So, What's anyone doing to comply with Rapid Shutdown?

[inetdog]

The rapid shutdown occurs over 10 sec. The additional heat transfer offered by fins over that time scale would be insignificant.

A possibly error-prone stab at this:
A 10 HP VFD could have 1000 microfarad of capacitors in the DC bank. 10 HP = ~7500 W, a large-ish PV inverter size. Energy = 0.5 * C * V^2, = 0.5 * (1000 * 10^-6) * 500 ^2 = 125 J.

125 J dissipated over the 10 sec allowed by rapid shutdown is 12.5 W... so if this analysis is in the ballpark, the power rating of the resistor isn't too hard to meet.

Remember that you could be looking at the optimizer side, not the inverter side, so the power will be more like 300W maximum. And the ripple filtering does not have to be as good as with a typical VFD. The VFD has to store power delivered in the form of 120 or 360Hz pulses regardless of what the working frequency of its PWM output stage is.

The combined effect is that the bleeder resistor power would be maybe one hundredth of what you calculate.
If the capacitor discharge has to happen at the inverter instead, you still do not need anywhere near the energy storage capacity that you calculated.

[sensij]

I see from the installation diagrams that my assumption that the bleed resistor circuit is only closed when shutdown is enabled is wrong. So, if power dissipation requirements were high enough, fins would help. It sounds like my stab at power dissipated is wildly high. I suppose that someone who has one installed could feel how hot the resistor is getting. Based on what inetdog has posted, it might be slightly warm, but probably not hot.

[J.P.M.]

The shrink wrap will make it harder to dissipate the heat when the resistor becomes the "dump load". If the wrap is not rated properly it could degrade depending on how many times the load is dumped and what the ambient temp is in the box.

I may be worried about nothing because the "load" being dumped is small and does not produce much heat.

I just know what dump load resistors look like and they have a lot of air around them to get rid of the heat quickly before it breaks down wire insulation.

What are some of the dimensions of the resistor likely to be ? Weight ? Specific heat ? If it's wrapped up, weighs a few grams and has a low sp. ht., ?? Just sayin'.

[sensij]

It turns out that my guess above might not have been quite so bad. I just opened up my 3 kW SolarEdge inverter and found 3 x 1000 uF capacitors, labeled for use at 500 V. 0.5 * 3000 * 10^-6 * (500)^2 = 375 J, and dissipated over 10 sec is 37.5 W.

[SunEagle]

It turns out that my guess above might not have been quite so bad. I just opened up my 3 kW SolarEdge inverter and found 3 x 1000 uF capacitors, labeled for use at 500 V. 0.5 * 3000 * 10^-6 * (500)^2 = 375 J, and dissipated over 10 sec is 37.5 W. Based on the discussion in this thread, that suggests that the rapid shutdown resistor is sized to dissipate a steady 37.5 W, which would otherwise have been available for conversion to AC.

I am still trying to understand how the "wiring and those resistors" can reduce the voltage within 10 feet of the array below the 2014 NEC requirement.

I understand if you had micro inverters at each panel they could be shut off with no output. I can also see how the resistors will bleed off the energy in the caps but how does the SolarEdge inverter tell those optimizers to lower their voltage during sunlit hours?

[sensij]

I am still trying to understand how the "wiring and those resistors" can reduce the voltage within 10 feet of the array below the 2014 NEC requirement.

I understand if you had micro inverters at each panel they could be shut off with no output. I can also see how the resistors will bleed off the energy in the caps but how does the SolarEdge inverter tell those optimizers to lower their voltage during sunlit hours?

That is a default feature of the SolarEdge system. Under any fault condition, each optimizer will reduce its output to 1 Vdc... the "rapid shutdown kit" (aka bleed resistor) is not required. The maximum string length is 25 panels, so as long as the optimizers respond quick enough, the < 30 Vdc criterion of the rapid shutdown requirement is met.

I don't know how the fault condition is communicated. My guess is that when there is no fault condition, there is a steady signal generated by the inverter at some frequency that indicates "safe". It could also be a periodic burst, some kind of heartbeat. If that signal is missing or disrupted, the optimizer knows to drop to 1 V... it doesn't need an explicit command to do so. Really, this is just speculation, but since we know that SolarEdge uses kHz+ signaling on top of the DC to carry out slower status communication, something like this would be consistent with the architecture.

What the optimizer can't do, I suspect, is lower the DC voltage in the short lines between each solar panel and its optimizer. Those would probably go to Voc or something close to it, but I think that is still compliant with the rapid shutdown requirements, since it is within 10 ft of the array.

In my own installation, since the inverter itself is installed within 10 ft of the array, I think it would be compliant with 2014 NEC even without the bleed resistor.

[sensij]

My guess doesn't seem to be too far off. Here is a link to the patent for their optimizer shutdown technology. It includes the following:

During normal operation, a line communications transmitter 503 superimposes a keep-alive signal, for instance between 1 kilohertz to 100 Megahertz on direct current (DC) input lines 410 and 412 attached to serial string 523.

The patent is fairly non-specific and also talks about other ways to trigger the optimizer shutdown... rapid current change being one.

[SunEagle]

My guess doesn't seem to be too far off. Here is a link to the patent for their optimizer shutdown technology. It includes the following:



The patent is fairly non-specific and also talks about other ways to trigger the optimizer shutdown... rapid current change being one.

That does clear up my questions. Thanks.