PV Generation Output Levels Over Time

I will attempt to answer the question of "how much power is the system generating right now"? I'll expand on this to summarize how often the 4.9 kW system is generating at different levels. This is very interesting once you see the results. For example, during a day with sun and cloud the output may be less than 1 kW much of the time, but when the sun finally appears for an hour the output may be 4 kW. So, how much money do I make during cloudy periods and how critical are bright sunny days? Let's look!
The Questions :

• How often is the PV system generating at different levels [in watts]
• How much money is earned at each output level
• How often are the panels generating near maximum output or above

The Data :

• Fronius Datalogger Pro installed on PV system noon 2010-10-31
• Wattage output averaged and reported by data logger every 15 minutes
• Continuous monitoring when inverter putting power to the grid [greater than 15 W]
• 2011-06-07 17:15 is last measurement event
• Data points n = 9658

The Method :

• 15 minute interval data downloaded from datalogger and imported into spreadsheet
• Watts used to calculate kW, revenue at $0.802/kWhr, and watts generated per wattage of panels [20 SolarWorld SunModule poly panels at 245 kW each]
• Frequency distribution curves generated by pivot tables function

The Results :

Raw data

Figure 1 shows the raw data used for the analysis. You can see the output levels for each 15 minute period of the day when the system is operating [it shuts down automatically at dust when the panels generate less than 15 watts]. During November the maximum output is around 4000 watts. This drops to 100-200 watts during December other than a few days when it rises above 500 watts when there is no snow on the panels. A similar pattern is seen at the end of January 2011. From February to June there are many days when generation is over 4000 watts, even over the 4900 watts which is the maximum rated output of the panels. The maximum level seems to decrease back down to 4000 watts after March. This could be partially explained by the warmer temperatures during the Spring daytime compared to the colder February and March air temperatures. Power output decreases significantly when the panels are over 30 degrees Celsius.  Of course the panels only output their maximum daily limit for a short period. The morning and afternoon output is much lower and this is a larger duration of the day compared to the peak generation time period. Cloudy days also decrease the output level.

Figure 1.

Power Generation Levels

The shape of the power generation frequency distribution curve should come as no surprise when you consider the many points in Figure 1 that are less than 1000 watts. Sixty two per cent of the time the PV system is generating less than 1000 watts. Snow shading, mornings and afternoons, and clouds do not yield a lot of solar energy. Considering our house consumes about 300 to 400 watts continuously [1200 when the pool pump is running], the data show our house is a net power exporter much of the daytime over the year. The similar frequency of output from 1000 watts to 4500 watts is interesting. It suggests to me that once the sun rises there is quite a lot of time when the clouds are partially blocking the sun. Also, if we assume that the system can generate at least 3000 watts once the sun is up [i.e. 9 am to 6 pm] then we can see that Guelph has had a fairly cloudy period. There are 1327 15 minute intervals when the output is over 3000 watts, or 331 hours of "bright sunshine" out of 2415 hours of generation time! Let the Beatles some "Here Comes the Sun" come true this summer! Also:

• 45 minutes was 5000 watts to 5130 watts over this period - end of March. 5130 watts is 4.7% over the panel rated output, and that does not include the derating factors such as DC power line losses and inverter efficiency losses

Figure 2.

Revenue by Output Level

As frequent as <1000 watts may be, is it important for overall revenue, or what about the sunny periods. This really shocked me initially and now I know why I am so happy when it is sunny outside! Before I blab on any further, here is Figure 3. And this figure shows some very important information that relates to solar PV designs. I'll no doubt discuss this figure in future postings.

Figure 3.

Revenue resulting from 3000 watts and more represents the highest revenue levels! And our 4.9 kW rated system produced over its rated limit for 2 3/4 hours and earned $11.25 over this time.
Additional observations :

• total revenue to date generated is $2,229
• most revenue is earned at <3000 watts, but not by much
• about $100 was earned when revenue was greater than 4500 watts. If we use a common solar industry practice to oversize the panels relative to the inverter by approximately 10%, we would have lost approximately $100 in potential revenue. This undersized inverter is rationalized that a larger inverter may be considerably more expensive. In fact this may not be the best economically decision based on the data presented in Figure 3. [Of course, sizing the inverter and panels is more complicated than that]
• my initial solar configuration was going to be 4.9 kW of panels and a 4 kW inverter. Lost revenue would have been close to $400 [assuming that inverter efficiency across the power level range is equal for the inverter sizes]
• the generation frequency distribution in Figure 2 is opposite to the frequency distribution curve of Figure 3
• the frequent low sunlight levels is financially less important than sunny periods

Initial Conclusions

• your solar system should be designed to maximize the higher sunlight levels. Shade, ventilation, panel orientation, inverter sizing are all very important factors that must be explicitly accounted for when you have your PV system installed
• sizing your panels and your inverter is critical. Use the inverter manufacturers' software to properly design the number of panels, types of panels, the number of strings
• ensure that the inverter is sufficiently large to handle the high end of your panel output levels. Under-sizing the inverter by over 10% should be seriously rationalized unless there are upper limits to amount you are permitted to generate
• for example, the Ontario Power Authority's micro Feed-In Tariff Program has a 10 kW upper limit. You could maximize the amount of low sunlight level generation by increasing the number of panels, thereby shifting the mode of the Figure 3 revenue chart to the right. But this can only be done after using the inverter manufacturers' system configuration software. Failing to do this may result in lost power generation or damage to the inverter
• the 5100 watt inverter on our system is able to take advantage of the few periods of very intense sunshine yet this has earned me significant revenue. Over the period of 10-20 years of our microFIT program this will add up to a sizable income cohort
• choosing the location and orientation of the panels should try to maximize the bright sunlight periods of the day

Undoubtedly there are many more things I'll see in these data. My next posting will evaluate the power generation frequencies relative to panel size. This will partially answer a question I faced when I was choosing what type and brand of inverter to install [string inverter vs. micro-inverter - hint : I think i made the right decision].