Meeting of the Minds: Where Process and Discrete Manufacturing Converge

Meeting of the Minds: Where Process and Discrete Manufacturing Converge

When it comes to continuous improvement, discrete and process manufacturers may have more in common than they think. A close look at the two industries shows they have opportunities to learn from each other.

What could a plant manager from a process manufacturer that makes yarns and fabrics for industrial applications possibly learn from an automaker such as Toyota? Conventional wisdom says companies should benchmark against similar industries to gain knowledge relevant to their operations. But as we've seen with businesses ranging from medical institutions to governmental agencies adopting lean principals, one industry may have best practices that can be tailored to fit a completely different work environment.

The same could be said for process and discrete manufacturers. Generally speaking, process industries are characterized as businesses that make products in bulk quantities, such as chemicals, pharmaceuticals, gasoline, beverages and food products, which often undergo a chemical conversion. On the other hand, discrete manufacturers produce or assemble parts or finished products that are recognizable as distinct units, such as automobiles or computers, capable of being identified by serial numbers or labeling products and measurable as numerical quantities rather than by weight or volume.

Even with these stark differences, process and discrete manufacturers have a history of glomming on to one another's improvement methods. As lean and Six Sigma gained in popularity throughout the 1980s and '90s within discrete operations, process manufacturers began taking note, says Peter Martin, vice president of strategic ventures at automation technology provider Invensys Process Systems. Initially, the process industry's forays into the continuous improvement trend didn't fare so well. When process manufacturers tried to implement statistical analysis methods which are used primarily in discrete operations, such as Six Sigma, they didn't work because they focused too much on defects, Martin says.

"In the process industries we don't tend to do defect-oriented manufacturing," he explains. "[For example], if you charge a little too much pigment, you can just put a little more base in and fix it. The mindset in the process industry is direct, real-time control. The mindset in discrete manufacturing is after-the-fact statistical analysis to get continuous improvement. So the mindsets are very different."

Process manufacturers finally realized value from continuous improvement programs when they stopped applying methods that focused on statistics. "When the process industry started looking at what discrete was doing in continuous improvement, they started saying, 'We should be able to use different techniques, maybe not statistical techniques, to continuously improve our critical performance variables like contribution margin, or energy cost or production value,'" Martin says.

Lean Adaptations

Lean became a reality for several Milliken & Co. plants when the privately held textiles and chemicals manufacturer applied lean manufacturing techniques to optimize lead time and align itself with customer demand variations, says Chris Glover, Milliken Performance System practitioner. Like many other process manufacturers, the company dabbled in lean throughout the 1980s and '90s, but initially didn't fully understand how to gain significant improvements from the methodology, says Glover, who has worked in a variety of leadership roles within Milliken since 1985.

"We went to Japan as an organization and benchmarked many Japanese companies in the early '90s and [lean] was visible inside the Toyota Production System," he explains. "It was just part of their operating system, but when we brought it back we weren't really sure how to incorporate it in our activities, partly because we were process and couldn't figure out why you had to measure every millisecond a person was running a process machine. So we struggled a little bit with that type of diagnosis."

In discrete industries, manufacturers can deploy material requirements planning (MRP) systems to manage raw materials that require long lead times. But process manufacturers deal with materials that usually need to be moved quickly, notes manufacturing systems consultant Jim Ranallo.

"You have inherent variability in the raw materials coming in, so the processes that are built to handle that variability are very different than discrete, which is more assembly of manufactured product that is tightly controlled and very well defined in a build of materials," Ranallo explains.

Lean can be used to replace MRP to manage shorter lead-time materials using a kanban signal, he continues. "Say I'm ordering an ingredient and I have a supplier that responds in two weeks. I can use a two-bin system where I have two pallets of materials in a warehouse. When one of those pallets is consumed, I can place an order to refill that order in the warehouse because I know I'm going to get an order in two weeks."

Milliken had worked with value stream maps with the hope of realizing significant waste reductions in specific areas. But gains were minimal because the plants didn't make lean part of their enterprisewide culture, Glover says. The results started to manifest after the company spoke with customers who requested wider-scale lean adoptions.

