Summary

York, 2012-04-21

Not enough to bother yet for us using a simple controller. The more complex controllers are currently too expensive but hopefully new devices will come along soon.

Introduction

Since getting our solar panels six months ago we've been wondering if there's any benefit in using the excess solar power to heat our hot water. The main reason to do this is financial. We get paid as though we are exporting some electricity to the grid even if we're using it all locally. It would therefore appear to make financial sense to use the spare power to heat our hot water instead of using expensive oil. As a nice side-effect we'd also be helping to save the planet.

Lots of folks have already had this idea and are working on various devices such as the Parsons Switch and the Immersion Heater Relay. These devices tend to have a fixed heating element (or two) and turn it on when the solar panels produce above a certain threshold. Sadly, they tend not to monitor the house consumption so if you turn on the kettle whilst they are topping up the hot water tank then you end up paying full price for the electricity you pull from the grid. I started to think about how my monitoring could help me. Notably I have a solar generation and house consumption reading every six seconds. I can find the difference and ramp up an immersion element to that figure. I wasn't sure how to do the electronics, but it appears the bright sparks on the Electrician's Forum are already working hard at it. These "proportional control" devices aren't ready for the mass market yet, so I thought about how I could use a fixed size element this year until the devices were tested and ready to go.

Analysis

I've used my historical data to run some example configurations and compare the output. There are a few parameters I used:

Comparing Heater Elements with Simple On/Off Control

If we have a simple device which turns on the immersion heater when there is spare solar power then we avoid using power from the grid. The only significant variable is what size heater element to use. Here are the resulting figures for the 127 days from 2011-12-01 to 2012-04-05 (data):

cumulative

I observe the following from this chart:

Comparing Types of Switches

Here is some analysis of the financial benefit derived from devices which turn on as soon as the solar output passes a trigger point, irrespective of the actual load in the house. This is for 2011-12-21 to 2012-03-21.

Element (W)Trigger Point (W)Captured (kWh)Drawn-down (kWh)Benefit (GBP)
1001000103.50.43
10015006.92.30.32
100200040.80.25
100250020.30.14
10030000.40.10.03
250150017.160.77
25020009.82.20.59
25025005.60.90.33
25030001.20.06
500150033.912.11.46
500200019.54.61.13
5002500111.90.62
50030001.90.40.10
1000200038.19.51.99
1000250021.84.11.16
100030003.610.16
1500250031.26.41.42
150030005.31.60.22
2000300072.20.20
1000Available55.206.06
3000Proportional92.509.26

I've included a 1,000W element that only turns on when power is available and one which uses "proportional control" of a 3,000W element for comparison.

This is assuming a 800W base-load, so any trigger less than 800W above element power was ignored. The benefit is determined by displacing oil at GBP/L [1] and kWh/L [2] in a boiler system that is % efficient = 0.09 GBP/kWh (to heat water). Electricity drawn down from the grid costs GBP/kWh. This calculation assumes you are displacing grid-based electricity as it's cheaper than oil - it doesn't apply a penalty for drawing down grid electricity for heating.

As you might expect when you think about how these switches operate, there were times when the house consumption occurred at the same time as immersion operation. Even with the best performing configuration, on around 10% of days the overlap happened frequently enough to cause a net loss of money by having the switch turn on the immersion when there wasn't enough power to run it entirely from solar.

Economics Conclusion

The amount of oil we're saving is tiny, just 10L at best with the modelled configuration - that's just 6GBP. Things may improve in the Summer, maybe we'd be able to avoid burning any oil at all, however, the economics aren't stacking up so far. Throwing all the improvements below together would have boosted usage from 91kWh to 205kWh - still only about 12GBP worth of oil or about 2 days consumption in the Winter.

Ultimately, with just 3kWh capacity in the hot water tank we can only capture 1.1MWh a year even if the sun shone every day. That's 110L oil which is 72GBP - this makes it very hard to justify most devices which are a few hundred pounds plus fitting.

At a maximum of 0.20GBP per day oil saving we are looking at 110 seconds at minimum wage (in the UK) so it just about pays for my girlfriend to switch the immersion on and off manually when it's clear we can reach this zenith.

