A Spatio-Temporal Attention Mechanism Based Approach for Remaining Useful Life Prediction of Turbofan Engine

Table 1

Data set introduction.

Sub-datasets	FD001	FD002	FD003	FD004
Engine units in the training dataset	100	260	100	249
Engine units in the test dataset	100	259	100	248
Fault modes	One (HPC degradation)	One (HPC degradation)	Two (HPC degradation, fan degradation)	Two (HPC degradation, fan degradation)
Conditions	One (sea level)	Six	One (sea level)	Six
Training samples	17731	48819	21820	57522
Test samples	100	259	100	248