Ewis Installation Quality

EWIS Installation Quality (manufacturing-quality/assembly/ewis-installation-quality)

Use when the task is verifying the physical installation quality of an electrical wiring interconnection system (EWIS) bundle during aerospace assembly: checking that the conductor bundle does not overfill its conduit or clamp cross-section, that the round-trip voltage drop of a power or signal run stays within its percent-of-bus limit, that no conductor bend is tighter than the minimum radius factor allows, and that the bundle keeps the required separation clearance from nearby hydraulic or fuel lines and structure. This leaf implements the four checks in pure Python, stdlib only; the wire gauge, resistance per meter and current are inputs, and the geometry, fill, drop, bend and separation verdicts are the outputs. It is the second leaf of the assembly pack and pairs with manufacturing-quality/assembly/fastener- installation-quality, which owns the structural side of the same acceptance step. EWIS design tasks that select the harness routing or splice layout are out of scope here.

Domain quick reference

• Wire area: A = PI * (d / 2)^2 for a round conductor of diameter d, with PI = math.pi. Areas add across the bundle.
• Bundle fill ratio: fill = sum(A_i) / A_conduit, the summed conductor area over the conduit inner area. New-install limit FILL_LIMIT = 0.40; the check passes when fill <= limit (fill exactly at the limit passes), margin = limit - fill.
• Round-trip resistance: R_rt = 2 * R_per_m * L. The return path doubles the one-way resistance of a run of length L.
• Voltage drop: drop_V = I * R_rt and drop_pct = 100 * drop_V / V_bus. Default acceptance limit VOLTAGE_DROP_LIMIT_PCT = 2.0 percent of the nominal bus voltage; margin = limit - drop_pct.
• Bend radius: required = BEND_FACTOR_DEFAULT * conductor_diameter with BEND_FACTOR_DEFAULT = 6.0 (typical shielded bundle minimum, a factor of the conductor diameter). Verdict margin = actual / required - 1; a negative margin fails.
• Separation: margin = actual / required - 1 for the distance between the bundle and a fluid line or structure; a negative margin fails.
• Fill, drop, bend and separation all compare with a <= style pass (margin >= 0 passes), so the four verdicts combine into one overall_pass with the failing checks named.
• Units are consistent mm for diameters and radii, m for run length, ohm/m for resistance per meter, V for voltage, A for current, mm for distances, percent for the drop.
• AS9100 frames the manufacturing acceptance context; the relations above are standard engineering methodology, summary-only.

Workflow

1. Gather the run inputs: bus voltage, the conductor diameters of the bundle (wire_diameters), the conduit inner diameter, resistance per meter, run length, load current, actual bend radius, and the actual and required separation distances.
2. Compute the fill with bundle_fill_ratio and get the verdict from fill_check against fill_limit (default FILL_LIMIT).
3. Compute the round-trip resistance with round_trip_resistance, feed it to voltage_drop with the bus voltage and current, and compare the drop_pct output against drop_limit_pct in the report.
4. Check the bend with bend_radius_check(conductor_diameter, actual_bend_radius, factor): the largest single conductor in the bundle drives the required radius (worst case).
5. Check the separation with separation_check(actual_distance, required_distance).
6. Aggregate everything with ewis_installation_report in one call when the inputs are complete; it returns the four verdicts, the failing_checks list and overall_pass.
7. Confirm the deterministic anchors with the contract test scripts/test_ewis_installation_quality.py.

Worked example

A 28 VDC EWIS run: 12 conductors of 2.0 mm diameter in a 12 mm conduit, R = 0.008 ohm/m over 15 m, current 5 A, actual bend radius 8 mm, separation from a hydraulic line 60 mm actual against 150 mm required.

Module outputs (scripts/ewis_installation_quality_logic.py):

• Wire area: wire_area(2.0) = 3.1416 mm2; bundle 37.699 mm2 against a conduit area of 113.097 mm2.
• Fill ratio: bundle_fill_ratio([2.0] * 12, 12.0) = 0.3333, fill_check margin = 0.40 - 0.3333 = +0.0667, PASS (0.333 <= 0.40).
• Round-trip resistance: round_trip_resistance(0.008, 15.0) = 0.24 ohm exactly (2 * 0.008 * 15).
• Voltage drop: drop_V = 5 * 0.24 = 1.2 V, drop_pct = 4.2857% of the 28 V bus (assert 1e-9 relative), FAIL against the 2% limit, margin -2.2857 points.
• Bend radius: required = 6.0 * 2.0 = 12 mm, actual 8 mm, margin 8/12 - 1 = -0.3333, FAIL.
• Separation: margin = 60/150 - 1 = -0.6, FAIL.

