Saturday, 28 September 2019

The surprising cost of a pilot light (waakvlam)

We have a low energy bill. This is the case because we've done quite a lot of work in our home to improve the insulation so that our central heating rarely comes on, and we've tackled our electricity consumption by installing solar panels. However, we've not yet done anything to the gas equipment in our home which was already here when we moved in 12 years ago, in part because until now it's not been easy to tell which piece of equipment used most gas so should be targeted first.

While we've had a smart electricity meter for almost a year now, and I've used a plug in measuring device for much longer to check which appliances had higher than expected consumption, our energy company didn't install a smart gas meter until a week ago. The old meter was not easy to read for small levels of usage. But the new meter has made it easy to find out something which I had long wondered about: How much of our not very high gas consumption was wasted to no effect.
The new gas meter. Since installation we've burnt 5.725 cubic metres of gas.
The gas water heater
How much gas does a pilot light (waakvlam) use ?
Our house has three devices which run on gas: The gas hob in the kitchen, the central heating boiler and a separate water heater which heats water only for the shower, bath and bathroom sink.

The water heater is really old. Old enough to use a pilot light (waakvlam) instead of starting itself with an electronic igniter whenever hot water is required.

If you're unfamiliar with what that means, there is a very small flame which burns continuously, 24 hours a day, 365 days a year, just waiting for someone to turn on the hot tap so that it can be used to ignite a much larger flame to heat water.

In the past I've asked several people who work for the gas company, or otherwise seem to know about gas appliances how much gas is used by such a flame and I've always been re-assured that it's "next to nothing", "unmeasurable" or "similar to a mobile phone charger", but I was never quite convinced. The new gas meter has allowed me to measure how much gas is being consumed and the result is surprising.

The pilot light. It's small, but any gas burnt here is wasted.
Meten is Weten. It costs how much ?
One day this week we took readings from the gas meter while avoiding using any gas appliance for 18 and a half hours so that period passed with only the pilot light burning gas. Over 18.5 hours, the meter showed that 0.283 cubic metres had been consumed. That equates to 0.366 cubic metres per day or 134 cubic metres per year.

134 cubic metres of gas isn't insignificant at all ! In fact, it turns out that in summer months our gas usage is dominated by the consumption of the pilot light, which consumes more than we use for hot water and cooking combined. Over the whole year it consumes rather more gas than we use in February to heat our home. It's an appalling waste not only of gas but also of money: That pilot light costs us nearly €90 a year to run.

Like a phone charger ?
The comparison made previously with a mobile phone charger is particularly absurd as phone chargers genuinely do consume an unmeasurably small amount of electricity when they're not in use (unplugging them is something that some people do in an obsessional way because it looks like it'll save energy, when actually the effect is almost nothing at all). But this pilot light consumes a very measurable amount of gas. 134 cubic metres of gas is equivalent to about 1340 kWh of electricity. If a phone charger used that much it would certainly be measurable. It would also add somewhat more than €100 a year to the electricity bill and the charger would be rather hot rather than cold to the touch.

The next step
Obviously this old water heater has to go. That has long been the plan because actually we'd like to get rid of gas altogether. It's not happened yet because we prioritized insulation and electricity first. But discovering how much this thing wastes has given new urgency to the plan. At the very least we need to be rid of this water heater. It appears to be possible to buy an instant electric heater for about the annual cost of the gas for this, and an electric heater would effectively cost nothing to use because it would operate on the excess electricity from our solar panels which we currently export to the grid and for which the electricity company pays us very little. So I expect to change this quite soon.

Monday, 8 April 2019

Seven years of solar power: How valuable is our solar energy ?

I've been interested in solar energy since before I can remember. My first practical experience beyond things like solar powered calculators was in the mid 1980s when I put a square meter of surplus panels on the roof of my parent's home. These provided 12 V DC which I used, amongst other things, to charge the battery in my then quite new laptop computer. It wasn't until seven years ago, though, that we installed a large system on our home and the seventh anniversary of that system passed on April 5th.

