Ford Motor is increasing the variety of sensors in its assembly lines to turn Big Data into Even Bigger Data while democratizing access to the information that’s collected. “We’re enriching the data the equipment generates by adding sensors like infrared monitoring and vibration sensors to complement traditional data [collected], like cycle time and pressure, to create a richer pool of data,” said Mike Mikula, director of manufacturing for vehicle programs. “We can now build smarter analytics around the contributions of those signals to the quality of the product, the efficiency of the process and the health of the equipment.”
Helping Ford in its effort are “most of the bigger IoT platforms,” he said. Ford and one of those IoT platform companies—Google—announced a partnership this year for industrial cloud data management services and analytic software applications, among other services.
To push use of the resulting information to the shop floor, Ford is building an IoT data platform that gives production workers access to analytics via low-code and no-code apps that require no or little knowledge of computer languages.
The newly devised apps can then be scaled to other users in the enterprise who are dealing with similar processes and equipment, Mikula said.
Meanwhile, Ford is also selectively favoring human brain power over software to analyze data and turning more and more to in-house coders than applications vendors.
“The solution will be dependent upon the application,” Mikula said. “Sometimes it will be software, and sometimes it’ll be a data analyst who crunches the data sources. We would like to move to solutions that are more autonomous and driven by machine learning and artificial intelligence. The goal is to be less reliant on purchased SaaS [software as a service].”
Ultimately, Mikula said, the efforts are intended to decrease the overall cost of manufacturing—savings that can then be passed along to consumers.
Two of Google’s competing data giants, Amazon and Microsoft, now offer cloud hosting and software solutions to manufacturers, too.
“There are a lot of players who’ve been outside of industrial IoT like Google, Microsoft and Amazon that are recognizing the potential for them to leverage their software strengths to displace some of the traditional Industrial Internet of Things incumbents,” Mikula said.
The tech giants comprise a New Big Three, similar to the Big Three reference of yesteryear to Ford, General Motors and Chrysler. The tech giants’ industrial software for the vehicle makers and others help with components and objectives of Industry 4.0—digital twin, predictive maintenance, machine vision quality checks, autonomous operations and more.
While these tech giants are leveraging their expertise with data and software to be part of Industry 4.0, they also have domain knowledge from making things.
An ‘insane’ supply chain
“We are probably one of the largest manufacturing companies in the world,” said Dominik Wee, global managing director, manufacturing and industrial at Google Cloud. “Google has an insane, extremely deep supply chain. We have a supply chain that’s as complex as any global manufacturing company.”
The company makes the computing hardware in its data centers and designs its own computer chips. On the consumer side, it makes phones, as well as Chromecast dongles that add smart functions to a television.
While Google has been a manufacturer since the company started longer than 20 years ago, it began investing heavily into its manufacturing services for others when Google Cloud President Thomas Kurian was hired in 2018, Wee said. Kurian previously worked at Oracle for 22 years.
As manufacturing becomes digitized, Google’s methodologies that were developed for the consumer market are becoming relevant for industry, said Wee, who previously worked in the semiconductor industry as an industrial engineer.
“We believe we’re at a point in time where these technologies—primarily the analytics and AI area—that have been very difficult to use for the typical industrial engineer are becoming so easy to use on the shop floor,” he said. “That’s where we believe our competitive differentiation lies.”
Wee said what Google has done for its technology for manufacturing under Kurian mimics what it did previously on the consumer side: make it so easy to use that you don’t even notice you’re doing so.
Quality inspection using Google Cloud’s Vision Inspection is a good example, he said.
“Because machine vision is very advanced, it is used a lot [for quality inspection] and so we’ve made it very easy to deploy machine learning in a shop-floor context and use very few images to do this. You don’t have to be a programmer or a machine learning specialist to do it,” Wee said. “It’s literally point and click.”
He pointed out the ease with which companies can pilot Vision Inspection and then scale it: “To transfer the methodology, you don’t need to bring in anybody from Google or bring in Company XYZ to do it for you. The people inside the factory can do it. This is the big unlock—where machine learning moves from fantasy to being in widespread use in manufacturing.”
Ease of use is a competitive advantage for the company, Wee said.
“We talk about pilot gap,” he said. “A lot of companies have tried machine learning and augmented reality and predictive maintenance, and they did it at one point in one part of their global manufacturing footprint and it was a lot of work. Very specialized people were needed but they were never able to scale it out.”
In addition to being easy to use, Google Cloud uses open-source software, which keeps manufacturers’ options open, Wee said.
Google also asserts that the strength of its analytics and AI is unmatched: “Being a company that has data processing at the core of its 20-plus-year existence, we would claim that nobody understands that better than we do,” he said. “If you have a huge amount of data to tackle in any context including from the shop floor, we would argue that we are the right company to do that.”
Transforming the workforce
Indranil Sircar, Microsoft’s CTO for the manufacturing industry, would probably respectfully disagree with Wee.
“Microsoft Cloud for Manufacturing will help customers reimagine their corporations, building more agile factories and creating more resilient supply chains as well as transforming their workforce,” he said.
