Back to AccuPredict Services Pte Ltd
Profile photo

Published at Monday, October 18, 2021 7:44 AM on the AccuPredict Services Pte Ltd organization's page

Predictive Maintenance in Industry 4.0

It is a truism to say that the pace of innovation is accelerating across industries with the rapid introduction of new technologies. The term ‘exponential change’ stated as Moore’s law, initially applied only to the number of semiconductors embedded on a chip. Increasingly however a similar pace of change is being seen in several fields. A new term ‘Industry 4.0’ has entered the lexicon. It refers to the combination of technologies that are transforming the world of manufacturing.

Revolutionary progress of Technologies:

The world of machines is being transformed by technological progress in a number of fields. Four that stand out are:

  • Electronics: Progress in electronics means we now have miniaturized components with growing capabilities.

 

  • Communications: WiFi capabilities have improved and transmitters can be incorporated into miniaturized components. Ample high speed bandwidth at low cost is available these days in most parts of the World.

 

  • Information Processing: With advent of Cloud, ample storage and processing capacity is available from a number of providers.

 

  • Artificial Intelligence/ Machine Learning: Progress in Computer Science/ Mathematics enables the creation of powerful algorithms that can quickly analyse data and project trends into the future.

Let’s take a slight detour before we delve deeper into how these four come together to enhance manufacturing capabilities.

Evolution of Maintenance Strategies:

Since the start of Industrial Revolution, manufacturing has been under pressure to improve Asset Utilization and thereby improve the bottom line. Maintenance has been focused on keeping machines running while avoiding actions that cause costs to spiral out of control. Three distinct approaches to Maintenance developed as technological capabilities evolved.

  • Reactive Maintenance: In the initial days, neither was equipment very reliable nor did operators have the data on failure rates. This led to equipment being run till components failed leading to the term ‘Breakdown Maintenance’ being also used to describe this mode of operations. This of course led to missed manufacturing schedules, higher cost of spares and reduced equipment life.

 

  • Preventive Maintenance: As data on component wear was compiled, operators began to adopt a strategy of replacing components in anticipation of failure. While this ensured equipment did not fail, maintenance costs went through the roof as a result of components being replaced prior to end of operating life. Today this remains an acceptable operating strategy only for mission critical equipment like aircraft engines.
  • Predictive Maintenance: Technology improvement enables accurate understanding of the deterioration of key components; its cause and the Mean Time before Failure (MTBF).  We can also monitor various rotating & reciprocating components 24X7 from a remote location ensuring machines run forever within operating conditions.

How does Predictive Maintenance work?

With the improvements in the four technologies we spoke about earlier, we have today miniaturized sensors that can be installed on key components of equipment. These come with WiFi built in and hence can transmit the vibration signal to a remote location via the internet. Cloud based servers can process the incoming information from all the sensors to project the condition of the equipment into the future and predict approaching failure with a high degree of accuracy. In turn that enables the operator to take corrective action in time to prevent the failure from happening.

Two approaches to Predictive Maintenance:

There are two emerging approaches to prediction:

  • Pattern Matching: Algorithms combine the vibration signals from the numerous sensors on equipment to create a vibration ‘pattern’ that characterizes the initial state. As components deteriorate and approach failure, the signals change. Machine Learning algorithms are able to pick up this change and highlight an emerging failure. This methodology does not demand deep understanding of component behaviour. However it also means that the method lacks the ability to pinpoint exactly which component is failing and what the corrective action needed is.

 

  • Individual Component Trending: This method looks at the individual components and the trend in their vibration. Machine fundamentals define how a component approaches failure - the vibration pattern and the frequency at which this vibration occurs. It is therefore possible to predict not just the time to failure, but identify the failing component and cause of the failure. By picking up these emerging patterns early, it is possible to anticipate failure much ahead of its occurrence and with a high degree of reliability. The biggest challenge in adopting this method is the deep technical expertise needed for adoption.

Benefits of Predictive Maintenance:

Whichever strategy a manufacturer adopts, one thing clear is that life on the shop floor will never be the same again! Here’s how predictive maintenance benefits operations:

  • Improves equipment productivity: By accurately anticipating emerging failure, operators are able to keep machines running within operating limits without experiencing unplanned downtime. In addition, machines that need to produce high quality components, experience a significant reduction in rejections as a result of eliminating excess vibrations.

 

  • Saves investment in emergency spares: Predicting upcoming failures in time enables engineering teams to order spares as needed and thereby reduce the need to stock up on spares to meet ‘what if’ scenarios.

 

  • Reduces need for skilled engineering talent: Hiring & retaining skilled talent to work on the shop floor is becoming a serious challenge particularly in developed countries. Eliminating emergency actions, enables a planned reduction of engineers needed on the floor as tasks are offloaded to the backend.

 

  • Enhances equipment life: Equipment operating within its design limits enhances its life. In effect Predictive Maintenance is a key contributor to an Organization’s ESG strategy by eliminating the need for early replacement of equipment.

 

So what are you waiting for? Take the plunge! Make a break with the past!! Predict the Unpredictable!!!