Applications of Machine Learning Techniques in Asset Management of Engineering Systems: A Bibliometric Analysis

Abstract

This bibliometric analysis systematically maps the scientific landscape of machine learning (ML) applications in asset management for engineering systems. It uses a corpus of 1,158 publications from 2005 to 2025. The study's methodology is based on a comprehensive asset lifecycle perspective (ALC), aligning findings with the strategic clauses of the ISO 55001standard. The analysis reveals exponential growth in publications since 2018, confirming the field's rapid maturity and relevance. Keyword co-occurrence analysis identified seven thematic clusters, with the intellectual core overwhelmingly concentrated on the operational phase (ISO 55001, Clause 8) and specifically on predictive maintenance (PdM), fault detection, and diagnostics (Clusters 2 and 3). Citation burst analysis reinforces this focus, highlighting systematic reviews on PdM as the most influential literature defining the current research frontier. The study also identifies significant research gaps in the holistic application of ML across the entire ALC. strategic phases such as planning, acquisition, operation and disposal (including value recovery and end-of-life management) are underrepresented in the core literature. These findings suggest that, although ML provides a highly developed set of tools for technical operations, future research must shift its focus strategically to support the strategic objectives of the ISO 55001 framework. This will facilitate a complete, value-driven ALC management system.