Optimising my inverter turbine curve

 The inverter that connects my Powerspout to the grid does more than just convert DC power to AC; by the load it imposes on the alternator, it determines the speed at which the shaft turns, and the speed at which the shaft turns is critical if a pelton is going to extract the maximum power from the head and flow of water available.

I run my inverter in what is called 'turbine mode'**; run like this, the load the inverter imposes on the alternator is determined from a 'mathematical table' stored in memory in the inverter, that tells the inverter how much load to apply depending on what DC voltage it is receiving from the alternator.

The 'table' is really an equation relating DC input voltage to AC output watts, and whilst a new inverter is delivered with a table chosen by the manufacturer, the table can be changed by the user.

A user might want to change the table because the factory parameters may not suit the performance characteristics of his hydro site, in particular the inverter may not allow the pelton to rotate at its optimum speed; the table in the inverter will work with the factory settings but it will almost certainly not work to produce the very best efficiency.

In a small hydro installation where power outputs are so small, tweaking efficiency to the best possible is a good idea to get the installation to be as productive as possible, and it has been for this reason I've been playing around with optimising the table in my WindyBoy inverter.

Without going into more detail than is necessary to explain how a table relating DC voltage to AC watts is constructed, some familiarity with Microsoft Excel is required.

The relationship between these two parameters is described by a single equation relating x and y, where x is DC voltage and y is AC watts.

The factory default table uses the equation: y = 0.0001x^3 - 0.0306x^2 + 2.92x - 83.

When this is plotted on a graph with x values chosen to be in the range of the DC voltage produced by my Powerspout, it produces the green curve in the graph:

Last year, I spent time carefully investigating what voltage produced the most watts output, for each of the nozzle sizes I operate my Powerspout with over the course of the year.

As a result of that exercise, I came up with a revised table, giving a revised curve, which is shown by the orange line. You can see it is a flatter curve than the green one and that at low power outputs (100 to 150 Watts) when the DC voltage is around 170 volts, the line starts to curve upwards without ever reaching the x axis.  The effect this had in practice was that at this low level of power output the inverter failed to connect to the grid, - presumably because the look-up table confused the inverter regarding the load it should be applying.

Last year therefore, I could not continue to generate with my smallest nozzle (0.53 litres/s) using the inverter with this table in its memory and I had to swap to another inverter which still had the factory default table.

This year, with summer progressing toward the driest time of year in September / October, I wanted to find a curve which would be able to work at this low flow. Playing around in MS Excel has yielded the blue line in the graph which I hope will do the trick.

You can see it almost overlays the orange line for most of its course, and that is good because it means it will make the inverter operate at the optimal points determined by my investigations last year, and crucially it does not start to curve upwards when power outputs are low.

Time will tell if my playing around has worked. Already in the few weeks I have been using this revised curve, when flows are still around 1.5 litres/s, there is a marginal improvement in power output compared to the orange curve, lifting water to wire efficiency by just one half of a percent! (53.3 to 53.8 %).

Every efficiency gain counts !

Here's a picture of loading the revised formula to the inverter from a laptop, running SMA's WindyBoy set up programme, and using SMA's USB Service Interface cable:

 

** Most inverters used for Powerspouts are solar PV inverters which are programmed to operate in Maximum Power Point Tracking mode; MPPT works for a hydro; but in this mode the inverter is constantly seeking the maximum power point, which for a water turbine is not changing, and so the output is seen to be constantly fluctuating when it could be a nice straight line.

Earth fault ? - no problem !

Addenda to original post: see end for latest situation.

Lately, the inverter connecting my Powerspout to the grid has been signalling an error message.  I've got to the bottom of it now and cured the issue but the journey has been an educational one which I thought might be helpful for others to know about.  The inverter is an SMA Windyboy, which is the same as the SunnyBoy, and what I describe only applies, I think, to these two SMA inverters.

