Society & Economicspreprint2026-08-09

Organisational Trajectory Classification in PRONOSTIA Bearing Degradation: An Exploratory Phase-Space Analysis

Open access0 citations

Abstract

Version 2 — Retrospective reassessment and corrected interpretation This version records the outcome of a retrospective reassessment of the exploratory PRONOSTIA bearing-degradation analysis published in Version 1. The original analysis identified reproducible differences in derived trajectory and feature-space behaviour within the PRONOSTIA archive. Subsequent testing confirms that some of those numerical and descriptive differences are genuine properties of the analysed data. However, the retrospective reassessment does not support the stronger interpretation that these differences establish distinct underlying degradation-state or organisational trajectory classes. In particular, the original move from observed feature-space separation to claims about what kind of physical or organisational states those differences represented was not independently established by the evidence. The reassessment tested the historical separation against bearing identity, operating condition, temporal progression, amplitude and texture information, resistance-label construction, within- versus between-bearing variation, and prospective held-bearing discrimination. The resulting evidence supports a narrower interpretation: measurable differences and separations exist within the PRONOSTIA data, but their physical or organisational meaning is underdetermined by the available evidence. The original classification language should therefore be treated as exploratory rather than as an established degradation ontology. Version 1 is retained unchanged as part of the scientific record. Version 2 preserves the reproducible empirical observations while correcting the interpretation placed upon them. This version is provided to maintain a transparent audit trail between the original exploratory analysis and the subsequent controls used to reassess its conclusions.

// Source

View paper (DOI)Open access versionOpenAlexZenodo (CERN European Organization for Nuclear Research)Published 2026-08-09

Authors: Robert McElhinney