Wang Kang, Wang Aimin, Wu Long
Digital Manufacturing Institute, Beijing Institute of Technology, Beijing 100081, China.
School of Mechanical and Electrical Engineering, Shandong Jianzhu University, Jinan 250101, China.
Sensors (Basel). 2024 Dec 19;24(24):8107. doi: 10.3390/s24248107.
Accurately predicting tool wear during the machining process not only saves machining time and improves efficiency but also ensures the production of good-quality parts and automation. This paper proposes a combined variational mode decomposition (VMD) and back propagation (BP) neural network model (VMD-BP), which maps spindle power to tool wear. The model is trained using both historical and real-time data. To improve accuracy, the internal power data from the machine tool are used to calibrate the model's input data. Data collected from milling experiments are used to test the model, with sensor-collected power being compared to the model's predicted power. The average error was 1.1256%, which confirms the reliability of the model. In practical applications, the model enables the real-time monitoring of spindle power, helping prevent excessive tool wear during machining. This offers significant guidance for actual production processes.
在加工过程中准确预测刀具磨损不仅可以节省加工时间、提高效率，还能确保生产出高质量的零件并实现自动化。本文提出了一种结合变分模态分解（VMD）和反向传播（BP）神经网络的模型（VMD-BP），该模型将主轴功率映射到刀具磨损。该模型使用历史数据和实时数据进行训练。为提高准确性，利用机床内部的功率数据对模型的输入数据进行校准。通过铣削实验收集的数据对模型进行测试，将传感器收集的功率与模型预测的功率进行比较。平均误差为1.1256%，这证实了该模型的可靠性。在实际应用中，该模型能够实时监测主轴功率，有助于防止加工过程中刀具过度磨损。这为实际生产过程提供了重要指导。