The Lean Embrace

Among Milliken's products are fabrics used in various automotive applications. Steeped in lean from its earliest beginnings, the company's auto manufacturing customers pushed Milliken to understand the continuous improvement process better. The company responded by establishing what it refers to as its Lean Enterprise system in 2006 at several plants in Georgia and South Carolina. Through value stream mapping, the company identified waste points, what was driving lead times and the Plan For Every Part (PFEP) system.

Invensys' Peter Martin sees more opportunities for discrete/process knowledge sharing in the future.
Typical PFEPs provide visibility into inventory by charting characteristics of each part, including part numbers and dimensions, and measuring supplier performance metrics such as delivery time. That helped the company determine its replenishment points, safety stock volume and materials runs. "It was especially critical last year with the increase in most of the chemical costs with the oil prices rising," Glover explains. He estimates the company's lean initiatives have helped reduce customer lead times 30% to 60% across multiple plants.

The Dow Chemical Co. developed a lean simulation tool relevant to the process industry to help management teams in its plants address inventory problems, according to Dow global supply chain process consultant Martino Fernandes. The initiative started in 2002 with lean simulation technologies and Lego models for management to gain buy-in from them.

"Lean can definitely drive value in the process manufacturing environment," Fernandes told AMR Research in April 2008. "The key is to demonstrate how so that management believes it. Experiential learning through simulation is very successful in helping to reshape traditional paradigms."

As of 2008, the average range of improvement for teams participating in Dow's lean exercise are 10% to 15% increase in fill rates, 30% to 40% reduction in cycle times and a 10% to 20% inventory improvement yielding a 5% to 15% storage space requirement, according to an AMR report.

Discrete Conversations

Conversely, some discrete manufacturers are keeping a close eye on trends in the process industry, particularly when it comes to wireless technology, according to analyst firm ARC Advisory Group. Discrete manufacturers are showing interest in the development of wireless process communication standards such as ISA 100 and the highly addressable remote transducer protocol, according to a recent ARC study. Potential uses include the implementation of wireless sensing technologies in automotive applications where robotics are used to reduce cable failure in moving equipment and enable the monitoring of information processing and devices.

In addition, wireless sensors can be used to establish predictive maintenance schedules through the collection of vibration data, says Ralph Rio, ARC research director. In process environments sensors are often used to monitor cooling fans, whereas discrete manufacturers might use sensors to observe the voltage characteristics of a motor.

ARC predicts the worldwide market for wireless devices in discrete manufacturing will grow at a compounded annual growth rate of 16.2% over the next five years. But growth will be limited until standards are developed, notes Chantal Polsonetti, ARC vice president.

"While the business drivers are in place, including wireless' status as the ultimate fieldbus from the perspective of wiring reduction, the lag in technology and standards development suitable to meet discrete industry requirements will contribute to an ongoing fissure in growth prospects for discrete versus process industries over the next five years," Polsonetti notes.

There may also be opportunities for discrete manufacturers to implement control devices used in process industries, says Invensys' Martin. He points out the frequent use of vision systems in the pharmaceutical industry for quality control as one such technology that's making its way to discrete operations. However, the greatest potential for information sharing may exist in mixed process and discrete environments, such as mining and metalworking or foods, says Martin, who explains that process manufacturers utilize mathematical controls to drive continuous improvement while discrete is more focused on logistics.

"There are examples of plants like an oil refinery that's 99% continuous and an automotive plant that's 99% discrete, but there are a lot of industries in between those two extremes that combine continuous and discrete processes right within the same plants," he says. "In those cases you've got a huge mix of logic and mathematical approaches, and I think there's a lot that can be learned by both sides looking at how on the process side you can apply logistics to do better production scheduling, to do better demand scheduling, or on the discrete side using mathematics to do more deterministic control."

OTHER RELATED ARTICLES

Lean Manufacturing Visionary Jim Womack On Frontiers Of Lean Thinking

Preparing for ERP with Manufacturing Best Practices

Top 5 Tactics to Support Lean Across the Entire Enterprise

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish