Manufacturers who have deployed the digital or smart factory have put down their pencils, found new uses for their clipboards and closed their spreadsheet programs in favor of using real-time data gleaned from condition monitoring of their machinery.
Plex Systems’ Anurag Garg said there are dozens of solutions that will enable data collection easily, but the real value is in contextualizing that data.
In addition to their factory machinery, some producers are monitoring the condition of their products via the Internet of Things (IoT).
Using the IoT to monitor products in the field has enabled manufacturers to also become service providers.
In plants, sensors, data collection and analytics are helping manufacturers squeeze the most efficiency out of their operations, whether their product is paint, industrial foam, scrubbers and sweepers, tires, high-tech equipment or something else entirely. This enables a plant to adjust operations to:
The evolution of the very nature of manufacturing is driving the adoption of condition monitoring, a feature of the smart factory in Industry 4.0.
“The rules of manufacturing have changed: It used to be that success was measured by how many widgets you have made, how quickly and how cheaply,” said Anurag Garg, VP and head of analytics & IoT at Plex Systems. “Now, success is measured by your responsiveness and ability to optimize metrics. So, you’re forced to look for efficiency.”
The biggest opportunity is not just the data collection, he said. There are dozens of solutions that will enable data collection easily. But the bigger challenge is contextualizing that data. What do they mean?
A&K Finishing, a plastics paint solution provider, worked with Plex to digitize its quality-control sheets and checklists, and found its old, manual data entry was taking up 80% of some employees’ time.
Plex got A&K to a point where its executives have consistent and reliable quality data—and are able to use the contextualized insights gained from that data to halve their scrap rate.
In recent decades, third-party IT firms have offered, and manufacturers have invested in, disparate systems that make it easy to capture data. But there’s no centralized place to store it.
Systems like Plex can help by aggregating data in one central location.This ensures that all users have access to the data they need in one place as a single system of record so the right user can have the data at the right time with the right context.
“There is one data source whether you’re the CEO or a shop floor operator,” Garg said. “Then all the problems around context and synthesis, application and dissemination fall into place.”
This also helps ensure that various plants in a global enterprise learn from each other. Otherwise, individual locations can operate in information silos, whether out of neglect or competition. One factory may not want to share performance-enhancing data with a sister facility because it gives the innovator a competitive advantage when it comes time for layoffs or plant closure.
Condition monitoring also enables predictive maintenance.
“Manufacturers have told us for years what they really care about is ...‘Can you give me predictive insight on machine health and reliability, part quality?’,” Garg said.
Plant owners want predictive maintenance information for several reasons. For their immediate needs, unscheduled downtime is costly. One estimate is that unplanned downtime costs six times as much as planned downtime, and ranges from $20,000-50,000 per hour due to lost production, wasted labor costs, decreased quality and increased scrap. This is especially true in food and beverage and pharmaceuticals manufacturing, where a whole batch can be tossed.
Longer term, predictive maintenance helps the CFO with budgeting for new machinery acquisitions.
Predictive maintenance is still in the future at Creative Foam Corp., but it realized other benefits when it implemented Plex’s Cloud ERP at its seven factories in Colorado, Indiana, Michigan and Tennessee.
Staff at the plant had been documenting all shop-floor processes, including machine uptime, amount of scrap and number of pieces produced manually with pencil and paper. They had no standard tracking procedure to follow.
With the new Plex-generated reports, plant managers are finally able to track overall equipment effectiveness, or OEE. They can also drill down into the reports to get details about:
The reports allow the team to calculate overall equipment efficiencies based on the percentage of scrap for the day compared against the company’s standard rate. It also gives the uptime numbers, which is the number of actually produced parts minus the downtime. Such drilled-down data enables the plant managers to quickly make strategic calculations and informed decisions.
Not only that, six of seven Creative Foam plants are 100% paper-free.
When Yokohama Rubber, which makes high-performance tires and operates in 120 countries, relocated its production bases overseas from Japan, the company saw a need for global factory optimization.
Until then, Yokohama had worked toward optimization, but in each factory individually.
“Twenty years ago, we were still making our tires in factories in Japan, so we considered the tires that were being manufactured at our overseas factories as stock for inventory adjustment,” said Tadashi Suzuki, managing corporate officer for Yokohama Rubber. “But now our production ratio is higher overseas, and we are beginning to see the need for global optimization to have a better grasp of our global production status.”
Not only that, but a single factory makes a huge variety of tires, for compact cars and sports cars, as well as high-powered sedans and large construction equipment. It was possible to monitor the production status on individual factories. But when Yokohama tried to make cross-sectional comparisons of all their factories’ production status, it was taking too much time.
