Nyquist does show up here. You need to sample more than twice the frequency to figure out the amplitude and phase. Much more is better - to push the aliasing fold up to where the attenuation of the harmonics times the rolloff of the low-pass filter stomps them down well below the signal you're after. Synchronized with the switcher may also be good - to avoid systematic distortion from a beat between the sampling rate and the clock of the switching waveform making it back through the inductors and caps and biasing your measurements.
That's off the top of my head. Don't have time to examine the thesis for how it's dealt with there.
= = = =
Sampling is done of V and A within the wild AC. Then the instantenious power is calculated. This is then filtered. Wikipeadia told me what a first-order low-pass filter is doing and even gives me an explanation how it can be digitally simulated ...
Which is how you get the measurement. But what you do with it is a tad more complicated.
Adding load to the genny raises the current and lowers the voltage. But the higher current adds load to the blades that slows them. That changes the working TSR, which changes the angle of attack on the blades, which changes the available power, ...
The max power point will be at a particular TSR for pretty much all wind speeds. This means the RMP (and the GENERATED voltage, before resistive drops in the coils and transmission wiring) at max power will go up with the wind speed. But the available power goes with the cube of the wind speed. So the torque (and the current) will go up with the wind speed squared.
Some max power point trackers work by "hunting" for and tracking the max power point. They modulate the load slightly, at a frequency below the response time of the blade speed, measure whether there's more power at the high or low end of the modulation, and use this as feedback to move the average of the load to chase the max power point as the wind speed changes. That works well when the wind is reasonably steady. But it is problematic when the wind is gusty or the mill is furling.
This one (according to the introductin) uses an apriori approach: It tries to keep the relationship between voltage and current right (current proportional to square of GENERATED voltage) to get the the most out of the mill. That means it has tuning parameters (probably one or two of 'em) that form an estimate of the size and impedence of the mill (diameter, TSR, turn count, field strength) and the resistance of the coils and wiring. This works even when things are gusty. But if the tuning parameters aren't set right it tracks at some offset to the actual max power point and you don't get the most out of the mill. Also: If it overcorrects for the wiring resistance you effectively have a negative resistance, so the system will oscillate. So it must be set either to undercorrect or not correct at all, missing the target somewhat.
[ Parent ]
