the grid during daylight hours. So in addition to any decrease from utility-scale PV there will be an increase in demand from customers with rooftop PV installations.
Consider a major shopping mall that has a rooftop PV capacity of 4,75 MW. On an overcast day the max production could drop to as low as 1,2 MW, placing an additional load on the grid of 3,55 MW during peak production. The mall is not alone in having rooftop PV. Medical centres belonging to a health care group are all equipped with rooftop PV, as are many other large and small businesses in Gauteng.
It could be argued that the grid has the capacity to cater for overcast periods as solar power ceases at night anyway, but the majority of rooftop PV loads occur during the day and will drop at night as well as the PV production. That also ignores the fact that the grid in future will rely heavily on the increase in wind production during the evening, that is not available during the day. During day time PV loss will introduce an additional peak load on the grid that will have to be met somehow.
Energy storage may supply an answer to the problems of short periods of overcast weather, but the cost of a system that could ride through seven days of bad weather would be exorbitant for commercial and industrial users and out of the question for residential properties. Storage certainly can aid in smoothing out short term variations in output, but, with current technology, cannot provide an economic or affordable solution to long periods of overcast weather.
Consider the situation with five days of overcast weather. If we allow the daily contribution of solar to be 100%, then over five days 500% is required. The production over five days at 20% capacity is only 100%, leaving the storage to be capable of holding 400% of the daily generation. In the case of the shopping mall, the average daily generation is estimated to be 21 MWh. To carry the mall through five days of overcast weather, the battery would have to store approximately 80 MWh of energy, at a cost that probably far exceeds that of the solar PV. This does not take into account the extra equipment required to recharge or replace the stored energy.
Future plans for the grid need to take the effect of overcast conditions on rooftop solar into account, as these are likely to form a significant portion of the generation portfolio, and are relocated in areas with a much higher probability of consecutive overcast days than areas where utility systems are located.
 M Suri, et al: “Cloud cover impact on photovoltaic power production in South Africa.”SASEC 2014.
 AREP report, 2019.
 CM Cewen: “Spatial processes and politics of renewable energy transition: Land, zones and frictions in South Africa”, Political Geography 56 (2017).