PWM converter effects in a generator 2

[flkris]

All,

I've been reading IEC standard 60034-18-41 on PD testing type I insulation machines fed by voltage converters. It talks about all the hazards (for the electrical machine) coming with the utilization of these drives (PWM converters), and how there may be voltage overshoots at the terminals of the machine and this will stress the insulation.

The thing is: for everything (not just this standard, also NEMA std and lots of papers) I've read, it seems like they are only speaking about MOTORS as the "machines", for the PWM converter will be connected to the motor side. Instead, for a generator, as I understand this is the configuration:

(wind) generator - rectifier - inverter (PWM) - line filter - grid.

So my question is: in this case will there be voltage overshoots and all the trouble that's specified for motors, mainly variable speed induction motors? Is it right to apply the same test procedures as described on IEC?



To enter in detail, the generator I'm interested in is a variable speed CDSG (Converter-Driven Synchronous Gen) for wind power application. Forgive my ignorance, I'm new on the subject and the only thing I can find is either about the hazard for motors or the effects on the grid (in the case of generators) due to converter operation.

 

[LionelHutz]

The system you are investigating actually has back to back inverters. I would expect that the generator can see the same effects as a motor.

 

[flkris]

Thanks for answering LionelHutz, but what do you mean by back to back inverters? The only thing I was told about the type of converter was that: Active switching IGBT (AC-DC-AC), Leading PF control f: 2-4kHz, dv/dt < 1200V/µs, 1 converter/system.

So should I believe this dv/dt value may occur at the generator terminals? Even if so, IEC 60034-18-41 only defines the test procedures according to the "stress category"* of the insulation system, so I guess I should contact the Converter Drive manufacturer in order to discover it, is that right?
*Notice that the stress category is defined based on the overshoot factor (peak voltage by the DC bus voltage) and the impulse rise time.

If someone experienced in this matter could explain me why standards like NEMA only speaks about voltage overshoots (due to the converter) and dv/dt for inverter-fed MOTORS, I would be really grateful. Also, I must say I understand the definition of dv/dt according to NEMA (MG 1), but I realized this is not mentioned by IEC (they only define 'tr', the rise time). I mean, isn't it all a mess?!

 

[LiteYear]

Back to back inverters means that there is a PWM converter on both the input and the output. The alternative is to only have one on the output and use a bridge rectifier on the input. From all you've provided you could still be going either way - back-to-back inverter or rectifier and inverter. In the back-to-back version there switching elements connected to the generator, while in the rectifier-inverter version the switching elements are effectively isolated from the generator by the DC bus.

Now the reason you're seeing motors mentioned everywhere is simply because they're established and common. Generators and power electronics are coming on the back of those developments. The concepts are very similar, you just need to draw out the components.

For example, if you really do have a rectifier on the generator side, then there are no switching elements on the generator side - all the switching will be effectively isolated by the DC bus between the rectifier and inverter. There'll be harmonics and power quality issues on the generator side, but not the dv/dt or voltage doubling issues you're considering. If you do have an inverter on the generator side, then all the motor issues apply. There's be operational differences because the predominate power flow in the other direction, but I think you probably don't need to concern yourself with those at this stage.

dv/dt and rise time are two ways of describing the same thing. dv/dt means the rate of voltage increase per unit time, and rise time is the time taken per unit voltage increase. They're measuring the same thing. It's only a mess because it's new to you - it always seems this way at first!

 

[flkris]

LiteYear: thanks a lot. All of this is important in order to evaluate the necessity of PD testing the generator, and for that I wanted to follow IEC standard 60034-18-41, which is meant for "converter-fed machines" - so I conclude I can only consider my generator as a converter-fed machine for this practical purpose when I do have switching elements on its side. That was exactly my doubt, and you solved it, thank you.

If it is not too much to ask, do you know where can I find material and literature concerning the operational differences you mentioned in the case there's an inverter on the generator side?


Still, about the rise time, of course you are right about new things that usually only seems messy at first, but I believe it is not just that in this case. The way I've read the definition of rise time, it is not the time taken per unit voltage increase, but the "time for the voltage impulse to go from 0% to 90% of its final value". It is closely related to the dv/dt definition, but they don't describe the same thing, in a way that when I have one, I don't necessarily have the other (so I could have, for example, dv/dt = 1000 volts/microssecond and a t_r = 2.5 microsseconds and a peak voltage of about 3125 V, considering dv/dt ~ 0.8*Vp/t_r). Also, the reference value of the voltage for the t_r definition is different in NEMA and IEC, the former being the steady-state value (= DC bus voltage) and for IEC it is the actual peak voltage, so it will be bigger.

 

[LionelHutz]

The generator would have a rectifier if it runs at a fairly constant speed. In that case, it'd be more efficient to just directly line connect the generator and get rid of all the active electronics. This is why I expect the generator to have a converter on it. That way, the generator can track the maximum power point of the turbine as the wind speed varies.

The committie working on NEMA MG1 decided to write the section for motors. I have had no involvement so I can't tell you why.

You do know that a motor can become a generator by injecting power into the shaft?

 

[flkris]

Yes, I do, LionelHutz. What I understood from "operational differences" mentioned by LiteYear was that there may be something different about the voltage overshoots and pulse rise time to be considered, other than being exactly the case for motors. If the changing in the power flow direction has nothing to do with it, then why bother thinking about it? (I mean in order to investigate the effects of the inverter in the machine).

Thanks for the explanation about using rectifier or back to back inverters.

 

[LionelHutz]

The direction of power flow doesn't have any effect on the voltage overshoots and rise times. The switching of the IGBT modules combined with the cable and generator characteristics basically determine those.

The converter is being used for a very specific application. I would believe the generator and wiring and converter would be standardized for a specific wind turbine so it's possible the system was designed to limit the overshoots and/or rise times.

 

[LiteYear]

Yep, the voltage overshoots and rise times are dependant on the switching of the IGBTs. Further, the intended direction of power flow dictates the switching of the IGBTs. So indirectly the voltage overshoots and rise times are dependant on the direction of power flow.

Unfortunately the "operational differences" have a lot to do with the particular strategy employed by the manufacturer. The reasons for active switching instead of diode rectification are complex and subject to optimisation strategies that differ in each application. As a conservative guess however, I think it would be safe to assume the same worst case scenario as the drive/motor standard specifies.