Digitize To Deliver On-Demand Production
EP Editorial Staff | May 14, 2020
Companies that provide data visibility are 30% faster at taking product designs to production and 23% better at managing safety and risks.
By Joe Gerstl, GE Digital
Traditionally, successful manufacturing production has relied heavily on skilled personnel. Experienced employees installed equipment and implemented processes, relying on their expertise and intuition to keep everything running at peak efficiency.
Times have changed and the baby boomer generation of engineers is retiring. To combat this loss of knowledge, manufacturing companies are driving data analytics as a top priority for their company. Using this data, companies are able to “retrain” a new generation of engineers by taking operational data and using tools such as artificial intelligence (AI) and machine learning. Combining these technologies with accumulated data makes the new generation of workers more effective than their predecessors by providing them with an opportunity to react faster and be more productive. Failure to implement robust digital capabilities puts companies at a distinct competitive disadvantage.
It’s no secret that data analytics have the power to shift the competitive landscape. In fact, the number one fear of companies unable to implement a data-driven strategy is that competitors will gain market share at their expense. Today, it’s the refined collection, analysis, and application of data that optimizes efficiency and productivity. Data-driven predictive asset maintenance can save as much as 12% of scheduled repairs, reducing overall maintenance costs by 30% and breakdowns by as much as 70%.
Manufacturers suffer from excessive costs related to materials, labor, packaging, and shipping. For the most part, these added expenses are due to waste resulting from rework, unscheduled asset downtime, maintenance, and late shipments. Predictive analytics have the power to help reduce all of these costs. In fact, one performance chemical company reports increasing capacity by nearly 20% after adopting a predictive-analytics model.
Demand and Data Deluge
Production managers are tasked with keeping costs and risk under control while producing what the market wants, when it wants it, and at increasingly higher quality—even when that means quickly changing lines and entire facilities over to new products. The demand for change is increasingly common as consumer preferences shift almost daily. The result is a rapidly eroding time buffer between market demand and production lines.
Near term, manufacturers risk not having the flexibility to make the goods people want to buy today, as they don’t have the big-picture visibility they need with respect to equipment, people, processes, and materials. In an environment where overall manufacturing growth is stagnant, new revenues can only come from taking market share away from competitors.
To improve operations, companies have automated select activities, primarily by purchasing point-product software solutions. However, production data is often gathered manually—isolated in siloed systems and different formats. Some manufacturers are taking a holistic, data-driven approach, but not every organization has taken that step.
It’s difficult to start and stay ahead. Most organizations have equipment from multiple suppliers in plants of various sizes and ages, some of which they built themselves and some they acquired—and no two facilities are the same. To realize benefits across the enterprise, organizations must create a common system for gathering and analyzing data.
According to LNS Research’s (Cambridge, MA, lnsresearch.com) Patrick Fetterman, “Industries are quickly moving beyond machine or process monitoring to problem solving and are aspiring to closed-loop processes, where the analytics inform machines how to prevent downtime and quality problems before they occur.” Rather than spend time and resources building your own technology infrastructure, you need proven skills and tools, including applications and hosting services, to make the transition to a data-driven model a practical reality.
The journey begins with standardized connectivity to track data from plants, equipment, materials, and people. You can then begin to measure and compare equipment metrics and operator practices to identify the reasons for production losses—without making costly investments in additional IT equipment or staff.
To become a truly digital company you need the enterprise-wide ability to deliver data from the source equipment to the cloud where it can be shared and analyzed. The resulting insights enable you to reduce production losses and drive additional efficiency.
It’s time to take action if your current system doesn’t provide the visibility. If, over the years, you’ve installed separate best-of-breed systems for inventory, production, downtime, and quality—and systems vary from plant to plant, or if data is manually relayed from one system, with corresponding delays and the potential for transposition errors, some critical data might not be captured at all. As a result, no one in the plant can obtain a single up-to-date view of the entire manufacturing process or analyze the situation from a historical context.
When a problem arises, it becomes difficult to identify the cause. You’re forced to rely on the intuition of your most experienced staff—a resource you’re steadily losing to retirement. Unable to optimize, downtime and scrap production increase at your busiest plants and excess capacity sits idle elsewhere. You need to extract the right equipment and quality data from your operations and share it with the right people across your organization in an easy-to-consume format.
Without complete, uniform performance data from equipment, it’s impossible to make valid comparisons. A trusted digital-transformation partner should be able to help you determine exactly which metrics to track across all of your assets—and how to ensure data is always consistent, regardless of the equipment generating it. Standard data will allow you to easily roll up production results across all your lines and plants to give you a single source of the truth.
When you can define the relationships between each element of your manufacturing process, you can create a digital representation of your plant—a digital twin. The next time a machine goes down, you can quickly spot the other equipment, operators, and processes that are affected and redeploy resources. Or, you can use unexpected downtime as an opportunity to perform much-needed maintenance. With real-time data and analysis, along with specific-data models, your workforce can make informed decisions on the fly.
According to LNS, a typical data model spans an enterprise’s assets, processes, and systems to holistically optimize performance. Data models play an essential role in any IIoT solution and are especially important for industrial analytics, including AI and machine learning. Using sophisticated data analytics, world-class companies have developed finely tuned market-sensing capabilities to stay on top of changing purchase patterns. In a study from LNS Research entitled, “Benchmarking the Value of Enterprise Data Visibility,” companies that use IIoT analytics and provide data visibility are 30% faster at taking product designs to production, 24% better at managing product and supplier quality, and 23% better at managing safety and risks.
For many companies, their existing facilities were not originally intended for today’s small runs and rapid changeovers. Complete visibility into all of your plant’s operations makes rapid output shifts possible. Plus, you’ll have the information necessary to generate a new bill of materials, reroute components, change equipment schedules, and re-assign employees—allowing you to keep pace with today’s shorter demand cycles. EP
Joe Gerstl is Director of Product Management for Manufacturing Software at GE Digital, San Ramon, CA, ge.com/digital. He’s worked in the software industry and manufacturing for more than 30 years, in various roles including engineering, sales, and product management.