Unsupervised condition change detection in large diesel engines

This work presents a new method for unsupervised change detection, which combines independent component modeling, and probabilistic outlier detection. The method further provides a compact data representation, which is amenable to interpretation, i.e., the detected condition changes can be investigated further. The method is successfully applied to unsupervised condition change detection in large diesel engines from acoustical emission sensor signal and compared to more classical techniques based on principal component analysis and Gaussian mixture models.