“Don’t fix what ain’t broken,” the cliché runs. But anyone with expensive assets to maintain – from precision tooling machines to a fleet of jets – knows that you don’t fix only the things that are broken. To minimize downtime, it’s important to regularly check your assets’ condition, tune them up periodically and replace the supplies they consume.
The problem is, routinely checking asset conditions on a pre-set schedule doesn’t necessarily respond to the conditions at hand. Some machines need more frequent tuning and adjusting, while others just run and run. And wise equipment owners will plan accordingly, focusing maintenance efforts where they’re most needed without wasting time on unneeded routine checks.
That kind of insight used to be embodied in a seasoned field engineer or a parts clerk who kept careful records and could predict failures. But with today’s digital manufacturing processes, where a single chemical plant could have thousands or even millions of controls, and sensors all connected by the Internet of Things, neither “just fix what’s broken” or scheduled maintenance can keep up.
Predictive maintenance helps you focus on what makes a difference
Asset performance optimization with predictive maintenance is the art of collecting the digital output from your machines and processes and analyzing it to track their performance and predict future needs using artificial intelligence. Effective asset performance optimization will also analyze unstructured data, such as written field notes or the text of maintenance manuals, to add relevant context to your analytics findings. And it’ll report the findings in real-time dashboards that are easy to check at a glance.
This kind of AI-enriched insight delivers:
- Smoother, more predictable operations because problems can be addressed pre-emptively, not after a disruption
- More uptime, which equals more revenue opportunities
- Happier customers, since there’s less risk products will have production flaws or be delivered late
- Lower repair and maintenance costs and labor overhead
- Higher safety levels, as the need for service can be detected before a failure occurs
- Less risk of litigation or penalties for unscheduled failures
In fact, predictive maintenance typically reduces machine downtime by 30-50 percent and increases machine life by 20-40 percent.1
Recent estimates suggest it can save 15-20 percent of maintenance costs for industrial machinery and transportation. For cases such as engine oil changes, it could reduce the overall cost of failure by up to 67 percent. Considering that industrial machinery repair costs $36 billion a year just in the U.S., predictive maintenance offers large potential savings.
To unlock these savings and other advantages, you need the right technology solution: OpenText™ Magellan™, the powerful, flexible AI-powered analytics platform.
Asset performance optimization (APO) from OpenText leverages the Magellan platform to deliver predictive insights derived on a holistic knowledge base, with the ability to improve those insights over time as assets are used, maintained or taken out of service. It powers IoT with AI capabilities to acquire, merge, manage and analyze Big Data and Big Content, including data from sensors, EIM systems and external sources. It comes with a range of APIs that easily integrate with sensor inputs and any back-office manufacturing or ERP software a customer may use. And end users can customize its built-in visual interface to provide exactly the measurements and insights they’re looking for.
How it works
Here’s one example of how asset performance optimization works in a steel “mini-mill.”
Sensors at each step of the steel production process stream information from IoT sensors via the OpenText cloud to Magellan. Let’s say there’s a problem at Point A with the rollers that press a slab of freshly poured steel into thin, flexible, precision-cut sheets; the sheets are coming out too irregular for use.
OpenText software can check the data streams from the relevant components (Point B), ranging from the crucible melting and pouring out the steel to the laser-guided blades that trim the steel roll, or microscopic variations in the roller height. It can compare the current condition against historical performance records and associated data.
Magellan can digest and aggregate a wide range of data inputs, including:
- Running hours for a given piece of equipment
- Historic roller failure rates
- Inspection schedules
- Supplier quality records
- Text of operations manuals.
It discerns patterns that lets it ask questions to guide users towards making well-informed, timely decisions, such as:
- Should we shorten the inspection intervals?
- Did a contaminated batch of scrap for melting slip through inspections?
- Can the design of the rollers be improved to reduce future failures?
And then it makes a predictive conclusion based on these patterns:
- Raw materials from a certain supplier turn out to have a chemical composition that changes the steel’s melting point, making it prematurely cool and turn brittle.
That’s where people step in to make the fix (Point C). The engineers can view a digital model of the production line hosted in the OpenText Cloud and determine how to reconfigure it, or negotiate with suppliers for better-quality metal. The same way, if the sensors spot a part that’s about to wear out, the mill manager will know to order a replacement right away.
This broad-based data-gathering and predictive analysis can be applied to nearly any machine-intensive industry. Instead of a steel mill, picture a city bus or a power plant. Asset performance optimization can deliver value in any situation.
Asset performance optimization at any scale pays off
Where you have large numbers of low-cost assets in heavy use, such as espresso machines, it saves money by focusing maintenance techs’ attention on the machines closest to failure, rather than performing unneeded inspections on machines that don’t need it. This improves uptime, boosts revenue opportunities, and increases customer satisfaction.
Asset performance optimization from OpenText also works for public and nonprofit institutions such as hospitals, utilities, and mass transit agencies, which generally have fewer assets but more expensive ones (from MRI machines to jets to natural gas pipelines).
It can more thoroughly highlight maintenance needs and target the components that most need attention, reducing the risk of outages. People are happier when they can rely on their buses or flights leaving on time, their medical appointments taking place as scheduled, and the lights staying on.
Want to learn more? Read an IDC Industry Brief on maximizing operational output with OpenText Magellan. This intriguing ebook provides a guided tour through the world of manufacturing and other industries dependent on heavy equipment and expensive production assets and explains how the Magellan APO solution can help.
———————————————
1 McKinsey & Company, Manufacturing: Analytics unleashes productivity and profitability (2017)
