As a market, the smart factory is a freight train with a full head of steam and plenty of businesses hopping aboard. Global Market Insights’ research anticipates that the smart factory market will grow at a CAGR of more than 9% from US $110 billion in 2022 to US $280 billion in 2032.
That means a lot of executives in the boardroom have been convinced to begin the smart factory journey. It’s a journey that can take years and may never really be finished, but can also yield benefits after a single step. And there’s never been a better time to start. Manufacturing competition is intensifying as customers demand more customized products and sustainable production methods. Combined with uncertainty in the supply chain and difficulties with on-shoring, there’s a dire need for the efficiencies that a smart factory’s data-based visibility can provide.
A smart factory is a facility that enables the production aspect of smart manufacturing, which is the digitalization of all areas of manufacturing including product design, production, the supply chain, distribution, and sales. In practice, a smart factory is based on continuous improvement of processes based on digital data. Sensors embedded throughout the smart factory collect, send, and receive streams of data via a cloud computing platform. There, artificial intelligence (AI) models and machine learning algorithms analyze the data for actionable insights that inform decision-makers within the organization and/or automatically improve processes within robotics and automation systems in real time.
The smart factory and smart manufacturing are key components to Industry 4.0, or the Fourth Industrial Revolution, where the Industrial Internet of Things (IIoT), AI, big data analysis, cloud computing, and other technologies drive the digital transformation of industry. The previous industrial revolutions all introduced levels of manufacturing automation by way of the steam engine, assembly line, and computing power. Industry 4.0, however, is characterized by intelligent automation—automation that improves on its own based on a feedback loop of operational data.
The key components of a smart factory work together to collect and analyze data, then act on the insights from data analysis. When all of these components work together successfully, a smart factory can reduce costs while quickening the pace of production, as well as making operations more sustainable and safer for the workforce.
Smart factory data collection starts with an Industrial Internet of Things (IIoT) network, a system of devices and sensors equipped to send and receive data transmitted to and from a central location—usually a cloud computing system. The IIoT components may be integrated into modern factory machinery, or may be standalone devices and sensors installed onto industrial machines, vehicles, and other devices. All IIoT components communicate with each other by generating and transmitting real-time data. Some IIoT sensor and input types include ultrasound, radar, lidar, force sensors, and computer vision cameras.
An IIoT network enables the benefits of Big Data analytics, which include the real-time connectivity and seamless sharing of insights across systems and teams. This silo-breaking connected data improves collaboration, informed decision-making, and productivity—all in the service of innovation and customer satisfaction.
Big Data would not be such a big deal without artificial intelligence (AI) algorithms on cloud computing platforms to analyze it. AI data analysis translates the raw material gathered from the IIoT into valuable, actionable insights that can save time, resources, and money. Machine learning, an algorithmic subset of AI that empowers software to automatically improve from experience, can improve quality assurance using computer vision and high-powered GPUs to search for product defects at higher-than-human accuracy. For example, machine learning object-detection algorithms fed from ultrasound data could detect cracks in material better than people could.
Machine learning can also improve supply chain management—using computer-vision inventory management and data to optimize logistics routes and inventory space. It could also predict demand patterns to help align production volume with demand.
However, machine learning’s greatest boon may be its predictive maintenance ability. This works when potentially thousands of IIoT sensors on smart factory equipment record data on the equipment’s condition and machine learning algorithms predict from the data when a machine will need maintenance. Manufacturers can then plan the maintenance around periods of low demand or arrange for an alternate production line. Either way, the business avoids a costly shutdown from equipment failure.
PricewaterhouseCoopers’ (PwC) “Digital Factories 2020” reported that 66% of manufacturers planned to use predictive maintenance by 2025. And that effort seems to be worth the trouble, as Deloitte reports that establishing a predictive maintenance system in a smart factory can reduce maintenance planning time by 20–50%, increase equipment uptime and availability by 10–20%, and reduce overall maintenance costs by 5–10%, among other benefits.
