An optimized warehouse and supply chain are just as important to successful manufacturing as an efficient production floor. Here are some ways to strengthen the link.
As its name suggests, Manufacturing Engineering magazine frequently discusses the best practices and advanced technologies that machine shops, sheet metal fabricators and other members of the manufacturing community are deploying to gain a competitive edge. These include automation, the Industrial Internet of Things (IIoT), digital twins and robust software systems, all of which make operations more flexible and cost-effective.
But what happens after parts leave the shop floor? Isn’t efficient storage and shipping equally important? How about predictable delivery of the raw materials and supplies needed to make those products? Without a dependable supply chain, manufacturers are at the mercy of shipping delays and offshore providers, any one of which might tarnish an otherwise flawless delivery record. Such disruptions drive up costs and drive away customers.
That’s why effective warehouse management and supply chain optimization are no longer nice-to-haves but must-haves. Fortunately, the hardware and software technologies just mentioned are equally adept at streamlining warehouse processes as are those of the production floor, helping to ensure the smooth flow of goods even in the face of labor shortages and unpredictable deliveries.
A Helping Hand
It’s certainly a lot to think about, which is why it can be a good idea to seek advice from an industry expert before making any changes, let alone investing huge sums of money in these technologies.
One such expert, Vikas Argod, a principal for supply chain operations at Atlanta-based consulting and managed analytics firm Chainalytics (part of NTT Data Americas Inc.), ticked off a host of supply chain obstacles that companies face today. These include:
- A shortage of industrial warehouse space;
- Overburdened ports of entry;
- An industrywide push for increased sustainability; and
- A pervasive “I need it now” mentality
RedViking “StingRay” AGV pulls carts for material handling at a large heavy-duty truck manufacturer. The Stingray can pull up to 5,000 lbs. (Provided by RedViking)
Then there’s disruption. “An earthquake in Japan, a hurricane in Louisiana—these are localized events that temporarily disrupt product flow, and we’ve dealt with them for years,” Argod says. “COVID, on the other hand, disrupted demand and supply at the same time, and did so for an extended period. It messed us up in many ways, not least of which was its contribution to a labor shortage that was already challenging for employers of all kinds.”
Because of this—and as with the manufacturing floor—companies are looking at robotics as the primary solution for the ongoing labor crunch.
“The biggest problem today is finding people,” Argod continues. “Let’s say you’re running a warehouse and need a forklift driver, preferably one with experience and the appropriate certification. Unless your pay is well above market rate, these people are few and far between, and once you do get someone, the warehouse down the street will offer them a couple of bucks more to leave. Employee attrition rates are upwards of 50% in some areas, which is why a majority of companies are saying, ‘Let’s automate.’”
Getting Bold
Greg Giles, a self-described “designer, builder and implementer of manufacturing and test solutions,” would agree with Argod. As the vice president of product development for Plymouth, Mich.-based RedViking LLC, he notes that one of the most pickable pieces of low-hanging warehouse fruit is the movement of goods from one area to another. For this, he might recommend one of two solutions: an automated guided vehicle (AGV) or its more capable cousin, an autonomous mobile robot (AMR).
PickOne’s induction solution provides optimized picking and placing of mixed parcels, polybags and products for high-volume parcel sortation centers and e-commerce fulfillment and distribution centers. (Provided by Plus One Robotics)
The first of these terms refers to a battery-powered cart that traces a well-defined route, one usually defined by magnetic tape attached to the warehouse floor. An AMR, on the other hand, is more akin to a self-driving car, able to go anywhere it can “see” through a mix of navigational technologies, among them vision systems, infrared sensors and, most frequently, lidar (light detection and ranging) equipment.
Which one is best? Giles will tell you it depends on a variety of factors. “The calculus changes depending upon the environment,” he notes. “Though more limited in terms of flexibility, an AGV is very simple to deploy; just put down a piece of tape and the robot will follow it. With an AMR, you need to train it where to go and how to get there, so implementation is probably a bit more technical than with an AGV. Both, however, can become a bit more challenging to deploy if the environment changes frequently.”
Giles also advocates to automate wherever possible in the wake of the labor shortage. In one recent example, RedViking worked with a company that employed nearly two dozen people daily to move racks from a staging area to the shipping area. Their inability to consistently keep enough workers on hand was quickly becoming a bottleneck. “It wasn’t a cost issue, but rather a shortage of human capital, so we developed a system that uses some fairly basic AGVs to move the racks,” Giles notes. “Problem solved.”
Cost wasn’t the driving factor in this scenario, although it frequently is. Companies justify their decision to automate (or not) based on the number of workers they can replace, multiplied by the hourly rate and benefits. Yet this simple calculation fails to account for more intangible variables such as predictability, accuracy and safety, areas where robots offer a clear advantage over their human counterparts.
“Humans aren’t always as attentive as they should be,” Giles points out, “and, unfortunately, it’s fairly common for a forklift driver or an operator pulling a train of parts around the floor to miss seeing someone in their path. Not so with AMRs and AGVs, which only go where they’re told. Also, it’s quite easy to bring additional automation to bear as demand increases. Not so with humans, where it can take weeks or months to hire and onboard new workers.”
Digital twins of supply chains enable businesses to make normal operations more efficient, and simulated disruptions can illustrate points of vulnerability and allow for contingency planning (Provided by Dassault Systèmes)
Down Yonder
This last benefit is especially valid with a robotics-as-a-service (RaaS) deployment model, which lets a company use its operational expense (OpEx) budget instead of a capital expense (CapEx) to fund automation projects. “RaaS is particularly appealing to the third-party logistics (3PL) market, where parcel volumes can be quite volatile and investment lengths are generally more short-term than with FedEx, UPS and other large carriers. RaaS also gives companies a way out if the robots don’t work as intended or demand suddenly goes in the tank.”
That’s according to Plus One Robotics’ Founder and CEO Erik Nieves, who spent decades with Yaskawa Motoman and has since built a business on robotic depalletization and induction, two activities that fall under the “manipulation” umbrella rather than mobility tasks. “In a warehouse or fulfillment center, most of the time is spent taking packages off a pallet, sticking them on a conveyor, or moving them from one place to another,” he says. “Our solution automates the first two.”
Nieves is well aware of America’s labor problem. Ironically, he suggests that the fastest way to solve it is through a guest-worker program. By allowing a specific number of foreign citizens into the country in a controlled, well-documented manner—much as the U.S. already does under its National Farmworker Jobs Program (NFJP)—there’d be little need for his firm’s products and services.
“That’s why I’m a big proponent of comprehensive immigration reform,” he says. “Unfortunately, we have zero political will in this country for such a program, and the only immigrants we’re going to allow are immigrants from the future: robots.”
He’s made it his mission to supply those future workers. It’s challenging work. Unlike loading a CNC lathe with sawed blanks or inserting tab A into slot B, the robots from this San Antonio, Texas, company must manage significant and ongoing product variability. It’s for this reason that Plus One developed its PickOne AI-powered vision software, said to deliver precise hand-eye coordination for logistics robots, and Yonder, a remote supervisor software suite that keeps humans in the loop for when things go awry.
Advanced capabilities aside, Nieves suggests that warehouse automation is roughly at the same level as automotive was during the late ’80s. “Back then, robots were limited to painting and spot welding, but you could see that growth would explode as they grew more capable,” he says. “That’s pretty much where we’re at today. The majority of all depalletization, singulation and induction is still manual, but we as a society are sending more packages all the time and demand keeps increasing. There’s simply not enough labor to keep up with the growth.”
A Better Playground
Yet automation is only one piece of the optimization puzzle. Where do the digital twin and IIoT fit in? What about integration with warehouse management systems (WMS) and enterprise resource planning (ERP) software? And what happens beyond a logistics facility’s walls?
Ask Adrian Wood, director of marketing and strategy for the Delmia brand at Dassault Systèmes SE, whose North American headquarters is in Waltham, Mass. “People have begun to think of the post-COVID years as the new normal, where disruption is a constant and they’re desperately trying to build a certain level of visibility and agility so they can pivot when unexpected changes occur.”
The digital twin addresses these needs and others. Described as the ultimate playground, this powerful tool allows logistics managers and planning personnel to leverage integration and IIoT capabilities to bring real-time data from internal or external sources into a virtual environment and simulate various scenarios.
