add power-flow static VAR compensator in SP domain

[leonardocarreras]

Voltage-controlling shunt reactive device modelled as a PQ bus, to be driven by the SVC outer control loop. No active power, no admittance stamp; participates via bus classification and reactive setpoint. Reuses the SynchronGenerator per-unit/attribute pattern (V_set_pu, Q_max_pu, Q_min_pu).

The solver-side outer loop that makes this usable is the pull request above it in the stack.

[github-actions]

DPsim LLM review

Claim vs. code: matches the description.

Found 5 high, 10 medium, 1 low (0 anchored to lines below).

🔵 Optional / low-confidence (16)

• Add virtual destructor to polymorphic base class [high · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:16
• Missing validation of input parameters in setParameters [high · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:49
• Missing override specifiers on virtual interface methods [high · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:61
• Missing scheduler dependency for updateReactivePowerInjection [high · 34% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:99
• getApparentPower returns a value that is never consumed by the power-flow solver [high · 28% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:113
• Unnecessary flush after logger call in constructor [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:24
• Provide default argument for ratedVoltage in setParameters [medium · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:49
• Incorrect comment for member variable mBaseVoltage [medium · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:20
• Use DEBUG level for component creation log [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:23
• Demote noisy INFO logs in setParameters to DEBUG [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:38
• Demote power-flow section header log to DEBUG [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:56
• Use DEBUG level for constructor log [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:23
• Parameter‑setting logs should be DEBUG [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:38
• Per‑unit calculation logs should be DEBUG [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:56
• No unit test exercising SP::Ph1::SVC behavior [medium · 30% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp
• Duplicated per‑unit conversion logic [low · 20% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:15

Claim vs. implementation

• Claimed: Add a power-flow SP-domain static VAR compensator model as a PQ bus with reactive setpoint/limits, plus docs and solver base-voltage handling.
• Done: Introduced a new SP::Ph1::SVC class with PQ-only power-flow behavior, reactive setpoint/limits and per-unit conversion, wired it into the solver, pybind, build, and documentation.
• Difference: none

How this review was produced

13 specialized finder passes raised 63 findings over the diff and the full changed sources. After de-duplication, 61 were re-checked against the current file and the base-class / interface headers it inherits (code as truth), escalating survivors to a stronger model: 45 refuted as unsupported, 16 kept (16 tentative).

Refuted by verification:

• Incorrect extraction of reactive power from Complex injection (dpsim-models/src/SP/SP_Ph1_SVC.cpp): updateReactivePowerInjection intentionally stores powerInj.imag() as the reactive power setpoint, and the header documents Q_set as reactive power injection
• Apparent power representation does not match power-flow convention (dpsim-models/src/SP/SP_Ph1_SVC.cpp): getApparentPower() is documented to return apparent power with P always 0, and the implementation matches that contract
• Incomplete explanation of the power-flow SVC boundary condition (docs/hugo/content/en/docs/Concepts/Models/network-injection-and-compensation.md): The added section already explains that fixing reactive power uses constant-power equations and an outer loop, which is the PQ-bus modeling choice.
• Incomplete documentation of the power-flow SVC interface (docs/hugo/content/en/docs/Developer Guide/Model Implementations/network-injection-and-compensation.md): The added power-flow section explicitly says it stamps nothing into the admittance matrix and that updateReactivePowerInjection is the setter that moves the injection.
• Incorrect apparent power return type for SP::Ph1::SVC (dpsim-models/src/SP/SP_Ph1_SVC.cpp): the SP header says P is always 0, so returning Complex(0., **mSetPointReactivePower) is the intended convention
• Use DOUBLE_EPSILON instead of ad-hoc 1e-6 tolerance (dpsim-models/src/SP/SP_Ph1_SVC.cpp): the file explicitly uses Math::isFinite and a positivity check; the complaint is about a different tolerance policy, not a defect
• Validate base apparent power before division in updateReactivePowerInjection (dpsim-models/src/SP/SP_Ph1_SVC.cpp): updateReactivePowerInjection checks mBaseApparentPower for finiteness and positivity before dividing
• Per-unit division by base voltage lacks explicit non-zero check (dpsim-models/src/SP/SP_Ph1_SVC.cpp): the guard already rejects zero and negative base voltage with mBaseVoltage <= 0
• getApparentPower returns Complex with zero real part (dpsim-models/src/SP/SP_Ph1_SVC.cpp): returning zero real power is documented in the header as correct for an SVC
• Attribute usage: ratedVoltage parameter is unused (dpsim-models/src/SP/SP_Ph1_SVC.cpp): ratedApparentPower and ratedVoltage are intentionally logged only, as the header comment states
• Inconsistent guard against non-finite base apparent power (dpsim-models/src/SP/SP_Ph1_SVC.cpp): the downstream division is guarded by the same finiteness/positivity check in updateReactivePowerInjection
• Guard against division by zero in updateReactivePowerInjection (dpsim-models/src/SP/SP_Ph1_SVC.cpp): the code checks mBaseApparentPower before dividing by it
• Use Math::isPositive instead of manual comparison (dpsim-models/src/SP/SP_Ph1_SVC.cpp): manual comparison mBaseVoltage <= 0 is already the explicit positivity check
• Use Math::isPositive for base apparent power guard (dpsim-models/src/SP/SP_Ph1_SVC.cpp): manual comparison mBaseApparentPower <= 0 is already the explicit positivity check
• Remove redundant semicolon after setTerminalNumber (dpsim-models/src/SP/SP_Ph1_SVC.cpp): the semicolon after setTerminalNumber(1) is present in the source but is harmless style, not a defect

_{Automated, non-blocking review. May be wrong. Models: find mistral-small-4-119b-2603, gpt-oss-120b → verify gpt-5.4-mini → final gpt-5.5.}

[github-actions]

DPsim LLM review

Claim vs. code: matches the description.

Found 1 critical, 9 high, 20 medium (0 anchored to lines below).

🔵 Optional / low-confidence (30)

• SVC per‑unit parameters never computed during initialization [critical · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:54
• Incorrect reactive power injection extraction from Complex argument [high · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:108
• Incorrect apparent power representation in SP domain [high · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:113
• getApparentPower returns Complex with zero real part instead of RMS3PH convention [high · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:113
• Add virtual destructor to polymorphic base class [high · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:16
• Missing override specifier on virtual method [high · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:61
• Expose SVC per-unit attributes to Python [high · 35% confidence · unconfirmed] in dpsim/src/pybind/SPComponents.cpp:166
• Missing validation of ratedApparentPower and ratedVoltage in setParameters [high · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:32
• Inconsistent guard between updateReactivePowerInjection and calculatePerUnitParameters [high · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:100
• Missing override specifier on virtual methods [high · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h:55
• Inconsistent description of power-flow SVC boundary condition [medium · 35% confidence · unconfirmed] in docs/hugo/content/en/docs/Concepts/Models/network-injection-and-compensation.md:67
• Inconsistent per-unit parameter calculation description [medium · 35% confidence · unconfirmed] in docs/hugo/content/en/docs/Developer Guide/Model Implementations/network-injection-and-compensation.md:71
• Unnecessary semicolon after constructor initializer list [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:27
• Inconsistent switch-case indentation [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:82
• Missing override specifier on virtual method [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:99
• Redundant check of mBaseApparentPower positivity [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:100
• Inconsistent comment style for SVC base voltage handling [medium · 35% confidence · unconfirmed] in dpsim/src/PFSolver.cpp:334
• Inconsistent comment style for SVC base voltage propagation [medium · 35% confidence · unconfirmed] in dpsim/src/PFSolver.cpp:456
• Use DEBUG level for component creation log [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:23
• Use DEBUG level for rated parameter logging [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:38
• Use DEBUG level for per-unit calculation log [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:56
• Incomplete validation of powerInj in updateReactivePowerInjection [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:104
• Add new SP_Ph1_SVC.cpp to the build list [medium · 35% confidence · unconfirmed] in dpsim-models/src/CMakeLists.txt:182
• Include new SP_Ph1_SVC.h header in Components.h [medium · 35% confidence · unconfirmed] in dpsim-models/include/dpsim-models/Components.h:25
• Expose SP::Ph1::SVC to Python bindings [medium · 35% confidence · unconfirmed] in dpsim/src/pybind/SPComponents.cpp:166
• Add a test or example exercising SP::Ph1::SVC in Python [medium · 35% confidence · unconfirmed] in docs/hugo/content/en/docs/Developer Guide/Model Implementations/network-injection-and-compensation.md:58
• SVC attributes not exposed to Python [medium · 35% confidence · unconfirmed] in dpsim/src/pybind/SPComponents.cpp:166
• Missing finiteness checks for set‑point attributes before per‑unit conversion [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp:71
• Missing dedicated model documentation page for SP::Ph1::SVC [medium · 35% confidence · unconfirmed] in docs/hugo/content/en/docs/Concepts/Models/network-injection-and-compensation.md:61
• Missing unit test for SP::Ph1::SVC power‑flow behavior [medium · 35% confidence · unconfirmed] in dpsim-models/src/SP/SP_Ph1_SVC.cpp

Claim vs. implementation

• Claimed: Add a power-flow static VAR compensator in the SP domain as a PQ-bus, zero-active-power, no-admittance-stamp component with per-unit reactive setpoint/limits and solver-driven base-voltage handling.
• Done: Adds a new SP::Ph1::SVC class, exposes it to Python, integrates it into build/docs, and makes the power-flow solver assign its base voltage from the resolved node zone.
• Difference: none

How this review was produced

13 specialized finder passes raised 31 findings over the diff and the full changed sources. After de-duplication, 31 were re-checked against the current file and the base-class / interface headers it inherits (code as truth), escalating survivors to a stronger model: 1 refuted as unsupported, 30 kept (30 tentative).

Refuted by verification:

• SVC source not listed in TEST_SOURCES for CI test builds (dpsim-models/src/CMakeLists.txt): The file only adds SP/SP_Ph1_SVC.cpp to MODELS_SOURCES; there is no TEST_SOURCES mechanism in this CMakeLists.txt to satisfy the claim

_{Automated, non-blocking review. May be wrong. Models: find mistral-small-4-119b-2603, gpt-oss-120b → verify gpt-5.4-mini → final gpt-5.5.}

[codecov]

Codecov Report

❌ Patch coverage is 22.22222% with 70 lines in your changes missing coverage. Please review.
✅ Project coverage is 72.69%. Comparing base (5564723) to head (8326843).
⚠️ Report is 3 commits behind head on master.

Files with missing lines	Patch %	Lines
dpsim-models/src/SP/SP_Ph1_SVC.cpp	0.00%	65 Missing ⚠️
dpsim-models/include/dpsim-models/SP/SP_Ph1_SVC.h	0.00%	2 Missing ⚠️
dpsim/src/PFSolver.cpp	77.77%	2 Missing ⚠️
dpsim/src/pybind/SPComponents.cpp	92.85%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #656      +/-   ##
==========================================
- Coverage   72.72%   72.69%   -0.03%     
==========================================
  Files         512      514       +2     
  Lines       33219    33278      +59     
  Branches    18017    17986      -31     
==========================================
+ Hits        24158    24193      +35     
- Misses       9000     9084      +84     
+ Partials       61        1      -60     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sonarqubecloud]

Quality Gate passed

Issues
4 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
6.4% Duplication on New Code

See analysis details on SonarQube Cloud