Monitoring of Machining Process of Machine Tool based on Artificial Neural Network using Serial Data

Recently, research on the application of knowledge-based smart technologies such as unmanned, optimization, and monitoring is increasing in the machine tool industry. Traditionally, the processing conditions of machine tools are very complex depending on various processing conditions, workpiece material, tool characteristics, and chip evacuation characteristics, so it was not easy to predict. In addition, despite numerous studies on monitoring the machining process of machine tools, it is very difficult to monitor the state of the machine tool and deal with it unmanned due to the complex characteristics of the machining process. In the actual field, basic monitoring technology is applied and warning and notification functions work when problems occur, but the operator must take action to solve the problem. Therefore, in order to enable unmanned operation, a technology that predicts the state of a machine tool in advance is required. It is possible to develop a technology that predicts the state of a machine tool by finding the relationship between problems occurring in the machine tool and the resulting state using an artificial intelligence algorithm and combining it with monitoring technology. In order to monitor the processing state of the machine tool, this paper measures the signal of the machine tool and analyzes it using an AI-based analysis model to determine the relationship between the processing state of the machine tool and the signal during machining. Collects signals that fluctuate according to the processing state of the machine tool, and an AI model for multi-class classification was designed and trained by combining CNN, LSTM and multi-layer perceptron (MLP). The monitoring of machine tool processing status and tool status was performed by the learned AI model using the serial data of machining process of the machine tool.