Here to Stay: Why Manufacturers Need to Integrate AI Solutions

Staff
By Staff
8 Min Read

While we’re still in the relatively early stages of its implementation, artificial intelligence has cemented its place in modern manufacturing as we continue to develop new applications for it. 

Understandably, some manufacturers are hesitant to dive head-first into AI; they rely on established processes and introducing any artificial learning element represents an unnecessary unknown. Still, the benefits of AI far outweigh the drawbacks, and by further understanding its uses, end-users can better leverage its integration, which can ultimately lead to operational success.

At its core, AI is about managing data. Manufacturers need to start there by having a firm grasp of what data they need to collect, store and use for actionable insights. 

It’s not enough to receive data regarding performance metrics; it needs to be leveraged as a proactive tool to guide future efforts. Moreover, the data that’s being monitored needs to be accurate and processed in real-time, as conditions constantly change within manufacturing. 

Inaccurate or outdated data can be both detrimental and costly to operations. However, without the proper infrastructure to digest and present it concisely, manufacturers can be flooded with information that they don’t know what to do with. 

There is a myriad of software tools that can help translate it, so strategy is everything when approaching AI implementation. In fact, many end-users are creating dedicated teams that manage the influx of data, which speaks to its importance regarding a company’s bottom line.

Cybersecurity is equally as important when considering AI, as threats are evolving daily, both for IT and OT. With an increasing number of interconnected components incorporating AI software, there are more touchpoints that attackers could compromise if the proper safeguards aren’t in place. 

End-users need to implement firewalls, encryption and a robust data layer to protect against these multi-level attacks. The data layer acts as a filter in identifying faulty data and threats while protecting the various applications that are operating in the system. 

Regarding OT security, access control is a critical component of AI usage, as a variety of measures are available. Facial recognition is an emerging component that can be utilized both in HMI settings and even on remote access for phones, which many of us already use on our personal devices. 

Additional security concerns include USB drives on computers, which are becoming even more common in factories amidst a general shift to digitalization. Attackers could simply insert a thumb drive to steal sensitive information, so it’s important to have the necessary firewalls and multi-factor authentication installed to protect against such risks. 

Cybersecurity is such an important part of the future of manufacturing that regulations continue to be developed, including the Cyber Resilience Act, which is already being enforced in Europe and is expected to be adopted in the United States soon. 

Manufacturers will need to prove they have the necessary safety measures in place to respond to cyberattacks and provide clear audit reports of operational processes. Additionally, suppliers will need to do the same, and we at Bosch Rexroth are prioritizing staying Cyber Resilience Act-ready.

AI integration relies on strategic communication between IT and OT components, like Bosch Rexroth’s ctrlX CORE shown above, which helps streamline controls and manage data that’s critical to machine learning applications.Bosch Rexroth

Security and safety go together, which brings up another unique component of AI-integration: cobots. Like AI, they’re being utilized more often in manufacturing, and they, too, need to be managed efficiently with proper safeguards. 

Controllers can be programmed so that a cobot recognizes and reacts to human presence on the floor and subsequently operates at a safe speed. Cobots can perform some of the more strenuous or dangerous tasks within manufacturing, which further keeps employees safe and allows them to focus on tasks better suited for humans. 

Cobots are also valuable in their flexibility, which is a noteworthy change from traditional robots that were typically programmed for a singular task. As manufacturers’ needs shift, cobots can adapt along with them, which reduces the need to stop processes, reprogram them and resume production. 

That’s especially helpful in industries like consumer-packaged goods, which often have multiple types of products on a single line. Additionally, smart vision systems that are integrated with cobots are an efficient quality control measure. 

Cobots can learn to identify flaws in products earlier in the process and help reduce scrap, which obviously saves the manufacturer in supply costs and boosts efficiency.

Another essential use case for AI is in preventive maintenance. Machine learning software can be programmed to detect anomalies in the machine itself by monitoring key sensor information like temperature, vibration, pressure and even oil analysis for hydraulic systems. 

AI can determine what “normal” operating conditions are and what an acceptable variance is before notifying managers. This can extend the life of the machine by recommending slight preventative adjustments before a larger malfunction occurs. 

Manufacturers can also input historical data into their software system to teach AI what their typical MTBF (Mean Time Between Failures) has been. This helps customize the machine learning so that the AI isn’t relying on the stock recommendations from the machine manufacturer, while the end-user can better predict how long that machine will last under their unique conditions. 

Finally, supplementary information regarding the machine, like manuals, help files, etc., can be uploaded into the AI model to further inform future recommendations.

Speaking of the future, AI usage will only grow in the coming years as the benefits, as noted above, directly impact productivity. AI is also involved in demand forecasting, advising manufacturers if they are producing the right number of products for a typical season, as well as providing a digital twin-like simulation of what a process footprint might look like under certain conditions. 

It’s important to emphasize, however, that while it’s easy to get caught up in AI possibilities, manufacturers must base their strategies on clear goals of what they want to monitor and how they will protect that valuable information from ongoing threats. 

AI technology isn’t worth the investment if it’s not accurate, relevant to the manufacturer’s needs, or properly programmed. Nevertheless, it’s an exciting time to imagine what manufacturing will look like in the near future, as a new wave of intelligence helps guide daily decisions.

Garrett Wagg, ctrlX AUTOMATION Product Manager, Automation & Electrification, Bosch RexrothGarrett Wagg, ctrlX AUTOMATION Product Manager, Automation & Electrification, Bosch RexrothBosch Rexroth

Share This Article
Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *