Summary |
One innovative “Intelligent Predictive Maintenance System.” has been developed and integration. It contains a Smart Vision Module, an Edge Computing Device(hardware), and an Intelligent Prognosis Software called Prolepsis. The System is to perform the in-situ on line predictive maintenance of equipment using the state of arts techniques corresponding with machine learning, embedded smart visioning, edge computing, and cloud-based industry IOT, etc. It provides such precise-and-robust predictions of imminent failure to assist engineers to decide the best-time-of-repair, and thus it enables to increase the availability for the equipment and production yield. In a case of commercial wafer handling robot arm, we have demonstrated the embedded intelligent predictive maintenance core system enables to predict in-situ the maximum eccentric quantity in the next minutes to prevent the wafer-drop failure during the handling period. |
Scientific Breakthrough |
One innovative “Intelligent Predictive Maintenance System.”has been developed. It was successfully integrated to perform the in-situ on line predictive maintenance of equipment using the state of arts techniques corresponding with machine learning, embedded smart visioning, edge computing, and cloud-based industry IOT, etc. It provides such precise-and-robust predictions of imminent failure to assist engineers to decide the best-time-of-repair, and thus it enables to increase the availability for the equipment and production yield. We have successfully demonstrated the embedded intelligent predictive maintenance system to the case of commercial wafer handling robot arm to predict in-situ the maximum eccentric quantity in the next minutes to prevent the wafer-drop failure. |
Industrial Applicability |
A Predictive Maintenance System has a Smart Vision Module, an Edge Computing Device, and a Prognosis Software called Prolepsis. It uses machine learning, embedded smart visioning, edge computing, and cloud-based IOT techniques to perform on-line predictions of imminent failures to decide the best-time-of-repair. It can be applied to most of manufacturing-based companies corresponding with mechanical, chemical, electric, semiconductor, optoelectronics, etc. It was successfully demonstrated to a wafer handling robot arm to prevent the wafer-drop failure. |