Paired vs non-paired UL9540 requirement

[zanydroid]

I have seen a few posts here and there saying that some jurisdictions accept non-paired UL9540 systems.

What is the provenance of these claims?

Is there a primary source for this, IE which IRC or NEC revision / code section contains authoritative enabling text allowing the looser version of UL9540? Or broadly accepted interpretation like Mike Holt

I am specifically interested in California but also interested in other states for my general edification

 

[pvgirl]

The 2016 version of UL9540 allows batteries to be listed on their own. The 2020 version of UL 9540 requires a battery and inverter combination to be listed. Requirements for UL9540 are in the IRC, Section R328 Energy Storage Systems, which states the requirement for UL9540 listing, in Chapter 44 Referenced Standards, states which version of UL9540 is referenced by that version of the code. California has amended the IRC to require 9540 Edition 2—2020, thus requiring a listed battery inverter combination. California also has requirements in it's energy code beyond the IRC.

IRC 2021 as published by ICC and adopted by many other states only references the 2016 version of the code. Other localities may their own amendments that may affect the requirements for energy storage systems. Also the listing for the product in question will also state which components are part of the listing, that is if a system is listed as a battery and inverter combination only that combo may be used, only in case of battery being listed on it's own, can it be used with any compatible UL 1741 listed inverter. Also some localities haven't yet adopted IRC 2021, and with out any local amendments, UL9540 is not yet required in these locations.

 

[zanydroid]

. California also has requirements in it's energy code beyond the IRC.

Thanks, so for California it is pretty clear cut. What are the additional California requirements?

IRC 2021 as published by ICC and adopted by many other states only references the 2016 version of the code. Other localities may their own amendments that may affect the requirements for energy storage systems. Also the listing for the product in question will also state which components are part of the listing, that is if a system is listed as a battery and inverter combination only that combo may be used, only in case of battery being listed on it's own, can it be used with any compatible UL 1741 listed inverter.

Is this paired 9540 vs battery only 9540 distinction (2016 standard) unambiguous, or might there be some room for interpretation?

What were the extra requirements on batteries for battery-only listings?

 

[pvgirl]

Thanks, so for California it is pretty clear cut. What are the additional California requirements?

CEC(California Energy Commission) has a list of approved solar equipment, I am not familiar enough with California's regulations to know where or what are the regulations that require only using equipment on this list.

Is this paired 9540 vs battery only 9540 distinction (2016 standard) unambiguous, or might there be some room for interpretation?

What were the extra requirements on batteries for battery-only listings?

The listing certificate, which is commonly available on the website of producers of UL9540 equipment, will clearly state what equipment is covered by the listing, as well as what version of UL9540 it is listed under.

I don't know of the actual requirements for listing under the different versions of UL9540, as these standards aren't publicly available and quite expensive.

IRC 2021 section R328.6 Electrical Installation, covers the requirement for UL1741 listing of inverters or listed combinations.

I've attached 2 listing certificates, one from KiloVault showing a 2020 version listing showing what battery, inverter combinations are listed by this certificate, and one from Fortress showing a 2016 battery only listing.

 

[zanydroid]

How does UL9540 and building code work if you add a SCC, AC charger, or gas generator onto the DC or AC bus of a listed UL9540 ESS? I have not seen a detailed discussion (or any discussion) yet on that.

 

[pvgirl]

How does UL9540 and building code work if you add a SCC, AC charger, or gas generator onto the DC or AC bus of a listed UL9540 ESS? I have not seen a detailed discussion (or any discussion) yet on that.

Under 9540:2020 Ed 2, anything added to DC bus of the ESS, solar charge controllers, chargers, DC generators, would be in scope of the 9540 listing, and to be completely compliant would need to be included in the listing. On the AC side it's a little different as energy going in to storage has to go thru a charger that is part of the ESS and would not be in scope of 9540. As always the final decision is made by the AHJ

The newly released UL 9540 Ed 3, introduces the concept of a DC ESS, which would allow the listing of a module just includes protection(BMS, etc), and storage(battery), and doesn't necessarily limit use to just a select set of products. Although we will have to see if the market produces these products.

We could see server rack batteries, become listed as DC ESS, and be usable with what ever listed components we want in the future. If these DC ESS, also come with proper 9540A testing, the 20kw and 3ft spacing rules in the IRC can be reduced to more reasonable values.

FYI if you register for an account at https://www.shopulstandards.com/ you can get free access to read the UL standards for yourself.

 

[zanydroid]

Thanks. That jives with what I assumed — UL9540 compliance is contradictory to building modular solar setups, and I don’t know if there’s any modular components (SCC, AC charger, inverter) included in a UL9540 listing.

