How to estimate the energy generated by a solar system, given its location on earth and the system design, as well as validate that any estimate is reasonable.
Different places on the planet get different amounts of solar energy per square meter each year depending on things like their distance from the equator, average cloud cover, atmospheric pollution, altitude etc.
Organisations such as NASA have measured solar energy reaching ground level for most positions on the planet, over many years. For example NASA have made that data available on their website, where you can enter the coordinates of any location, to access a wealth of useful data about how much solar energy normally reaches the ground, based on historic data, at that location. Because these data-sets represent real readings spanning long periods (eg 25 years), they can yield quite reliable results.
In addition, there are software packages available (such as PVSyst), that use this data to simulate the expected energy you could generate from a solar PV system design. The software also needs other information, such as details about the exact solar panels used and the installation design (size and orientation of the solar array etc).
Solar PV panels still have relatively poor efficiency ratings (17% is a normal operational rating at the moment), which means that a relatively small proportion of the incoming solar energy hitting the PV panel is converted into electricity. In addition, installation factors such as the orientation of the panel may also adversely affect operational efficiency. Operational temperatures also play a part (most panels operate best at lower temperatures).
Software simulation tools like PVSyst take all of that into account, to yield fairly reliable estimates of the actual useful energy you could expect from a particular solar installation, given the location on earth and other system design parameters.
However, like all software simulations, the results are only as good as the inputs and assumptions used. The danger is that overly optimistic parameters are used to get a favourable result. This unfortunately happens some times, because its easy to “tweak” the simulation to get a more “competitive” result, and most people will not be able to easily check this.
One way to validate the results is to use “Equivalent Hours per Day at Max Solar Power” (sometimes referred to as “Solar Hours”), to validate the estimates of energy generated by the solar system over any period.
The graph below shows this concept:
The blue line shows the normal shape of power generated by a solar system on a nice sunny day. You can see it has curved shape (sinusoidal), which starts to generate power in the morning a few hours after sunrise, rises to a peak at noon, and then drops back down to zero again a few hours before sunset.
The area under that blue line represents the total energy (kWh) that will be delivered by that solar system, on average, each day. The peak power delivered at noon should equate to the peak power rating of the system (under ideal conditions). If one multiplies the area under that graph by 365, that would give you the total energy expected from the system per year.
However, because its difficult to asses the area under a curved line, its easier to use a simplification which is depicted by the red rectangular profile. That rectangle has exactly the same area under the red line as the area under the blue line, but now its very easy to understand, its simply the number of hours (width of the rectangle base) of “effective sun”, at the max power rating of the system (referred to as “kilowatt peak” denoted by kWp). This will give you exactly the same estimate of expected energy generated per day from the system. One can thus reduce the entire simulation result down to one number which is independent of the size of the system. That is the number of “Equivalent Hours at Max Solar Power” per day. This number is dependent only on two things; the location of the system on earth, and the efficiency of the system design. It thus provides a good way to compare system designs for the same location. It also highlights suspect simulation results which may be too optimistic.
For example, in Johannesburg one can expect, from current technologies, about 4.1-4.8 hours per day of Equivalent Hours at Max System Power, on average over the year. This means that what ever your installation peak power rating is, you can simply multiply that by 4.1, to get a conservative estimate of what to expect in terms of average energy output (kWh) per day, over the year. Any simulation that comes up with numbers much higher than that, should be queried, because that would be higher than normal. That situation may well be possible, but it would require something special to make it likely.
In the event that you receive an estimate of annual energy generation expected from a solar system (in kWh), you can work out the Equivalent Hours at Max System Power as follows:
Equivalent Daily Hours at Max Power = Annual Energy Generated (kWh) / (365 x Max Power rating of system (kW))
Some locations like the Northern Cape, receive very high levels of incoming solar energy. In those cases one can expect about 6 hours of effective maximum power out of a solar system. But in locations that do not enjoy such plentiful solar energy, one cannot expect numbers as high as that. Certainly if a solar simulation is predicting that sort of result in Johannesburg, one should start asking questions.
When comparing quotes for systems, which will all be mounted on the same roof, using similar technology – then this number should be the same, or very close, because if the location is the same and the technology is the same, and the different solar arrays (regardless of their sizes) will all be mounted on the same roof, at the same angle and orientation etc, then there can be no real difference in the effective hours of generation at max power per day. Thus any significant differences need to be queried.
Ultimately, when making an investment in solar power, you are paying for the energy that the system is expected to produce over time, and thus understanding this issue becomes crucial. This one number is a very useful measure that enables you to make a sensible judgement.