Date applied: 2026-05-18
Reviewer: Claude Opus 4.7 (machine), pending Korbut human confirmation
Status: PROPOSED — not certified. Please verify against your
working calculation file before treating Outputs_corrected.xlsx as the
new reference.
The Tier 2 discharge row in the original Outputs.xlsx is character-
identical to the Tier 2 suction row (50.49 m total equivalent
length, 0.094 bar section ΔP). But the input fitting sets in Test Case_v1.xlsx differ materially between the two lines:
| Fitting | Suction count | Discharge count |
|---|---|---|
| Nozzle (vessel outlet) | 1 | 0 |
| Elbow 90° R=1.5D | 1 | 0 |
| Globe valve (straight pattern) L/D = 340 | 1 | 0 |
| Y-strainer L/D ≈ 250 | 1 | 0 |
| Contraction d/D = ½, K = 0.25 | 0 | 1 |
| Check valve (swing) L/D = 100 | 0 | 1 |
| Gate valve full-bore L/D = 8 | 0 | 1 |
| Nozzle (vessel inlet) | 0 | 1 |
Physically, the discharge ΔP must be roughly a third of the suction ΔP (the globe valve alone dominates the suction). Equal values can only arise from a copy-paste / stale-link error in the certified workbook.
Two independent runs (Claude Opus standalone + the TIROS skill chain
ls-author-line + ps-author-pump) produced the same physically-
correct discharge value of 0.039 bar vs the reference's 0.094 bar.
| cell | original | corrected |
|---|---|---|
| Total equivalent length — fittings | 27.39 m | 10.26 m |
| Total equivalent length — pipe | 23.1 m | 11.0 m |
| Total section length | 50.49 m | 21.26 m |
| Frictional pressure loss of section | 0.094 bar | 0.0392 bar |
The discharge friction feeds directly into pump discharge pressure → pump ΔP → TDH → hydraulic power → shut-off ΔP. With the corrected discharge friction substituted (and Korbut's suction pressure and elevation convention preserved):
| cell | original | corrected | change |
|---|---|---|---|
| Discharge pressure | 2.095 bara | 2.044 bara | −0.051 |
| Pump differential pressure | 1.002 bar | 0.951 bar | −0.051 |
| Pump differential head (TDH) | 10.26 m | 9.73 m | −0.53 |
| Hydraulic power (100% η) | 0.343 kW | 0.326 kW | −0.017 |
| Shut-off differential pressure | 1.203 bar | 1.141 bar | −0.062 |
- All Tier 1 values (line sizing) — both lines correctly select 2½" STD; suction = discharge for these as expected (same mass flow, same fluid).
- All Tier 2 suction values — match within tolerance (50.49 vs 50.27 m total; 0.094 vs 0.093 bar). Cosmetic split between "fittings" and "pipe" columns differs but totals match.
- Tier 3 suction pressure (1.093 bara) — preserved per Korbut's elevation convention. The standalone Opus run computed 1.123 bara using a 1.7 m static head (liquid surface = 2.3 m above grade); Korbut's value implies ~1.4 m static head. This is a definition question (where in the source vessel does the "2 m elevation" refer to?), not a math error. Worth a separate conversation.
- Liquid head (10.23 m/bar) — depends only on density, unchanged.
- NPSHa (11.22 m) — depends only on suction side, unchanged.
- Rated volumetric flow (12.33 m³/h) — input-derived, unchanged.
P_discharge = P_dest + ΔP_static_discharge + ΔP_friction_discharge + ΔP_equipment + ΔP_CV
= 1.10 + 0.195 + 0.0392 + 0.01 + 0.70
= 2.044 bara
Pump ΔP = P_discharge − P_suction
= 2.044 − 1.093
= 0.951 bar
TDH = ΔP_pump × liquid_head
= 0.951 × 10.23
= 9.73 m
Hydraulic power = ΔP_pump × Q_rated
= 95100 Pa × (12.33/3600) m³/s
= 0.326 kW
Shut-off ΔP = 1.20 × ΔP_rated (HI 1.3 typical for centrifugal absent vendor curve)
= 1.20 × 0.951
= 1.141 bar
If you (Korbut) intended the discharge values to equal the suction
values for some reason not visible in the case spec (e.g., a
deliberate convention that "both sides use the worst-case section"
for a first-pass estimate), let me know and I'll discard
Outputs_corrected.xlsx. Otherwise this should be the working
reference until you re-certify.
Outputs.xlsx— your original certified workbook, UNCHANGEDOutputs_corrected.xlsx— proposed corrected version (this note's diff)Test Case_v1.xlsx— the case inputs (unchanged)Test Case_v1_filled.xlsx— standalone Opus's filled answer setgrading_report.md— per-cell grading of the standalone run vs original Outputscorrections_note.md— this file