This is part one of a three-part series covering 4IR, data, and the future of manufacturing.
The Fourth Industrial Revolution (4IR) sets itself apart from other eras of industrialization in that it is concerned with more than optimizing productivity. This new era implements cross-disciplinary advancements to catalyze systemic changes in manufacturing and global economies. Sparked by the fusion of technologies such as the Internet of Things (IoT), machine learning, robotics, AI, and others, 4IR ushers in a digital transformation that will transform manufacturing from the supply chain to the workforce.
Many of the technologies driving 4IR rely upon vast amounts of data to be implemented at scale, data that needs to be stored and heavily analyzed. Seeing an opportunity to leverage its industry-leading technology and expertise in data, Western Digital is implementing 4IR technology across its own manufacturing operations.
Data from the factory floor
The technologies in this new era of industrialization are driven by data and lots of it. Western Digital’s factories, spanning the globe and manufacturing both HDD and flash products, were the perfect place to collect data—and act on it. Factories from the US to Thailand were equipped with hundreds of IoT sensors and cameras that collect data constantly. Every machine part, every test, and every worker on the floor corresponds to a data point stored for analysis.
“We are mostly collecting data and analyzing it to turn it into business value,” said George Ng, senior global director of data science and digital transformation at Western Digital, about his work at the company’s HDD factory in Thailand. “That’s 15,000 data points before assembly, then 15,000 after assembly, added to the 7 petabytes of data collected across all global operations.”
Ng worked primarily on the digital transformation of the HDD factory in Thailand before transitioning to lead a global analytics team focused on digital transformations. Capturing data that was previously lost on the factory floor is key for Ng to improve analysis and, thus, performance.
“Compared to 10 years ago, sensors are less expensive, and cameras have also gotten more advance and cheaper,” Ng said, speaking to the differences between 4IR and previous eras of industrialization. “When we have a sensor for every machine, we are able to collect more data to build bigger models and have more effective machine learning.”
“It’s impossible for a human to figure all of these things out, but, suddenly, all of these technologies empower us to do that,” said Ng. “It’s allowed us to achieve product yield that we’ve never seen before.”
From procurement to delivery
Machine learning is at the core of Ng’s work, and another part of the 4IR technological symphony. IoT sensors and the data they collect train machine learning algorithms that process the data faster than a human could ever hope. The insight they provide covers the entire factory, including the base materials needed for manufacturing.
“These data can drive material planning, for example,” he said. “If we can use less material to make the same product, that can reduce the amount we have to procure and lower costs for the end user. That, in turn, can impact how often we ship out units, which impacts our trucks and warehouses. Pretty much everything from procurement to delivery is impacted.”
Dávid Gyulai, director for advanced analytics at Western Digital’s advanced analytics office, focuses on optimizing transportation for the company’s supply chain. His focus is on minimizing the transportation needs for products and optimizing what can be a very expensive part of manufacturing.
“Our first project was to optimize cargo pickup times based on historical data and we used machine learning to predict cargo travel times. That way we decided the best day to ship out certain cargo based on the data,” said Gyulai. In global supply chain management, time impacts everything. Faster shipping means less money spent on fuel, fewer emissions in the air, and more resiliency in the event of a disruption.
“Now, we are currently working on forecasting. We’re trying to predict the future, essentially. We are looking at demand and logistics data, and more to figure out the best way to get our products delivered,” he said.
Gyulai explained how Western Digital’s supply chain is very complex and requires a lot of transportation among network nodes, including factories, warehouses, distribution centers, and lots of customer around the world. “Some of the biggest costs can come from transportation, so we build models to make the supply operations faster, more flexible, efficient, and sustainable,” he said.
Tech takes over the factory floor
New technologies also create new techniques and operations in the factories, changing the workflow for workers and machines. Carmen Ng Soo Fun, director of data analytics engineering, global operations at Western Digital, specializes in smart factory implementations at the company’s global manufacturing sites.
“Our team is trying to implement the latest technologies in our factories,” Fun said from Malaysia during a video interview. She explained how she harnesses such technology in the factories to improve operations.
“It’s about detecting emerging issues and responding quickly,” she said. “We are inferencing and checking in real-time. We are monitoring the machines and parts in real-time and responding as needed.” This streamlines the operations of the factory, minimizing or eliminating downtime, and frees up workers to take on less repetitive, more meaningful tasks in the factory. Computer vision is one example of that process.
“Computer vision is like a pair of digital eyes. Whatever humans can do with our eyes, computers can do too. Inspections, for example, can be done by computer vision,” said Fun. Instead of an operator staring at the factory line for hours, trying to discern where, if any, malfunction is occurring, a computer can do it for them. The workers are then freed to tackle more important, complex tasks, like equipment repairs.
Some of Fun’s work sounds even more futuristic. Augmented reality is also on her radar, where she sees opportunities to reinforce learning and safety with real-time visual cues on equipment and machinery.
“When you use AR, the information can pop up in front of you,” she said, “meaning you don’t have to look at a PC for part information or training. This improves the safety and efficiency of our labor.” Instead of shutting down or slowing down the line, for example, AR can be used to show the inner working of a machine, translating complex knowledge intuitively without slowing down production.
A blueprint for the future
These 4IR advancements across Western Digital’s factories are vital for the company’s future. The foundation of automated operations and data-driven decisions drive value for the business, while simultaneously bolstering its products and people.
“Normally, operators are doing repetitive work which can be replaced by automation,” Fun said. “So, workers can then be up-skilled to work in new fields. But, humans are always needed to maintain everything.”
As Fun knows, automation doesn’t destroy jobs. It transforms them. There is always a need for skilled labor. Western Digital continues to invest in upskilling its workforce and preparing them for a 4IR future, providing a blueprint for the future of global manufacturing operations.
Driven by technology and changing everything from daily work to macroeconomics, the Fourth Industrial Revolution is already sculpting a new path forward. Optimizing global operations and supply chains is no small task, but 4IR’s capabilities continue to grow. Western Digital is just one example of how 4IR is building the future, one data point at a time.
In the coming installments of Building the Future, we will cover microfactories and the role 4IR will play in sustainability.