That’s only the beginning of what AI models in a smart factory can do. AI analysis fed by sensor data can determine whether people are using equipment correctly. It can optimize manufacturing processes and make them more reconfigurable and flexible to changes according to fluctuations in demand, changes in the supply chain, and so on. Optimization may involve altering the shop floor layout or the sequence of manufacturing processes for a more efficient option. It may even determine that outsourcing a component or a process could be more efficient or that branching out to manufacturing parts for another industry would be expedient. AI models can even monitor environmental conditions, such as humidity, and alert staff if they are affecting production.
AI combined with auxiliary smart factory technologies, such as extended reality (XR)—encompassing virtual reality (VR), augmented reality (AR), and mixed reality (MR)—can further optimize assembly-line processes. For example, workers with MR headsets could see the entire assembly process visualized with enhanced guidance that improves their speed and precision. The headsets could also show overlays of assembly instructions and give workers helpful prompts based on sensor feedback.
The final segment in the chain that begins with digital data and ends with physical output is smart factory automation and robotics, which put the recommendations from AI data analysis into action. Armed with machine learning and AI’s data-crunching insights, robotics systems can take over complex jobs that are overly repetitive and/or dangerous for human workers, as well as reduce human error, avoid unscheduled downtime, and inspect for quality consistently—all contributing to efficiency and a healthy bottom line.
Industrial robots are part of a smart factory’s IIoT network; there may be thousands of sensors generating terabytes of data per day in a robotics system. As that data goes back into AI analysis, which continuously analyzes it for optimization opportunities, it ideally creates a feedback loop of continuous improvement.
An ideal smart factory workflow would put materials in and get parts and products out, with sensors and AI monitoring and refining every part of the process. People would control production but work within the environment only as much as necessary. Or alternatively, people could work more with cobots that are increasingly better at interpreting their surroundings. Either way, people will be freed up to work on creating new innovations in the design and manufacturing spaces, while automating the repetitive work.
As the smart factory approaches that scenario, the human workforce takes on a role of managing the robotics and automation. They design, control, and direct the connected technologies using creative problem-solving; businesses need to prepare their workforce to be comfortable working with data analytics, AI, and robotics. In Autodesk’s 2024 State of Design & Make: Spotlight on Digital Factories report (executive summary), 84% of manufacturing industry leaders surveyed said upskilling is important to their company, while 78% said they were investing in training programs for digital skills (PDF, pg. 8).
Tying all these smart factory components together are digital twins—dynamic digital models that can represent any physical object, even one as large and complex as an entire smart factory. In the case of smart factories and other buildings, an optimized digital twin reaches its full potential from accurate building information modeling (BIM) data. The digital twin tracks real-time data from building systems, as well as from the robotics systems and other machines within the digital factory model, in order to calculate such important figures like the smart factory’s total energy consumption and emissions output. AI algorithms also analyze that data to inform factory staff about where to seek efficiencies and make other improvements.
Digital twins offer unparalleled visualization of a complex facility’s operations. A smart factory’s digital twin can be the focal point for all kinds of computer-aided facility management, such as predicting when machines will need maintenance and scheduling it with minimal downtime, as well as numerous other types of risk mitigation and resource optimization. The flow of data through robotics, AI, and digital twins brings the manufacturing industry closer to the possibility of an autonomous smart factory.
Platforms like Autodesk Tandem cloud-based digital twin software provide a user-friendly solution for creating a building’s digital twin and harnessing and connecting all its data created throughout its lifecycle, beginning with planning and designing. This kind of platform makes it easier for operators to navigate a complex building model and to connect building systems so that they can monitor and optimize the asset lifecycle of a smart factory.
Together, all components of a smart factory comprise a cyber-physical system—a system where computational and mechanical elements interact with each other and are tightly integrated, so that both complement each other, making the whole system greater than the sum of its parts. That said, the cyber and physical elements don’t need to be located near each other to operate. Other examples of cyber-physical systems include smart energy grids and autonomous vehicle systems.
