Understanding Wind Curve for MPPT

[romekgm]

Hi
I'm trying to understand how the Wind Curve was build, couldn’t find information in the web or the manual.
I read some articles about MPPT control, and found 3 methods, for MPPT, so now I have a theory and I want to ask you if its right.
There are 3 main MPPT algorithms:
1. Keeping optimal TSR by measuring wind vel , rotation vel , and then changing the rotation vel (by load control),to achieve optimal TSR (the optimal TSR table stored independently) , these is not the case with the classic ,as it have no input of wind vel.
2.Keeping optimal TSR by gradient descent algorithm , measurements needed are power extracted and rotation speed, no extra data is necessary , these is not the case with the classic ,as we have Wind Curve that we need to use
3.Keeping optimal TSR by steeking to optimal power curve , power curve is OPTIMAL POWER Vs RPM for specific wind  ,so for each wind we have one RPM and one Value of Power that is best for power extraction (so for many winds we have a table). It can be rewritten as OPTIMAL POWER Vs VoltageFromTurbine, OPTIMALPOWER = Vbattery*Ioptimal , so it can be also rewritten as Ioptimal Vs InputVoltageFromTurbine, that’s fits the Wind Curve the classic have.

So my conclusion is that the MPPT controller in the Classic works with Optimal Power Tracking by method 3, and each value in the Wind Curve fits the Current Output Voltage from the Turbine (RPM) Vs Amperage (Optimal Power/BattteryVoltage)
These is my theory, but is it correct...?

[boB]

Yes, you are correct.  Method number 3 is the one and is exactly why the voltage vs. current (power) curve
is used in the Classic.

It really should be battery (or turbine)  "power"  on the one axis, but was originally done with current.
The disadvantage of using current rather than power is that you have to modify the curve for different
battery voltages.

Other than this, it works very well.  The hard part is finding the proper data points to make a good curve.
If we had a wind tunnel for every turbine, this could be extremely fast and easy to do.

One of these days, we want to put "wind learning" into this so over time, the controller can optimize
the power curve.

boB

[romekgm]

thank you boB for the help
I hope you'll "wind learning" algorithm will be implemented , as it's a remarkable and very needed feature
all you need is one Engineer with some "machine learning" knowledge and you'll upgrade your product to a new level.
These is also can help with difficulty you mentioned achieving Wind Turbine Curve for Turbines that were not tested in a Wind Tunnel
If you know the profile of the rotor , so you know the CL and CD Vs angle of attack of these profile , you than can calculate the Torque and Force acting on the blade for specific Wind, and from there, the  Power extracted from the Wind . These in not as accurate as Wind Tunnel Calculation but gives you some Theoretical  Power Curve . Than if you have the "wind learning" algorithm it staring curve will be the "Theoretical Curve" and the algorithm will  correct all mistakes in the theoretical model or changes in blade dynamics due to time\use.

[romekgm]

Hi, after understanding the MPPT mechanism, there is some logical conflict that I can't comprehend, I hope someone can explain it to me, electrical power extraction is not intuitive for me at all :-(
Let’s assume that the output of the Midnite is connected only to a battery (no dump load, relays, controllers) .For one hand in MPPT control we are trying to pull power from the turbine to achieve optimal RPM (for optimal Power extraction) but from the other hand, we need to push into the battery, current that is suitable to the battery charging state and its health.
Let me describe a problematic example:
Let’s say that we have a Wind Turbine that works in optimal RPM and optimal Load , currently no adjustment needed(power extraction in  its peak) ,the Turbine  is rotating in 500RPM, under a Torque of let’s say , 20N*m, that’s mean that the power that available for extraction is ,by the eq:  Power = Torque*AngularVelocity = (500*2*pi/60)*20=52.33[rad/sec]*20 = 1047Watt ,and we have a Load , a 48 battery , that’s absorbs these power of 1047W. That’s mean that’s our Midnite Classic is transforming the 1047W to 48V in 21.8A as Power=Voltage*Current (in reality I think the charging voltage is bigger, but it’s not important for the example). All is good, but our Battery is slowly charging, and its absorption rate is decreasing (battery resistance Is climbing), and it cannot absorb these 21.8A in these Voltage, and here is the problem:  If we continue to extract Optimal Power of 1047W from the turbine, we need to continue to provide 21.8A to the batteries, maybe by increasing the voltage(does that what happens?) , or we need to lower the current , but if we do so , we extract less power , the Turbine will accelerate, and then the  MPPT will kick in ,and will try to increase the load ,but it can’t ,the battery can’t accept that current , so we lose the Optimal Power extraction as soon as the current cannot be absorbed by the battery…

From my very long example, two questions arise:
1.   MPPT works only in charge stages where maximum current from turbine cannot exceeds the absorption rate of the battery? If so, that’s mean, that when selecting batteries, its best co select one that has a very high current absorption rate?
2.   If the battery charge current depends only in what Power we need to extract, then optimal Battery charging algorithms (for health and long endurance of the battery) cannot and are not implemented here?
Thank you for the patience of reading these very long post…
Roman

P.S
I atached a graph, of charging current and voltage, in Lithium Battery

[dgd]

Let me describe a problematic example:
...
and here is the problem:
  If we continue to extract Optimal Power of 1047W from the turbine, we need to continue to provide 21.8A to the batteries, maybe by increasing the voltage(does that what happens?) , or we need to lower the current , but if we do so , we extract less power , the Turbine will accelerate, and then the  MPPT will kick in ,and will try to increase the load ,but it can’t ,the battery can’t accept that current , so we lose the Optimal Power extraction as soon as the current cannot be absorbed by the battery…

And this is what the Clipper is for. As the Classic gets the battery bank full and less charging current is needed then the Classic will use one of its AUX controls to communicate with the Clipper which PWMs load resistors into the turbines AC outputs.
This allows the turbine to keep working or start to slow doen under the resistor load and most importantly prevent the turbine from overspeeding.

dgd

[romekgm]

thank you dgd for the quick reply
but there are system that do dont use Clipper , but use a battery controller, in my case TrisStar Controller, with a dump load
The controller checks if the battery is full ,and if so , opens a circuit to a Dump Load , it prevents overcharging but is not part of the MPPT logic .
I atached a sketch of the system , so you can see what I mean

[dgd]

I understand what you are doing now - from your diagram.
If the DC output from the turbine control was connected directly to the battery then the turbine would continue trying to send as much current as possible to the battery. The dump load and TS60 would prevent battery overcharging
This type of dump load cannot directly control the turbine.
When you introduce the Classic 250 between the control box and battery then the Classic will work using mppt to maximise current to the battery and as battery gets full it will simply reduce the current.
The TS60 and dump load will not be used as the Classic now controls charging current.
So you need to use the Classic to control a dump load  on the output from the turbine before it gets input to the Classic.
The resistors can either be in the three AC lines from the turbine or after the rectifier and in the DC line that goes to the Classic.
This is what the AC and DC Clippers do, along with a bunch of other things to protect your turbine.

So to solve your problem you either need a Clipper to work with the Classic or design some other before the Classic dump load system to use the energy from the turbine - or turn the turbine off.

dgd

[boB]

The full 3 stage charging works with MPPT.  When the battery voltage goes to the voltage
regulation stage,  (Absorb, Float or EQ), the input (turbine) voltage will rise to keep the
battery voltage constant.  That is when the battery can not take the power without
going over the voltage set point OR the controller has reached its current limit.

A battery side dump load can help keep the battery from going over-voltage but if
the wind and turbine is strong, the controller may current limit and still raise its
input voltage.  At this point, the turbine has less load on it and will tend to
spin faster RPMs.  If the turbine can limit its RPMs and thus its output voltage,
then all is well as long as that output voltage is below the maximum voltage input
limits of the MPPT controller.

Bergey assumes that their XL.1 48 is going to limit its RPMs in high winds to keep
the voltage at the input of the Classic below 250 volts.  This works most of the
time but NOT always.  That's when they run into trouble.

A clipper can not only act as a fail-safe for many different turbines but can also
be commanded to keep the turbine loaded down just enough where it is offering
enough power to help keep the system at the constant voltage regulation that is
needed for any particular charge stage the battery is in at that time (Absorb, Float, EQ).

This commanded clipping can also help reduce noise from the turbine as well as
bearing wear because the RPMs can be kept lower.

The Clippers also have their own fail-safe voltage limit set point that is usually set
higher and out of the way of the Classic's clip voltage command set point.
This fail safe voltage is usually set for a voltage that is just under the maximum input
voltage rating of the controller and can help out if the Classic should for some reason
stop commanding the clipper to clip. 

The Clipper also has its own internal power supply powered from the turbine that
can run the clipper circuitry as well as cooling fans for the dump load resistors if
they should get that hot.

[romekgm]

Hi
Bob , I have a problem with your first paragraph:

“When the battery voltage goes to the voltage
regulation stage,  (Absorb, Float or EQ), the input (turbine) voltage will rise to keep the
battery voltage constant”

In Absorb stage, the voltage of charger (Midnite Controller), will stay fixed (will not rise! that’s the definition of absorb stage), while the current will decrease slowly, and the battery voltage will increase slowly to Absorb Voltage.             CurrentFromMidnite = (VoltageMidniteOut â€" VoltageBattery) / BatteryResistance. In these case, the power that the system draws from the Turbine is: Voltage from Midnite * Current from Midnite, it will not be equall to Power from turbine , so MPPT can’t exist here and the Turbine will accelerate or decelerate.

I found in the Midnite manual, the next sentence:

Absorb â€" Constant voltage charge stage to fill the batteries. The controller is regulating so maximum power will not be seen at this time. The Absorb timer is also counting down to the switch to Float

“The full 3 stage charging works with MPPT,  (Absorb, Float or EQ)”
I don’t understand how it possible for the Midnite Controller to control two different things , one is the current it needs for load control(MPPT) and the second is the current needed for absorb stage, those are two different things, that require different current .

I’m sorry for my repeated questions, I probably miss something basic, but just can’t understand where I go wrong…

[boB]

Hi
Bob , I have a problem with your first paragraph:

“When the battery voltage goes to the voltage
regulation stage,  (Absorb, Float or EQ), the input (turbine) voltage will rise to keep the
battery voltage constant”

In Absorb stage, the voltage of charger (Midnite Controller), will stay fixed (will not rise! that’s the definition of absorb stage), while the current will decrease slowly, and the battery voltage will increase slowly to Absorb Voltage.   

If you look a bit closer you'll see that's what I said.  The input voltage is what changes (the turbine),
not the battery voltage.   The input voltage is raised by the controller to keep the battery voltage
constant OR the output current below maximum current limit.
The closer the controller's input voltage is to the energy source's open circuit voltage (at any instant),
the less power is taken from the source (turbine or solar etc)

CurrentFromMidnite = (VoltageMidniteOut â€" VoltageBattery) / BatteryResistance. In these case, the power that the system draws from the Turbine is: Voltage from Midnite * Current from Midnite, it will not be equall to Power from turbine , so MPPT can’t exist here and the Turbine will accelerate or decelerate.

I found in the Midnite manual, the next sentence:

Absorb â€" Constant voltage charge stage to fill the batteries. The controller is regulating so maximum power will not be seen at this time. The Absorb timer is also counting down to the switch to Float

“The full 3 stage charging works with MPPT,  (Absorb, Float or EQ)”
I don’t understand how it possible for the Midnite Controller to control two different things , one is the current it needs for load control(MPPT) and the second is the current needed for absorb stage, those are two different things, that require different current .

The power (and current) necessary to keep the battery right at the Absorb, Float or EQ voltage is
going to be either right at the maximum the energy source can provide at any one time or the
power (and current) necessary to keep the battery at that Absorb, Float or EQ voltage will
be LESS than what is necessary (the controller will NOT be in MPPT mode at that time).

The current into the battery PLUS the other DC loads will draw whatever they need AT that voltage.
The charge controller can not dictate what current they require at any particular voltage.  The controller
can only try best it can to keep the voltage at the regulation set point, otherwise it has to go back
to MPPT to keep trying the best it can....

I’m sorry for my repeated questions, I probably miss something basic, but just can’t understand where I go wrong…

No problem at all.  That's what we're here for.

[romekgm]

thanks boB ,
now it makes sense