基于BP神经网络的深孔切屑形态预测模型

BP-neural-network-based Chip Form Prediction Model for Deep-hole Drilling System

针对内排屑深孔加工系统排屑难、效率低的问题,建立了切屑形态预测模型,确保深孔加工的顺利排屑和加工过程的顺利进行。通过对切屑形态进行整理和分类,利用十进制编码将其转化为数据结构。首次利用BP神经网络系统,对45#钢、灰铸铁两类加工材料建立切屑形态预测模型。应用效果表明,该模型可根据深孔加工的切削用量和冷却液参数对切屑形态进行基本预测,为切削用量的合理选择及优化提供理论依据。该模型的建立对提高深孔加工效率和降低加工成本有很高的实用价值。

Aimed at the difficulty in chip removal and the low efficiency in a deep-hole drilling system, a chip form prediction model was established to ensure good chip removal and smoother hole-making. Based on their morphological characteristics, chip forms were analyzed and classified into different categories, and then transformed into data patterns by using decimal encoding. It is the first time that two chip form prediction models for 45# steel and gray cast iron were estab- lished by using the BP neural network knowledge. The experiment results show that the model can approximately predict a form style according to the cutting and coolant parameters in a deep-hole processing system. The model can also provide a theoretical basis for choosing and optimizing cutting parameters. The establishment of the model has a high application value in enhancing machining efficiencies and reducing processing costs in deep-hole drilling.