Industry 4.0 & Smart Manufacturing
How organizations optimize factory productivity with predictive maintenance
Keith Higgins, vice-president of digital transformation at Rockwell Automation, discusses what's held manufacturers back when it comes to predictive maintenance
July 9, 2020 by Keith Higgins
In 2016, PwC reported that over 70 per cent of manufacturing companies planned to significantly increase investment into digitization efforts over four years – a combined financial commitment of over $907 billion, or roughly five per cent of revenues.
Today, we’re reaching a critical inflection point for wide-scale rollout of digital transformation in the manufacturing segment, as organizations are progressing from pilot or proof-of-concept IIoT projects to scalable IIoT deployments, according to the Global IoT Decision Maker Survey from IDC.
Over 30 per cent of those surveyed said they’ve already launched IIoT solutions, and over 40 per cent said they’re looking to deploy solutions in the next 12 months.
As companies are moving into this next phase of digitization and looking to benchmark their state of digital operations, we are observing a shift in the perception of digital transformation projects. Beyond exploring the benefits of data-driven solutions, organizations are moving toward understanding how these projects can be leveraged to most optimally scale smart factory initiatives.
To realize the most compelling business outcomes, enterprises are prioritizing which use cases are most critical to overall digital transformation success and deliver the highest ROI.
While many different use cases have been identified and put into production, predictive maintenance has emerged as one of the most common. Here are some important questions about predictive maintenance.
What can predictive maintenance accomplish in an industrial setting?
Survey data from Information Technology Intelligence Consulting highlights that a single of machine downtime can cost organizations between $100,000 and $5 million, depending on the industry and use case scenario. Unplanned downtime costs industrial manufacturers about $50 billion annually – and poor maintenance strategies reduce overall factory productive capacity up to 20 per cent.
To prevent costly and disruptive downtime, industrial organizations are leveraging predictive maintenance to identify potential issues, reduce the occurrence and length of equipment interruptions, and get the most value from assets and budgets.
Predictive maintenance capabilities help inform operators how and why a machine is degrading, allowing operators to conduct necessary, specific maintenance rather than reacting to machine failures or wasting time on unnecessary repairs. Predictive maintenance equipment studies patterns that precede the downtime events identified in an organization’s maintenance history, then trains agents to recognize those same patterns in the future.
As new data is generated, machine-learning agents offer around-the-clock tracking of all live sensor data, looking for the trends identified. Additionally, agents can watch for atypical patterns that may represent new failure modes to be investigated.
The benefits of predictive maintenance solutions are palpable. What held us back in the past from achieving it?
Manufacturers are under constant pressure to keep up with increasing product complexity while reducing costs. Unplanned downtime and unexpected capital expenses add to this pressure and are very disruptive and costly. Mechanical components degeneration over time can drastically alter machine performance and lead to sometimes catastrophic failures.
Unfortunately, gradual deterioration is challenging to track manually. Without the ability to anticipate equipment failures and identify root cause, you lose significant amounts of time and money.
With the rise of IIoT adoption over the last few years, data from connected equipment, lines, processes and facilities is pouring into factories. The information locked in these data streams has transformed how organizations manage operations, solve issues and adapt to change.
To implement successful predictive maintenance initiatives, machines, devices and their operators must communicate seamlessly. OT and IT data has proven to be much less effective when placed into silos; however, the industry lacked purpose-built capabilities to marry IT/OT data and users accordingly – until recently.
Modern digital transformation solutions now connect and contextualize data between OT and IT and provide industry organizations with an improved understanding of overall factory operations, leading to proactive factory maintenance, rather than reactive.
Integration between OT and IT, made available through increasing IIoT implementations, is the key to deliver new levels of predictive insight for achieving productivity gains at scale.
Can you give us an example of a predictive maintenance scenario?
Let’s highlight a real-world story that the Rockwell Automation team sourced from an employee at a tissue manufacturer. The organization used an extensive coordinated drive system with high-power drives to run a tissue machine.
At the end of the line, the tissue was wound into a large roll. After about a decade, they started experiencing drive modules faulting, and in some cases, failing.
When a drive on the machine faulted out, and the roll was no longer being driven, the tissue ripped, creating a mess of tissue until the machine was stopped. Then the tissue had to be cleaned up, and the machine rethreaded – slowly restarted by pulling the tissue through all the rolls on the machine.
The faulty drive module had to be troubleshooted and replaced. This caused several hours of unplanned downtime, lost production, and a significant, unexpected expense.
Instead, by implementing predictive maintenance capabilities, organizations can monitor and analyze asset condition to get alerts about its productivity levels, power consumption, health status and internal wear. Manufacturers can minimize production defects by predicting when the number of defective products is likely to exceed a threshold percentage and provide the root causes for the expected failure.
Companies can leverage predictive analytics to prevent significant unplanned downtime. For example, a connected system can give the maintenance team a notification to replace a drive module during the next scheduled outage, avoiding unscheduled downtime.
What do you see the future of predictive maintenance becoming?
Unscheduled downtime is one of our customers’ top threats to maximizing revenue. Machines equipped with predictive, as well as prescriptive analytics capabilities, help manufacturers avoid this critical risk through improved maintenance.
Even more valuable than knowing when something is about to fail is knowing what you need to do to fix it. Prescriptive analytics dashboards use historical failure data and trends to tell workers what corrective action to take.
This information can help make sure the right steps are taken to avoid asset failures, keep production on target, maximize quality, and more.
Predictive analytics dashboards can be customized to deliver the specific information that workers need to solve different problems. For example, a worker in a food manufacturing plant can gain insight into mixing, filling, and packaging assets on a line all in one place.
In summary, industrial organizations are implementing predictive maintenance into their digital transformation strategies to lower service costs, maximize uptime, and improve production throughput.
With the IIoT expected to add $15 trillion of value to the global economy by 2030, we can expect industrial enterprises to adopt increased digital transformation efforts throughout this new decade to improve efficiency and implement practical solutions to industry challenges such as factory productivity.
Keith Higgins is vice-president of digital transformation at Rockwell Automation.