If we take a look at the training data variables (acoustic_data and time_to_failure):
we can see that there are 15 segments; each segment contains a fault (where time_of_failure reaches 0).
The following images show the segments of the training data; each segment in an image. Note that all images have the same y-axis which makes comparison easier.
