Digitizing manufacturing: how companies are using data to improve production
By Rebecca Guenard
- The internet of things (IoT) and data analytics are coming together to create a new, digitized way of manufacturing known as Industry 4.0.
- Companies are beginning to implement new technology platforms that integrate big data and artificial intelligence (AI) to create virtual factories that act as a proxy for the real thing.
- The latest production innovations are not just optimizing manufacturing processes; they bestow an agility on manufacturing that helps companies adjust to changing markets.
There are certain tasks once deemed too delicate for computers, like blending whiskey (https://tinyurl.com/y6mfd4jm) or picking strawberries (https://tinyurl.com/y5z66rqd), but computers are fast learners; they achieved the ability to do both. Computers are also learning to predict a cow’s milk production (https://tinyurl.com/y63albyh). They are monitoring soil nutrients, regulating moisture levels, and eradicating weeds. They are preventing crop loss by scrutinizing the weather. In fact, farmers are one of the fastest adopters of technology that combines data-collecting sensors with machine learning to automate and optimize the food production process (https://tinyurl.com/yxtotf9r).
The agricultural industry is rapidly adopting automation, connectivity, and digitization to optimize its use of resources (https://tinyurl.com/yxtotf9r). The trend mirrors a similar overhaul in manufacturing across every industry sector, from plastics to pharmaceuticals, from cosmetics to beverages. Consumers expect technology to make their lives convenient. Unlocked cars ping cellphones so that owners can instantly remedy the oversight. These same technological conveniences are now being applied on a large scale to improve the efficiency of product development and production.
“The ubiquity of data is what is driving everything,” says Jim Gavin, corporate accounts manager for food and beverages at Siemens, a global company specializing in digitization products headquartered in Munich, Germany. Like a Fitbit, which monitors and collects information about the human body to improve users’ health, data collected during a manufacturing process is applied in a feedback to improve the process. Manufacturers can make or change a product faster and sell it to their customers sooner. The goal is to incorporate the convenience of modern technology into the manufacturing plant. “Everyone will be in the digital age before too much longer,” says Gavin. “Like it or not, everybody is going to get there at some point.”
The first industrial revolution began when man and animal power were replaced with steam. Production improved significantly by mechanizing processes that moved under steam power. Factories saw another bump in productivity nearly a century later when assembly lines running on electrical power led to mass production. Computers and automation brought the next wave of production enhancements, speeding up manufacturing and reducing errors. Industry 4.0 represents industry’s fourth revolution, in which computer networking and cloud computing enable a new level of comprehension for every aspect of a manufacturing process. With this level of knowledge, manufacturers hope to learn how their processes can be improved, how to maximize output, and how to make changes most efficiently.
Technology giants, like Google and IBM, are at the forefront of this revolution (http://www.prnewswire.com), but traditional and smaller companies are taking interest too. Industry 4.0 allows companies to link raw materials, components, manufacturing, and logistics using the computational power of AI and the networking capabilities of the IoT. Companies are not blind to the potential for profitability this technology could achieve (https://tinyurl.com/yxma7qdn). Fats and oils businesses are no exception. Siemens offers digitization products specifically designed for agribusinesses and edible oil producers that are prone to low profit margins and could benefit from the technology. And manufacturers that incorporate agricultural raw materials and edible oils into their processes, such as Procter & Gamble and Unilever, are also looking at ways in which digitization can reduce their bottom line.
FIG. 1. Industry 4.0 indicates that manufacturing is undergoing its fourth revolution. First there was steam power, then electricity and computers. Now there is big data, AI, and the IoT.
“There is lots of potential in Industry 4.0 to provide continuous feedback and improvement loops,” says Gavin. Siemens has linked their process control systems with an edible oils process library so that manufacturers can apply real world data to realize better plant performance. They have also designed products that assist with the handling and milling of grains. Sensors measure the amount of grain in silos and the weight of grain transported to hoppers and trucks. Digitizing this information can help oil and grain manufacturers track their inventory more efficiently. “That helps customers make decisions quickly and helps them make production changes they need in a timely manner that helps them perform better,” says Gavins.
Mark Meili is the director of modeling, simulation, and digital innovation at P&G in Cincinnati, Ohio, USA. He is responsible for P&G’s data infrastructure strategy for both R&D and product supply chain. Meili says his company is in various phases of implementing Industry 4.0. “Like many other businesses, we have some areas that are actively doing things today, a few places where we have been doing similar things for a number of years, and then some other areas where we are not very far,” he says. “We are certainly getting more intentional about it, so there is more work under way than there was 12 or 18 months ago.”
According to Meili, P&G’s business areas are asking the questions that any industry should when considering Industry 4.0: What do we really need to measure? Can we measure that? Can we analyze the information from that measurement in real time? “We are doing a little bit of So what? analysis,” says Meili. “We should not do this just because we can. We should do this because we know that we can use it to drive important decisions that enable the business.”
Meili offers the example of P&G’s surfactants business, where they mostly use their own chemical feedstocks to make products using a vertically integrated process. Industry 4.0 gives them the flexibility to consider how to utilize a raw material like palm kernel oil most profitably, and how to run processes in a way that takes optimization into account for a variety of feedstocks. “By collecting enough data, over long enough time, the information from the process gives me confidence that I actually understand all these interactions,” says Meili.
Though optimization can lead to cost savings, another big buzz among manufacturers is how AI can use large data sets to teach them about their processes (https://tinyurl.com/yxma7qdn). AI tools, like machine learning, use processing data collected by plant instruments to develop a virtual version of the manufacturing operation. This faux factory is known as a digital twin. The Wall Street Journal predicts that in 2020, every major manufacturer will be using at least one digital twin (https://tinyurl.com/y67w8ehh).
