CONTRACT MANUFACTURING
AI Applications in SMT Electronics Manufacturing
Joe Booth, CEO, Altus Group A
rtificial Intelligence (AI) has become a buzzword across industries, yet
understanding its real-world applications often lags behind the hype. Despite this, AI is already significantly impacting sectors like electronics manufacturing. While some envision entirely autonomous “lights-out” factories, I believe that this level of automation is unlikely to be fully achieved. The most substantial benefits will come from the first 90 percent of automation, with the final 10 percent being the most challenging and resource-intensive. Over the next decade, AI will continue transforming manufacturing, solving current challenges and enabling process improvements once considered impossible.
The Evolution of Smart Manufacturing
The “Smart Manufacturing Vision Pyramid” from 2022 provides a framework for understanding the progression of manufacturing technologies. Aspects of Smart 1.0, such as basic automation and data collection, are now taken for granted, while Smart 2.0 elements are increasingly common in modern factories. Moving forward, several key stages are critical to realising the full potential of smart manufacturing: 1. Data Acquisition: The foundation of
smart manufacturing is the ability to collect accurate, repeatable and high-quality data. Without reliable data, the subsequent stages of analysis and automation are undermined. 2. Data Insights: Statistical analysis is used to derive insights once data is collected. These insights enable manufacturers to identify trends, optimise processes and make informed decisions.
3. Digital Twin and Simulation: Using real data, manufacturers can create digital twins; virtual replicas of physical systems. These digital models allow for simulation and experimentation in a risk-free environment. For example, before building a new production facility, a digital twin can test different layouts and process flows, optimising efficiency and minimising potential issues. 4. Prediction: Predictive capabilities are where AI truly begins to shine. By combining real and simulated data, manufacturers can forecast future outcomes, such as when a machine is likely to fail or when maintenance is needed. This allows for proactive interventions, reducing downtime and increasing productivity. 5. Autonomy: The ultimate goal is autonomy, where machines communicate with each other, make decisions and adjust processes in real-time without human intervention. This could involve autonomous process tuning,
automatic calibration or even self-healing systems that correct errors before they impact production.
Complete autonomy may still be years away, but many of these capabilities are already available, thanks to advancements in AI for Surface-Mount Technology (SMT) manufacturing.
Real-World Applications of AI in SMT Manufacturing
Koh Young, a leader in SMT technology, has been at the forefront of integrating AI into their processes, particularly in
measurement-based inspection technologies like Solder Paste Inspection (SPI) and Automated Optical Inspection (AOI). These systems generate vast amounts of data, which AI can harness to improve manufacturing processes.
One common challenge in SMT
manufacturing is developing robust inspection programs, especially during New Product Introduction (NPI) when data is limited. For example, with only a few product units during NPI, the inspection program may not have enough data to optimise programs to accurately detect defects, increasing the risk of faulty products slipping through or false positives requiring unnecessary review. AI
20 OCTOBER 2024 | ELECTRONICS FOR ENGINEERS
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54