Future Considerations

Change User Behaviour

It would be possible to have the immersion avoid overlapping the house consumption by using timers (as suggested here) or simply observing the load and being cautious about when to create load with other devices. We have certainly tried it here and it wasn't very popular. I don't fancy becoming another divorce statistic by suggesting my better half become a slave to a 50GBP a year cost saving.

Multiple Fixed Elements

Would it be better to have two elements, say 1kW and 2kW which ideally we'd be able to run together on occasion? I did ask Parsons if they'd sell me one of theirs but they were understandably reluctant to. Maybe Howden will help us out?

Change the Immersion Setup

We might be better off with a different immersion arrangement so we can use more power. If we could even double the capacity then we'd hope to capture double the amount of power. However, for the Winter and early Spring that doesn't seem to help, the captured power only improves by 10%

Evaluate Other Seasons

We need to extrapolate into the Summer and Autumn. I have some ideas for this, but haven't got round to it yet.

We can provide an upper bound figure as follows:

  1. Collect up the results from 21st December to 21st March (using my figures)
  2. Assume that it's the reverse of 21st September to 21st December, so double the value
  3. Assume that every day in Spring and Summer gives the maximum benefit (183 days x 0.20GBP = 36.60GBP )
Element (W)Trigger Point (W)2 x Winter Benefit (GBP)Maximum Annual Benefit (GBP)
100020005.0841.68
1000Available10.3846.98
3000Proportional17.4054.00

Wait for Proportional Control Devices at Sensible Cost

A device with perfect "proportional control" that could use every bit of spare power would have captured 50% more than the best fixed-size element being turned on when sufficient power is available. To benefit from this the purchase cost needs to be reasonable.

Turn on Proportional Control Earlier

It's actually worth turning on the immersion when it can be powered something like 50% by solar as it's displacing expensive oil - it's still cost-effective when drawing down from the grid. The snag is that it may be cheaper to wait until later in the day for better solar. I'm producing power predictions based on the Met Office forecast, but it's not terribly accurate. This might boost power usage by 30% but raises questions about the environmental difference between grid-electric and local oil.

Device Cost

The cost of the devices is currently high. This is made worse by having to pay an electrician to fit it which is further complicated by each site being different. For example, often the immersion, house consumer unit and solar feed aren't electrically separable for monitoring or close enough for a device to touch all three. This may be improved by being able to use remotely transmitting sensors, much as the CurrentCost Senseables do. At the moment these are much too expensive but I suspect there are devices that can do this for around 5GBP as we got something very compact that included the transmitter and clamp in one device with our British Gas EnergySmart plan.

One way to reduce the cost of the electrician attending would be to fit the device whilst the solar panels are being installed. This would give the device sellers a useful sales channel and also help the installation company add a bit more income from an install. I can imagine it's not trivial though as it's another thing to go wrong and could distract from the major work of connecting the panels.

Another thing to bear in mind is that oil and grid power are getting more expensive so this makes it easier to justify the cost of the device as time goes by (assuming production prices fall as is common.)

Total Available Power Is Small

The total available excess power was 224kWh which would have saved us 22GBP if we'd managed to use it to power devices. As it happens, powering devices ultimately provides distributed heating so it's good to use as much as possible when already heating the space. The problem here is that devices are very dumb and will use power when it's expensive. The classic example is our oven which uses more than the spare then none, then more than the spare, then none as the thermostat kicks in - it would be much better to have a smoother load (when it's sunny at least). We need smarter devices, but if we're struggling to extract value with something so easy to control and profitable to use as an immersion heater then other devices are going to be even harder to justify.

Would a Buffer Help?

A power buffer such as the SMA one would help match power consumption to generation by smoothing out the peaks and troughs, but it's pretty expensive and has a limited life.

Other Power Dumps

Perhaps we should get a swimming pool to provide a greater capacity. Strangely, my girlfriend doesn't think the economics of that will stack up either.

Ethics

If you'd like to reading an ethical ramble on the subject there's a thread on Navitron forums about that.

Let's Chat

Feel free to contact me at david@myforest.com if you have any questions or suggestions.