A physics-informed recurrent neural network for long sequence remaining useful life prediction of rolling bearing with implicit function

Key Points

• The physics-informed recurrent neural network significantly improves remaining useful life prediction accuracy for rolling bearings.
• Key evidence shows a reduction in prediction error by 25% when using this model over traditional methods.
• Assessment using a physics-informed modeling approach highlights the effectiveness of incorporating physics in neural network structures.
• Further validation in real-world settings is needed to generalize findings beyond the initial dataset.