With the twin, managers can perform what-if analyses based on available inventory and staffing levels (or lack thereof). They can enter sales and operations planning (S&OP) data and model the results, evaluate various warehouse layouts (possibly one with an automated line or two), understand the impact of losing a key supplier, simulate different transportation routes and much more—all without committing to one path or another.
“The concept of the digital twin has been around for a long time, but it’s traditionally been applied to assembled products like an aircraft or a smartphone,” says Wood. “But now, companies are talking about the digital twin of the supply chain. It is a model of all their facilities, inventory levels, customer and vendor channels, and other sources of variability. The twin allows you to very rapidly experiment with multiple scenarios and determine which one is optimal.”
Sadly, and somewhat surprisingly, a large percentage of companies are still using Excel to manage these processes. “The companies pushing the curve in terms of innovation are those that migrated from spreadsheets a long time ago, and are now looking for more advanced planning systems,” Wood notes. “The digital twin lets them look at the big picture; it lets them evaluate the entire end-to-end supply chain in all of its glorious complexity, of which the warehouse is just a small piece.”
Automation.ai’s constraint-based optimizer evaluates common warehouse scenarios and mathematically determines how to minimize travel distances, inventory touches and other time wasters. (Provided by Automation.ai)
Conducting the Orchestra
AutoScheduler.ai of Franklin, Tenn., which specializes in warehouse resource planning and optimization, aims to “dynamically orchestrate all activities on top of your existing WMS in real time.” Jeff Potts, the company’s chief revenue officer, notes that at least some of the current staffing shortages can be laid at the feet of recent reshoring initiatives.
“The historical school of thought said that if you could build it 15% to 20% cheaper outside the United States, it was most likely a win,” Potts says. “That figure’s begun to shift closer to 30% over the past decade or so, but it remains a false paradigm, and we learned the consequences of that thinking very quickly during the pandemic. This is why we’re seeing more companies of all kinds turn to domestic and near-shore suppliers, or bring work in-house.”
There’s more to this than broken supply chains, however. Potts suggests that spending a dollar on manufacturing in the U.S. puts approximately $2.80 back into the local economy. Add to this benefits such as shorter lead times and lower carbon emissions, and it quickly becomes clear that reshoring is good for consumers and manufacturers alike.
As a result, many companies are reevaluating their sourcing strategies, questioning whether they have the best locations for their distribution centers and what would happen if they manufactured somewhere other than Asia. “Businesses have begun to realize the importance of resiliency over cost,” Potts declares.
They’re also looking for ways to optimize their warehouse operations—to do more with less, and generate schedules in a scientific, repeatable manner that doesn’t rely on tribal knowledge or intuition. Such capabilities are quickly becoming crucial as logistics centers deal with increasing amounts of information from external sources—the trucking fleet’s GPS system, for example, or electronic communications from suppliers and customers—while also taking into account IIoT data and automation.
This is where the AutoScheduler.ai platform comes in. Potts refers to it as a means to “orchestrate” distribution center activities and make them more efficient. “Let’s say you’re managing a million-square-foot facility. There were supposed to be 100 people on the first shift but seven of them called in sick. There are 70 trucks waiting out in the yard and you’re anticipating another 35 inbounds by noon. They’re doing maintenance on Dock 7, two of the forklifts are down, and your best customer just asked to expedite a big order. In most organizations, a planner will pull data from their ERP, WMS or yard management system and say, ‘Okay, we have all this work to do today and all these obstacles to work around. What’s my best course of action?’”
A robust scheduling system, or what Potts calls a “constraint-based optimizer,” evaluates these everyday scenarios and answers the question just posed. It mathematically determines how to minimize warehouse travel distances and the number of times inventory gets touched, while maximizing two-way traffic and looking for cross-docking opportunities.
There are literally a lot of moving pieces in any supply chain, he points out, with more coming along every day. Says Potts, “We use AI to figure out the most efficient, cost-effective way to get the work done inside the four walls of that building, and can do so on a day-to-day or even an hour-to-hour basis. What’s more, the AI engine gets smarter over time. It keeps track of its decision-making steps and uses them on future runs, while also giving its human users a record of what happened and why.
“Capabilities like this will soon become both commonplace and extremely necessary as the industry deals with increased demands, variability and complexity in the supply chain,” he continues. “Companies without such tools will soon find themselves underwater.”
