Evidence-Driven Reproducibility Audit for Asset-Level PV Assessment in Distribution Networks

Archived distribution studies are difficult to reuse when source calculations, public asset records, processed outputs, and proprietary model files support different levels of inference. This study develops a claim-specific audit for incomplete distribution network archives. Three evidence gates distinguish documentary consistency, asset attribution, and decision-specific adequacy; four archive stages separate preservation and integrity documentation from executable import and numerical reproduction. The method was applied to three retained operating points of a 13.8 kV feeder in Ambato, Ecuador. Eight roofs associated in the worksheet with T54 (code 21076) yielded 205.364 kWp from 1506 m² under the retained sizing assumptions. A public EEASA GIS record corroborated the code, 125 kVA rating, and 13.8/0.22 kV nominal voltages, while the physical-to-model correspondence remained unverified. In the maximum-PV state labeled 1.5825 MWp, the archived feeder-head export was 0.670 MW, and the archived T54 loading metric was 156.915%, triggering a nameplate-based review without establishing thermal overload. Only T54 satisfied the four-element documentary attribution rule; model execution and numerical reproduction were not tested. This contribution is a domain-specific synthesis of established provenance and reproducibility principles that makes claim boundaries explicit. It remains a single-case demonstration, without empirical evidence of improved analyst agreement or decision quality.