Abstract
In the past, most of traditional master craftsmen always adopted the acoustic actions to recognition the situation of machine. Along with the development of time and technology, the mode of industry has changed with the Fourth Industrial Revolution (Industry 4.0). The long been known for the mother of industry, mold industry, has been inevitably impacted by Industry 4.0. This research stems from the structure of the six-level IoT model, through Internet connecting sensors, data collection, and the appropriate implementation of human and machine interface to intellectualize the injection molding machine. This research has collected 130 times of audio frequency, and there were 53 effective data sets, in sum there were 34,030,640 datasets. There were 5 manufacturing actions of petroleum molding machines that were successfully identified. Due to the low accuracy of one of the manufacturing actions, the training of audio frequency is based on the other four. In the end, there are 93.64% of accurate AI intelligent identifying models. Concurrently, through labeling the audio characteristics of different manufacturing parameters, the model recognizing audio characteristics from injection molding machines under different injection speed and rotation speed parameters is successfully trained. It is expected that in the future, other researchers can use this research as a reference to further strengthen the correlation between audio characteristics and injection molding machines to engage a more in-depth and diverse application of this topic.
JEL Classification: C80, C88, C90.