The primary advantage to all the data gathered from a smart factory is the ability to improve anything based on the clearest possible picture of all systems.
With IIoT devices connecting assets like machines and facilities throughout a smart factory, the data gathered and analyzed can reveal performance and maintenance issues that need corrective action, avoiding equipment failures and downtime. A smart factory also optimizes production capacity while reducing time to market. Deloitte’s 2019 smart factory study reveals that companies implementing smart factories experience gains up to 12% in manufacturing output and labor productivity. It also predicts that by 2030, smart factories will achieve 30% higher net labor productivity than traditional factories.
Smart factory data and technologies move operations away from reacting to circumstances and toward predictive and prescriptive actions based on analytics that make supply chains more resilient and the workforce’s efforts more productive. For example, more accurate demand forecasting enables optimized, just-in-time inventory management.
Like a digital transformation of a business, a smart factory transformation is a journey without a clear endpoint. You can see benefits after the first steps of implementation and continue to refine performance even after a smart factory is “finished.” The journey also differs depending on an organization’s circumstances. A clear first step is assessing needs with an initial systems audit, then planning a strategy around those needs. This should include executive buy-in and establishing clear goals, budgets, and desired outcomes.
The next phase requires integrating the technology for the “brains” of a smart factory: comprehensive data collection and the modern database, management, and analytics software systems to derive insights from the data. The technology to implement includes a secure IT/network system; an IIoT infrastructure to supply the devices, sensors, and computation to gather relevant data throughout the factory; and cybersecurity to make the system resilient. That technology then has to be integrated into the system so that the information technology (IT) and operation technology (OT) work together instead of independently.
With the data collection system in place, you’ll need the database and analytics software installed to derive intelligent insights for data-backed decisions. Most smart factories lean heavily on cloud connectivity, machine learning, and other AI algorithms to send data wirelessly to software like Autodesk Fusion Operations to track and analyze operations on the shop floor.
Businesses should undertake all these changes with the help of a diverse array of specialists from every department in its workforce, early on in the process. Greater employee involvement in the transformation will help the changes take effect, and you should also factor in efforts to retrain and upskill staff to work with any new equipment and systems. The size of a smart factory’s workforce may or may not change, but the skills its employees need to monitor systems and work with new data streams certainly will.
A smart factory transformation may happen rapidly or over time, but either way, it ushers in a great deal of change to an organization. Practice change management—where any changes to manufacturing processes are documented, reviewed, and approved before implementation—to keep the journey on track.
Finally, all of these strategies for implementing a smart factory are meant to be repeatable at dynamic scales. If, for example, a manufacturer transforms one of its many facilities into a smart factory, the lessons learned from that experience should apply to additional smart factory transformations of larger facilities or to the entire organization.
Support from the boardroom can determine how far along you’re likely to be in smart manufacturing. Based on this metric, there are three types of companies:
First is the trailblazer company, where the chief technology officer (CTO) takes the point position. This CTO will have the CEO, chief information officer (CIO), and chief financial officer (CFO) on board and will actively seek input and guidance from production/operations on how to best adopt smart manufacturing for the unique needs of their factory floor and workflows.
Next is the explorer company, usually led by production/operations staff who have identified a need to further automate processes on the floor and must convince the higher-ups it’s a good idea. Their closest ally is the CTO, who will bring the necessary clout to executive colleagues and will be prepared with input about how smart manufacturing will affect every other business unit, from risk assessment and cybersecurity to the plant manager.
Third is the follower company, again led by production/operations staff and with the CTO, CIO, and a few other department heads on board, but without the requisite case studies from other departments about how the smart factory will integrate with business goals at every level.
The difference among the three smart-factory leadership styles is both quantitative and compelling. Trailblazers (see infographic) spent up to 65% of their budget on the initiative and enjoyed a KPI increase of 20%. Explorers spent 19% for a KPI increase of 10%, and followers spent 13% for a gain of only 8%.
