Smart manufacturing made smarter with MIT solutions

Needless to say, IIoT or industrial IoT has become a major trend, utilized by more and more manufacturers across the globe. In this regard, Taiwan solutions providers have a lot to offer. This note looks at some of the advanced, cutting-edge MIT smart manufacturing solutions, which were on display at Secutech 2023.
 

Predictive maintenance

 
Predictive maintenance has become a term we hear a lot these days. Indeed, machine failure can lead to production shutdowns, which can be costly to manufacturers. Today, thanks to IoT and AI, equipment that are about to fail can be detected or predicted, and dealt with accordingly.
 
As an example, Harbor displayed their AI-enabled vibration sensor that can detect abnormalities in vibration and predict impending failure.
 
“Our sensor is placed close to the vibration source, for example the motor of a machine. While the motor turns, the sensor gathers raw data which is then ingested into an edge analyzer that can build various models for outcomes. If one day the sensor determines that the machine is heading towards the fail outcome, it will issue an alert, and the operator can act accordingly,” said Ricky Cho, Project Manager at Harbor.
 
According to Cho, the sensor also has applications in other fields. “In civil engineering or smart city applications, the sensor can be placed near bridges. For buildings, it can be placed near water tanks/pump. But for now, the primary application is still smart manufacturing,” he said.
 
Indeed, AI now plays a key role in predictive maintenance. Profet AI, another Taiwan company, brought to Secutech its AI that also has predictive maintenance applications.
 
“One of our major clients is a cement company with lots of cement mix machines that simply can’t stop. So our software is deployed at their factory, and it can predict when maintenance is needed. The company feels very good about it and has collaborated with us for quite some time,” said Richard Tu, Special Assistant to GM at Profet AI.
 
Tu further cited other smart manufacturing applications. “For example, gold plating is a key process for making printed circuit boards. You can tell the AI the thickness you want, for example 0.3 cm, and it will tell you how to set the parameters,” he said.
 

Machine vision

 
Machine vision is also key in smart manufacturing, whereby minute, miniscule defects can be detected by AI-based software instead of human eyes, which have limitations and can get easily fatigued. In this regard, Foresight has AI solutions.
 
“The AI inspects the image of the prototype, looking for defects. If defects are found, the prototype will not proceed to production. Use scenarios are quite varied, ranging from solar panels to IC packaging. Solar panels, for example, may have tiny cracks on the surface, and the AI can detect it,” said Yeh Yao-chih, Director at Foresight.
 
He explained the benefits. “Before you might have three to five inspectors. Some may see defects and some may not. Some can see details and some cannot. With AI, inspection can be much enhanced. This can help save manpower, not to mention reducing eye fatigue,” Yeh said.
 

Integration

 
Needless to say, there is a wide range of smart manufacturing AI solutions out there. An open platform, therefore, is vital in integrating different AIs to meet users’ different industrial needs and requirements. During Secutech, Nuuo displayed their open VMS platform, which supports a range of AI solutions including those made by Solomon, whose AI can convert analog meter video into digital data.
 
“When the temperature or humidity crossed the threshold – for example the humidity threshold is 58 percent and the meter reading shows 70 percent – the AI will detect it and issue an alert on the VMS, or remotely to the operator via his mobile app; the operator will immediately know what and where the problem is. In addition, if integrated with the access control system, this sudden surge in humidity will also prompt the door to open immediately for safety,” said Simon Hsu, Sales Manager at Nuuo.
 
To make this possible, openness and integration are key. “In addition to Solomon’s AI, we also support door readers, facial recognition and other AIs from different brands. Really, with openness and integration, we can achieve more with less,” Hsu said.