OpenEnergyMonitor Community

Micro wind turbine monitoring

Has anyone used an Emonpi to monitor output from a wind turbine?
I currently have a system set up to monitor household consumption (CT1) and output from my solar PV (optical pulse sensor on generation meter), along with derived export/import. I am planning to add a 3.5kW wind turbine which will be off-grid and just used to supplement heating.
The turbine won’t be metered - is there any reason why I can’t merely attach a CT sensor to the output cable from the inverter and log this as another feed in Emoncms?
In fact can I suggest an additional app could be written that adds self-generated wind power to the existing My Solar app?
Most off grid systems - my system will only be partially off grid - combine solar panels with a wind turbine and batteries so there is a good market for monitoring these hybrid systems.

The biggest problem that I see is the sensitivity of the emonPi inputs - 100 A full scale. While you had your Pi open, did you notice a pair of holes in the pcb behind the CT sockets? Those are there to allow you to fit a different (in this case higher value) burden resistor in place of the SMT one, which will give you a more sensitive input. A good choice would be 120 Ω, the same as the burden resistor on the high sensitivity input on the emonTx, which would give you about 18 A full scale. (I’m assuming 240 V here.)

There will be no harm done if you don’t modify the input, but you won’t have the best accuracy at low outputs.

With an off-grid system, the generated waveform won’t be in synch with the AC voltage source which is supplying your existing EmonPi system. So standard software, which calculates V * I to get power, won’t give correct results for this channel.

Thanks for your very rapid response Robert!
You’ve identified something I never would have thought of which now seems
obvious, and it seems a very elegant solution. The output will certainly be
lower than my PV system which can approach 20kW or more than 80A - I think
it’ll be around 220V (the inverter is Polish) which equates to about 16A.
In fact I won’t be needing the two inputs to run simultaneously as I am
going to “export” my Emonpi to another site to monitor a 50kW PV array with
a GSM dongle for communication.
In the meantime I am “upgrading” to a Solar Log for my home system, which I
picked up for a bargain price (£200 plus postage from Latvia) on eBay.
When I get the turbine up and running, is there an Open Energy Monitor
setup which would make more sense to use just for monitoring the inverter
output, or would it be easier for me just to get another Emonpi, bearing in
mind the burden resistor?

Thanks Robin, I had already figured as much. I would probably still be content with the fact that the readings would come with a large health warning attached to their accuracy! Unless I could plug the voltage sensor into a socket wired off the inverter so that at least when there was some generation it would be related to the right voltage? This is assuming the Emonpi wouod only be for monitoring the wind turbine as per my previous post.
Or am I barking up the wrong tree?

I could suggest an emonTx for the turbine, reporting to your emonPi. Then you’d have the option using the ac adapter to give you real power, and the high sensitivity input.

1 Like

Hi Robert and thanks - so I can use a voltage sensor to read the real inverter voltage. Given my previous comment about taking the Emonpi to another site, I assume there’s no need to have an EmonTx and an Emonpi just for wind turbine monitoring though?

I don’t know what your other system is capable of, and I wasn’t sure what you meant by “another site”. If it’s out of radio range of your Pi, then you’ll need something to record the data, so another Pi (modified) instead of the emonTx would be what’s needed, unless of course your other system can somehow receive data from the emonTx.

I am planning to add a 3.5kW wind turbine which will be off-grid and just used to supplement heating.

I encountered precisely this situation in a diverter system that I supplied for use with a grid-tied wind turbine. When the turbine output reaches a certain level, any excess spills over to a non grid-tied inverter which is just used to heat water. Because the “overflow” load is resistive, my approach was to regard the consumed power as being proportional to the current squared. The resulting data is transmitted to emonBase as an extra field in each routine data message.

Changing the algorithm to calculate power as I * I rather than V * I for one of the CT channels requires only a minor change to the code.

Robin, I understand this up to a point…
Firstly, I’m sure you’re right, I’m just trying to make sense of it given
my very basic knowledge of electricity.
Your example does indeed sound like my situation, from the point of surplus
power reaching the non grid tied inverter.
If the consumed power is proportional to I*I for resistive loads presumably
you have to apply a factor to the equation to give an actual power figure.
How is this calculated? Is the actual power therefore independent of the
voltage or is this taken into account elsewhere in the calculation?
If I can use a voltage sensor to track the voltage from the inverter with
an Emonpi, combined with a current sensor on the cable feeding the
heater, I think that’s the simplest solution for me…

The standard algorithm evaluates real power by making lots of instP = V * I calculations. These values are summed together over the recording period and then divided by the number of contributions to find the average power. To convert this to Watts, a calibration value is applied.

The standard current-only sketch evaluates apparent power on the basis that the voltage will be fairly constant, e.g. 230V AC. With your non grid-tied inverter, I have no idea how the voltage will behave when there is insufficient power to drive your permanently connected load. But if the load has a constant resistance, a calculation of I * I should give a reasonable indication of the power.

For calibrating such a system, you will just have to adjust the cal value until the reported value looks to be about right.

OK that makes sense. But I thought the voltage sensor on the Emonpi meant
that an accurate voltage figure could be used?

The AC voltage that is sensed by your EmonPi basestation is phase-locked to your mains supply. That’s fine for V * I calculations when measuring current that is supplied from this source. But it’s no good when the current that’s being measured is asynchronous.

Hope I’ve not misunderstood your meaning …

Robin, the ac adapter for the Pi is independent of its power supply, so with the 5 V supply on the mains and the ac adapter on the inverter output, the Pi should read correctly whatever the voltage.

It would need t be another Pi as the other site is about 12 miles away…

Yes that’s what I thought…

OK. I thought the Pi was already in place for conventional grid-tied measurements, and the wind-turbine channel was to be added using a spare CT input.

Nope, just looking to monitor output from a single off grid wind inverter. Solar at home will be monitored using Solar Log. Solar at remote site 12 miles away will be monitored (hopefully) with my Emonpi - currently monitoring home solar - with added GSM dongle.
Sorry for the confusion.

If it’s of any relevance I have recently set-up an emonpi to record both PV (2.5kW) and Wind (6kW) in an off-grid set-up with standard CTs attached after the respective meters (effectively inverter output). The inverters are both grid-tied models set up for off-grid and AC coupled into an Island Grid system. This works fine and values correspond very nicely with both meters and Inverters (there is an off-set between these two anyway) and inverter standby draw values also look accurate - thus far anyway!

Hi Adrian. Nice to hear your setup works and it must be interesting to see the traces of solar v wind and whether they complement each other well during daylight hours, ie if you have more solar power is there often more wind power and vice versa.
It makes me wonder why the Emonpi is not promoted more as a device for monitoring all renewable electricity generation not just solar.