Diagnosing short-term pit–crusher decoupling from routine production reports: A source-audited field study

Routine production reports separate pit haulage, equipment hours, material classes and crusher feed, potentially obscuring the source of short-term losses. We developed a source-audited diagnostic workflow using 19 consecutive daily report packages from an anonymised open-pit operation. Reconciliation provided 38 pit shifts, 18 daily fleet-hour observations and 34 matched pit–crusher shifts. Same-shift total loads and crusher tally were weakly associated (Pearson r = 0.108, p = 0.543); ore-based proxies showed similarly weak relationships. Threshold screening identified five discordant shifts, three of which met a stricter median-absolute-deviation (MAD) criterion. In the strongest upstream case, low fleet-hour coverage coincided with near-median crusher feed, while loads per active truck-hour remained comparable to reference shifts. In three downstream cases, reduced crusher tally coincided with maintenance/STOP annotations while pit haulage continued. These annotations provide internal documentary concordance, not independent validation. A nine-day single-truck outage offered asset-reliability context without a clear fleet-hour or night crusher-feed collapse. Under hypothetical input changes of up to 20%, four events retained their threshold classifications, although the upstream MAD-core event lost that status at 10%. The workflow identifies recorded pit–crusher mismatches without establishing inventory buffering, material rerouting or other causal mechanisms.