hems/fnn: productionnominalmax carries Wp (module peak) while EEBus defines ProductionNominalMax as W (AC real power)

[andig]

Follow-up from the wording review of #31948, which renamed the FNN config key maxcurtailpower → productionnominalmax to align with EEBus ProductionNominalMax.

Nuance

EEBus CS-LPP documents the field as active power in W:

SetProductionNominalMax: nominal maximum active (real) power the Controllable System is allowed to produce … value: nominal max power production in W

(eebus-go usecases/cs/lpp/public.go; SPINE characteristic powerProductionNominalMax / contractualProductionNominalMax.)

The FNN templates describe the same value as Wp:

• templates/definition/hems/fnn-eebus.yaml, fnn-gpio.yaml, hemspro-*.yaml
  • en: Installed generator power (Wp) / help Rated generator/module power (Wp)
  • de: Installierte Generatorleistung (Wp) … Bezugsgröße nach § 9 EEG

Wp (DC module nameplate / § 9 EEG installed generator capacity) ≠ nominal AC real power (W): inverter AC max output is usually below DC module Wp. So the EEBus-named key carries a Wp value per the help text.

Impact

Not a functional bug. For § 14a curtailment the value is just the 100% reference that the curtailment percentage scales (hems/fnn/fnn.go setProductionLimit / MaxProductionPower), and the installed-generator figure is the intended reference either way.

Options

1. Clarify help text that the value is nominal max AC power in W (tight EEBus alignment), or
2. Keep the § 9 EEG Wp convention (domain-correct reference for § 14a) and document the intentional divergence from the EEBus W definition.

🤖 Generated with Claude Code

[github-actions]

Investigated and can confirm the discrepancy is accurately described.

• hems/fnn/fnn.go decodes the productionnominalmax config value into MaxCurtailPower, which is used purely as the 100% reference that the curtailment percentage scales against (fnn.go:214, setProductionLimit).
• The FNN templates (templates/definition/hems/fnn-eebus.yaml, fnn-gpio.yaml, hemspro-*.yaml) document this value as Wp (installed generator/module nameplate power, § 9 EEG reference).
• The EEBus side (hems/eebus/eebus.go:148, SetProductionNominalMax) forwards the same config value to CsLPPInterface.SetProductionNominalMax, which per the EEBus CS-LPP spec is nominal AC real power in W, not module Wp.

So the same config value is fed into both an EEBus field specified in AC-W and template help text that documents it as Wp — the two conventions genuinely diverge, as the issue states.

This is a documentation/naming nuance rather than a functional defect (the § 14a curtailment behavior is unaffected either way), and choosing between the two options requires a domain/product decision:

1. Reword the help text to describe nominal max AC power (W), tightening EEBus alignment, or
2. Keep the Wp/§9 EEG convention and explicitly document the intentional divergence from the EEBus W definition.

Given this needs a maintainer decision on which convention to standardize rather than a code fix, I'm leaving this for discussion rather than opening a PR.

🤖 Generated with Claude Code

[premultiply]

Option 2 is correct. Law beats outdated EEBus bloat spec.

[andig]

Two observations: certain inverters settings operate on AC power only (whatever Wp you may have) according to @CiNcH83 and the law is specific to Germany. What result to we want for wording- Wp on the HEMS side and W/Wp on the inverter side as required?

[CiNcH83]

• The FNN templates (templates/definition/hems/fnn-eebus.yaml, fnn-gpio.yaml, hemspro-*.yaml) document this value as Wp (installed generator/module nameplate power, § 9 EEG reference).
• The EEBus side (hems/eebus/eebus.go:148, SetProductionNominalMax) forwards the same config value to CsLPPInterface.SetProductionNominalMax, which per the EEBus CS-LPP spec is nominal AC real power in W, not module Wp.

Both would be wrong for Huawei at least. It uses max. AC power as reference.

Here are my solar system specs with a SUN2000-8KTL-M1:

• Generator Power: 9.315 Wp
• Nominal (Rated) Power (Nennleistung): 8.000 W
• Max. Active Power: 8.800 W (cos φ = 1)
• Max. Apparent Power: 8.800 VA

When I set my SUN2000 to 30%, it feeds-in @ 2,6 kW (=30% of 8.800 W):

SunSpec looks to be the same. WMaxRtg is the reference to WMaxLimPct.

[andig]

Both would be wrong for Huawei at least. It uses max. AC power as reference.

So- however we name it its wrong. We can then either: use the regulatory name (which is correct in some sense, independently of any device) or make up a new description. Or we leave it as-is each in its own "standard". I'm open to a concrete PR that makes the desired changes across the board. Happy to review there.

[CiNcH83]

Problem is that it can't be generalized IMHO. It is an inverter implementation detail as to what it uses as reference for calculation, either generator DC power, inverter nominal AC power or inverter max. AC power. I think that many inverters don't even know the connected generator power. For Huawei, this peak power only exists in the cloud for efficiency metrics and fault diagnostics.

Best would be to use an API where the power limitation passed in Watts by the operator is passed 1:1 to the inverter. Huawei has such an interface. Seems like the %-interface is more common though, also being covered by SunSpec (which in fact uses inverter max. AC power as reference).

[andig]

Best would be to use an API where the power limitation passed in Watts by the operator is passed 1:1 to the inverter. Huawei has such an interface. Seems like the %-interface is more common though, also being covered by SunSpec (which in fact uses inverter max. AC power as reference).

We could add a second interface. Let's keep this in the back of our minds. Atm it would be better to stabilize what we have, the workaround is clear.