Hi @rahulmen
Welcome to the community! Very good question.
Upfront: here a general outline which might be worth to take a look at:
see also: Deep learning is a small part of ai - #6 by Christian_Simonis
So I guess you have an algorithm in mind, I guess something like a threshold check or so, right?
My understanding is:
Anomaly detection serves a use case which you can often solve with AI resp. ML like in this early warning system.
But in general, there are also alternative methods like using classic (physical) domain models or simple threshold checks. The suitability of the solution often depends on the complexity of the use case, the availability of (training) data and also the deployment scenario among others.
Hope that helps.
Best regards
Christian