Feat add battery rate max charge dc for hybrid inverter dc solar charging by F17HEH · Pull Request #3580 · springfall2008/batpred

F17HEH · 2026-03-16T09:21:44Z

Add DC-coupled solar charging support for hybrid inverters

Problem

On hybrid inverters (e.g. Sigenergy SigenStor) the solar panels connect
directly to the DC bus. During ECO mode the battery charges from solar
without the power passing through the AC stage, meaning it is not subject
to inverter_limit (the AC throughput cap).

Predbat was applying inverter_limit to all ECO mode battery charging
regardless of inverter type. This caused two compounding inaccuracies:

Charge rate — the model underestimated how fast the battery fills
from solar on bright days, because it capped DC charging at the AC
inverter ceiling rather than the battery's true DC charge rate.
Round-trip loss — the model applied battery_loss (AC charging
efficiency, typically 4-5%) to DC solar charging, which bypasses AC
conversion and is more efficient (~2%).

The knock-on effect of both is that the optimiser sees more predicted
solar clipping than will actually occur and compensates by planning a
higher overnight grid charge target — costing money unnecessarily on
Agile-style tariffs.

Additionally, has_reserve_soc was incorrectly set to False for
hybrid inverters, preventing Predbat from using the reserve SoC control.

Solution

`apps/predbat/fetch.py`

Read battery_rate_max_charge_dc (watts) from apps.yaml and convert
to kWh/min. Added else clause so the attribute is always set (None
when unconfigured), preventing AttributeError at runtime.
Read battery_loss_dc as a loss fraction, defaulting to battery_loss
if not configured so existing installs are completely unaffected.

`apps/predbat/prediction.py`

In ECO mode when inverter_hybrid is True and pv_now > 0, compute
a fresh charge curve using battery_rate_max_charge_dc as the ceiling
instead of charge_rate_now_curve capped by inverter_limit. This
correctly models the DC bus charging path.
For the SoC update step, detect DC solar charging (battery_state == "e+"
on a hybrid with solar) and apply battery_loss_dc instead of
battery_loss. This gives an accurate model of the higher DC charging
efficiency.
The __init__ copy block already copies both new attributes from base
into the prediction instance so threaded runs have access to them.

`apps/predbat/config.py`

Set has_reserve_soc to True for hybrid inverters. This was
incorrectly False, preventing Predbat from using the reserve SoC
control on systems that support it.

`docs/apps-yaml.md`

Added ### battery_rate_max_charge_dc section under Inverter control
configurations, documenting the setting, when to use it, how to
calculate the value, and the companion battery_loss_dc setting.

No behaviour change when

inverter_hybrid is False (AC-coupled systems)
battery_rate_max_charge_dc is not set in apps.yaml (falls back to
battery_rate_max_charge)
battery_loss_dc is not set in apps.yaml (falls back to
battery_loss)

Example configuration (Sigenergy, 2 × SigenStor BAT 10.0)

# Each BAT 10.0 is rated at 4600W DC — two units = 9200W
battery_rate_max_charge_dc:
  - 9200

# DC solar charging bypasses AC conversion (~98% efficient vs ~95% for AC)
battery_loss_dc: 0.02

Tested on a Sigenergy system (6kW hybrid inverter, 18kWh battery, 7.2kW
panels) on Octopus Agile. Previously the plan showed the battery charging
at the AC-limited rate during solar peak and planned higher overnight grid
charges than were needed. After this change the predicted charge rate
correctly reflects DC bus capacity and overnight charge targets are reduced
on days with good solar forecast.

No regressions observed on overcast days or overnight grid charging windows.

the battery to charge from PV at a rate higher than the AC inverter limit. Previously Predbat applied inverter_limit to all battery charging including ECO mode solar, causing it to underestimate how quickly the battery fills from solar and potentially over-planning overnight grid charging. Changes: - fetch.py: read battery_rate_max_charge_dc from apps.yaml (watts), convert to kWh/min, store on self; falls back to battery_rate_max_charge if not set so existing configs are unaffected - prediction.py: in ECO mode charging when inverter_hybrid and pv_now > 0, use battery_rate_max_charge_dc instead of charge_rate_now_curve capped by inverter_limit; skip AC inverter clamp for DC solar charging path - docs/apps-yaml.md: document battery_rate_max_charge_dc under inverter control configurations No behaviour change for AC-coupled inverters or when the config key is not set.