The error message was "Earth Fault". Since the supply from my turbine is not intentionally grounded this meant a possible fault in the turbine, or possibly on the cable coming from the turbine to the inverter; fortunately the inverter continued to operate so I felt in no hurry to get to the root of the problem.
In this situation, SMA's trouble-shooting guide instructs you to exclude a genuine earth fault by inspection and testing. If this first-off approach doesn't reveal what's wrong, the next step is to test the 2 varistors housed within the inverter because, the guide says, a failed varistor can cause the Earth Fault warning to be displayed.

In an SMA inverter, the varistors look like this:

Each has 3 wire tails and when the insulating shroud is removed, each is revealed to be a composite of two components: a varistor connected in series with a thermal fuse.
The trouble-shooting guide says to test for continuity between B and C, ...if there is no continuity the assembly needs to be replaced; if continuity exists, ...look for a fault elsewhere in the inverter.
It was only sometime later, after I had removed the insulating shroud and seen there were two components beneath it, that I realised testing in this way only tests for continuity across the thermal fuse; it cannot test the functionality of the varistor itself, which is the blue disc-shaped component.

When I did the continuity test, both the varistor-cum-fuses showed continuity, and so I concluded, as the SMA literature had led me to believe, that they were OK.  But a week later, having exhausted all other possible causes of the fault condition, I replaced both with new ones and, hey presto, no longer was Earth Fault displayed: I had got to the bottom of it 😊. (** - but see post scripts below)

So what's the 'science' behind all this ?  The purpose of a varistor is to eliminate voltage surges which might damage the equipment the varistor is protecting; they are 'sacrificial' devices, meaning they can be destroyed by the excess energy they absorb, and they are also 'wear' components, meaning they gradually lose their function from the cumulative effect of absorbing energy from lesser voltage surges which are not great enough to destroy them.

In extreme situations, varistors can catch fire, either because of the magnitude of the energy passing through them or because of the duration the energy flow exists for.  In this situation they pose a fire risk which might destroy the very device they are meant to be protecting and to mitigate this, a thermal fuse is sometimes included in series with a varistor, - a fuse which will 'blow' and terminate the supply through the varistor if its temperature exceeds a set point.  This is the arrangement of the two components found in what SMA call 'their varistors'. 

SMA state that 'their varistors' are specially manufactured and are not commercially available, - except, of course, from SMA. The cost (in 2017) is €15 plus shipping and VAT (total €29.75), for a pack of two (part code SB-TV3, with insertion tool).

The two components that make up 'their varistor' can however be found on the open market, and with a soldering iron to connect them in series, they can be made up more cheaply. The varistor is an Epcos S20K320 (Vrms 320v, VDC 420v, Imax 8000A, Wmax184 joules Pmax 1 W) and costs 0.55p from Farnell (order code 100-4305).
The thermal fuse is more difficult to find but I managed to track down 6 on eBay: it is a Tamura E3F 250v, 3A~, 115℃ and each one set me back €2.20, inclusive of p&p. But it can also be purchased direct from China for less than this.

Why should the varistors have failed in my inverter to cause the Earth Fault warning? I figure that with 12,492 hours of continuous operation, at ~300v DC, the natural 'wearing' process going on in a varistor is accelerated and probably accounts for their failure. After all, an inverter handling power from a water turbine operating all the time sees a lot more use than one connected to a wind turbine or PV array. 

12,492 hours is about 17 months and so I'm thinking this is the interval at which I should expect to see the Earth Fault warning recurring, - and therefore that I should have enough stock of the varistors and thermal fuses to meet that sort of replacement frequency.

It was nice to have gained a little more understanding of how the technology was designed to work.

** note added 3rd September 2017: the fix turned out to be not long lasting. Within 2 months, the inverter was displaying Earth Fault again, initially intermittently and then permanently.  Replacing the varistors yet again, with new SMA ones, failed to get the fault light to extinguish***.  I conclude that either there is truly an earth fault somewhere or the sensing circuitry within the inverter has gone wrong.  I think the latter more likely.  It being an issue which does not affect performance, I plan to do no more about it.

