We report that continuous monitoring of a thermodynamic stability metric, Phi = I × rho - alpha × S, detects bearing failure risk substantially before end of life in all 10 tested XJTU-SY run-to-failure trajectories. With a fixed coupling constant alpha = 0. 1 and critical threshold Phic = 0. 25, the metric crossed the critical threshold at an average of 86. 1 percent of bearing lifetime remaining, with a minimum of 73. 2 percent and a maximum of 90. 4 percent, across lifecycles spanning 42 to 2, 538 measurement files. No per-bearing training, no failure history, and no threshold adjustment were used. We also report two narrower extensions. First, the same formula and threshold correctly classified a cross-domain subset of 11 systems across mechanical, aerospace, and geophysical domains with no threshold adjustment between domains. Second, we derive and validate an inverse design constraint from the stability condition on the same 11 systems, with the caveat that only one stable case was available for the inverse design validation. All experiments use real published datasets with no synthetic data. Negative results and boundary conditions are reported throughout. The methods described in this paper are the subject of U. S. Provisional Patent Application No. 63/960, 829 (filed January 15, 2026). No license to implement or commercialize the described methods is granted by this publication. All rights reserved.
Shawn Barnicle (Fri,) studied this question.