Rethinking Maintenance at Sea: From Data to Decisions — Rule-Aware Prescriptive Maintenance for Explainable and Compliance-by-Design Tanker Fleet Operations

This working paper presents a practitioner-led conceptual framework for rule-aware Prescriptive Maintenance (RxM) in ocean-going tanker fleet operations. It addresses the practical maintenance-management problem of converting PMS records, defect history, condition evidence, predictive analytics, and operational constraints into maintenance decisions that are technically justified, operationally executable, and defensible to classification societies, Flag State, vetting stakeholders, insurers, and internal assurance functions. The paper argues that maritime maintenance analytics should not be evaluated only by prediction accuracy. In tanker technical management, a useful maintenance recommendation must also be explainable, reviewable, auditable, and suitable for compliance-driven operational environments. The article, therefore, positions RxM not as an autonomous replacement for marine engineering judgement, but as a structured decision-support layer connecting condition evidence, predictive insight, asset criticality, human oversight, PMS execution, and regulatory traceability. The paper introduces the Prescriptive Recommendation Traceability Index (PRTI) as a conceptual metric for assessing the quality of AI-supported maintenance recommendations. PRTI is proposed to evaluate evidence traceability, rule traceability, action rationale, human review, and compliance closure. The metric is presented as a research concept and is not yet validated as an industry standard. Further validation through PMS case analysis, expert review, and operational testing is required. This working paper is written from the perspective of tanker technical management and is intended for marine engineers, technical superintendents, tanker operators, maintenance planners, classification society stakeholders, maritime researchers, and digital maintenance technology providers interested in explainable, compliance-by-design maintenance decision systems.