***yet another note added 12th March 2018, - following my further replacement of the varistors, which failed to extinguish the earth fault warning, I did nothing.  After a while, I'm not sure how long, perhaps 2 weeks, the earth fault warning DID disappear, - and it has remained off ever since up to the time of writing this note. Perhaps new varistors just take time to 'bed in'.

****an even more recent note added 4 Aug 2018, - the warning light remained off as related above, until 3rd July 2018; from then it has been on constantly; so today I first replaced the varistors again: no change, - and then removed them completely: still no change.  It would seem therefore that the varistors are not the cause, as was surmised above, but if the warning light extinguishes in the next few days, that hypothesis will not be right.  I'll post another update if that is the case.

*****this saga runs on and on ! - and since I see from Blogger stats that quite a few people visit this posting, I have written this latest update.- this one is dated 22 Oct 2018. 

• As recounted above, on 4th Aug I removed the varistors completely and the earth fault persisted.  Being somewhat fearful of operating without the varistors, I re-installed them sometime late in August.  Earth fault was still displaying. 
• I then changed out the entire inverter for a brand new one (also a Windy Boy 1200) and the earth fault STILL displayed.  So at this point I had to conclude there truly WAS an earth fault in the turbine or transmission cable.
• Wanting to keep the brand new WB 1200 with no/few hours of use, having seen that it too displayed earth fault, on the same day as having put it in, I changed it out back to the original WB 1200. 
• I then re-started packing the 'dry-side' of the turbine, i.e. the alternator side, with silica gel. I had stopped doing this some months before because the bags never lasted more than 3 weeks before needing re-charging.  But this time, I resolved to stick at keeping the dry side below 50% RH.  And what should happen after about 2-3 weeks, to my astonishment, the earth fault disappeared.  I suppose that must have been about 5 weeks ago (i.e. mid Sep) and the fault has not re-appeared to the date of writing this. No doubt there will be a follow on in due course !!

****** further note added 14 Jan 2019: from mid Sep 2018 until now earth fault has NOT been displayed 😊.  In this time, I have kept the silica gel bags doing their drying action and have also instituted steps to minimise water entering from the wet-side (see here).  I have also changed the dump load element for a new one because the original was beginning to show pitting of the chrome plating.  Which of these interventions, if any, was instrumental in stopping the earth fault displaying I am at a loss to say.

******* further note added 28 Jun 2019: earth fault has not displayed since mid Sep 2018  😊  😊. In this time, use of silica gel has not been constant and I now do not use it at all. My conclusion is that it never played a part in solving the earth fault issue. 
The dump load heating element was changed on 15 Nov 2018, as mentioned above. The earth fault warning had extinguished before this change was made so it can hardly have been instrumental in curing the problem, - unless the problem was an intermittent earth fault which has been eradicated by putting in the new element.  This is the most credible explanation to my mind.  Some time I'll get round to measuring the insulation resistance of the old element to see if it falls below that of a new one.

******* further note added 27 March 2020: the earth fault warning has not returned.

******** further note added 30 Jan 2022: earth fault has not re-appeared since a new dump load element was put in 15 Nov 2018; there can be little doubt the previous element was the cause of the earth fault. Problem solved !

'Low flow matters' revisited

Owners of small water turbines usually get excited in the wet months of a year. That's when their turbines start to perform at maximum capacity.  Yet last December, just when my turbine was increasing its output like everybody else's, I found myself writing in this blog that I couldn't wait for the next dry season.  What on earth was I on about ? - and now that the next dry season is happening, why was it that I was so keen for these dry months to arrive?

The development that prompted my comment was when I started using a WindyBoy instead of a SunnyBoy inverter and came to see that the WindyBoy, quite unlike the SunnyBoy, caused the turbine to operate at a lower voltage at times when flows were small.

