Our approach
- Verification: the code implements the method correctly. Checked against hand calculations and worked examples from the standards.
- Validation: the method fits the real problem. Checked against established tools and measurements (in progress, see below).
- Each reference case stores its input network, the expected quantities with their tolerance, and a source note with the derivation or the page and table it comes from.
- The cases are deterministic and run automatically with every change to the calculation engine. Every result bug we fix becomes a new regression case.
- Tolerances are never loosened to make a case pass: a disagreement is traced to an assumption, a method difference or a bug.
Methods
| Study | Method | Scope | Verification |
|---|---|---|---|
| Load flow | Newton-Raphson power flow (pandapower 3), balanced three-phase, positive sequence | Bus voltages and angles, branch currents and loading, losses. Loads as constant power scaled by the study case. | Initial cases passing |
| Unbalanced load flow | Three-phase Newton-Raphson power flow in symmetrical components (pandapower 3 runpp_3ph) | Phase and line voltages, voltage unbalance (VUF, NEMA), phase and neutral currents, loads on their phases. Generators as balanced injections; no three-winding transformers. | Initial cases passing |
| Short circuit | IEC 60909-0:2016 equivalent voltage source at the fault location (pandapower 3) | Three-phase, line-to-line, line-to-ground and two-phase-to-ground faults: Ik″, ip, Ib, Ik (with generator λ factors), Ith, Sk″ per bus and branch contributions. The minimum case and the IEC TR 60909-4 example networks are not yet verified; the λ closed form is pending review against IEC 60909-0 Figures 18/19. | Initial cases passing |
| Short circuit (ANSI/IEEE C37) | IEEE 551 / C37.010 first-cycle, interrupting and 30-cycle networks with machine reactance multipliers and separate R and X reductions (engine's own impedance-matrix solver) | Three-phase, line-to-ground, line-to-line and double line-to-ground faults: Isym, X/R, Iasym, Ipeak, interrupting multiplying factors (local and remote, NACD), 30-cycle current and contributions. Verified on hand-calculated cases; the IEEE 551 ch. 10 example is pending. | Initial cases passing |
| Voltage drop, device duty, cable, transformer and protection sizing | Checks and sizing rules over the load-flow and short-circuit results (NEC and IEC 60364 rule sets) | Verified on hand-calculated cases (V-03, V-05 to V-09). Normative tables and limits are pending legal and transcription review; offered to pilot organizations. | Initial cases passing |
Reference cases
17 cases pass today. They verify the load-flow and IEC 60909 and ANSI/IEEE C37 short-circuit pipelines end to end on small networks solved independently by hand.
| Case | Network | Reference | Tolerance | Status |
|---|---|---|---|---|
| V01a | Utility bus alone, three-phase faultShort circuit (IEC 60909-0) | Independent hand calculation from the IEC 60909-0:2016 formulas (500 MVA, X/R 10 at 20 kV) | ±0.5 % on Ik″, ip, Ith, Sk″, Rk, Xk | Passing |
| V01b | Utility, 630 kVA transformer (with KT), LV busShort circuit (IEC 60909-0) | Independent hand calculation from the IEC 60909-0:2016 formulas, including the network transformer correction factor KT | ±0.5 % on bus and branch currents | Passing |
| V01c | V01b plus an LV cable to a second busShort circuit (IEC 60909-0) | Independent hand calculation (cable resistance at 20 °C for maximum currents) | ±0.5 % | Passing |
| V01d | V01b at 60 Hz with a 0.1 s fault durationShort circuit (IEC 60909-0) | Independent hand calculation of the frequency dependence of κ, ip and Ith | ±0.5 % | Passing |
| V01e | Two-bus radial feeder with a constant-power loadLoad flow | Closed-form load-flow algebra for a single series branch | V ±0.00001 pu, angle ±0.001°, P/Q/I/losses ±0.5 % | Passing |
| V01f | Three-bus radial network through a tapped Dyn11 transformerLoad flow | Closed-form load-flow algebra (ideal tapped transformer, 30° phase shift) | V ±0.00001 pu, angle ±0.001°, P/Q/I/losses ±0.5 % | Passing |
| V04a–g | Line-to-ground, line-to-line and two-phase-to-ground faults, fault impedance, breaking and steady-state currents (generator λ factors), branch contributionsShort circuit (IEC 60909-0) | Independent hand calculations with symmetrical components from the IEC 60909-0 formulas | ±1 % on Ik″, ip, Ib, Ik and contributions | Passing |
| V02a | IEEE 14-bus test systemLoad flow | Published IEEE 14-bus solution (bus voltages and angles); generator reactive power is not compared because the published values are inconsistent with the same data | V ±0.001 pu (bus 4: ±0.0015 pu, documented), angle ±0.05° | Passing |
| V02b | Industrial network: utility, transformer, motors and loadsLoad flow | Independent hand calculation | V ±0.001 pu, P/Q ±0.5 % | Passing |
| V17a–b | Unbalanced load flow: single-phase and line-to-line loads behind a Dyn transformer and on a grounded LV feederUnbalanced load flow | Independent hand calculations with symmetrical components (phase and line voltages, VUF, NEMA unbalance, phase and neutral currents) | V ±1e-6 pu; unbalance and currents ±0.01 % | Passing |
| V-03 | Voltage dropVoltage drop | Worked examples of the NEC handbook and the IEC 60364 application guide; hand calculationPassing on independent hand calculations for NEC (V03a) and IEC (V03b); the handbook worked examples are still to be added. | ±0.05 percentage points | Passing |
| V-05 | Device duty: breakers at the limit, X/R correctionDevice duty | Constructed cases with known answersConstructed cases V05a–b: breakers (IEC and UL ratings with X/R correction), Ics, busbars and cable thermal withstand. | Exact pass/fail; margin ±0.5 percentage points | Passing |
