6.48 mm diameter nozzle delivering 0.91 l/s to the runner which is rotating at 1084 rpm and generating 225 watts into the grid at an overall efficiency of 47%.

Monday 25 April 2016

A new approach to dry season running

Changing nozzles to track diminishing flow can, at this time of year, be something which needs to be done rather frequently.  Last year between May and October I needed to change nozzles 11 times.

All 'direct-to-grid' connected hydro turbines need to operate at a fixed rpm: the frequency and voltage of their ac output demands it.  But a Powerspout is different because its output is fed to the grid via a rectifier and an inverter. These 'condition' the electricity produced by the SmartDrive to turn it into perfect mains quality electricity, and a consequence of being grid connected in this way is that operating at a fixed rpm isn't necessary.  However 'dirty' the output from the SmartDrive, meaning whatever the frequency and voltage, both of which are determined by rpm, the rectifier will 'clean' it up and the inverter will make it indistinguishable from mains ac. 

This being so, changing nozzles to keep track of changing flow becomes less of an imperative. The one purpose in doing it is to keep head at the design height so turbine rpm will be kept at design rpm. 

If slavish adherence to a fixed rpm isn't essential for a Powerspout, why not, as flow diminishes, let the header tank empty instead of changing to a smaller nozzle? - let a new head level establish itself somewhere down the penstock: - the place will be where the flow through the nozzle(s), reduced by the reduced head, equals the inflow to the tank; and let the rectifier and inverter do their job of making good the 'dirty' electrical output caused by the pelton turning at a reduced rpm.

In short, instead of aiming to generate as a fixed head, fixed speed installation*, think instead: variable head and variable speed.

I have not yet tried this way of operating.  In a test to see what would happen if the tank was allowed to empty, the observations were supportive of the theory, so I intend to give this new approach a try. (The test is reported here in a Dec 2014 blogpost.)

Instead of the 11 nozzle changes of last year, I reckon I might get away with just 2 or 3.  The big downside will be a loss of efficiency in the system because it is operating away from its design parameters, so that for a given flow, less will be generated compared with the conventional, fixed head approach;  and there will be the downside too of silt being carried into the penstock from the bottom of the empty header tank. But on the plus side, every drop of water at this time of year when water is less will be used for generating, instead of some being lost because the header tank is full and overflowing.

We'll see how it works out ! - but I suggest others don't follow in case it leads to unforeseen disaster. I'll report back.

* with a Powerspout the operational speed is actually never quite fixed, it varies with flow / power output, so referring to it here as a fixed speed installation is slightly inaccurate. In theory, if the head is constant, the optimum speed will be constant, so in that sense it is a fixed speed installation.

Friday 15 April 2016

Two 'water years'

At the half way point in my 'water year' (1st Oct to 30th Sep) output to the grid is holding up better than in previous years, a reflection of the wet time we've been having here in Wales (the green line is this year's plot):




Perhaps this'll even be the year I'll reach that dreamed of total of 4 MWh. 

The end of March marked the end of a different water year, the one set by the Environment Agency for water abstraction returns to be made (1st Apr to 31st Mar).

My abstraction licence limits me to 49,982 m³ and the close of their year saw my actual abstraction coming in at 46,365 m³. So that was alright then !

When I tried to submit this information on-line in the way that was possible last year to the 'English' Environment Agency, there was a problem: during the year abstraction returns had been devolved to the 'Welsh' Environment Agency, aka Natural Resources Wales.

Unfortunately, and not only unfortunately but also un-surprisingly, NRW had not prepared for receiving abstraction returns at the due time.  A phone call to head office to ask how it should be done; a promise to call back with an answer which never happened ... what is one supposed to do in the face of such bureaucratic inefficiency?

The dead-line for a return is threateningly stated to be 28 April, - just 2 weeks away.  What enforcement action could they impose I wonder.

Then, - oh the relief! - I found I could submit my return to the 'English' Environment Agency in the same way as I did last year.  Too bad that it has gone to the wrong agency.  Let them sort it out.  I beg to submit, your honour, that I did send my return in, on time, by the only way I found to be possible !


Monday 4 April 2016

Measuring windage.

Windage is defined in the dictionary as "the air resistance of a moving object ... such as a rotating machine part". For a pelton therefore it is the resistance encountered by the cups of the runner as they turn through the air in the casing. What is in the casing though is not just air but the more dense medium of water droplets suspended in air. There is a veritable fog of spray created by the water jets striking the cups.  

Because windage is a resistance to rotation, it causes power to be lost from the power available for generating electricity.  The loss is unavoidable, though it can be mitigated by designing the pelton casing so that spent water and spray are kept away from the rotating runner.   In small turbines like the Powerspout this is almost impossible to do.

Normally, power lost to windage is very difficult to measure and can only be estimated mathematically. But in a 'Eureka moment' recently I realised that two features of my installation provide a unique opportunity to actually measure windage: the turbine is situated outside so removal of the transparent glazing wouldn't matter, and secondly I have a record of power output captured continuously and visible on-line*.

So this afternoon the experiment was done: at a precisely noted time, the glazing was removed whilst the turbine was operating; the turbine was left to run for just under 15 mins with the glazing off and then, without stopping the turbine, the glazing was put back on.

Here are some pictures and below them, the resultant power trace captured from the on-line record:












The power trace unequivocally shows a jump in power output coinciding precisely with when the glazing was removed and replaced.  There can be no doubt about its authenticity.  The actual gain in power amounts to 20 watts (~705 to ~725 watts).

One can speculate that the total power lost to windage will actually be double this because, ...on the assumption a jet strikes the splitter ridges dead centre and is split exactly in half, ...there will still be the windage loss arising from the half bouncing back from the bulkhead.  So it's fair to say that windage loss probably totals 40 watts, which is just under 6% of the power output at the flow used for the experiment.

The flow used was 2.78 l/s.  So in the 15 minutes of the experiment 2.5 tons of water sprayed out (2,500 litres): not an experiment for those whose Powerspouts are housed in smart enclosures !

It was fun to do and pleasing to get the result I sort of expected.  My thanks to Paul Jones, my neighbour, for helping with it and being the one who got wet.

*You can view the actual record here but scrolling back to 4th April 2016 will be tedious for people who read this even a few days after the day of the experiment.