…led solar charging On hybrid inverters solar panels connect directly to the DC bus, meaning the battery can charge from solar without passing through the AC stage. Predbat was applying inverter_limit to all ECO mode battery charging regardless, causing it to underestimate how fast the battery fills from solar on bright days. The knock-on effect was the optimiser seeing more predicted solar clipping than actually occurs and compensating by planning a higher overnight grid charge target, costing money on Agile-style tariffs. Changes: - fetch.py: read battery_rate_max_charge_dc (W) from apps.yaml and convert to kWh/min; add else clause so battery_rate_max_charge_dc is always set (None when unconfigured). Read battery_loss_dc, defaulting to battery_loss if not set. - prediction.py: in ECO mode when inverter_hybrid and pv_now > 0, compute a fresh charge curve using battery_rate_max_charge_dc instead of the AC charge_rate_now_curve capped by inverter_limit. Use battery_loss_dc for the SoC update on the DC solar charging path instead of battery_loss. - docs/apps-yaml.md: document battery_rate_max_charge_dc and battery_loss_dc under Inverter control configurations. No behaviour change for AC-coupled inverters or when the config keys are not set.

…arge_dc-for-hybrid-inverter-DC-solar-charging

Copilot

Pull request overview

Adds configuration and modelling support for hybrid inverters where PV can charge the battery via the DC bus (bypassing the AC inverter limit), and updates Sigenergy support/docs accordingly.

Changes:

Introduces battery_rate_max_charge_dc and battery_loss_dc configuration handling and uses them in the prediction model for hybrid ECO + PV charging.
Adjusts Sigenergy inverter capabilities (has_reserve_soc) and updates templates/documentation to describe the new settings.
Expands docs/examples for Sigenergy SigenStor DC-coupled charging configuration.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
templates/sigenergy_sigenstor.yaml	Documents new DC charge rate/loss settings in the Sigenergy template.
docs/inverter-setup.md	Adds Sigenergy-specific guidance for configuring DC charge rate/loss.
docs/customisation.md	Documents `battery_loss_dc` behavior and how/when it applies.
docs/apps-yaml.md	Adds full reference docs for `battery_rate_max_charge_dc` and `battery_loss_dc`.
apps/predbat/prediction.py	Implements hybrid DC-bus charge rate modelling and DC loss handling in ECO mode.
apps/predbat/fetch.py	Parses new config options from apps.yaml into runtime attributes.
apps/predbat/config/apps.yaml	Updates the sample apps.yaml with commented examples for the new options.
apps/predbat/config.py	Enables reserve SoC support for Sigenergy inverter definition.

You can also share your feedback on Copilot code review. Take the survey.

apps/predbat/prediction.py

apps/predbat/fetch.py

apps/predbat/prediction.py

…r-DC-solar-charging' of https://github.com/F17HEH/batpred into feat--add-battery_rate_max_charge_dc-for-hybrid-inverter-DC-solar-charging

Copilot

Pull request overview

This PR improves Predbat’s modelling for hybrid inverters where PV can charge the battery via the DC bus (bypassing AC inverter throughput limits), and enables reserve SoC support for Sigenergy systems. This aims to reduce plan inaccuracies (solar clipping/overnight charging) on DC-coupled hybrid setups while keeping existing non-hybrid behaviour unchanged.

Changes:

Add config parsing for battery_rate_max_charge_dc and battery_loss_dc and propagate them into the prediction engine.
Update prediction ECO-mode charging to use a DC-bus charge ceiling (and a separate DC charging loss/efficiency) on hybrid inverters with PV.
Enable has_reserve_soc for the SIG inverter type and document the new configuration options (docs + templates).

