Live power feed

I find it useful having a display on my computer of how much power my Powerspout is producing.  Many times a day, when doing other stuff on my desktop, I'll check to see what the output is. It saves a walk downhill to the turbine or uphill to the header tank, - though I do those as well !

The data stream I view is the same as anyone can view by clicking on this "Live power" link.

Viewing it often as I do has thrown up observations which make me curious as to what I'm seeing. In this diary entry I want to write about these.  I am, after 3 years, still reaching for explanations for some of them.  

The output should in theory be absolutely constant. And mostly it is, as this record shows, captured just now as I started writing.  The variation of 1 watt either side of 216 W simply represents the limit of the system's ability to resolve and display the true reading; the base line is dead flat for 7.5 hours:

But the record is not always a flat line like this. Earlier in the year, I captured this record showing how a change in temperature in the SmartDrive housing, created artificially by blocking or allowing ventilation, made the power output anything but a flat line. As can be seen the baseline dropped 20 watts with the compartment warming up and rose again when it cooled down:

Discussion about this 'temperature effect' can be found on these blog pages.

Recently, I think I've come to understand what's going on in another type of record I see almost every day.  Unlike the fluctuation caused by temperature, which is a genuine change in the power output of the Powerspout, this one is only an apparent change.  It arises from the way in which the monitoring system works and not from a real variation in the power being put out.

An example can be seen in this trace from last week:

What is noticeable here is that output is dead level between midnight and 06.00 hrs, and dead level again between 20.00 hrs and the following midnight. In between these times it appears to jump around all over the place.  This in-between time is day-time, the time when activity in the house is causing most electricity to be used. It is also the time when my solar panels are working, putting their output into the grid via the house consumer unit.  Could it be that the level of energy consumption in the house and the output from my solar panels have an effect on the display of hydro generation?

I think the answer is yes and the explanation is twofold. It lies in understanding the way the hydro monitoring gets its data and what happens to mains voltage when household consumption is high or the solar panels are generating.

The way the live power feed gets its data is from the frequency of a pulse taken from the installation's kWh meter.  The meter is calibrated to give 1000 pulses per kWh.  The parameter which the meter actually measures to give these pulses is current.  It doesn't measure voltage, which is the other parameter necessary to give an energy reading.  It just assumes that voltage will be constant, and for my Elster meter it assumes 230v.

Now in reality, house voltage is never quite constant at 230v. It varies slightly up and down from this value as a result of fluctuations in voltage in the main grid network; in addition to these fluctuations there are further fluctuations contributed by voltage drop over the supply cable from grid to house, and these will be dependent on how much power is being taken moment by moment.  So the voltage seen in the house actually varies significantly through the day, going down as power consumption goes up, but going up if there is much solar input.

What this variability in voltage means for the display of live power is that what is, in reality, a constant power output now appears to be not constant; and it mostly appears to be not constant in daylight hours because that is the time when the factors causing disturbance to voltage are most prevalent. It appears not constant because when the prevailing voltage is less than the meter's nominal 230v, the current the meter will measure will be higher by as much as the voltage is lower; since the meter measures only current, and this current measurement alone is the source of the frequency of the pulses on which the power display relies, the result will be an apparent upward blip on the power trace.

Conversely, transient or sustained voltages above 230v will cause downward blips and troughs.

It is only because the sensitivity of the system which is measuring live power is so good that such fluctuations in voltage can be seen to have any effect.  For all practical purposes, they matter not at all !

Thoughts about water storage

Some considerable time before I installed my Powerspout, when I was still trying to decide if it was a scheme worth proceeding with, a friend who had much experience with hydro came to walk the ground with me.  He advised creating a bit of water storage.

At the time, I wasn't very sure what he meant. I assumed he meant digging a sizeable reservoir.  The nature of the terrain made this an impossible proposition so I was left wondering about his advice, - when he was with me, I didn't like to show my ignorance by asking him to explain !

Gradually I came to realise why storage is valuable, - not to provide a huge volume in the hope of extending generation into the dry months but just a small volume to facilitate day-to-day operation by buying time before action needs to be taken when outflow and inflow don't match up.