This was a startling difference and the plot below shows it.  As can be seen, with diminishing power the SunnyBoy's operating voltage progressively rises, whilst the WindyBoy's progressively falls.

The significance of this was great: up until this time I had always operated with a SunnyBoy using MPPT mode and it had caused me problems in the dry months.

In the first summer, from beyond the middle of July, the operating voltage rose to be too high to keep the turbine running. In the following year, having obtained an 18 pole stator to use in place of the 42 pole one, voltage was kept down and this enabled generation to continue all through the second summer.  But there was the hassle of needing to change stators. My excitement about the WindyBoy came from anticipation that I might in future be able to keep operating without need of having to change the stator.

So how are things working out ? - not quite as promisingly as hoped is the short answer!

The issue centres on speed.  Whilst WindyBoy does keep voltage lower it does so by imposing such load on the SmartDrive that shaft speed is reduced significantly. This effect can be seen in the plot below:

The problem with this reduction in speed is that it makes the pelton operate at an rpm where maximum energy is not extracted from the head and flow available.

One way around the problem is to re-program the 'table' in the inverter, by which I mean re-program the algorithm it calls upon to work, to make it draw less current from the pma* for a given voltage when it is working at the lower end of its range.  This can be done, I think, and I'm beginning to explore how to do it.

Another way which is simpler is to tinker with the SmartDrive.  By placing a packing washer on the shaft so that the rotor is made to stand off a bit when the knob is fully tightened, the operating rpm is lifted back toward the speed at which the pelton functions most efficiently.



Doing this is not, strictly speaking, increasing the air gap as is commonly done with axial flux pma's to adjust their output.  Rather, in a radial flux, "brake drum" type pma which the SmartDrive is, it is degrading the flux linkage between the rotor field and the poles of the stator**. But it has the same effect as increasing the air gap, which is to reduce the voltage generated at a given rpm, or, as happens when feeding the output to an inverter in turbine mode, giving the same voltage at a raised rpm.  

A few days ago when I put in the O-ring illustrated, the rpm was lifted from 829 to 897 and with this, the ac watts output from 449 to 462.  No other change was made which might have confounded these results, and in particular the hydraulic parameters powering the pelton were identical before and after fitting the O-ring.

When I went on to put two such O-rings in, the effect was not beneficial at all: ac watts became worse than with no O-ring despite rpm having risen to almost exactly the theoretical optimum speed for a Powerspout pelton on my site. The explanation for this is that the gain from the pelton operating at its 'sweet spot' speed was outweighed by poorer performance from the SmartDrive.   So the method is not straightforward.  It involves trial and error to find how much packing will benefit pelton performance before that benefit is cancelled by degraded performance from the SmartDrive.

Summer is still in its early stages here in Wales.  It's still too early to say whether my excitement 6 months ago about operating with WindyBoy in the dry months was premature.  I'll see how things go as flows diminish further and continue to play around with rotor packing. 

The experiment I have running in parallel, of operating with variable head, seems to be going OK. It allows a much more 'hands off' approach to changing nozzles as flow diminishes, and significantly, it allows use of all the flow.

The stimulus behind all these efforts is to squeeze as many kWh's out of this 'low flow' time of year as I can, - all in the hope I'll hit my 4 MWh total for the year!

*pma = permanent magnet alternator
** more exactly it is reducing the area of the gap field

Thoughts about inverters: Part 2

This post is best read after having read Part 1.

When an inverter starts accepting power from a Powerspout and feeding to the grid, an electrical circuit is completed.  In this circuit, the source of power is the SmartDrive alternator and the inverter is the load (resistance)*.  

In common with all circuits, the resistance in the circuit will determine the current. But in this circuit, unlike in other circuits where the source of power is a source which gives constant voltage, eg: a battery, a change in resistance here will not only determine the current but also the voltage.

The reason for this is to be found in the behaviour of permanent magnet alternators (PMA's).  With a SmartDrive PMA, the voltage it puts out is affected by two factors: rotational speed and load. Thus:

• the voltage output is directly proportional to the speed of revolution (rpm).
• the voltage output per revolution (v/rpm) is inversely proportional to the load in the circuit.

