How to draw an Owl or implement Predictive Maintenance
Arun Gowtham
Asset Management Consultant | Reliability Engineering | Predictive Maintenance
I was recently reading Seth Godin ?? The Practice when an idea in the book struck a chord. It’s the classic meme instructing how to draw an owl in just two steps. Step 1: Draw two ovals and a line. Step 2: Owl.?It's frustratingly funny!
It pokes fun at the hidden complexities & lack of clear instructions to do a task successfully. An old comic book instruction manual is cited as the reference, but the truth of that meme is still prevalent in all sectors. Take equipment Predictive Maintenance, for example. I wanted to learn how we can leverage Machine Learning (ML) to accurately predict equipment failure, in advance, for maintenance planning. As I spend more time researching this topic for my work at Owtrun, the more questions I'm left with. Reference texts explain the theory of failure mechanism but not the practice steps for a maintenance technician. Academic papers go into the mathematical models of failure and are specialized for a novice reader to understand equations & citations. Training classes are prosaic with few default exercises; scaling up the application requires more training classes. Commercial products & services are marketed with promising headlines that wall off the minutiae behind the sale. Some resources offer structured PdM frameworks which are very useful. These encounters made me wonder, why do I see patterns of the 2-step owl meme? Some of the possible reasons:?
领英推荐
For the case of ML-based Predictive Maintenance, all the above cited are reasons that we do not have a thoroughly detailed guide. In my opinion, the order of reasons above is also ranked from most probable cause to least. Seems like experience is the only way to learn, but taking help can accelerate the process. My work is now focused on helping others to (draw an owl) implement Predictive Maintenance. This step-by-step guide details How to implement an effective PdM program. It may not be granular for some readers. Hoping the above-hashed reasons offer some clue.