Yokohama implemented the Fujitsu Intelligent Dashboard, one of several smart factory products, to work on visualizing the production status of its 15 main factories in Japan and overseas. First up was the vulcanization process, which is the final step in the tire manufacturing process and involves heating the tire under pressure with steam.
After Fujitsu completed the system in January 2017, its use gave everyone from the management team to factory employees the ability to compare production results, operating rates and achievement levels among factories in real time.
The digital technology is used to view processes and solve bottlenecks in real time, while delivering improvements in efficiency, reduced inventory and increased throughput. It provides one single view in real time of Yokohama’s entire global operation, complete with management reporting and drill down to machines and processes that support root cause analysis if something goes wrong.
Suzuki added: “The data used to be separate pieces of information, and the way we amalgamate and display this data in real time will be a key factor. Continuous improvement will result from sharing the success stories, as well as from evaluating our progress in overcoming our weaknesses.”
The final goal of Yokohama Rubber is to establish a cycle in which the factories understand problematic areas and come up with improvement plans that produce noticeable results immediately.
To achieve this, it’s imperative to gather data and create a structure in which the cycle can be put to use.
Ray Russ, senior director of Industrial IoT and Smart Factory at Fujitsu, is a 10-year veteran of smart factories.
“We found a lot of companies would get a platform for condition monitoring or integration to their shop floor systems and they were doing a lot of proofs of concept,” he said. “A lot of times, they found it was great but it would never get rolled out to the rest of the enterprise. A lot of companies have got sensors on all their machines but they’re not doing anything with the data they’re collecting.”
So, a few years ago Fujitsu built up its smart factory framework services. It’s technology agnostic and is based on analyzing a company’s data and focusing on certain domains—engineering, shop floor integration, plant maintenance, quality, etc. The Fujitsu experts pick a few domains, drill down into their processes, their infrastructure and their people, and then lay out a framework or a roadmap for them to become a smart factory.
“Plant maintenance is a good example,” Russ said. “You’re going to go from time-based maintenance to condition monitoring to predictive analytics.” With predictive maintenance software, a plant will probably not get results until at least six months in, and the more data plant managers collect the more accurate the results will be.
“The first few times that machine breaks down, you know why, and you can go back and look at your data and say, ‘Here’s the point where it broke down’,” Russ said. “The more times that happens, the more accurate your data becomes and the better your algorithm gets—and then you’re starting to shut down before it hits that breaking point.”
Part of what Fujitsu includes in its assessment is a maturity model. It works with the research firm Gartner and its manufacturing maturity model to gauge where the company will be at in relationship to its competitors and other companies of its size at the end of the smart factory roadmap laid out for them.
Condition monitoring (CM) isn’t just for machinery in a factory. It can also include operations and the products made in that factory that are part of the IoT.
“More and more companies are trying to build that IoT via sensors into their equipment so they can monitor and provide more value to their customers,” Russ said.
A Fujitsu customer that makes buffers and sweepers started putting sensors in its equipment several years ago because it wanted to build a warranty program for its customers.
The sensors monitor fluid levels and brush levels so that a customer can log into a portal and look at all their equipment. The customer can find out when maintenance is due, if they’re low on spare parts or need any of the products their machinery uses.
Condition monitoring of both products and operations is where software like PTC’s ThingWorx Industrial Innovation Platform comes into play.
“A system like ThingWorx excels at pulling and modeling information from disparate sources and types to allow for monitoring, analytics, and response at both the operations/component and the system level,” said Ryan Cahalane, digital transformation director for PTC. “And the impact of anomalous conditions can be understood in real, business-level key performance indicators that enable better decision-making.”
Colfax, a diversified industrial technology company, adopted ThingWorx and Microsoft Azure IoT in two of its companies to create an IoT infrastructure for its products: Howden, which makes precision air and gas handling equipment, and ESAB, a welding and cutting equipment producer.
“Where condition monitoring was often limited to only high-value assets in an overall production process, today the scope of application has widened to include whole systems and even value chains,” Cahalane said.
He was VP of digital growth at Colfax before joining PTC. There, he led Colfax’s Data Driven Advantage initiative, which was focused across the enterprise to identify and unlock digital opportunity, in products and operations.
Howden builds ThingWorx into its compressors to gather and analyze critical equipment data that help optimize its performance and operate efficiently. Because the equipment’s condition is monitored, customers can easily identify or be alerted to abnormalities that might result in machine failure and avoid unscheduled downtime.
At ESAB, ThingWorx has led to improved productivity, documentation and asset management for its customers. As an additional benefit, the data collected automatically by ThingWorx is easily assembled and transformed into reports for regulatory bodies.
Colfax’s next step is to incorporate ThingWorx and Microsoft Azure into its own production operations, which will lead to better efficiency, quality and other typical IoT benefits.