Starting Small with Automation: Advice from Industry Giants
POSTED 01/08/2024 | By: John Lewis, A3 Contributing Editor, Tech B2B Marketing
Factory automation is hardly a new discipline, although it has eschewed a great many electromechanical components in favor of increasingly more digital ones. Implementing automation solutions offers many benefits: more can be done with less, errors can be minimized, and quality can be improved. But where should a manufacturer start? We spoke with a number of industry leaders in factory automation to get their advice for how to start small and set the stage for success.
Avoid Common Pitfalls
For those preparing to dip their toes in the automation waters, there can be tremendous apprehension. Some mistaken perceptions about automation can make it a challenge for companies to begin. For instance, leaders at many manufacturers fear automation because they believe it will somehow result in lost jobs — something they are often reticent about addressing with workers.
“Fear of job displacement should be acknowledged and taken seriously,” said Fredrik Ryden, CEO of Olis Robotics. “But the reality is robots and automation are poor replacements for people. People have a commonsense understanding of the world and the laws of physics that far exceeds the abilities of any industrial robot available today.”
Paul Anderson, director of field application engineering for Omron Automation Americas, agreed. “Concerns about job displacement are based on the belief that a person will be eliminated rather than repurposed,” he said. “There is published data showing that, in practice, adding robots to plant floors results in stable or increased headcounts.”
Manufacturing and industrial workforces are drastically changing, according to Damon Sepe, N.A. OEM segment manager-heavy industry equipment at Rockwell Automation. “Across the globe, older workers are moving toward retirement and taking decades of experience and knowledge with them, while a limited talent pool of younger, less experienced workers takes their place,” he explains. “Automation should be viewed as means to address these workforce challenges and create new opportunities to deploy and support these systems in the field.”
Another worry about automation is that it will be complex, expensive, and time-consuming.
“While there are certainly complex applications for machine vision, there are also many simple and easy-to-set-up products that can be put into production in days — not months — and won’t break the bank,” said George Gauthier, senior manager for product marketing at Cognex.
Chris Liu, product manager for Siemens, said it’s not necessary to automate everything everywhere all at once. “At the very beginning, start with some pilot processes or pilot projects and demonstrate the benefits that can come from automation at a smaller scale. Scale up only after achieving success on the smaller automated processes.”
At the same time, it’s critical not to underestimate the amount of work that needs doing — an error Patrick Varley, product marketing manager (robotics) for Mitsubishi Electric, sees often. “This can come from a variety of origins: inexperience, listening to the wrong advice, and not understanding the process that you are automating as well as you thought.”
“Automation should be applied only after a thorough evaluation of available solutions has been conducted to ensure proper scale in terms of cost and size, as well as capabilities for today and in future have been considered,” said Sepe. “It's important to partner with an automation supplier with a long history of applying and supporting automation technology.”
Developing an automation strategy can help address management fears while ensuring the implementation process goes smoothly. Nonetheless, Anderson said he often sees companies add automation without a plan in place.
“Without a clearly defined strategy, it’s easy to fix small problems rather than looking broader in scope to work towards identifying larger problems or opportunities,” he said. “A cohesive automation architecture allows companies to begin working on small, manageable tasks, while still allowing them to make improvements that are consistent with an overall automation strategy.”
Identify Automation Candidates
Part of developing an automation strategy is identifying those manufacturing processes best suited to automation. But what exactly makes a process “suitable?”
According to Gauthier, costly and mistake-prone jobs are a good place to start, including “inspections that if missed lead to expensive product returns or rejections, processes that produce excessive amounts of waste, or critical production steps that can lead to problems later.”
Anderson agreed that error-prone processes are often worth automating. He added that other good candidates are processes often causing bottlenecks and operations requiring significant operator interaction.
“Removing bottlenecks and errors improves throughput and targeting points of significant operator involvement reduces repetitive work and potential sources of human error,” Anderson said.
Along those lines, Gauthier said that automating processes that are “dull, dangerous, or dirty” often makes sense. Ryden added that automating tasks that take a toll on workers’ bodies is a great way to improve ergonomics. “Even the most skeptical worker will buy in to the project once they see how much better their working lives will be with robot assistance,” he said.
According to Sepe, “A good place to focus is on high value business outcomes such as increasing machine throughput, reducing change-over time, and improving equipment uptime.”
Chris Liu, product manager for Siemens, said that automation is a good choice for processes that rely heavily on data analysis or manipulation. “They can actually benefit from those automation technologies,” he said. “For example, if you would like to perform local data manipulation or data management, you can use edge computing.”
