Need help with MPPT curve for 48V Bergey XL1 paired with Ginlong inverter

[MICHAEL FEERICK]

Hi,

Joined this forum over two years ago, but this is my first posting !

I've had a Bergey XL-1 24V battery charger setup running here in Ireland since 2010, but after the Power Center controller and batteries died 2 years ago, I thought it would be a good idea to try and rejig it for Grid Tie operation.

So, to that end, the 24V stator has been replaced with a higher voltage 48V one (supplied by Bergey), and a Ginlong GCI-2G-W Wind Grid Tie inverter has been purchased (it's a 2014 model: wind inverters for small turbines seem to be quite rare these days !)

The Ginlong has the option to define a user MPPT table, and I downloaded the spreadsheet share by Rob (Becker) in the post "Power Curve for Scratch Made Hugh Piggott Turbine", but haven't figured out how to use it yet :unsure:

Given that the rated power of the XL-1 is 1000 watts @ 490 rpm, in a wind of 11 m/s, what would be a reasonable starting table for this turbine ?

I haven't connected the turbine to the Ginlong just yet, but measuring the open-circuit DC voltage output in a variety of windspeeds, we see voltages varying from 80V up to around 160V. The tech support guys at Bergey say it can go as high as 240 volts in a strong wind, which is still well within the inverter maximum input voltage of 500V.

However the maximum input current at the Inverter DC terminals is 10 Amps.

Would it be reasonable to assume that the XL1 output current would stay well below 10 Amps - maybe in the order of 6 to 7 Amps maximum, given the relatively high output voltage of the turbine ? (high compared to the old 24V winding, I mean)

The Ginlong manual does emphasize the need to calculate the current at each MPPT point in the curve, to ensure it stays below 10A

Any help or advice on this topic would be warmly appreciated :bigsmile:

Cheers

Mike

[Rob Beckers]

Hi Michael,

If you could please send me the following info:


The exact diameter

The target TSR (if you know it)

The RPM-to-voltage ratio, or the open/unloaded AC Voltage vs. some RPM as you measure it

The phase-to-phase resistance (Ohm)


And you mention the current limit is 10 Amp, so I'll take that into account. With those numbers I should be able to calculate an MPPT curve.

-RoB-

[MICHAEL FEERICK]

Hi Rob,

Thanks for taking a look at this so quickly - appreciate that :-)

In terms of the info you asked for.....


Rotor diameter = 2.5 meters
Target TSR I believe is 5 (though calculating it based on the rated RPM of 490 in a 11m/s wind suggests TSR is 5.8 )
The phase-phase resistance is 3.1 ohm


But regarding the RPM to voltage ratio - not sure about this. The turbine is still rectifying the wild 3-phase output in the nacelle, so it's DC coming down the tower (Bergey could not supply the 3-phase slip rings)

And unfortunately the turbine is actually back on the tower at the minute, so it's difficult to get any AC phase measurements from it

Would it help you if I made a table of measurements of turbine open-circuit DC voltage versus RPM ?

I've attached a pic of the XL-1 performance graph and dimensions for battery charging, in case that helps.

Many thanks

Mike

[Rob Beckers]

Hi Michael,

Alas, that Voltage vs. RPM number is crucial. It's not possible to generate anything remotely meaningful without it. Just one (unloaded) value is fine, this should be linear; twice the RPM is twice the Voltage when there is no load. So one single RPM and Voltage pair is all that's needed.

-RoB-

[MICHAEL FEERICK]

Thanks Rob,

Just be sure here - a DC output voltage and its corresponding RPM - correct ?

Getting the voltage reading is no problem, and for the RPM, I've ordered up a bicycle speedometer to measure that, as per the hack detailed at https://www.instructables.com/Tachometer-made-from-a-bicycle-speedometer-cycloc/, so it could be a week or so before I can get back to you.

Thanks

Michael

[Rob Beckers]

Michael, AC phase-to-phase Voltage is better, but I can work with DC Voltage if that's all you have. Just make sure it's unloaded. Loading up the output also adds the effect the windings resistance into the mix and makes it hard to separate the two.

-RoB-

[MICHAEL FEERICK]

Hi Rob,

I was having problems measuring the rpm with the cycle computer, so changed approach - maybe you can confirm (or otherwise) if this is a valid way to measure RPM in this instance......

I stuck a scope on the turbine DC output, as follows:


Ch A - DC coupled, 50V per division, to get a feel for the DC voltage.
CH B - AC coupled, 10V per division, to look at the AC component in more detail.



I was expecting to see the classic AC textbook waveform of 6 rectifier pulses per rotation of the rotor, but there seemed to be many more pulses than that.

The rotor has 12 magnets around its circumference, so I'm working on the assumption that the AC component would have a frequency of 72Hz (6 x 12) at 1 revolution per second.

One can see the DC voltage increasing, and the frequency of the pulses also increasing, as the RPM of the turbine increases ( and vice-versa).

If the above is correct, then the dc voltage to RPM ratio is coming out at around 0.63, +/- 0.01 .....does that sound like a correct value ?

For example, at 113V, the RPM is 178 rpm => 0.63 V/rpm

I've attached a sample screenshot from the scope for this particular measurement.

BTW, I'm using a separate Fluke multi-meter to measure the DC voltage (the scope measurement was consistently 4 volts less than the Fluke value - and I'd trust the Fluke on this one !! )

Let me know what you think.

Thanks

Michael

[Rob Beckers]

Hi Michael,

Why look at the DC when you have AC available? The AC side should give you a frequency (most scopes will auto-calculate that for you as well), just looking between any phase-to-phase, and with 12 poles the RPM is:

RPM = Frequency x 60 / (Poles / 2)

In other words, if you see 10 Hz coming out, it would be 60 x 10 / 6 = 100 RPM.

Does that work for you? Or do you only have DC coming out (is the bridge build into the turbine?)? If you only have DC available it gets complicated. Converting single-phase frequency to the peaks you see 3-phase isn't straight forward. I'd have to first wrap my head around that one.

-RoB-

[MICHAEL FEERICK]

Hi Rob,

The DC output is all I have access to at the minute - the 3-phase rectifier is built into the nacelle, and to get at the AC terminals, the fiberglass cover needs to come off, which entails removing the tail assembly.....it's possible to do that, just not quickly. I'll see what I can do.

Thanks for sharing the formula for the rpm - that's good to know.

What would be a ballpark 'volts-per-rpm' for a turbine of this type - are we talking 0.1 to 0.3, typically ?

Best regards

Michael

[Rob Beckers]

OK, done some head-wrapping: With each of the 3 phases (A-B-C) rectifying against each other you should end up with 6 little DC variations (not quite sinewave) for each period (AB - AC -BC - BA - CA). Now, those 6 DC "humps" would be interpreted by a scope as 6 periods of a (poor) sinewave, so you get 6x the frequency of the the original 3-phase.

So, if your scope is seeing 41.3 Hz (LeCroy, nice stuff but expensive!) that would have been 6.88 Hz 3-phase. That should in turn correspond to 68.8 RPM. And you say you measured 113V RMS (semi-DC, unloaded) at that RPM? That would make 1.642 V/RPM DC-unloaded.

That seems high.... Typical numbers for a turbine of this size, assuming reasonable TSR, and cut-in for battery charging at something like 3 m/s, it should be more around 0.4 or 0.5 V/RPM.

That only gets you so far, as another important parameter is the phase-to-phase resistance of the windings. Once the alternator produces power there will be more than a little Voltage drop over those windings purely due to resistance, and ultimately that also determines where you have to cut things off in terms of current to avoid overheating the alternator.

To be honest, all this is somewhat reading the tea-leaves. To get any reasonably reliable numbers you'd have to actually get the thing down and measure RPM, Voltage, and resistance. Even with those numbers as accurate as you can measure it's still a crude approximation to make an MPPT table out of that.

You could determine the MPPT curve via trial-and-error, but that would require an anemometer somewhere near hub-height (so you can plot power vs. wind speed). By changing the curve up/down you could iterate to the best one.

-RoB-

[MICHAEL FEERICK]

Hi Rob,

Firstly, thanks for taking the time to work through the analysis - it's much appreciated :)

The figure of 41.3 Hz you mentioned - that's the reciprocal of the delta time between the two measurement cursors on the screen (which are quite hard to see - they're the faint vertical dotted lines) - but those measurement cursors are controlled by the user, and aren't an automated measurement. I should have turned them off - my bad.

The actual frequency of interest is the 214.6 Hz measurement for CH B - the maroon-coloured trace.

When we plumb that number into your calculations, we get the following:


35.66 Hz 3-phase
RPM = 356 rpm
DC Voltage/RPM ratio = 113/356 => 0.317


That last figure seems a bit more respectable :bigsmile:

The per-phase resistance of 3.1 ohms was measured by Keith in Bergey tech support, so I'm reasonably confident that's an accurate value.

Is this enough data to generate an initial MPPT table ? The Ginlong inverter table can have up to 32 points, but from what I read in similar posts on this forum discussing MPPT tables, it sounds like 8 points would be quite sufficient.

yeah, the LeCroy scope is nice - tbh, I'm more au fait with the Tektronix scopes, and the difference in UI between the two does confuse me at times, when using the LeCroy....too many soft keys !

Best regards

Michael

[Rob Beckers]

Michael, looks like Bergey targeted a TSR of 5.5 for the turbine. If I target 6, and use the numbers provided, I get this as the MPPT values:


Wind (m/s) RPM Freq. (Hz) Heat (W) Eff. (%) Idc(A) DC (V) P (W)
5 229 22.9 20 86% 1.2 61 73
6 275 27.5 40 83% 2.1 71 148
7 321 32.1 73 80% 3.1 79 246
8 367 36.7 123 78% 4.2 87 370
9 413 41.3 195 75% 5.5 95 521
10 458 45.8 294 73% 6.9 101 699
11 504 50.4 427 70% 8.5 106 903
12 550 55.0 599 67% 10.2 111 1134
13 596 59.6 819 65% 12.1 115 1390
14 642 64.2 1093 62% 14.1 118 1669


The last two values, DC Voltage and output power, are what goes in the table. The 5 m/s value is not a bad one to start with; the turbine will spin up and produce a little before that, but there's just not much power in those lower wind speeds (I get 20 Watt out at 51 Volt DC, at 4 m/s wind speed). The power values take the inverter loss into account, so this is output power (which is what's used for the MPPT table). Just a handful of values is all you need, it's fine if the inverter linearly interpolates between.

You can see the heat dissipation rises rapidly for the stator, and frankly I wouldn't go beyond the 11 m/s value. At that point there's over 400 Watt in heat going into the stator, even with the wind cooling the thing that is a lot!

Given that Bergey claims 1kW out at 11 m/s for battery charging (which is typically less efficient vs. driving an inverter at MPP), it could be that the Volt/RPM value is a little higher than what's used here (I've converted it back to unloaded AC RMS from your DC value, since that's what I use in my spreadsheet, maybe it's me). If you feel the turbine is running at too high an RPM with this curve you can increase the power values some.

Let us know how it goes, and good luck!

-RoB-

[MICHAEL FEERICK]

Many thanks, Rob - I'll program this table into the Ginlong, and hopefully we'll get some wind this week to see how it performs - will certainly let you know how we get on.

That's an eye-opener on the amount of power dissipated in the stator !!

Best regards

Michael

[MICHAEL FEERICK]

Hi Rob,

We had some usable wind yesterday, and the MPPT table seems to be doing its job :bigsmile::bigsmile: - we harvested about 2kWh over a 24-hour period where the wind was between 11 and 13mph.

I needed to tweak the power curve somewhat to play nicely with the Ginlong software, as it expects power values to be a multiple of 10, as are the voltage values ...so this is what I ended up with:

Volts DC Power (Watts)
60 70
70 140
80 240
90 450
100 690
110 1000
120 1000
130 1000
140 1000
150 1000

The last 5 rows in the above table are to keep the DC input current below 10 A, which is the maximum allowable input value stated in the Ginlong manual.

Do you know if this maximum input value is controlled solely by the power table - or do I need to provide some overcurrent protection circuitry as well, like a circuit breaker, at the DC input to the inverter ?

There are some stronger winds forecast for this Thursday (15mph steady, with 30mph gusts), so I want to be sure that the inverter is working within its limits.

I've built the overvoltage protect circuit that you designed (the one that uses the Omron overvoltage relay controller) which is needed anyway in the event of the grid going down...so maybe I can use that in a pinch.

Its good to have the Bergey running again and earning its keep !!

Best regards

Michael

[Rob Beckers]

Hi Michael,

Glad to hear it's working! :whoo:

As long as the inverter is working the current should indeed be controlled by the MPPT table. Current into the inverter should be slightly larger than the calculations would suggest, because the power is defined as 'output' from the inverter, on top of that come inverter losses (so 1000 Watt out, means about 1040 Watt going in).

Keep in mind that the grid can go down, and therefore the load on your turbine. I don't know if the Bergey is rated to run unloaded, or if that leads to self-destruction. If it's the latter you need to put something in place to prevent that (dump load or such).

-RoB-