Skip to main content

Hello all,

A lot of development happening in MEMS-based low-cost sensor, wireless technology, edge processing and AI/ML-based analysis. There are some products and solutions that talk about automated fault diagnosis (even prognosis) of rotating machines. Most common is the vibration-based solutions. We know that vibration signature analysis helps us in finding machinery faults, but can we automate the diagnosis process ?

I have following queries related to new developments in asset health monitoring area:

1. If you have used any automated fault diagnosis/prognosis system, what was your experience on accuracy ?

2. Can we set similar rules/logics for different machine configuration ? I guess, number of rules will be proportional to number of machine types.

3. Wireless vibration sensors do not (not sure) provide phase information. Can we still perform fault diagnosis ?

I posted similar question on AI and Machine Learning forum, but did not get any response.

Looking forward for your views.

Tags: Machine learning, Automated fault diagnosis

Replies sorted oldest to newest

These systems are very expensive and almost for no use, I can tell you that they cause more problems then before. Sellers will tell you the opposite of course. AI level, wireless capabilities, storage capabilities, high price, educating staff, staff resistance.....20-30 years too soon as electric cars are. Now we see the car batteries are charged with diesel generators. This is no more then commercial fraud. I am writing from my own experience but won't go to details as company names etc.

Last edited by Registered Member

Add Reply

Link copied to your clipboard.