Of course, a smart factory is not a miracle solution that comes into being at the snap of fingers. Smart factory conversions come with high upfront costs that factor in new or upgraded equipment, workforce training, and cybersecurity. The smart factory market analysis report from Grand View Research estimates that manufacturers average two to five years to get a return on their smart factory investments—give or take some time depending on the scope of the smart factory transformation.
Fortunately, legacy manufacturing equipment—even if decades old—can usually be brought into an IIoT network without excessive hassle by adding gateway devices that send and receive digital signals.
Cybersecurity techniques are also keeping up against threats to smart factories; blockchain applications especially come into play for restricting access to machines and other assets, securing systems and keeping accurate records with “smart contracts” that can, for instance, track goods throughout the supply chain.
Organizations across the world and across product categories have succeeded in streamlining operations with smart factories. Deloitte’s The Smart Factory report (PDF, pg. 3) has aggregated improvements of 10–20% in asset efficiency, 10–30% in product quality and reduced waste, and 20–30% in reduced labor and inventory cost for manufacturers investing in smart factories.
The automotive industry has been particularly fertile ground for smart factories. Porsche built a smart factory for its first fully-electric sport car, the Taycan, with the goal of achieving a zero-impact facility that considers the emissions, resource consumption, and waste for the entire site. The Stuttgart, Germany factory is considered a model of sustainable, flexible, and smart manufacturing.
Another German electric carmaker, e.GO, used integrated factory modeling to save an estimated 35% on overall costs for building its smart factory. The company says it was the first automaker to integrate 5G connectivity throughout its factory, which enables stable, reliable, dedicated IIoT communication that allows real-time data to optimize operations.
Heating and air conditioning specialist Viessman utilized digital factory planning and augmented reality (AR), 3D visualization, and laser scanning to envision its heat pump smart factory, a carbon-neutral facility focused on efficient, sustainable production. Viessman created a digital twin of the factory during construction to inform the sustainability targets before completion and continues to leverage the digital twin’s insights during production for computer-aided facility management and predictive maintenance.
In building its digital smart factory for producing pulp and paper mill machinery, Austrian company ANDRITZ also relied on 3D modeled digital twins to optimize production processes virtually before putting them into practice. Clustering algorithms evaluate the data from the thousands of sensors in a pulp mill and reviews with their customers the potential for maximizing the mill’s operation. Such practices have resulted in saving many millions of euros for ANDRITZ’s customers.
The technologies of Industry 4.0 that power and augment smart factories—cloud computing, the IIoT, 5G, AI and big data analytics, robotics and automation, XR/VR/AR, digital twins—are maturing. People are already starting to talk about Industry 5.0, which emphasizes the harmonious interaction of humans and machines to achieve greater resilience, sustainability, and efficiency.
Smart factories and the technologies powering them have helped the profitability of manufacturing businesses through efficiency and productivity, but have also helped the industry cut waste for resources, energy, and materials; lower emissions; and make inroads to developing a circular economy by making products that can be reused or recycled. Finally, smart factories are generally better for workers. First off, they’re safer. Travelers Insurance conducted studies that suggest that automation and cobots within factories could reduce workplace injuries by as much as 72%. Not only that, but the retraining and upskilling programs often necessitated by smart factories can give employees higher-level, more marketable skills.
For all those reasons, smart factories benefit not only a manufacturing business, but also its workers and the world at large. The smart factory is here to stay, and it will only get smarter with time.
This infographic was originally published in January 2021. Drew Turney & Missy Roback contributed.
Markkus Rovito joined Autodesk as a contractor six years ago and joined the team full-time as a content marketing specialist focusing on SEO and owned media. After graduating from Ohio University with a journalism degree, Rovito wrote about music technology, computers, consumer electronics, and electric vehicles. Since his time with Autodesk, he’s developed a great appreciation for exciting emerging technologies that are changing the world of design, manufacturing, architecture, and construction.
Emerging Tech
Executive insights
Emerging Tech