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
templates/sigenergy_sigenstor.yaml	Adds example configuration for DC charge rate/loss on Sigenergy.
docs/inverter-setup.md	Documents Sigenergy DC charge rate configuration guidance.
docs/customisation.md	Documents `battery_loss_dc` behaviour and usage.
docs/apps-yaml.md	Adds `battery_rate_max_charge_dc` / `battery_loss_dc` to apps.yaml documentation.
apps/predbat/prediction.py	Implements DC-path hybrid ECO charging curve + DC charging efficiency.
apps/predbat/fetch.py	Parses new apps.yaml options into internal (kWh/min + efficiency) fields.
apps/predbat/config/apps.yaml	Adds commented template entries for the new settings.
apps/predbat/config.py	Sets `has_reserve_soc` to `True` for SIG inverter type.

You can also share your feedback on Copilot code review. Take the survey.

apps/predbat/fetch.py

docs/apps-yaml.md

apps/predbat/prediction.py

templates/sigenergy_sigenstor.yaml

apps/predbat/config/apps.yaml

docs/inverter-setup.md

apps/predbat/prediction.py

docs/apps-yaml.md

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

ward0 · 2026-03-17T20:22:09Z

Could this also working for Huawei inverters? Can i help testing?

F17HEH · 2026-03-17T22:27:04Z

Could this also working for Huawei inverters? Can i help testing?

Yes should work on any DC coupled invertor.

F17HEH · 2026-03-19T14:14:34Z

@springfall2008 this is all ready for review again.

Copilot

Pull request overview

This PR extends Predbat’s hybrid-inverter modelling to better represent DC-coupled solar charging behavior (e.g., Sigenergy SigenStor), adding new configuration options for DC max charge rate and DC charging efficiency, plus enabling reserve SoC support for Sigenergy.

Changes:

Add battery_rate_max_charge_dc and battery_loss_dc configuration parsing and documentation.
Update prediction ECO-mode logic to allow DC solar charging to bypass AC inverter_limit when configured, and apply a DC-specific charging efficiency.
Enable reserve SoC capability for Sigenergy inverter type and add config parsing tests.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
templates/sigenergy_sigenstor.yaml	Adds example config entries for DC max charge rate and DC loss.
docs/inverter-setup.md	Documents Sigenergy-specific DC charge rate setup.
docs/customisation.md	Documents `battery_loss_dc` behavior and rationale.
docs/apps-yaml.md	Adds new apps.yaml option documentation for DC charging modelling.
apps/predbat/tests/test_fetch_config_options.py	Adds parsing/validation tests for the new config options.
apps/predbat/prediction.py	Implements DC charging rate ceiling and DC efficiency application in ECO mode.
apps/predbat/fetch.py	Parses `battery_rate_max_charge_dc` and `battery_loss_dc` from config.
apps/predbat/config/apps.yaml	Adds commented example configuration for the new settings.
apps/predbat/config.py	Enables `has_reserve_soc` for Sigenergy inverter profile.

templates/sigenergy_sigenstor.yaml

docs/apps-yaml.md

apps/predbat/config/apps.yaml

apps/predbat/fetch.py

apps/predbat/prediction.py

apps/predbat/fetch.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

F17HEH · 2026-03-20T17:03:30Z

@springfall2008 - lets hope this is the last one :)

springfall2008 · 2026-03-21T10:07:52Z

Sorry just started to look at this.

I can see why it makes sense to add battery_rate_max_dc if its higher than the ac limit.
But why add a new loss value, we have battery charge loss setting already?

springfall2008 · 2026-03-21T10:09:05Z

apps/predbat/prediction.py

                    battery_state = "e-"
                else:
-                    battery_draw = max(diff, -charge_rate_now_curve_step, -inverter_limit, -battery_to_max)
+                    if inverter_hybrid and pv_now > 0 and battery_rate_max_charge_dc is not None:


Why if pv_now > 0, it could be 0.00001 and trigger this condition?

springfall2008 · 2026-03-21T13:53:34Z

Looking at this further there are all sorts of issues with this change, for example what happens if you are force battery charging and you have Solar and the solar is therefore doing DC charging while some part is AC from the grid?

Currently there is no battery loss dc as we use battery loss for that and inverter loss is ignored for the DC portion.