Well, Dcbel and the V2H solutions on the market today probably have the listing for the add-on CCS connectors but that’s not exactly what most people are interested in.

It is good to hear that there is some work to go towards modular systems. Curious what is being added over 1973, I thought that already required protection.

At some point I will have to learn how to read through those standards efficiently.

 

[houseofancients]

Thanks. That jives with what I assumed — UL9540 compliance is contradictory to building modular solar setups, and I don’t know if there’s any modular components (SCC, AC charger, inverter) included in a UL9540 listing.

Well, Dcbel and the V2H solutions on the market today probably have the listing for the add-on CCS connectors but that’s not exactly what most people are interested in.

It is good to hear that there is some work to go towards modular systems. Curious what is being added over 1973, I thought that already required protection.

At some point I will have to learn how to read through those standards efficiently.

Of course it is contradictory, as it was heavily influenced and lobbied by the power-co and us based man. of solar equipment...it is in both their interest to keep installations small and highly expensive

 

[Hedges]

It would make sense for listing to cover cells and whatever electronics controls voltage and current to them.
A BMS can't really do that, except by communication with the inverter/charge controller.

The LG Resu-H battery which was 48V (or so) worth of cells and bidirectional boost converter to 400V would make sense listed on its own.
For a 48V battery to be listed by itself and allow arbitrary sources and loads, I think it would need a DC/DC converter. Having that, multiple batteries could be paralleled, because that would not involve shorting voltage sources together.

Sunny Boy Storage has three HV battery inputs, 10A each, or they can be paralleled for one larger battery. With a similar idea but the DC/DC converter built into battery, we could have HV (or LV) DC bus allowing mix-n-match.

 

[zanydroid]

I’m not sure a power converter should be prescribed. A ton of UL9540 HVDC inverters support a range of input voltages. Having a converter on the other side doubles the hardware cost of this path.

Why not a standard communicating/smart cable interface. Where the smart cable first negotiates the voltage and only closes the power connections once they come to agreement.

It can also communicate the precharge specs.

Btw those smart power cables of another sort are in NEC2023 for high powered applications.

 

[zanydroid]

Of course it is contradictory, as it was heavily influenced and lobbied by the power-co and us based man. of solar equipment...it is in both their interest to keep installations small and highly expensive

When we look back on this in 15 years hopefully they only had this monopolistic model in place for a 5 year transition period.

 

[pvgirl]

It would make sense for listing to cover cells and whatever electronics controls voltage and current to them.
A BMS can't really do that, except by communication with the inverter/charge controller.

In the context of safety, the BMS and other protection systems can detect unsafe current or voltage conditions, and disconnect the battery from the malfunctioning, or mis configured equipment, and remain safe.

 

[zanydroid]

I wonder if there is a relevant SunSpec standard being worked on for this (modular/cross vendor interoperable UL9540 combinations)

 

[Hedges]

I’m not sure a power converter should be prescribed. A ton of UL9540 HVDC inverters support a range of input voltages. Having a converter on the other side doubles the hardware cost of this path.

Rather than doubling, delete DC/DC from inverter and implement DC/DC in battery. It is then told by inverter what voltage to deliver to DC/DC rail.

Just as GT PV inverters have multiple MPPT for multiple PV strings, AC coupled HV battery inverters have multiple battery inputs, which I think are DC/DC converters.

Transplanting that DC/DC to the battery makes it a stand-alone ESS with 100% control over voltage and current of battery. Not just a disconnect mechanism.

Why not a standard communicating/smart cable interface. Where the smart cable first negotiates the voltage and only closes the power connections once they come to agreement.

I think you still need to test the pair of battery and inverter together for UL compliance.

 

[zanydroid]

Rather than doubling, delete DC/DC from inverter and implement DC/DC in battery. It is then told by inverter what voltage to deliver to DC/DC rail.

Just as GT PV inverters have multiple MPPT for multiple PV strings, AC coupled HV battery inverters have multiple battery inputs, which I think are DC/DC converters.

Transplanting that DC/DC to the battery makes it a stand-alone ESS with 100% control over voltage and current of battery. Not just a disconnect mechanism.

But then you need to pay for a bidirectional DC/DC for every battery, there is then minimal hardware savings over say Enphase or PowerWall architecture that are AC coupled with a built-in inverter on every battery (even if not needed for the desired performance level). In fact you can today just buy a lot of AC coupled batteries and not worry about pairing, it's just expensive. In 5 years if there's a lot of companies in the battery space they might just switch to AC coupling to make listing & code compliance simpler, and if that's the way to do things then there's less market to introduce a cross-vendor DC approach.

The alternative is adding things like more current, voltage, temperature, shutoffs what have you to trigger BMS cutoffs. Those are sensing & telemetry types of circuits, not power handling ones, so they should be a lot cheaper.

Probably if you count number of transistors or compare optimizer to microinverter cost (basically the same thing). The DC-DC converter would be about 33% cheaper on the power handling side than the AC coupled battery. While the version with no power converter would be 90% cheaper and close to the current cost of a battery...

There's also the question of whether the requirement here is to aggressively prevent all possible battery damage or just catastrophic damage to the house. Maybe DC-DC or AC coupling would be better at the former while smart disconnect would be sufficient for the latter. And on average the expected solution cost might be better with the latter. EG if I paid $2000 per battery vs $1500 to knock off a 1% extra annual risk of the battery failing in a contained way, but the same risk of catastrophic uncontained failure, is that worth $500.

I think you still need to test the pair of battery and inverter together for UL compliance.

Post #6 says there's a new UL 9540 Ed 3 that allows "unit-level" compliance testing. IMO if the protocols and levels of self-protection are improved they can go back to separate testing.

Unless you think there's a fundamental engineering reason to persist with combined testing.

 

[Hedges]

If the battery can protect itself from fire, that is sufficient.
We know that BMS can't protect against excessive input voltage, like SCC which shorts PV input to battery. Either need specific listing requirements on the SCC/inverter, or high voltage isolation in battery.

DC/DC has to be sufficient for current from battery, and larger than battery's allowed current is wasted money.
The example of Sunny Boy Storage appears to be 3 separate DC/DC converters, which can be paralleled. Although potentially it is 3 separate DC/AC converters.
If DC/AC, you need tight integration when they are used for grid-forming, which includes delivering surge. Multiple DC/DC and single (or single brand) DC/AC would be easier to make work for backup and off-grid.

 

[zanydroid]

We know that BMS can't protect against excessive input voltage, like SCC which shorts PV input to battery. Either need specific listing requirements on the SCC/inverter, or high voltage isolation in battery.

That's where getting a rock solid communications & startup protocol comes in, & including conformance to this in the listing.

With a DC DC converter in the battery you still need to standardize or communicate the maximum voltage that the battery might be connected to.

 

[Hedges]

HV rail should one one-time set and forget. "Source/Sink to keep rail at 400V".
I think a complexity will be that HV rail needs to ripple at 120 Hz or 50 Hz or whatever rectified AC is. Battery will be expected to supply starting surge or absorb load dump for multiple cycles, typically a couple seconds more or less.

I would be nervous having string of cells get voltage current from some other manufacturer's electronics. Could be made to work, and it is now. But a switcher per string of cells seems like a good way to parallel multiple batteries, rather than a DC bus and relying on resistance to balance.

Multiple AC coupled batteries would be another approach. Data link with standardized protocol across multiple vendors. That is easiest for grid-tie, more difficult for grid-forming but should be quite doable. Simulated resistance (voltage sag proportional to current) is a way to share load. Periodic communication in terms of what percentage of available wattage or available watt-hours are being sourced/sunk could avoid having wiring resistance drive the balancing, and could help keep all batteries on-line from 0% to 100% SoC.

If we just parallel batteries on a DC bus, they will supply different fractions of current draw as they move through their curve. Also, all hit 100% at the same time, leaving none to absorb load-dump. On separate switchers (AC or DC coupled), they can each periodically do a run to 3.65V per cell for balancing, and then run down to an SoC that is good for long term life.

 

[zanydroid]

I think you can add a switcher FET for lower cost than a DC DC converter.

There are HVDC systems today designed to increase capacity (kWh and kW) by series stacking, so there is no balancing in that architecture. The inverters are designed to operate at variable voltage. Of course this has a capacity limit.

Is there a reason that voltage verification before making a connection will be harder to make work for HVDC batteries vs other applications? (there's production systems doing that for low voltage and line voltage). The batteries should be able to protect themselves to the voltage rating of the FETs used in their BMS so that comes for free. And both inverter and battery should start in open state on the power lines before negotiating a voltage. The other smart connection standards negotiate on every re connection (useful for handling power loss / restart states) or continuously (the continuous ones do that to disconnect on detecting faults, this could also be useful to detect if the other side might have lost power, and then conservatively disconnect the power lines).

 

[zanydroid]

Following up on this a bit.

I was trying to understand what the difference is between ESS and UPS, and what the state of code writing is to disambiguate this (given that ESS is defined expansively enough to include UPS, yet UPS have been installed for much longer and plenty of people still need UPS, including commercial sites that really really care about UPS which one would imagine is a strong forcing function for new code).

Has anyone been following this?