First complete month

... practically.

As regular readers will be aware, our solar panels were out of action for 5 days this month. (See "Hiccups").

But apart from that, I now have a complete month's worth of readings.

For the sake of calculating an average, I've accounted for the 5 days of missing readings by allocating the predicted average value for each missing day (4.97kWh).

I've spreadsheeted the results, which are as follows:

The purple line at the right hand side is the readings from the last few days of August, and the red line is the readings for the whole of September.

The predicted daily average is 4.97kWh per day.  I've not added this line to the graph, but you can see where it would lie from the flat section between 19th-23rd, where the panels were out of action.

I've put a dotted linear trend line in to show the underlying change in the actual kWh generated for September. Oddly, it climbs towards the end of the month. I was expecting it to go down as the amount of daylight diminishes towards winter. I suspect this is because the weekend at the end of the month was unusually sunny, thus skewing the trend line upwards.

I have put together another graph, which shows the predicted generation for the whole year, broken down into months.  I have divided up the predicted annual generation figure into 12 months based on the average number of hours of sunlight per month from the met office figures at

http://www.metoffice.gov.uk/climate/uk/averages/19812010/sites/blackpool.html

The predicted total for September was 149kWh, and the actual generated (including assumed average for 5 days as discussed above) was 157kWh.

Other stats I've pulled out of the readings: 

• The actual readings are 105.4% higher than the predicted
• the array is generating 0.37kWh per each hour of daylight (sun up to sun down, irrespective of clouds)

It is satisfying to see that the actuals are slightly higher than the predicted.  It will be interesting to see how October pans out.