With accurate inputs, we should be able to stop any shutdown
Published on : Monday 08-06-2020
Sharad Dhumane, Metallurgist and Technical Freelancer.
Industrial Maintenance is almost like a Pilot’s job. The plane would not take off without a Pilot. A well-resourced Plant will fall to its knees, in absence of a competent Maintenance crew. On the other hand, an ordinary Plant with old machinery, can turn out a commendable performance, if the crew is competent and dedicated. And it does not matter how insignificant a machine is, if it is part of an assembly line, where whole line comes to standstill, if any stage or process stops. Somewhere in the last few decades, Industries slowly woke up to the secret of effective Plant Maintenance. The well trained crew, stocks of spares, maintaining Breakdown records to know what breaks down frequently … all this became a Standard Industry practice for fighting Machine Shutdown. As Industries became more efficient, they realised that even with best Breakdown Maintenance practices and crew, they still lost valuable Production time. And time is money.
So people worked on reducing this Downtime, as it came to be known. Their work resulted in a better practice, where Machines were switched off at some intervals, preferably on weekends, to inspect, detect parts which were about to fail and replace them at a suitable time. Sometimes all it took was cleaning, greasing to prolong the failure. The crew started observing any abnormal signs, such as oil stains (indicating oil leaks), vibration (indicating wear), machine noise, etc. Also check points were planned based on the Root Cause of the Failure. This reduced the Downtime significantly and the practice became quite popular. It received a name “Preventive Maintenance”.
Conditions were now much better due to significantly low Downtime and improved levels of Confidence. However the Continuous Improvement Philosophy adopted by most well run Organisations kept up the momentum going. Failure prediction became quite accurate. So once you knew that the noisy bearing will last for next three months, you had all the comfort to plan that replacement. Maybe that long weekend vacation is a good time to do that. Couple it with the switches in furnace panel board. And there, you just did a nice little plan to avoid Breakdown of two machines, without disrupting anything. You also had the comfort of not stocking on the Bearings and switches. Since good amount of time is available, you could try to negotiate on the prices too! Predictive Maintenance was the new Strategy on the block!
But the folks didn’t stop here. Think of squeezing the lemon for that last drop? So Prescriptive Maintenance is a Technique, which collects data of an equipment, analyses it to come up with specialised recommendations. It can also predict a corresponding outcome, if this recommendation is implemented. In this aspect, it differs from Predictive Maintenance. For example, a Software predicts a Bearing failure in three month’s time, from its Temperature. It also suggests reducing the equipment speed to prolong the Life.
So if you reduce the Machine Speed, the Bearing will work for another 15 days extra. Weighing the pros and cons, you decide to do this, so that the replacement can go into the long festival vacations.
Which Strategy do we choose?
Most young companies can immediately hit the Preventive Option. The Predictive Technique would need a mature Data. So maybe after a couple of years, they can move into Predictive Technique. If a good Software is available, they can surely think of Prescriptive Maintenance at that time.
Why do Companies still opt for Corrective Maintenance?
It offers a simplicity and low Maintenance involvement. The root cause of the Failure is identified and an appropriate action is taken. The repetitive Bearing Failure is identified s due to a misalignment and is now corrected. The benefits of reduced Downtime are quite close to Preventive Maintenance and weighing pros and cons, the Plant may choose Corrective Maintenance.
Has Predictive Maintenance become easier in the new IIoT times?
It must. Smart sensors placed at the critical positions of the Machine keep sending critical machine condition information to your Computer/Mobile. So vibrations, noise, temperature data keeps coming round the clock. Smart charts let you know how these are trending and also alert you of an imminent failure. You truly can sit in a chair and know the Plant Machinery health in a glance.
So is a good Inventory Management still a requirement for a successful Story?
If the Prediction is trustworthy, then surely the Inventory Management can be diluted to a large extent. You just stock highly critical parts, which take long time to reach the Plant. Cost of the Inventory, amount of Management time it takes, space the inventory takes up in the Plant … everything gets reduced to a miniature level. Being able to work with a miniature inventory can be looked upon as an indicator of strong Predictive Maintenance capability of the Plant.
Can we aim for Zero Downtime with the aid of AI/ML?
It is logical that with accurate inputs, we should be able to stop any shutdown. Or shift the Shutdown to a timeslot, when Plant is not running. After all, a Machine shutdown is due to a specific cause and if that cause is accurately controlled, there should be no Shutdown.
Will the Maintenance Strategies change post Covid-19?
Covid-19, after its pandemic run is over, will become another run-of-the-mill virus. Its sting will no longer bite. When a vaccine gets on board, the Covid-19 will truly become a history and the World will get back on its feet. So it seems unlikely, given the way Technology has progressed, that Covid-19 will become a long time hurdle.
Sharad Dhumane is a Metallurgist (IIT Kanpur), and worked in Automotive and Pump/Valve Foundries for three decades. The responsibilities covered Operations, Product Development and Technology. Recently he has taken to Technical Freelancing to be able to spend time in areas of his choice.