PWM

Oh come on... don't be pedantic. Consider the context of the question. We all know what PWM stands for when discussing a piece of electrical equipment.

If you are writing about automotive alternators or similar, you may prefer to think of these as current controlled current sources. The field winding that modulates (controls) the output is a very low impedance, and the output is a current that is proportional ( e.g., 100 times input) to the input. The output voltage is limited by the load and the rotational speed of the armature (field). Pulse Width Modulation can be used, but without filtering you will get pulses of current, pulses of torque, and lower energy efficiencies.

My recommendation would be to rectify the outputs then PWM the unregulated DC to the voltage level you need (or current for battery charging applications). The rectified voltage will be a function of the RPM and the load. PM alternators are typically used on many motorcycle engines and they simply "burn off" the excess energy with shunt style (inefficient) voltage regulators. One thing you will want to watch however, if you go this route, is not to exceed the voltage rating of the stator windings if you are lightly loaded at higher RPM. I doubt this will be a problem but you will want to measure your worst-case scenarios.

Yes it can, lots of VR's (Voltage Regulator) use PWM (IPS comes to mind first) but they do it like Robin said, take a DC and then regulate it to correct current/voltage for the Exciter Stator coil per command from a VR module for the correct voltage. Only disadvantage I've found is if a diode shorts in the exciter ring it usually will fail the PWM module when the Electrician puts a big battery to flash the field trying to get his voltage up/stable and finely produces enough current in the exciter ring to blow the shorted diode open, then the generator becomes unstable after the failed PWM Module is fixed.

A SCR direct AC to DC exciter system (like Baylor) is more robust than a diode AC to DC controlled by a PWM system. Most VR's now are SCR since the precision and speed of a PWM VR controller is lost in the inverse cubic relationship between the time it takes for a field inside a generator to change in response to a change in the VR's output. So a PWM system won't produce any more distortion than a PMG or any other system used to excite a generator, PWM's response speed is nullified by the physics of a magnet field so it's hardly used anymore.

NOW if you want to talk about Speed Control, in relationship to load sharing, you'll find a lot of PWM being used now with electronic fuel injection that can benefit from the extra response speed of a PWM circuit.

Here's a more from an old IPS book on how they use their PWM:

PWM Module

The PWM's job is to excite the generator field to produce 600 VAC. Its input is directly from the Voltage Regulator Module, called the VR Fire which will be a variable signal of about 1.5 to 3 volts DC on a VOM. The Output of the PWM is called field excitation F(+) and F(-). Excitation voltage is approximately 81 VDC. This voltage is a combination of Residual Voltage and the excitation voltage. Residual voltage is approximately 50 VDC.

The AC voltage (from step down transformers T1, and T2 600 to 125, 63, 37 VAC) is converted to DC, at a constant voltage level by a full-wave, three phase rectifying bridge.

This constant DC is switched on and off by the transistors to regulate the average DC reaching the generator exciter. Transistor switching input (gating signal) is provided by the Voltage Regulator Module, the VR Fire.

The PWM Module has two basic sections, a 3-phase rectifier to produce a DC voltage for field excitation, and switching power transistors that control the average DC current output to the generator field winding. Insert PWM Module Drawing.