Tuesday, 24 January 2017

Trends in the generation

With five years of data from my solar panels, it should be possible to find the answers to some interesting questions. Like are these panels producing any less electricity as they get older?

I'd probably need to compare the generation each month with the actual weather data to get a proper answer, but as a first approximation, I tried to calculate a normalised generation figure for each month. I did this by first finding the average generation for that month, dividing that by the average of the generation for the same month each year, then multiplying by the average generation for every month. 

While I was about it, and since it's really easy to copy spreadsheet functions into different columns, I did the same thing for the consumption.  

And here is what it looks like:

Google sheets does its magic and produces a trend line, assuming that each of the points is dancing around a linear trend. And if we can trust this, there is a downward trend. Generation has gone down a little under 5% in five years. At the risk of digressing into a discussion of logarithms, let's say that's about 1% per year. 

Interestingly, there is also a downward trend in the amount of electricity we use. This has gone down almost 10% in the last five years. There are spikes and troughs for the months of March and August when we have been away, so I should probably work out a way of ignoring these. Another factor that will slightly affect both graphs is the slight difference in each month, since I'm using the month of the electricity bill rather than the calendar month. 

The 1% per year loss of generation is in no way surprising. On the other hand, it is a surprise If we are on a trend to use 2% less electricity per year, and it is a pleasant surprise. I've always imagined that low energy houses are at their lowest energy levels when they are new, and as they age, lose efficiency, and acquire extra gadgets they steadily use more energy. That does not seem to be happening here yet. 

The other question I'm sure the data can answer is how big a battery I would need if I went off grid. 

As a bonus, here is a chart of the daily generation and consumption that is not at all normalised.