If you’ve ever looked at my System Specs page, you’ll see that I’ve got a fair bit of west-facing roof I don’t use. I decided to get an upgrade.
First port of call was Natural Solar, of course.
The two existing arrays tie back to the inverter with one string each. The new array would be joined onto the array to the left of picture (western). This made for the simplest install as it didn’t require new wiring to the inverter.
It also provided the benefit of generating from the sun in the afternoon. This is useful in the warmer months where I am more likely to use air conditioning.
Additionally, for those days that have overcast mornings, but sunny afternoons, I’d see the most benefit. Particularly true given the house has a rising ridge line to the east, so doesn’t see much sun early in the day.
With the detail sorted out, we agreed on a date for Splice Electrical to perform the upgrade. James and Nick turned up, and with their usual friendly professionalism, got to it.
A few hours later I was the proud owner of another 1.5kW of panels!
— UnleashThePowerwall (@AuPowerwall) October 5, 2016
upgrade complete. Now what?
Naturally, with an upgrade of this sort, you’re going to expect some improved results. The system has had a size increase of 30% (1.5 / 5.0 = 0.3). Would I get similar generation increases?
The main factor in all this is still the inverter. The SE5000 in my system is limited to 5kW in any direction. Therefore, expected generation, even with 6.5kW of panels, is limited to 5kW maximum.
I was quietly confident I’d hit this high mark regularly, given I generated 4.9kW or more at points during February and March, and even in May!
However, the panels aren’t in the same orientation as the rest of the system, so what is the effect?
Setting the baseline
According to one source, a solar PV system in Sydney should produce 3.9kWh per kW of installed panels per day. That is under lab conditions.
My initial system setup should (on average) have produced about 19.5kWh of electricity per day. For the lifetime of the system at 5kW, the SolarEdge API reports the following figure:
I hasten to point out that the data is a little murky. The SolarEdge API consolidates “generation” from PV and battery, because of the way it monitors flow.
Therefore, that figure only works if we’re assuming the battery is filled and drained every day. This certainly isn’t the case 100% of the time, but its enough to show we’re in the ballpark.
Another part of this is loss due to inefficiency. The Powerwall is about 92% efficient, meaning I have to spend around 1.5kWh of the 19.5kWh figure on the power going into and out of the battery.
The other factor is the timeline; we’re looking at a period from mid-February to early October. This includes the shorter daylight hours.
With all these factors considered, I’d say this is actually looking like a decent marker, even with the accepted error margin (+/- 10%) in the SolarEdge API.
Effect Of Upgrade
An extra 1.5kW of panels should result in a generation figure of 25.35kWh per day, on average, for Sydney.
The first caveat for this figure: it has been spankingly good weather in Sydney for the last week.
The new panels are also a different orientation to the others, which may affect the figure.
It is also a very small sample in terms of days. In the interests of science, I’ll leave this chart here to update daily. You can check in on it any time you like. A rolling 7- and 28-day chart is also on the Statistics page.
Average just shy of 33kWh at time of print. Pretty good weather!
A better analysis might be to look at the curves being produced by the different panel setups. The figures aren’t as important as the shape of the curve.
— UnleashThePowerwall (@AuPowerwall) May 13, 2016
That was from May, without daylight savings. The curve peaks around midday and is generally uniform. Pretty much what you’d expect. Decent result, too.
Now we throw another 1.5kW of panels on the western side, and add daylight saving to shift the curve to the right by an hour.
— UnleashThePowerwall (@AuPowerwall) October 11, 2016
A small point: that is my record day so far. And its only October. Tee hee!
More importantly, we see the extension of the curve from the new peak around 1PM (daylight savings, remember), through until the late afternoon. We get a lot closer to sunset for generation as well.
The peak 5kW on this particular day hit at 1128 hours, and it stayed there until 1514 hours. The 5kW system infrequently reached 4.9kW, and then only for short stretches.
With the extra panels, not only is the generation figure much higher, but the long afternoon sun really kicks in.
As we move further into the warmer months, I expect the 5kW peak to be longer. Most likely, this will result in much more export until the point the ducted air conditioner is required on a regular basis.
I need to develop a strategy to mitigate that. Perhaps running the air conditioner in “continuous” mode on sunny days will help. This aims to keep the house cool, and the thermostat will lower the overall energy requirement.
I’m still finalising the change in payback time on the new array. Due to the move to TOU pricing, it is getting hard to keep all the facts and figures in order.
For the most part, I’m going to keep rolling calculations based on the single-rate plan offered by Diamond Energy. The rest of it makes my head hurt.