Abstract The performance change of Ethylene propylene diene monomer (EPDM) was studied by the accelerated aging experimental method, and its lifespan predicted. results showed that thermal oxidation EPDM is dominated crosslinking reaction. With extension time, total density tensile permanent deformation gradually decrease, while strength hardness increase gradually. life prediction formula unde...

Canonical deep learning-based Remaining Useful Life (RUL) prediction relies on supervised learning methods which in turn requires large data sets of run-to-failure to ensure model performance. In a class cases, is difficult collect practice as it may be expensive and unsafe operate assets until failure. As such, there need leverage that are not but still contain some measurable, thus learnable,...