Sircar has been at Microsoft almost a decade. Before that, he worked at Hewlett-Packard longer than 20 years.
While both Google and Microsoft help manufacturers collect and analyze data from their machines, Sircar said the workforce-related components of his company’s services—including AI and mixed reality using HoloLens 2—are the real differentiators from not only Google and Amazon but also the traditional industrial software solutions providers.
For example, Mercedes Benz USA uses Microsoft’s Remote Assist, which lets a person at a computer help someone wearing HoloLens remotely.
In a video on Microsoft’s website, Edgar Campana, centralized diagnostic technician at Mercedes Benz of Coral Gables, said, “I can just put it [HoloLens] on and get immediate support. They can literally point things out to me as I’m looking at it. They can circle it. They can draw lines. It’s hands-on: It’s literally right there. I can be talking with them, going through the vehicle in real time. It’s very intuitive.”
While Remote Assist is bi-directional, and lets users talk back and forth, Mixed Reality Guides Solutions is one direction only and allows trainees to interact with holograms alone or in combination with physical objects.
“Airbus is a great example,” Sircar said, noting that the company has been using “the Guides in their manufacturing line, bringing workers to very quickly see an overlay on top of cabling and how it needs to be fitted.”
The 3D environment can offer features that real-life training cannot, such as the ability to view elements in three dimensions from any angle.
Airbus designers can virtually test their designs to see whether or not they are ready for the assembly line.
Microsoft manufactures HoloLens 2 and Surface Hub, an interactive whiteboard for business. It sells physical products, including the Xbox video game consoles and Surface touchscreen personal computers.
Still, “manufacturing, as we speak, has been outsourced a lot but … we definitely manage that entire production line, all the way from the design and procuring the complements and assembly line testing,” Sircar said.
On the cloud data side, the company manages manufacturing of the infrastructure end to end.
Microsoft’s first application for industry was in 2002, with enterprise resource planning software called Dynamics AX, he said. Azure IoT became available in 2016.
Also in 2016, Microsoft was invited to participate in Plattform Industrie 4.0, the German government’s initiative, Sircar said.
Microsoft co-founded the Open Manufacturing Platform (OMP) consortium with BMW. The OMP promotes a common open-data model, which is a shared data language for business and analytical applications to use.
Using same tech as Amazon
Amazon.com is known more for selling than manufacturing. But it makes Kindle, Echo devices and other consumer products. It also makes a large percentage of the hardware that runs its infrastructure and its own computer chips.
AWS (Amazon Web Services) rolled out an industrial line in 2020, including products and services for IoT, AI, machine learning, analytics and edge solutions.
The line features “new and existing services and solutions from AWS and the AWS partner network, built specifically for [software] developers, engineers and operators at industrial sites,” said Douglas Bellin, global head of business development for smart factory and Industry 4.0. “Collectively, these bring a modular approach to allow data collection, storage, analysis and insights.”
The process for making those would-be improvements starts early on with collecting and scrutinizing data, and AWS has figured out several ways to tame all of those “0”s and “1”s.
“If you start at the software and data level, there are more than 350 different protocols that are used in the industry,” said Bellin, who joined AWS in 2017 after longer than 10 years at Cisco.
AWS Lookout for Equipment uses historical equipment data from existing sensors, along with information from historical maintenance events, and it builds a custom machine-learning model that tracks the normal behavior patterns of that machine. When operational data deviates from the known normal, Lookout for Equipment flags the deviation to appropriate users via alerts and dashboards.
Another product, AWS IoT SiteWise, establishes a single data source by simplifying the extraction of data from databases commonly found in industrial facilities, transferring the data either locally or to the cloud, and structuring it to make it easily accessible to users and applications. The application framework allows computation of common industrial performance metrics, such as overall equipment efficiency. It also monitors operations across multiple industrial facilities, analyzes industrial equipment data, prevents costly equipment issues and reduces gaps in production.
In addition to standardizing data, another common hurdle when creating a smart factory is incorporating legacy equipment. Any machine tool equipped with a PLC will have some data to give up, but it will be minimal compared with modern machinery. In response, AWS has partners that can add the necessary hardware to track some machine parameters. It also created its own low-cost vibration and temperature sensor for rotating equipment, called Amazon Monitron.
Amazon Monitron is also a machine learning-based equipment-condition-monitoring service that enables predictive maintenance by analyzing sensor signals from industrial equipment, such as motors, pumps and gearboxes.
It is a fully managed, end-to-end system that includes sensors to capture vibration and temperature data, gateways to automatically transfer data to the AWS Cloud, and a mobile app for setup, analytics and notifications of abnormal machine behavior.
“Amazon Monitron is based on the same technology used at Amazon, leveraging over 20 years of anomaly detection experience to further improve model accuracy,” Bellin said.
With Amazon Monitron, reliability managers can start tracking equipment conditions in a matter of a few hours—without any development work or specialized training required, AWS said.
Manufacturers can use Amazon Monitron to enable predictive maintenance, monitor equipment remotely and track the condition of inaccessible equipment, Bellin added.