Amber Biela-Weyenberg | Content Strategist | Oct 16, 2023
Smart factories use a network of machines, devices, and advanced technologies to automate manufacturing processes and fuel smarter decisions.
The vast majority of respondents to a 2023 survey of 1,350 manufacturers in 13 countries by Rockwell Automation said they use or plan to use smart manufacturing technologies—which include the Industrial Internet of Things (IIoT), AI-based data analytics, digital twins, and advanced information security—over the next one to two years. Among their goals: improve product quality, lower costs, boost profitability, and build a more adept workforce.
But twice as many manufacturers as in the 2022 survey said they worry that they’re falling behind competitors when it comes to adopting the latest technologies. One-third of the manufacturers surveyed said the wide range of systems and platforms to choose from leads to “technology paralysis.” What follows is a guide to help manufacturers avoid such paralysis in building out their smart factories.
Smart factories involve many technologies that must integrate with each other, so planning is essential. Manufacturers must evaluate which technologies they’ll use and develop key performance indicators (KPIs) to measure their impact. For example, to gauge the efficacy of outfitting a machine with sensors that would indicate the need for maintenance, manufacturers might measure overall equipment effectiveness (OEE). That KPI considers the amount of time a machine is available for production, the speed of its work, and the quality of its output. Tracking overall labor effectiveness (OLE), or the overall utilization, performance, and quality of a manufacturer’s workforce, can help quantify the value of adding robots to a production line.
Meanwhile, manufacturers will need to train employees or hire new people. For example, smart factories generate massive amounts of data, collected from sensors on myriad machines and devices. Employees must understand how to set up and maintain these sensors—including the software that controls them—and use the information they generate to improve OEE, OLE, capacity utilization, cost effectiveness, and more.
Manufacturing executives need to get buy-in from factory managers and IT professionals before building a smart factory or introducing “smart” technologies and processes to a current factory. Feedback from workers on the factory floor is also advisable. For example, a machine operator may suggest which metrics to track on a production line or how to reconfigure the line to increase performance output. Manufacturers need cooperation from all employees in testing new technologies and equipment before using them in day-to-day operations. In each stage of your smart factory initiative, manufacturers should rally support from stakeholders as early as possible to smooth the transition.
When transitioning a conventional factory to a smart factory, a phased approach makes the process more manageable and prevents downtime. Manufacturers may prioritize areas with the greatest return on investment, such as a high-value production line. They should set at least one KPI for each stage of the modernization to track progress. For example, one phase of smart factory implementation involves installing sensor-enabled, internet-connected machines and incorporating robots into production processes. At this stage, manufacturers might measure throughput to track progress toward speeding production. At a later stage, measuring machine downtime can gauge the accuracy of AI-based predictions about the need for machine maintenance and the production team’s efficiency in responding.
One of the first implementation steps involves the IT team, which must configure and secure the networks that ferry data between machines, objects, and software. 5G systems are becoming the network of choice for many smart factories, given their increased capacity and low latency. Machines, robots, and other network-connected “things” outfitted with sensors share data with factory floor decision makers or inform automated processes. For example, a software engineer could program a factory machine to shut down automatically if its sensors record a certain temperature, to prevent damage and employee harm. Managers could also use data generated by factory machines to track OEE and identify improvements to machines and related processes. And they could use IIoT data to measure energy usage and resulting carbon emissions across the factory floor to meet sustainability goals and comply with emerging reporting regulations.
Smart factories collect massive amounts of data, or big data, during every stage of production from machines, devices, robots, and the applications that manage manufacturing, project management, and other back-office processes. All that data needs to be stored, processed, and managed in a data warehouse, then fed into analytics systems so it’s readily available to manufacturing executives, managers, and supervisors.
Artificial intelligence and its subset machine learning (ML) correlate, interpret, and constantly learn from the steady streams of data they ingest to help companies make better decisions. In a manufacturing context, these technologies are incorporated into the business software used to inform product development, forecast demand, predict factory equipment breakdowns, identify the potential for product defects, reduce waste, optimize transport routes, and more. A smart factory’s robots and machines, for example, could automatically scale production of a particular item if an AI- or ML-based analysis predicts a rise in consumer demand. The smart system reviews data from the outcome and learns from it to make better decisions in the future.
Companies that combine data from factory floor machines with data from manufacturing, supply chain, finance, sales, HR, and other enterprise applications in an enterprise resource planning (ERP) system are better able to solve problems than those that don’t. Suppose a parts shipment from a supplier is running late. A smart factory’s interconnected systems can pull together data on supply chain management, inventories, and customer orders to reveal whether the delay will affect the manufacturer’s ability to fulfill orders on time and the need to get parts from another supplier or location.
Because smart factories connect a multitude of equipment, devices, and applications, sometimes on a single network, vulnerabilities in any single system can expose the manufacturer to a debilitating security breach. Types of attacks include malware (ransomware is a particular threat), theft of personally identifiable information and valuable intellectual property, and denial of service. Vigilant system patching, complex passwords, and user training on how to avoid phishing and other social engineering techniques are a must, augmented by the latest security systems. It’s one reason that manufacturers are turning to cloud computing for their back-office applications, managed by providers that apply the latest security patches, tools, and techniques built into every layer of the stack, including business software, servers, and other hardware.
A quick scan of smart factory open positions advertised on LinkedIn reveals the following titles: director of intelligent manufacturing systems, senior manager of strategic insights and analytics, robot teaming coordinator, digital transformation lead, and autonomous vehicle specialist. While manufacturers still need lots of welders, machinists, factory managers, and production line workers, the profession is clearly becoming more technical, requiring employers to upskill their people and hire new talent. For instance, it may be relatively easy to upskill a factory manager to become a smart factory manager, but a role such as digital twin engineer requires more specialized skills that most employers will have to find elsewhere.
Although automation is a big benefit of smart factories (improved productivity and safety and reduced costs and human error), highly skilled managers and supervisors are still crucial. For example, robots perform many smart factory tasks, but robotics professionals need to program and maintain them. Machine learning algorithms can analyze a huge amount of product data to gauge consumer demand, uncover flaws, and inform ways to improve product designs, but that’s just one set of inputs that warm-blooded product design managers consider.
Oracle Smart Manufacturing is a portfolio of cloud-based applications, many of them imbued with AI, for managing the Internet of Things, equipment maintenance, quality control, supply chain planning, and business analytics. Among its many benefits, the integrated portfolio helps manufacturers monitor production performance, adjust production schedules, improve product quality, and avoid costly downtime.
What makes a factory a smart factory?
Unlike a conventional factory, smart factories network machines, devices, and applications—in many cases leveraging AI and machine learning—to collect and analyze data, automate processes, and improve decision-making.
Which technologies does a smart factory use?
Smart factories generally include some combination of the following technologies: the Industrial Internet of Things (IIoT), advanced data analytics, AI and machine learning, digital twins, robots/cobots, 3D printers, 4G/5G wireless networking, cloud-based applications, edge computing devices, and advanced information security.
How do smart factories change jobs?
While smart factories still employ plenty of production line and oversight workers, automation is replacing some jobs and requiring more technical skills for others. Specialists in robotics, data management, data science, software engineering, AI/ML, project management, and network/system administration are in high demand, with an emphasis on critical thinking and problem solving.