Temperature-constrained pulse-power limits in parallel microchannel cooling with unresolved flow maldistribution

Numerical code and data supporting the manuscript "Temperature-constrained pulse-power ceilings from sampled flow allocations in parallel microchannels" by Sihan Wei, Inner Mongolia University of Science and Technology. This revision is identified by the exact-version DOI 10.5281/zenodo.22995227. It retains the three original archives from DOI 10.5281/zenodo.22986619 unchanged and adds manuscript_revision_corrections.zip. The correction addendum identifies superseded results and interpretations; reproduction of the revised work requires all four archives. Files in this release: - code_and_data.zip: original two-channel model, retained numerical results and benchmark inputs.- array_extension_code_and_data.zip: four- and eight-channel models, allocation scans, wall fields and numerical checks.- review_response_and_experiment.zip: review-stage sensitivity studies and comparison with digitized temperatures from a published copper test section.- manuscript_revision_corrections.zip: corrected calculations, expanded evidence, plotting and document sources, revised documents, a correction README and file hashes in CORRECTION_MANIFEST.json. The correction archive replaces the two-pitch heating calculation, which previously mixed axial source profiles, and the liquid-capacity ablation, which previously inferred material identity from heat-capacity magnitude. It also corrects the conductance-floor interpretation, sensitivity directions, calibration descriptions and measurement-value interpretation. The periodic exchange-closure calibration is a separate fit to the periodic uniform response, not a transfer of the continuous-case multiplier. The correction README names the affected records and gives reproduction commands. Additional evidence comprises 16 constrained steady optimization runs, numerical comparison bounds, and 37 retained field/history cases including transverse-conduction ablations and interior pulse probes. These studies support the stated finite-sample ceilings and conditional state-ordering mechanisms. An eight-channel interior pulse probe slightly exceeds the balanced response; the work does not claim an exact continuous global optimizer. The primary numerical arrays are retained. Figure redesign, rounding notes and DOI updates do not alter the numerical results. No new experiments were performed. The conjugate benchmark uses attributed published numerical curves; a separate comparison uses eight digitized temperatures from a published experimental copper test section. That comparison does not validate the silicon arrays or establish hardware ratings. Sensitivities and cross-model discrepancies are not a statistical uncertainty budget. Third-party full-text articles are not redistributed. The three original archives retain their historical documentation; the correction addendum specifies which records and interpretations are superseded. Manuscript, DOI and publication-status statements inside the correction archive are a pre-publication snapshot. The current submission cites DOI 10.5281/zenodo.22995227; its DOI update does not change the archived scientific evidence.