But also you change ignores the current charge rate setting in the inverter if there is any solar in eco mode, which seems odd.

I think this needs a rethink.

Maybe we can just not have the DC part of the charge count to the inverter limit, would that work?

F17HEH · 2026-03-21T16:23:29Z

Ill have a review next week and see what i can come up with.

springfall2008 · 2026-03-21T20:04:40Z

I've written my own proposed, simpler, version of this - please review: Fix bug where inverter limit was wrongly applied to DC charging in EC

springfall2008 · 2026-03-22T13:28:51Z

This is now released in .8, so we can close this PR. Feel free to raise a new ticket or PR if you have issues with the new code.

F17HEH and others added 3 commits March 13, 2026 22:15

Merge branch 'springfall2008:main' into feat--add-battery_rate_max_ch…

5968049

…arge_dc-for-hybrid-inverter-DC-solar-charging

springfall2008 requested a review from Copilot March 16, 2026 20:48

Copilot started reviewing on behalf of springfall2008 March 16, 2026 20:49 View session

[pre-commit.ci lite] apply automatic fixes

9aea633

Copilot AI reviewed Mar 16, 2026

View reviewed changes

apps/predbat/prediction.py Outdated Show resolved Hide resolved

apps/predbat/prediction.py Show resolved Hide resolved

apps/predbat/fetch.py Outdated Show resolved Hide resolved

apps/predbat/prediction.py Show resolved Hide resolved

F17HEH added 2 commits March 17, 2026 09:11

Addressed review comments

ec33f5c

Merge branch 'feat--add-battery_rate_max_charge_dc-for-hybrid-inverte…

7628416

…r-DC-solar-charging' of https://github.com/F17HEH/batpred into feat--add-battery_rate_max_charge_dc-for-hybrid-inverter-DC-solar-charging

springfall2008 requested a review from Copilot March 17, 2026 10:00

Copilot started reviewing on behalf of springfall2008 March 17, 2026 10:00 View session

Copilot AI reviewed Mar 17, 2026

View reviewed changes

F17HEH and others added 3 commits March 17, 2026 10:14

Potential fix for pull request finding

8e93990

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

small update for comments

f4e03a8

Updated gating for dc fallback and added tests

1521bca

springfall2008 requested a review from Copilot March 19, 2026 20:13

Copilot started reviewing on behalf of springfall2008 March 19, 2026 20:13 View session

Copilot AI reviewed Mar 19, 2026

View reviewed changes

F17HEH and others added 2 commits March 20, 2026 09:11

Addressing PR comments.

0439cca

Update templates/sigenergy_sigenstor.yaml

f7a5816

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

springfall2008 reviewed Mar 21, 2026

View reviewed changes

springfall2008 mentioned this pull request Mar 21, 2026

Fix bug where inverter limit was wrongly applied to DC charging in ECO Mode. #3617

Merged

springfall2008 closed this Mar 22, 2026

Uh oh!

Conversation

F17HEH commented Mar 16, 2026

Add DC-coupled solar charging support for hybrid inverters

Problem

Solution

apps/predbat/fetch.py

apps/predbat/prediction.py

apps/predbat/config.py

docs/apps-yaml.md

No behaviour change when

Example configuration (Sigenergy, 2 × SigenStor BAT 10.0)

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ward0 commented Mar 17, 2026

Uh oh!

F17HEH commented Mar 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

F17HEH commented Mar 19, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

F17HEH commented Mar 20, 2026

Uh oh!

springfall2008 commented Mar 21, 2026

Uh oh!

springfall2008 Mar 21, 2026

Choose a reason for hiding this comment

Uh oh!

springfall2008 commented Mar 21, 2026

Uh oh!

F17HEH commented Mar 21, 2026

Uh oh!

springfall2008 commented Mar 21, 2026

Uh oh!

springfall2008 commented Mar 22, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

`apps/predbat/fetch.py`

`apps/predbat/prediction.py`

`apps/predbat/config.py`

`docs/apps-yaml.md`

F17HEH commented Mar 17, 2026 •

edited

Loading