Turning the corner

Thanks to other things happening in life, this entry is late - nothing new then, but its timely too. The coalition government in the UK has announced a review of the Feed-in Tariffs scheme and I learned about it when a friend sent me a link. The site he sent me to was none too informative, primarily because it wanted me to cough up cash to learn more than the headline. A Google search yielded more and brought a little relief. Articles here and here indicated that the core of the review would be to drop subsidies for systems rated at over 50 kW-peak. My little system is only 3.5 kWp so hopefully I will not be affected.

Numerical fun.

Since the panels were brought online, I've been hoping that their output over the year will beat the predicted value of 2,480 units. Throughout these blogs I've stated that this would represent an average generation of 6.85 units daily. Duh! Checking it now, I find that 6.79 units is the true daily average and I think I got 6.85 by having divided 2,500 by the number of days in the year. The following discussion is based on my erroneous belief that 6.85 was my predicted generation average.

As my spreadsheet containing my data has grown, I've kept a track of my 30-day average and, more importantly for my number fixation, a total average. By 18 May, I hope this latter figure will be over 6.85. A couple of weeks ago, I tried to predict this with an experiment on the spreadsheet.

On the assumption that the average amount of sunlight (a quantity known as insolation) between the summer and winter solstices would be the same as between the winter solstice and the subsequent summer solstice, I took my historic data from the autumnal side of 21 December and plotted it into the appropriate spring days leading up to the 18 May anniversary. Each data point would keep the same time from the solstice and, clouds willing, would represent the same insolation.

I'll admit it is a shaky assumption but it is fun. The results of this exercise were as follows: It predicted that my total average would descend to 6.85 on 11 February - today! So what actually happened? It is currently at 6.79, having passed the higher waypoint three days ago. So there is less power being generated this side of the solstice, presumably due to the very poor weather we've had recently. Other predictions that came from my exercise were that the average would bottom out at 6.55 on 18 March as lengthening days would turn it around; that it would rise back above 6.85 on 21 April, four weeks before the anniversary; and that it would show an average of 7.18 units average on the anniversary which would represent 2,621 units overall, a decent enough figure, though reality is already slightly behind this prediction.

So to graphs and figures:

A number of interesting things are apparent. First, the 30-day average (orange line) has made a definite upturn and that's great to see. Second, yesterday, we had the first really clear day in quite a while and a thumping 7.5 units were generated, the highest since 24 October. I had hoped that the spikes produced by occasional clear days would show an obvious curve nicely symmetrical about the winter solstice but the Scottish weather seems to have interfered too much and I cannot discern it. Finally, the decline in the total average (yellow line) is slowly being arrested.

Meanwhile, I bought myself a toy!

I was in B&Q (a DIY store) where they had wireless energy monitors available. I had thought about these before the panels were installed and had decided not to bother when I saw that it would be easy to gather readings from under the stair. Nevertheless, I had hankered after a way of being able to watch the rise and fall in generation as the Sun rose and fell, and as clouds came and went. My family were known to giggle as I would dive under the stairs to satisfy an instantaneous craving to know how much power was being generated. Now with this toy I can watch all day if I want.

The unit I went for is an Owl CM160+USB model that I got from an Amazon third party seller. These units are meant to tell householders how much power they are using either in terms of kWh (units) or in monetary terms. It comes in three major parts. A sensor goes around the live cable feeding power into the house. This connects to a unit that turns its output into a digital signal that is wirelessly transmitted to a display unit sat somewhere handy to view - in my case, on my desktop beside me here. Users can enter details of their tariff for accurate cost displays and it will also inform the user of how many grams of carbon their electricity usage is causing to enter the atmosphere. Neither of these latter two features were of much interest to me but one that did make me go for this model was a USB connection that would allow minute-by-minute readings to be displayed on the computer as a graph.

All I'm really interested in is the power display and I want to use it for the solar panels so instead of the incoming mains, I place the sensor around the live feed that comes from the inverter to the mains. Happily, the unit arrived on the beautiful sunny day we had yesterday and there was plenty of power to exercise its abilities. I left it set to assume our voltage was 230 volts. I think the setting of this will have an effect on its reading of power as I guess it is essentially reading the current in the cable and multiplying the value by the voltage to determine the power. Later, I set it to 240V as I think that is nearer the mark.

The unit works a treat, its displayed kW value going up and down as the occasional cloud gets in the way of the Sun. A reading of 1.8 kW from the inverter was about 1.75 kW on the monitor and I'm assuming the low reading can be blame on the voltage setting.

The one shortcoming I've found is that after the Sun has gone down and the inverter has switched off for the night, I continue to get a reading. Despite no power coming from the panels, it still shows 50 or 34 watts passing through. Is this an inherent offset in the unit's calibration? Does it instead represent a current draw by the system to power the inverter in its sleep mode and keep the meter going? First I'll try taking the sensor off and see what reading I get... (goes off to carry out experiment). When taken off the cable, the unit says zero. Good. Next, I'll place the unit next to the cables but not clamped around the live to see if a stray magnetic field is to blame... (goes off to carry out second experiment). Still zero.

I think this shows that the solar power systems themselves require about 40 watts to operate, even at night. Taken over a day, this represents (40 x 24 = 960 watts) a full kWh or one unit of electricity! Wow! If half of that time, the power comes from light on the panels, then I have to 'buy' 175 units annually to operate the panels. That's about £17. What it needs is a light-operated switch that completely disconnects the panels and inverter from the mains when it's dark outside.

I like this unit.