Gavin explains that before digital twins were invented, if a company wanted to reformulate their product, they would likely have to start over from scratch by redesigning their manufacturing plant and then working out the necessary specifications by building a pilot plant. Now, his company provides simulation tools that analyze every aspect of a plant before it is built. “If a company can get a plant sized right and the flow of the plant modeled correctly in simulation, it saves them a lot of cost, it saves them a lot of effort,” says Gavin. “Ultimately, they can make a better product.” He explains that digital twins help companies adapt to changes more quickly.
Digital twins can be intertwined in what Siemens calls a digital enterprise. The digital twin of a product can be run through the digital twin of a manufacturing plant. The product formulation and the source of the raw materials can be collected in a cloud computing data base where there is a digital twin of every step in a process needed to bring the product to market. With this simulated workflow, if a company wanted to change a raw material supplier, for example, they can virtually change their digital twin first to identify any unforeseen production interruptions before committing.
Meili says that P&G would like to get to the point where they have digital twins across their businesses, because it makes sense. “We want to have as much information coming off of our manufacturing systems as we reasonably can,” he says. “We want to know how we can make our systems smarter. How can we make ourselves smarter by getting much more information out of our manufacturing systems than we have today?”
Internet of things
One way to get more information out of manufacturing systems is to have them connected. Just as smart devices have become common in the home, companies are looking to networks for smarter manufacturing. Refrigerators now realize when a household is low on milk and instruct smart devices like Alexa to make a purchase. No one is forced to abandon a bowl of cereal for lack of milk. Thanks to these networked smart devices, there is always an ample supply.
Similarly, Meili says, P&G would like to take advantage of this type of technology in their manufacturing plant. They would like to switch from time-based maintenance to condition-based maintenance. In other words, instead of shutting down production at random times to perform maintenance, they want their factory equipment to act as smart devices and notify an operator when they need service. “If we could have the manufacturing system running more of the time, it would be more productive,” says Meili.
The desire to have a manufacturing plant with this kind of connectivity and flexibility may seem decades away, but it is currently a reality. Unilever, based in London, United Kingdom, has eight digital twins of plants in North America, South America, Europe, and Asia (https://tinyurl.com/yxsutfzd). The company took three weeks to create its first digital twin of a facility in Valinhos, Brazil, that makes Dove soap. Sensors collected data, such as the temperature or motor speed of factory equipment, then sent the information to the cloud where advanced analytics mapped out the best operating conditions to produce Dove. The company reports that the digital twin saved them $2.8 million by highlighting areas where they could lower energy costs and increase productivity.
Unilever’s digital twins also give plant operators a means of testing quality without stopping production. Workers on site can track if process parameters are within specification using handheld tablets. If the production gets off track, they can quickly share data with other sites right from the production floor to save time as they and their colleagues work toward a solution.
Meili cautions that manufacturers will benefit most from Industry 4.0 if they truly understand their process. The technology should augment good science and engineering, but we need to remain in the driver’s seat—just as we do when Google maps sends us on a circuitous route when we know a faster way out of town. “There is no substitute for understanding what the manufacturing process is doing in the first place,” says Meili. “Then you can use these techniques to help you do it better.”
When applied to the agricultural sector, these digital technologies are referred to as agri-food 4.0. As the agri-food industry is increasingly forced to respond to a growing number of challenges, such as extreme, unpredictable weather patterns and an expanding world population, they turn to technology for solutions (https://tinyurl.com/y3qxhwsr). In 2011, the European Union established the sensing, smart and sustainable (S3) initiative to provide advice to industries on how to implement an efficient work process (https://s3platform.jrc.ec.europa.eu/s3-platform). The agri-food section of this initiative has funded the development of products to improve the food production process.
Just as companies like P&G and Unilever gather as much information from their manufacturing processes as they can, farmers seek similar data from their crops. Sensors located throughout fields and networked through the IoT assess molecular response of crops to water, fertilizer, and soil types. The data collected by the sensors not only signal when crops are running low on resources, but AI also evaluates the data to determine how to grow plants with a maximum amount of nutrients given their growing conditions. Imaging technology gathered by aerial drones assist farmers with monitoring their fields, but also provides AI with the directional data to instruct self-driving tractors and vegetable-picking robots. Farm technology is being designed to optimize every aspect of farm management. Experts believe finding better, smarter ways to grow food is essential for feeding the 9 billion people expected to inhabit the earth by 2050.
FIG. 2. A Thorvald vegetable-picking robot. © University of Lincoln
The changes that technology has brought to agriculture reiterates Gavin's point that eventually everyone will enter the digital age. As the world begins to acknowledge the limitations of its resources, it is relying on our computational capabilities to calculate how best to use them.
Rebecca Guenard is the associate editor of Inform at AOCS. She can be contacted at email@example.com.
- Sensing, smart and sustainable technologies for agri-food 4.0, Miranda, J., et al., J. Comp. Ind. 108: 21–36, 2019.
- How smart, connected products are transforming companies, Porter, M.E. and J.E. Heppelmann, Harvard Business Review, October 2015.
- Internet of things, big data, industry 4.0—innovative solutions in logistics and supply chains management, Witkowski, K., Procedia Eng. 182: 763–769, 2017.The role of additive manufacturing in the era of industry 4.0, Dilberoglu, U.M., et al., Procedia Manuf. 11: 545–554, 2017.
- Guest Editorial Special Section on Advances and Applications of Internet of Things for Smart Automated Systems, Mitsugi, J., et al., IEEE T. Autom. Sci. Eng. 13: 3, 2016.