| V-06 / V-07 | Ampacity and cable sizingCable sizing | Sampled cells of NEC 310.16 and IEC 60364-5-52 tables with correction factors; worked examplesTable cells are checked against published reproductions of the tables (not yet against the standards' own text) and are pending legal and transcription review; the cable sizing cases (V07a NEC, V07b IEC) are hand calculations. | Exact table value; ±1 % with factors; exact size and governing criterion | Passing |
| V-08 | Transformer sizingTransformer sizing | Hand calculationHand-calculated cases V08a–b. The forced-cooling factors are typical practice pending review. | Exact standard kVA | Passing |
| V-09 | Circuit protection sizingProtection sizing | NEC 240.4 / 430.52; IEC 60364-4-43Hand-calculated cases V09a–b. NEC Table 430.52 and the 430.32 overload percentages are not yet cross-checked against a second source. | Exact rated current | Passing |
| V-12 (hand) | ANSI C37 faults (3F, SLG, LL, LLG), separate R and X reductions, machine multipliers, NACD and device dutyShort circuit (ANSI/IEEE C37) | Independent hand calculations from the IEEE 551 and IEEE C37.010 / C37.04 / C37.13 formulasCases V12g, V12h1, V12h2, V12ih, V12j and V12k. The first-cycle currents are also cross-checked against OpenDSS (phase-domain fault solution) within ±1 %. | ±0.5 %; decision and margin exact on device duty | Passing |
| V-12 (standards) | C37.010 example interrupting factors and C37.13 Table A.1 multiplying factorsShort circuit (ANSI/IEEE C37) | IEEE C37.010-2016 example, Tables 16 to 18; IEEE C37.13 Table A.1Cases V12e and V12f, checked against the English originals. | ±1 % on the duty; ±0.02 on read-off factors | Passing |
Pending and planned
These cases are part of the plan but not green yet. A study is only released to everyone once its cases pass.
| Case | Network | Reference | Tolerance | Status |
|---|---|---|---|---|
| V-04 | IEC TR 60909-4 example networks (LV and MV, with and without motors)Short circuit (IEC 60909-0) | IEC TR 60909-4 published example resultsWill be built from the technical report itself, with page and table references; no value is reconstructed from secondary sources. | ±1 % on Ik″, ip, Ib, Ith | Pending |
| V-01 min | Minimum short-circuit case (cmin, conductor end temperature)Short circuit (IEC 60909-0) | Hand calculation against IEC 60909-0 Table 1 | ±0.5 % | Pending |
| V-17 ext | IEEE 4-bus and 13-bus unbalanced test feedersUnbalanced load flow | Published IEEE PES test feeder resultsNeeds the feeders' published data and results; the 13-bus feeder also needs line impedance matrices and regulators the network model does not carry yet. | V ±0.5 % | Pending |
| V-02 ext | IEEE 4-bus feeder and MATPOWER cross-checkLoad flow | Published IEEE / MATPOWER results | V ±0.001 pu, angle ±0.05°, P/Q ±0.5 % | Pending |
| V-12 (Violet Book) | IEEE 551 worked examples: half-cycle currents, interrupting factors and the ch. 10 industrial networkShort circuit (ANSI/IEEE C37) | IEEE 551-2006 sections 9.8 and 9.9 and ch. 10 (Tables 10-2 to 10-6)V12a to V12d reproduce, but their numbers were read from a translated copy of the book and must be confirmed in an English edition. The ch. 10 network (V12i) needs the English tables. The total-basis local curves are not digitized yet. | ±2 % | Pending |
| V-13 | IEEE 1584-2018 arc flashNot offered yet | IEEE 1584-2018 annexes and the official IEEE calculator | ±1 % on Iarc, incident energy, arc-flash boundary | Planned |
Cross-validation
We are comparing EngifyHub with an established commercial tool on three to five real, anonymized projects, run by engineers who hold those licenses on their own projects. Every difference above tolerance will be explained as an assumption, a method difference or a bug. Planned
Traceability
- Every study run records the engine version, the input format version, a hash of its input and a snapshot of the network model it solved.
- Every value the software assumed because data was missing is listed with its source in the run and in its report.
- The engine is versioned: any change that can alter a result raises its version and comes with a release note.
Known limitations
- The load flow is balanced; the unbalanced load flow solves each phase with single-phase loads on their phases, models generators as balanced injections and does not support three-winding transformers. Unbalanced faults (line-to-line, line-to-ground, two-phase-to-ground) are computed with symmetrical components.
- Short circuit follows IEC 60909 or ANSI/IEEE C37 (the default in NEC projects). Arc-flash studies are not available yet.
- No dynamic studies: motor starting, transient stability and harmonics are out of scope for now.
- Results have not yet been cross-checked against an established commercial tool on real projects; that comparison is planned with pilot customers who hold the licenses.
- Typical values fill missing equipment data. They are conservative engineering defaults with a cited source, listed in every run, and pending independent review; replace them with nameplate data before relying on the results.
Engineering responsibility
EngifyHub calculations are engineering aids. They do not replace the judgment of a qualified engineer, and they are not a design, a certification or a sealed document.
Every result must be reviewed and verified by a qualified professional engineer, who remains responsible for the design and for its compliance with the codes, standards and requirements that apply to the project, and who must sign or seal it where the law requires.
Results depend on the data entered in the project. Values the software assumed when data was missing (typical values) are listed with their source in each run and report; replace them with the actual equipment data before relying on the results.
The normative and typical data tables used by the calculations are pending independent technical and legal review. The verification status of each method is published on the validation page.
Full dossier
The complete dossier (each case's network, expected values, derivation and obtained deviation) is available on request: legal@engifyhub.com.