.png){kind=icon}
.png){kind=icon}
Return
Remaining useful life aware physics-informed Bayesian digital twin for safety-constrained control in robot-integrated battery manufacturing
.png){kind=icon}
PRE
AI
A
U
S
N
F
DOI：10.1016/j.rcim.2026.103318.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Robotics-integrated digital twin converts calibrated RUL into chance-constrained actions (continue/derate/abort). • Physics-informed prognostics (wear/thermal) with Bayesian UQ is converted into decision-ready hazard–survival RUL. • Shift-aware calibration (drift detection + conformal quantiles) maintains coverage under regime change. • MPC-guided RL with a conservative-quantile safety shield; edge-deployable with PLC/ROS interoperability. • Industrial evidence on an EV battery line: Performance gains in downtime, OEE, and maintenance cost.
Keywords:
Remaining useful life
Digital twin
Bayesian uncertainty quantification
Safety-constrained control
Robot-integrated manufacturing
Journal
R
IF:
11.4
Papers:
3.3K
Citations:
1.3W
Organization
Cited Papers