Portfolio Project

• Arezoo Alipanah

• www.arezooaalipanah.github.io

• 2018 , 2019

Bearing faults is one of the essential issues that may cause a big catastrophe in factories. It is imperative to detect bearing faults before they cause these catastrophes. To do so, bearing signals need to be collected and studied carefully to detect the size of the defect, its place, and its progress. Using machine learning to detect these faults is very helpful since mostly the faulty bearing's signals might follow the same non-linear function that needs to be detected and distinguished. In my thesis, this has been done, and with an accuracy of 95% using the random forest and other methods via Matlab, it is possible to predict bearing faults with much more confidence.