Downtime, or a halt to production caused by catastrophic machine failure has a very negative effect on any company’s services and finances. Also, it can have a knock-on effect on its reputation.
That’s why more and more companies are opting for predictive maintenance, a system that anticipates possible incidents and prevents them before they happen.
Thanks to the interconnection of systems associated with IIoT, Machine Learning, Big Data and Business Intelligence, this approach to maintenance is no longer a pipe dream, it’s happening, and it’s more efficient and cheaper than other systems, such as preventive or corrective maintenance.
Table of contents
What is Predictive Maintenance?
Remember ‘Minority Report’? The film starred Tom Cruise and featured people with the ability to foresee the future, who were used to predict crimes; that’s more or less what predictive maintenance does.
But this time round there aren’t any people with ‘visions’, this system is fed by a mass of data gathered from sensors, machinery, etc. to identify patterns and anticipate possible issues.
Smart sensors make up the first line of defence against system failures: their job is to constantly transmit information about how each device is working, using wireless communication protocols.
Data transmitted to specialised software enables you to feed an IT ecosystem that seeks and finds patterns and learns from previous incidents to detect possible failures long before they happen, and then activate procedures to prevent incidents.
As we explained in a previous article,the convergence of Operational Tech (OT) and Information Tech (IT) in end to end solutions enables a flow of comprehensible data to be produced that we can use to prevent catastrophic machine failures and so provide a more efficient production process throughout the value chain.
Origins and history of predictive maintenance
Maintenance strategies originated in the First Industrial Revolution, when corrective or reactive systems began to be used. These later evolved into preventive and predictive approaches.
Corrective maintenance
This system also goes by the name of ‘run to failure’, and consists of acting only when the machine breaks down. It’s the system that eats up the fewest resources in planning, but it’s also the most hazardous: it puts safety, production, product and service quality at risk, and can even jeopardise a company’s survival.
Preventive maintenance
This system schedules maintenance according to a set of theoretical calculations of how long components last. This particular approach is based on regular events: service times, kilometres run, tons produced, etc., and on the surveillance of specific conditions: pressure, temperature, speed, tension, vibration levels, etc.
It was born in the 1930s and puts a lot of emphasis on reliability. It sets out to reduce the likelihood of breakdowns or loss of performance in a machine or facility: it tries to make the whole thing more reliable.
The main failing with this system is that it can lead to a ‘better safe than sorry’ approach. For example, components can end up being replaced ‘because the maintenance schedule says so’, when we don’t actually know if it’s really necessary. And that can lead to unnecessary investments in spare parts and labour.
Predictive maintenance
The aim with this type of maintenance is to anticipate a failure before it happens: intervening at the first signs of any symptoms to prevent an incident.
It began to be used in the 1990s. The system is based on a constant assessment of a facility’s status, backed up by information technologies. The first CMMS (Computerized maintenance management systems) had close ties to predictive maintenance.
It’s not a new approach. But thanks to the capacities offered by IIoT it can enter a new dimension: it’s cheaper and much more precise.
How predictive maintenance works in IIoT
The key to predictive maintenance is compiling, analysing and interpreting data to find patterns before a potential failure happens.
All the devices involved in the product life-cycle form part of the IIoT system. They are all interconnected by digital threads while the starting point for gathering the data is smart sensors , enabling data to be gathered and entered into the system in real time.
Specialist software analyses and interprets the data and immediately sends information to maintenance supervisors when it detects a pattern that hints of a potential incident. It can even automatically intervene to correct system operations.
So now we can go two steps further: one is communication protocols established to really speed up response times, while the other is highly efficient intervention protocols. The outcome of all this is improved overall reliability.
Will I get to eliminate downtime with predictive maintenance?
Taking downtime out of the equation is well-nigh impossible. But what you can do is reduce unscheduled stoppages by 99.99%. And keep under control all the internal failures.
Because when you apply predictive maintenance strategies, you get to transform your plant’s maintenance SWOT: all the Weaknesses (negative internal factors) are turned into Strengths. What you will have to do is keep a close eye on the Threats (the negative external factors) so you can turn them into Opportunities.
Benefits of predictive maintenance
Optimised maintenance management in industry offers many competitive advantages within Industry 4.0.
Companies that have opted for digitalisation and evolved their production plants towards smart factory models already have ideal ecosystems for seamlessly integrating predictive maintenance strategies.
And when you use a predictive maintenance strategy adapted to production cycle needs, the benefits just don’t stop coming.
Benefits of predictive maintenance
- Improved overall reliability. Thanks to reduced failures in general, especially catastrophic ones.
- More productivity. Because machinery is more frequently available
- Cost reductions. In spare parts. In industrial insurance.
- Reduced maintenance interventions.
- Increased product and service quality. Take all this together, and you have the added plus of improved company reputation.
- Reduced number of accidents and increased plant safety. Thanks to constant monitoring.
- Edge over your competitors Or a loss of that essential edge if you don’t give it a try.
Lower investments and quick returns
Companies that haven’t moved forward with the digital transformation process but are still just getting by should get moving pretty quick. According to the “Predictive Maintenance” report by Deloitte ,“On average, predictive maintenance increases productivity by 25%, reduces breakdowns by 70% and lowers maintenance costs by 25%”.
And while downtime is a threat that we’ll always have to find ways to cope with, the report also says that “predicting failures via advanced analytics can increase equipment uptime by up to 20%”.
Companies that disregarded this type of investment because of the costs involved, ought to take another look at their calculations… and consider an adequate predictive maintenance strategy.
Because now, detecting an anomaly before it happens and so prevent incidents that can cause unscheduled stoppages isn’t a fantasy any more.
