Simulation-Based Evaluation of a Hierarchical Decision Rule for Standby Fire-Pump Condition Assessment

Standby fire-pump monitoring must distinguish supported abnormalities from insufficient measurement evidence. This simulation study evaluates a hierarchical decision rule that combines diagnostic probabilities with measurement-quality checks and compatible normal references. The rule returns normal, abnormal, or indeterminate states and records the domains supporting an abnormal decision. On 1800 fresh synthetic windows, it accepted 1365 with one observed binary error, compared with 1021 accepted windows and no observed errors under strict single-domain corroboration. Hydraulic evidence supported acceptance in 259 of 360 hydraulic-labeled windows (71.94%); the broader acceptance count included cases supported only by other domains. The main constraint was deferral of 240 of 360 normal windows (66.67%). At a hypothetical 5% abnormal prevalence, the expected deferral was 640.1 windows per 1000, representing unresolved windows rather than measured inspection time. The results show how relaxing single-domain corroboration changes acceptance and the reported evidence within the specified generator. Evaluation on synchronized pump measurements and recorded inspection outcomes is needed to determine whether the resulting verification requests are useful in practice.