Liu added that process variability is another way to assess a process’s candidacy for automation. “Those with low variability or with minimum expectations — those are actually easier to automate,” he said.
Understanding the Technology
Returning to pitfalls, Liu said it’s common for manufacturers to choose the wrong technology for their automation processes. Gauthier agreed, adding that it’s wise to understand the available technology and major vendors.
“One development we’re excited about is edge learning — a subset of AI that makes setting up a vision application much easier and faster by building the AI processing directly into the vision system or sensor and pre-training the algorithms for common applications,” Gauthier said. “You don’t need to be an automation expert or a programmer to get it working, and you don’t need thousands of training images for it to learn from.”
Similarly, Liu noted edge computing as a cutting-edge automation technology. “A lot of manufacturers, they want to own their data. They do not want to expose their data externally on the cloud to the public or third-party cloud suppliers. So that’s why we introduced edge computing, meaning you do everything associated with data — like data collection, management, storage, analytics — locally on your shop floor by your own automation engineers.”
According to Anderson, predictive maintenance tools are among the best available for automating processes. “Predictive maintenance tools monitor devices on the plant floor — motors, cabinet component temperatures, etc. — to identify potential causes of failure before an unplanned downtime event occurs,” he said. “The downtime costs on these unplanned events easily pay for the cost of the solutions, so they can be quick wins for businesses.”
Varley noted that the components making up an automation solution today are more robust and reliable than ever before. “This provides confidence that the automation will remain in top operating condition throughout and beyond the calculated ROI period,” he said, adding that all products are not created equally — making it critical that you work with suppliers having a reputation for quality.
Digital transformation is making a dramatic impact in today’s smart manufacturing environments. Access to manufacturing data is not a new concept, but what is changing is the amount of production data being captured and how it is being used to improve operations.
“Our digital transformation consulting team works with clients on a strategic plan that addresses priority use cases, business justification, change management, and an execution roadmap for technology implementation and support, all customized to their unique objectives and digital maturity level,” Sepe explained.
There are nearly as many ways of gauging an automation solution’s effectiveness as there are different kinds of automation solutions. Collecting standard metrics, such as throughput and uptime, is straightforward using software packages that make it easy to quickly determine whether the deployed solution is performing to specifications, explained Ryden.
Anderson said one way was simply taking the financial approach.
“ROI/payback periods can be calculated by looking at the total financial investment and the benefits in throughput compared to before and after,” he said. Anderson added that throughput in and of itself can be a measure of success.
“Throughput and quality improvements are just different views of metrics that can help explain how the automation investments are improving the plant,” he said. “These metrics often mean more to employees than the financial metrics.”
Gauthier identified metrics by industry. “Manufacturers should look at production speeds, quality metrics, and scrap rates. Logistics providers should look at throughput.”
According to Liu, there are several different kinds of measurements. “You can actually use those automation tools to track your overall productivity improvements,” he said. “And you can also show how those automated processes enhance the efficiency of your business operations.”
Learning from Successes
There’s no shortage of automation success stories across any number of manufacturing industries. Turning to them can instill confidence and offer guidance.
“One that comes to mind is Federal Package, a contract manufacturer in Minnesota that recently engaged with Cognex and was pleasantly surprised by how easy it was to get started and how much of an impact a machine vision system can have,” Gauthier said. “Thanks to edge learning technology, training the system was easy and they were able to get a new automated inspection system deployed in less than an hour. Now, they see over 99% defect detection, can easily train the system to inspect new products when needed, and are delivering the quality their customers expect.”
Varley said that Mitsubishi Electric customer World Wide Fittings in Illinois implemented a loading and unloading system for a CNC lathe. Sean McCarthy, president and owner of World Wide Fittings, said improving efficiency was the primary goal. "What it gets down to is going from a 70% utilization rate to 95%,” McCarthy said.
For Sepe, Sani-Matic, a producer of cleaning and sterilization equipment comes to mind. “Sani-Matic relies on automation to control complex processes and meet regulatory reporting requirements,” Sepe said. “The company is also a good example of how it’s possible to leverage new digital cloud and automation technology to improve the equipment they provide and offer new services to their clients.”
According to Anderson, Omron has a customer that makes specialty lighting equipment for the automotive industry. “Rather than looking to automate the main production lines, they looked at automating the new product lines, as these would be low-risk opportunities due to lower throughput,” he said. “They were able to develop a modular automation that was highly flexible for new products, and once they felt comfortable using that solution, they began expanding their automation to several robotic systems and continued growing.”