Verdict table from ewis_installation_report(28.0, [2.0] * 12, 12.0, 0.008, 15.0, 5.0, 8.0, 60.0, 150.0):

overall_pass is False with failing_checks exactly ['voltage_drop', 'bend_radius', 'separation']: the drop, bend and separation checks fail and the fill check passes.

Verification

• Confirm bundle_fill_ratio([2.0] * 12, 12.0) returns 0.3333 (bound 0.30 to 0.37) and that fill_check(0.40) passes (pass_bool is fill_ratio <= limit, so exactly at the limit is a pass).
• Confirm round_trip_resistance(0.008, 15.0) returns 0.24 exactly and voltage_drop(28.0, 5.0, 0.24) returns drop_V 1.2 V and drop_pct 4.2857% (1e-9 relative).
• Confirm bend_radius_check(2.0, 8.0) gives required_radius 12.0 mm, margin -0.3333 and pass_bool False, and that 12 mm actual passes.
• Confirm separation_check(60.0, 150.0) gives margin -0.6 and pass_bool False, and that 150 mm actual passes with margin 0.
• Confirm ewis_installation_report on the worked example returns overall_pass False with exactly the three failing checks named above.
• Confirm ValueError rejection: conduit_diameter <= 0, an empty wire list, any wire diameter <= 0, voltage <= 0, current < 0, resistance < 0, actual_bend_radius < 0, factor <= 0, required_distance <= 0 and actual_distance < 0 all raise ValueError.
• Run the contract test offline: python3 scripts/test_ewis_installation_quality.py (32 tests, deterministic).

Related leaves

• manufacturing-quality/assembly/fastener-installation-quality: the sibling assembly leaf for structural fastener grip, clamp load and installation verdicts at the same acceptance step.
• manufacturing-quality/special-processes/special-process-qualification: change control when an EWIS wiring process itself changes and needs requalification; installation checks here assume a qualified process.
• avionics/data-bus/arinc664-afdx: the data bus signaling context that the EWIS physical layer carries; protocol design is out of scope for this leaf.

Pitfalls

• Computing the voltage drop one-way: the run's resistance is round-trip (R_rt = 2 * R_per_m * L), so forgetting the return path halves the drop - the worked example drops 1.2 V / 4.2857% only because the 0.24 ohm round-trip value is used.
• Sizing the bend check against the average conductor: the largest single conductor in the bundle drives the required radius (worst case), so a bundle whose mean diameter passes can still fail on its biggest wire.
• Flagging a fill ratio exactly at the limit: all four checks pass on margin >= 0, so fill 0.40 against the 0.40 limit passes - only a fill above the limit fails.
• Mixing units across the inputs: the relations assume mm for diameters, radii and distances, m for run length, ohm/m for resistance, and V/A for the electrical inputs, so a cm conduit diameter silently corrupts the fill verdict.
• Reading overall_pass without the failing check list: the verdict aggregates four independent checks and names the failures (voltage_drop, bend_radius, separation in the worked example), so a bare pass/fail without the list hides which check needs rework.
• Treating this leaf as an EWIS design tool: it verifies the physical installation of a run under a qualified wiring process; routing and splice-layout design are out of scope, and a changed process itself routes to special-process-qualification.

Behavior contract (gate 3)

Run the deterministic contract test (stdlib unittest, offline):

python3 scripts/test_ewis_installation_quality.py

The test covers the worked-example contract (fill 0.3333 within the 0.30 to 0.37 bound, exact 0.24 ohm round-trip resistance, drop 4.2857% at 1e-9 relative, bend required 12 mm against 8 mm actual, separation margin -0.6, and overall_pass False with exactly the three failing checks), the fill boundary at exactly 0.40 passing, pass cases for all four checks, custom fill, bend and drop limits, the round-trip and area identities, and ValueError rejection of every non-physical input in the validation list.

Compliance

• Standards referenced, not reproduced: AS9100 is the aerospace quality management system standard that frames the manufacturing acceptance context; the check relations above are standard engineering methodology, summary-only per standards-map.yaml.
• compliance: STANDARDS-REF, gated: false.