As the era of 5G has arrived, more and more AI applications are being introduced. The trend requires not only chips of higher processing and computing performance, but also servo drives and control systems with matching capabilities. In the field of industrial automation equipment in Taiwan, Delta is one of the leading companies. Recently, it has developed an AI-based technology for motor noise inspection, which greatly improves inspection accuracy to nearly 100%.
Delta has applied this technology to its motor production lines to enhance the yield rate and quality of its motors. Jen-ming Chang, manager of Delta’s Design Department, said that this experience has advanced Delta’s inspection expertise which it can pass on to help more customers in related fields.
Experience-based Motor Inspection Can Now Rely on AI
Motors have always been an important power source for modern machines and automation equipment. Should an irregularity occur, the lifespan of the motor might be shortened, which could affect the operation of the whole system. Delta makes sure its motor products undergo strict inspection before delivery to customers from the factory.
Motor inspection focuses on current, vibration and noise. Noise detection is the most adopted approach, which is carried out either through experienced listening or with equipment. According to Chang, these methods rely heavily on experience, which limits the accuracy of defect detection. “If you want to improve the quality of motor inspection, you have to go for AI technologies,” he said.
Delta’s ‘Abnormal Noise Inspection Technology’ Turns Experience into Analytics
Standardized quality is the major benefit of using AI technologies. Noise testing is applicable not only to motors, but to all kinds of equipment errors as malfunctions can often make the equipment sound differently during operation. Analyzing failures, experienced technicians can discover problems and deploy preventive measures.
Relying on experts for troubleshooting ensures the products receive personal attention, but equipment providers still face difficult challenges. For example, human hearing cannot ensure consistent quality. No matter how well a technician specializes in noise testing, once a personnel change occurs, it is difficult to pass on the experience.
“In short, the problem is that you cannot quantify experience and perception,” Chang explains. Subjective perception and experience lead to varied inspection standards. “Even if the inspection is carried out by the same person,” he elaborates, “inspection standards can still be affected by auditory fatigue.”
Self-developed Solutions Embody Delta’s Emphasis on Continuous Product Optimization
As Delta’s core products include motors and fans, it has already adopted automated testing equipment for production lines that solve problems such as varied standards and auditory fatigue arising from human auditory-based inspection. The equipment overcomes the challenge of quantifying sounds, but for advanced inspection further development is still required.
“For now,” said Chang, “automated noise detection relies heavily on experience to configure parameters.” As a result, the method can only be applied to inspection motors with consistent rotation speed. Once the speed varies, the system will fail to tell which products are defective. The failure “will cause batch-to-batch variabilities, and therefore affects our brand reputation.”
As a result, Delta has started to apply AI technology to automated motor abnormal noise inspection system in recent years.
AI Algorithm Increases Inspection Accuracy and Reduces Workload of Re-Inspection Staff
An AI algorithm performs self-learning through database references. It can use certain logic to determine the result even if the conditions are complicated. Chang indicated, “This feature can be applied to the abnormal noise inspection of motors with different rotation speeds.”
The application allows Delta to elevate the speed and accuracy of motor inspection even more. In the past, the Company adopted automated inspection systems to quickly determine motor quality, and then products would be re-inspected by staff. Now, AI is added between these two processes for re-inspection, and the inspection staff conducts the final testing.
“This method affects our production lines the least, but can significantly improve the accuracy of motor inspection while reducing the workload of inspection staff,” Chang pointed out.
AI Shows Great Potential with Nearly 100% Defect Detection Rate
Delta has applied the AI-based motor inspection system to part of its production lines, and the defect detection rate, as expected, has increased more than ever.
The first pass yield rate of the non-AI system was about 80%. After adopting the AI-based system, the defect detection rate reaches up to 90%-95%. Chang mentions, “If we exclude minor defects and only take into account the moderately and heavily defective products, the detection rate can reach nearly 100%.”
In the future, the AI-based system can be applied to other production lines and commercialized for all machinery equipment providers. Delta will also continue to develop more functions for the system. The application of AI technologies offers great potential, serving as a powerful basis for many innovative ideas. “In the future, we can even use AI to identify all of the different motor noises. We can find out the reasons for malfunctions and optimize production processes to raise yield rates,” Chang believes.
Applying AI to Production Lines with Complicated Conditions
Delta puts its own solutions into practice based on its solid expertise in manufacturing and practical application. Currently, some suppliers are introducing similar AI solutions with simpler functions to the market. However, these solutions cannot meet the needs of production lines with complicated conditions. On the other hand, Delta tests its software and hardware products on its own production lines, and the products are commercialized only after passing specific tests.
“Delta has already solved the problems that users may encounter when adopting our products,” says Chang, “which allows our clients to use the products right away.”
Leveraging AI Technology to Convey Delta’s Green Spirit
In addition to increasing inspection efficiency, Chang highlights that the AI-based inspection system clearly reflects Delta’s corporate mission: “To provide innovative, clean and energy-efficient solutions for a better tomorrow.”
In the past few years, going green has become a global trend. As companies gain more resources and have a greater impact, changing their own products to allow greener manufacturing becomes a contribution that creates environmental benefits. The development of this AI-based system is just one example. Leveraging AI technologies, Delta has significantly reduced defect rates while extending the lifespan of motors.
“Using resources wisely can relieve the burden on our planet. Meanwhile, companies can avoid material waste and unnecessary costs. We continue to further our deployment in smart manufacturing to achieve our sustainable vision for a win-win outcome for developing the economy and preserving the environment,” concludes Chang, revealing the important goals that are driving industries to adopt AI technologies.
(Quoted from TechOrange: https://buzzorange.com/techorange/2020/12/09/delta-mortar-ai-qc/)