So here we have a circuit where the load (the inverter) is variable and can set the voltage by changing the resistance it places on the circuit.  As an aside, we should note that since one determinant of voltage is rpm, the inverter also has some control over the speed of the turbine.

The question to be answered now is: by what process of logic does the inverter decide what load it places on the circuit ?  And the answer is there are two control techniques which are possible, variously named and described as follows:

• MPPT mode (maximum power point tracking) aka: Iterative / adaptive / intelligent load control, -  primarily designed for optimising output from PV.  To quote : Solar cells have a complex relationship between temperature and total resistance which produces a non-linear output efficiency. It is the purpose of the MPPT system to sample the output of the PV cells and apply the proper resistance (load) to obtain maximum power for any given environmental conditions. Different methods are used to find the optimum combination of voltage and current which will provide maximum power. In the "perturb and observe" method, the controller adjusts the voltage from the array by a small amount and measures the resulting power; if the power increases, further adjustments in that direction are tried until power no longer increases. This method can result in oscillations of power output. From Wikipedia (abridged)

• Turbine mode aka table mode: primarily designed for optimising output from a rotating generator. In this method: The inverter regulates the input current by reference to generator voltage by using a 'look up' table.  This table, which can also be represented as a curve, defines the relationship which gives best ac power output for any prevailing DC input voltage.  The table, and also the curve, can be programmed by the user to best suit it to the particular turbine and alternator being used. From SMA WindyBoy literature (abridged)

As mentioned in Part 1, both of these control algorithms are to be found in SMA WindyBoy and SunnyBoy 1200 inverters.  Although I'm not familiar with other inverters, eg the EnaSolar range, I believe you can choose either mode in these inverters too.

In the 2½ years I've been operating my Powerspout, I've been keeping a record of the operating dc voltage, (sometimes called the MPPv or Vmpp, ie the voltage at the maximum power point).  For most of that time, I have had a SunnyBoy operating in MPPT mode as the grid interface, but for the past month I've been using a WindyBoy in turbine mode.  The turbine curve programmed into it is the default, factory one without any optimisation by me.

The difference in the way the two modes function is very clearly seen in the plot below of MPPv against ac power out to the grid.  I should add that all data points were taken using the same 42 pole stator: 60-7s-2p-star (which has a v/rpm of 0.509 v when tested in open circuit conditions).

It can be seen that:

• MPPv trends down as ac power rises for the SunnyBoy, whilst the opposite is true for the WindyBoy
• the scatter of MPPv for the SunnyBoy is wide, narrow for the WindyBoy

From these observations, it can be deduced that a WindyBoy holds the dc voltage much more constant and at a lower level than a SunnyBoy.  The plot also shows clearly why I had the problem I had last year when using a SunnyBoy, - the problem of MPPv rising to such a high level at low ac power output levels that it began to knock against the V Clamp's dumping threshold set at 378 v.  This was what prevented continuing generation at low water flows: - so much power got dumped, it wasn't worthwhile continuing. See earlier blog post here.

This year, I used a reduced core stator to get around the problem.  But it would appear that if I use a WindyBoy in turbine mode at low flow times of year, the issue will not arise. 

I never thought I'd hear myself say this: I can't wait for next summer's low flows to check this out !

*(addendum written 27/1/2016) Like all over simplifications, this statement, that the inverter is the load, compromises truth. All grid connected generators run in parallel with each other so that properly speaking the load is provided by the sum total of consumer load on the grid.  It follows that the inverter needs to behave as an 'open window' to the grid, transforming (from dc to ac) as much power to the grid as possible with minimal power being lost within the inverter. Nevertheless, the characteristics of the inverter (its impedance, its capacitance and its resistance) at any point in time have an effect on the SmartDrive output, and so this simplistic statement stands, but purely as a means of gaining understanding of how inverter and PMA interact.

Thoughts about inverters: Part 1

Those Powerspouts which are connected to a national electricity supply are unusual amongst small water turbines in that they interface with the grid through an inverter. I don't know of any other make of water turbine which connects in this way. Small wind turbines more often do.

In the UK, as in other countries, there are strict regulations about connecting a privately owned generating plant to the national electricity network.  In the UK, these regulations are written down in the document: Engineering Recommendation G83 Issue 2 (August 2012).  It is commonly referred to as just G83/2.

In the original version of this document which was called G83/1 and was issued in September 2003, a useful distinction was made (which has been dropped in G83/2) between micro hydros connecting via an inverter and those connecting directly to the grid.  The former were designated Type A, the latter Type B:

There is a clever thing about Type A connection and it is hinted at in the diagram above. It is that a lot of clever electronics have been squeezed into one box. These electronics, both hardware and software, perform two main functions: 

• converting power from dc to ac 
• managing the grid connection according to the requirements of G83/2.  

The development work for these conditioning and controlling functions has been perfected by companies competing in the huge global market for inverters for the photovoltaic industry.  An inverter is, therefore, a sophisticated bit of kit whose price has been forced down by fierce market competition.  For what it is, it's a bargain.

The market for Type B connections is, by comparison with the solar market, tiny. Within this small market it is difficult to develop a grid connection package cheaply: economies of scale are absent and also there is such a variety of rotating generators available for type B installations (induction motors-as-generators, synchronous alternators, 3 phase, single phase) that standardisation is impossible.  Each has to be specially made for its location.  The price is high.

So all in all, Powerspout's use of a standard PV inverter is an elegant and economic solution to satisfying the complicated regulations of grid connection. It is surprising that other small hydro manufacturers have not followed the same route.  

There is, however, a not-so-clever thing about using an inverter: the electricity generated has to be changed first from ac to dc, and then back again to ac, - and at each conversion power is lost, making Type A installations intrinsically less efficient.  Lower efficiency means lower productivity, - quite significant lower productivity over the entire life span of an installation, and that in turn means a return on investment which is not as good. Perhaps this is the reason why others have not followed the same route.

The G83/1 document of 2003 foresaw that inverters used in Type A hydro systems would "normally be an adaptation of a PV inverter".  Today in 2015, it is evident that as a prediction this phrase wasn't precisely correct: the inverters recommended for use with Powerspouts are not adapted PV inverters but standard ones. They operate in the same maximum power point tracking (MPPT) mode that was designed for solar inverters.  EcoInnovation provide on their website 'compatibility tests' for several different inverters and all of them, they say, should be operated in MPPT mode as if they were handling power coming from an array of solar cells.

This adherence to MPPT mode for a Powerspout is something I have wondered about.  There is no doubt that it works and there is every reason to expect, theoretically, that it should control the speed of the pelton to the point where maximum power will be extracted.  But in the two years of running my turbine in MPPT mode, I have noticed that the way the inverter controls the turbine is not always all that it could be. In particular, the control of dc voltage at different levels of power output has given me problems.

Just recently, I have obtained a Windy Boy inverter.  As the name suggests, this was intended for interfacing a wind turbine to the grid.   Its electronic architecture is absolutely identical to the Sunny Boy but the way the inverter is programmed suits it better to a rotating generator rather than a photo-diode. The mode it operates in is 'turbine mode'.

The two modes, MPPT and turbine, are both programmed into all Sunny and Windy Boy 1200's. If you have the right computer connection cable it is possible to re-configure which mode your inverter will perform in. Not having this specialist cable, nor the expertise for the job of re-configuring my existing SunnyBoy, I was happy to find a second hand Windy Boy on Ebay which was already programmed in "turbine mode".

In the second part of this post, I want to try to explain as simply as possible my understanding of how an inverter controls the voltage output from a Powerspout, and illustrate how the two modes end up causing the package of inverter plus turbine to behave quite differently.