Solar Inverter minimum load requirements (Powerwall frequency shifting)?

[woferry]

So I debated posting this on the original thread, but figured it might be too much of a derailment so I'd just start a new one. Given some recent discussions around things that might be impacted by the Powerwall's tinkering with the line frequency when the grid is down, I'm still trying to wrap my head around the WHY?

From the Powerwall lead time thread:

It's actually more than just that. When the grid is down, there is nowhere for the power from the array to go except to the house or the batteries. If the batteries can't consume what the house isn't using, there's a problem. This is why the Powerwalls have to shut the array down if they aren't at an SoC where they can accept all the extra power.

Now this is one thing I haven't understood, though I'd wonder if anyone on this forum could really explain it. I'm an electrical engineer, but my focus has always been on DC, I'll admit that AC circuits have a bit of 'black magic' associated with them in my mind, so there may be something I just don't understand here, but it certainly seems counter-intuitive to me. I'll ask some of my EE friends who have more background in AC systems as well.

In a DC system, there's absolutely no issue with a supply being larger than the loads it powers (we do it all the time, it's called having margin). It's true even for the bits of these Solar/Powerwall systems. Take the solar panels (DC). You can unplug the MC4 connector or turn off the disconnect switch at the inverter and the panels will be fine. Similarly you can size the panels for more power than the inverter will ever use, and you'll get clipping and some lost power that you could have gotten with a bigger inverter, but that doesn't cause any harm to the panels (in fact many people recommend this).

And then there's the Powerwall battery (DC). Again, it can provide much more instantaneous power than its 5kW inverter supports, that's not a problem for the battery (in fact higher currents do more to shorten the battery's life), it just takes longer to drain it because the inverter is the limiting factor.

And even the AC-side of the Powerwall seems to be fine with this, i.e. the Powerwall's built-in inverter can output 5kW continuous, but nothing requires that you have at least 5kW of load on the circuit for it to work, in fact if you have more than 5kW (per PW) it will likely not work because you've exceeded its output and it will probably be shutting down to protect itself. If there is a minimum load requirement for the Powerwall I haven't heard it, and I should probably be concerned as my household loads can get pretty low (<300W on a typical day), so if the Powerwall would have trouble powering my house during an outage that would kinda defeat the purpose of having it.

So why is the solar inverter any different? What is the problem with drawing less current than the inverter is capable of providing at any given moment? Is the solar inverter that poorly-designed compared to the PW's inverter, or is all this just FUD and there's no real concern with there being more solar production capability than load to consume it? I'd sure hope that Tesla isn't playing these tricks without good reason (just trying to understand what that reason really is), especially since it seems there could be risks associated with pushing the frequency as high as Powerwalls apparently can/do (temporary functionality of things like UPS and PLC devices, and I'd wonder if perhaps even safety-style risks for large/small appliances with AC motors that could end up running well above their intended/tested speeds), so if it wasn't necessary it would be better to not do it at all.

I will note that reviewing the manual/specs for my Solivia 5.2, I don't see anything indicating any minimum load requirements. I also discovered while trying to research this that gas generators do have minimum load requirements, but this appears to be more of a mechanical thing, that the engines can 'drool' or get carbon build-up when operating at light loads (they're not damaged per se, they just may require more maintenance or occasional operation at higher loads to recover), so it seems to me that this has more to do with the mechanical method they use to produce the energy, where a solar inverter should be non-mechanical DC to AC conversion, more like the Powerwall's inverter, a home UPS, or the inverters that provide AC outlets in cars and the like, none of which seem to have the same requirement/concern.

Anyone here know the real answer? Thanks!

 

[wwhitney]

Here's my somewhat incomplete understanding:

A standalone inverter (like the Powerwall in backup mode) is designed to operate as a voltage source. It can provide whatever current is required by the connected loads, up to its rating.

The I-V curve of a PV cell is more like a current source, and a grid-tie inverter acts like a current source. When synced with an AC voltage source, the grid-tie inverter raises its output voltage above that of the reference voltage as required to push out the available power (current).

So when you have a mini-grid with both a voltage source and a current source, you have a minimum power that must be dissipated.

Hope this helps.

Cheers, Wayne

 

[power.saver]

Wayne is correct, a solar inverter is a current source, and pushes out as much current as it can based on the solar DC input. Normally, this is not a problem when connected to the grid, as the grid is an infinite sink. But when islanded to a Powerwall, that infinite sink turn into a finite sink that rapidly decreases in ability to absorb the current when the batteries get full.

Newer off-grid inverters have the ability to curtail their output proportionally using a technique called Frequency Shift Power Control (FSPC) whereby the battery signals to the inverter to reduce output by increasing the frequency. This is what the Powerwall actually does, and in my testing, it ramps up the frequency correctly. But since my inverters are not FPSC capable, they shutdown at a fixed over frequency condition. The result is the same, solar production is stopped when the battery has no more ability to accept power.

 

[miimura]

To say it another way, I think it comes down to regulation. Not law type regulation but what devices are designed to and responsible for maintaining the system within the design parameters.

A battery powered off-grid inverter is designed to maintain frequency and voltage according to its own internal regulation. If the battery voltage drops too far, the voltage might sag. If the load is too high, it might exceed the instantaneous or average power rating and it will shut down to avoid damage due to exceeding the current carrying capacity or heat dissipation limits of internal components. Light loads may be inefficient, but they will be stable.

A solar grid-tied inverter is primarily designed to push as much power as it can from its panels into the grid. It uses Maximum Power Point Tracking to balance the voltage and current from the panels to optimize the panel production.

When the Powerwall is in grid-tied mode, it is a programmable current source & sink. The Gateway computer is taking in data from the Neurio Current Transformers and voltage monitor so it knows the power flows of the grid and solar. It uses that data and its operating modes and strategies to calculate how much the Powerwall(s) should be charging or discharging into the home and therefore the connected grid.

When the Powerwall is in backup mode, it is just like the off-grid inverter described above. It is the "master inverter" and it is trying to maintain the voltage and frequency of all the connected loads and generating sources. Each Powerwall can usually charge or discharge up to 5kW. Problems arise when loads demand too much current for doing things like starting an AC compressor. Since solar grid-tied inverters are designed to push all the available power into the grid, the Powerwall must absorb the excess by charging the batteries. If the batteries are too cold or too full, or the power is more than the 5kW the inverter can handle, it must try to curtail the solar. The only mechanism that exists today is to change the grid parameters of voltage and frequency to make the solar inverter curtail its generation. If you were always operating a micro-grid, you could configure the solar inverters to be more friendly and curtail generation proportionally as @power.saver says above (FSPC). However, the grid actually wants the solar inverters to ride through these small variations, so I think there is some conflict in choosing a configuration that is appropriate to both on-grid and off-gird operations. If there was some out-of-band signalling between the Powerwalls and the solar inverters to tell them whether they are on-grid or off-grid, the interoperability would probably be greatly improved.