Seven year summary
We have 16 panels each rated for 235 W output for a total of a 3760 W peak. In this part of the world it is usual to calculate the expected total output as the equivalent of 925 hours of full sun each year on the panels, giving an predicted output of 3478 kWh per year. In practice, over the last seven years our system produced 23601 kWh in total or an average of 3372 kWh per year.

Output per month over seven years. Red bars show our electricity consumption, blue bars show the production from our panels and the yellow line shows where we are in total now relative to where we began. Our production in is higher than our consumption on average, though obviously that is not the case in the winter. The kink in the yellow line from July 2018 is the period during which our inverter was not working (I repaired it myself)
Several things contribute to the slightly lower output relative to the estimate: Our panels are mounted at the angle of the roof and not at whatever the ideal angle might be, they face South West rather than directly towards the South. But the most damaging thing so far as the average is concerned is that our inverter failed last year and so we had no output at all for a few weeks during the sunnier than average month of July. Calculating out average output for the six years before the inverter problem we come to 3410 kWh per year which is within 2% of the prediction.

Return on investment
The monetary value of the electricity which we've generated is about 20 cents per kWh for that which we use ourselves and about 7 kWh for the excess that we export to the grid. We have exported about 1200 kWh in total for a value of €84 and consumed about 22400 kWh ourselves for which we would otherwise have had to pay about €4500. In total, then, about €4600 has been returned from our original €8000 investment. If the same rate of return continues then the system will have taken 12 years to pay for itself.

Our new smartmeter. When the photo was taken
everything electrical in our home was powered
from the panels and 2 kW was being exported
to the grid.
If the inverter fails again and this time we have to replace it then that will of course increase the repayment time. This might happen. However the panels themselves are expected to last much longer than the repayment period.

A smart meter
Something else which changed last year was that our electricity meter was replaced in December.

The new meter doesn't give us any real advantage over the old mechanical meter which span backwards just as enthusiastically as forwards for nearly seven years. The electricity costs us the same amount either way around.

Of course it does offer an advantage to the electricity company because they no longer have to ask us to read the meter or visit to do so themselves.

What the new meter does give me, though, is that it counts differently and therefore I have a little more information for future calculations. While we pay the same amount of electricity in peak and off peak periods, the meter displays them separately, for both inbound and outbound electricity. So now we can see how much electricity we consume at night time and we can see how much leaves our home rather than being consumed here.

Solar electricity is more valuable than average because it is
generated at peak times of consumption (source)
Perhaps after a year of recording this information I'll have something which which I can create an interesting graph. Thus far I have just three and a half months of this extra data, all of it from winter and spring. There's not much contrast to see.

What has long been obvious of course, and is also visible even in the small amount of data which I have now, is that our electricity is generated only in the daytime and then mostly on sunnier days. As it happens, solar electricity remains quite valuable in the Netherlands because this daytime generation corresponds reasonably well with peak usage. That means that the electricity which we export, most of what we generate, is almost certainly of use to someone else.

The first thing I made the prototype
hardware do was wiggle a GPIO and
then send serial data as shown here.
Software for a smart meter
I'm late to have a smart meter installed but as it happens, I spent a few months two years ago working on software for the prototype smart meter hardware of one of several competing manufacturers. I'm almost certain that the meter that I have now is not related to the manufacturer for whom I did the work, so my code is almost certainly not involved in reporting my own electricity usage, but as I'll never open the box of the meter and look inside, I'll never know for certain.

It was quite an interesting project for a while because it was like a return to the old days for me, when I worked on 8 bit processors and embedded software. This work was with what for me was a new processor, the Renesas RL78. It's a quirky 8/16 bit design. The RL78 assembler is styled so that the source code looks a bit like that for the Z80, but the processor is really completely different. In total I had 32 kB of flash and 4 kB of RAM to work with. These days that doesn't sound like much at all but that means this device has about the same amount of memory as the guidance computers which took Apollo to the moon so it's enough to do great things. In this case it's in a finger nail sized package which consumes micro-watts, and the little RL78 can of course compute many times faster than the AGC.

Power for a home and a business
We generate more than enough electricity to power both our home and our business, though of course I always point out that we're not actually doing so at night, or on darker days in the winter. Anyway, if you want a support a solar powered bicycle business which uses no powered vehicles then you can do so here: