 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Windmills : An inconvenient truth about PMA’s
| Author | Message | ||||
DaveP68 Senior Member  Joined: 25/11/2014 Location: New ZealandPosts: 292 |
Having completed testing of various F&P stators in different modes to determine operating efficiency some clear "facts" emerged relating to PMA's. It is that current flow is directly linked to how much torque is required to drive a PMA if the voltage is not able to rise proportionally with increasing RPM. If a short circuit is applied to the output of a PMA then at a low RPM maximum current with flow with very high torque required turn it. With no load connected to the output, the open circuit voltage rises in proportion to RPM with not much torque being required to drive the PMA even at a high RPM. My point here is when a battery is connected to the output of the rectifier this clamps the voltage rise with any increase in RPM above battery terminal voltage "ONLY" the current can increase above this point. This results in a "Decrease" of operating efficiency of the PMA. It can also so have a further detrimental affect to the blade TSR due to requiring a higher level of torque to drive the PMA. The "Power Coefficient" will drop and this results in an even greater loss of "overall" system efficiency. Even a "correctly matched" resistive load on the output of a PMA will keep it an efficient mode over a wide RPM range. MPPT correctly setup make the PMA operate in it's most efficient mode over an even greater RPM range than a "correctly matched" resistive load. The other big benefit of MPPT on a wind turbine is it will keep the Blade TSR at it's optimum "Power Coefficient" providing maximum "accumulated watt hours" of power production... I now put the challenge out there to prove this is not the case with other PMA's. There are realities if you do not accept, will lead to frustration because you will be spending time on wrong assumptions and the results cannot follow! The Dunning Kruger Effect :) |
||||
Madness Guru  Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
The is plenty of information on the net saying the same thing. Solar Mike is working on a DIY MPPT charge controller, this could also be adapted to a wind generator or micro hydro. By sensing the RPM by monitoring the frequency before the rectifier you can preprogram the load based on RPM. It may be possible to incorporate self-learning, what would need to be incorporated is a dump load to keep it from over-revving as battery charge gets near 100%. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| DaveP68 Senior Member Joined: 25/11/2014 Location: New ZealandPosts: 292 |
The main point I'm getting across here relates to comment's posted on this site that there's not much difference in accumulated power output between MPPT (including GTI) and charging batteries directly. That statement is misleading (IMO false) if no actual real world results (data) are published to back it up. Nice work by Solar Mike and thanks for pointing in that direction as starting to take a lot of interest in MPPT systems as you may have already picked up. There are realities if you do not accept, will lead to frustration because you will be spending time on wrong assumptions and the results cannot follow! The Dunning Kruger Effect :) |
||||
SparWeb Senior Member Joined: 17/04/2008 Location: CanadaPosts: 196 |
Dave, I'm sure you will really enjoy learning about DIY MPPT controls like solar Mike's doing. This is a domain where active electronic control offers a big advantage over passive electronics, or reliance on mechanical power-curve matching alone. Steven T. Fahey |
||||
| DaveP68 Senior Member Joined: 25/11/2014 Location: New ZealandPosts: 292 |
Hi Steven Yes looking forward to having a simple version of DIY MPPT working sometime this year. The MPPT control part is mostly finished, but have never very build an inverter before so more or less starting from scratch there. David There are realities if you do not accept, will lead to frustration because you will be spending time on wrong assumptions and the results cannot follow! The Dunning Kruger Effect :) |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
All very true Dave, but there are some fairly significant problems to overcome with mppt as applied to wind power. The first is that any feedback system has some delay involved. The output must be "wrong" before any correction is applied. And once you change the loading on the wind turbine, the turbine rpm changes. The correction needs to be applied VERY SLOWLY to allow the system to stabilize, otherwise it can start creating some wild swings by overcorrection. Now if wind speed is steady and constant, an mppt system can very slowly steer the system onto its most efficient peak. But if the wind is turbulent or gusting, good luck trying to tune an mppt FEEDBACK system. The solution to this is feed forward correction. Now everyone knows what feedback is, but feed forward correction is a new concept to some people. The idea of feed forward correction of a system, is to measure all the inputs, and decide from a pre programmed table, how the system must be varied to create the desired output. As a wind turbine has only one input, wind speed, feed forward correction is very easy. You only need to know the instantaneous wind speed from a fast acting anemometer, to set the pwm value of your controller to what it needs to be optimum. If the wind is gusting madly, the applied correction will be pretty much instantaneous, it does not wait for rotor rpm to readjust or the output to change. If the wind speed suddenly drops, the loading is quickly reduced. If a gust arrives, you can massively increase the loading just for the duration of the gust. The feed forward system needs a lookup table, so that for every individual measured wind speed there is a stored value of pwm duty cycle. This has to be initially programmed into the lookup table, at least approximately. A slow self learning algorithm may modify that, so it can self learn what it needs to do at each steady stable wind speed. But a simple system without self learning will still work. Automotive engine management systems are all feed forward systems. They measure engine rpm, engine load, coolant temperature, and ambient air temperature, and from internal tables set the ignition timing and the fuel injector period. Its very complex, because there are so many different input variables. As the engine warms up, and you are accelerating up through the gears, there is simply not enough time for a feedback system to work. All the required variables are stored in lookup tables, and at any instant the engine management system knows exactly what the required ignition and fuel settings should be. A wind turbine is very simple in comparison. Only one input wind speed, and one output, controller pwm duty cycle. The exact relationship will follow some kind of curve, maybe a complex curve. But its still just a curve that can be programmed into a two axis lookup table. Something really crude might work, like ten potentiometers with ten knobs, set up to control pwm duty cycle for every measured wind speed from 0 metres per second to 10 metres per second. A silly example maybe, but it should still work in principle. You go out and see what the wind speed is, oh its three metres per second right now, so you tweak knob three for maximum battery charging amps. If it drops to two metres per second you go out and tweak knob two for maximum. After a while you will have all your knob settings and then you are set ! Something a bit more sophisticated would work much better, but you get the general idea of how a feed forward system can be initially programmed. And once a feed forward system is set up, the big feature is IT CAN NEVER BECOME UNSTABLE in very rapidly changing conditions. The output my not be exactly right, but it is not going to start violently surging or over correcting the way a feedback system can. Cheers, Tony. |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Commercial MPPT grid tie inverters sense the RPM of the wind generator and changes settings accordingly, this overcomes the lag issue. Also the Midnite Classic charge controllers do the same charging DC to DC however they don't sense the RPM, but they do change the algorithm very quickly as the input voltage increases/decreases. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| DaveP68 Senior Member Joined: 25/11/2014 Location: New ZealandPosts: 292 |
Hi Tony Thanks for the question on MPPT. In the system I'm working on it only requires one parameter to be measured. That is instantaneous voltage from the F&P stator (PMA) proportional to RPM. How I'm able to calculate everything from that voltage is via a complex mathematical algorithm that takes into account 6 variables too keep the blade TSR in it's sweet spot. It is real time and can respond much faster than any change in voltage due to any small or sudden changes in wind speed. I talking over a 1000 samples a second! I've been working on this for the last year. That is why I recently set up a fully programmable dynamometer (enhanced VFD). On a slightly separate note, another project I have in mind relates to what is explained in this video A 50kW SiC Three-phase AC-DC Converter Design Can anyone guess what the application I'm thinking of for a Bi-directional AC-DC/DC-AC converter is? It doesn't relate to the high temperature operation. There are realities if you do not accept, will lead to frustration because you will be spending time on wrong assumptions and the results cannot follow! The Dunning Kruger Effect :) |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
[quote]Thanks for the question on MPPT. In the system I'm working on it only requires one parameter to be measured. That is instantaneous voltage from the F&P stator (PMA) proportional to RPM.[/quote] You may find that the instantaneous voltage coming from the stator also varies with the amount of electrical loading. Its not just rpm dependent. The whole idea of feed forward is to find the optimum electrical loading point that will keep the TSR right in its sweet spot for any given wind speed. There is only one input variable, instantaneous wind speed. And one output variable, final electrical power loading on the machine. Its then a case of actual physical testing of the very complex combination of blade geometry and machine, to establish an optimum operating curve. The shape of which will be quite unique for each different combination of parts. Cheers, Tony. |
||||
| DaveP68 Senior Member Joined: 25/11/2014 Location: New ZealandPosts: 292 |
I have taken into account how an increase in electrical load on the stator output decreases the voltage with respect to the RPM. It is a percentage drop from open circuit voltage depending on how much current is drawn. This parentage value decreases for the same current value as the RPM increases. Due to F&P stators being mass produced, these characteristics are very predictable and can be built into calculations. There are realities if you do not accept, will lead to frustration because you will be spending time on wrong assumptions and the results cannot follow! The Dunning Kruger Effect :) |
||||
 Print this page |