How valuable such a buffer proves to be depends to a considerable extent on the behaviour of the flow one is capturing, - and operating a small hydro certainly brings you into an intimate understanding of the behaviour of your particular water source. 

Such understanding comes slowly.  After three years I am still getting the measure of how, as summer progresses, the flow initially drops off quite quickly but then slows down in its rate of decline.  Plotted on a graph, I think it would be an exponential curve. 

Another fragment of this understanding is that a bit of rainfall in the summer makes little or no difference to the rate at which flow falls off.   Yet another fragment is that when winter rain starts, it takes several weeks before the flow begins to pick up, but when it does pick up, it can pick up very quickly indeed: in my 2014/15 water year, flow went from 1.18 L/s to my maximum of 3 L/s in the space of just one week.

With a source like mine which, when it is diminishing, changes only slowly, my 5 cubic metres of tank storage means I can wait several days, even a week, from the time when the tank stops overflowing to the time when I need to step in and change to a smaller nozzle. Only because of this behaviour am I able to operate as a 'variable head' site, as described in an earlier post.

But for a different site where the flow is more 'flashy', by which I mean the flow rises and falls rather quickly in response to rain falling, even my generous 5 cubic metres of storage would be inadequate to operate in such a way. It wouldn't buy enough time, wouldn't provide enough of a buffer, before a nozzle change were to become necessary.

So the merits of putting in a tank at the top of the penstock, and the size of it, really depend on the site hydrology.

EcoInnovation has recently published a very good document about constructing intakes for their Powerspout turbines. It can be found here.  In it, mention is made of the benefits of providing storage and it is well worth reading if you are at the stage of wondering about putting in a turbine.

To illustrate how a storage tank can be sited relative to the take-off point from a water-course, here are some pictures:

Rotor packing

The weather has been cool and cloudy here in Wales for what seems far too long. There has been rainfall too, more than usual for this time of year, but none of this has stopped flow to the turbine steadily diminishing.  Today was a day for reducing the size of nozzle to keep flow delivered matched to what is available.

Doing this gave an opportunity.  I'm working in previously uncharted territory trying to make the most of these smaller summer-time flows whilst using a non-MPPT inverter**, and today's nozzle change gave scope to experiment.  

In previous diary posts, I've outlined what the aim is, - to keep pelton speed up in order to have it operate near its optimum rpm where most energy is extracted from the head and flow available. Keeping speed up can be achieved by making the rotor stand off from the stator coils by placing packing washers on the shaft. But the method is one of finding by experiment how much packing produces the best output: too little makes rpm and power output low; too much makes rpm better but reduces alternator efficiency so the end result is just as bad.  Somewhere in the middle lies a best compromise and the challenge is to find it. These pictures show how it can be done:

For these photos, the camera was carefully aligned with the back of the turbine casing so the amount of rotor stand off would be captured along with the rpm on the tachometer.  As can be seen, 8 mm of stand off produced most watts at a speed of 876 rpm, so that was how I left it, but I can't help feeling a tad more, say 10 mm taking rpm just into the 900's, could be better.  

That'll be a job for tomorrow perhaps...***

** It is worth stressing that I am only experimenting with rotor packing because this inverter does NOT use Maximum Power Point Tracking (MPPT) to optimise the power fed to it.  Inverters that use MPPT should find optimum speed automatically and with greater accuracy. They should therefore always give maximum possible power so long as voltage is within the range specified for MPPT tracking.  My own installation probably proved this last year: then I was using an MPPT inverter and with the turbine set up precisely as it was today except for having ceramic bearings and a reduced core stator, the output to grid was 299 watts.  Allowing 10 watts extra output for the decreased rolling resistance of ceramics compared to standard greased bearings, this was 14 watts more than the best I could achieve today (275 watts) with experimental rotor packing.

***... and indeed it was a job done later: an extra 1.75 mm (one of EcoInnovation's special packing washers, specially made for the job) was added to the existing packing of two rubber O-rings. It meant that the 'knob' only turned 9 full turns instead of the previous 10 needed to fully push home the rotor. The effect was to raise rpm to 974 and watts output to 281.  OK, that's a tiny gain over 275 w.  What's important though is that it proves, and calibrates, the method. Very satisfying !