Development of an intelligent deep-hole boring tool

通过对加工条件的监测，即赋予深孔刀具视觉、听觉、触觉、味觉、嗅觉等五种感官功能，实现了对L/D*1000型超深孔的优质加工。明确了以下三项内容。(1)对老式激光导引深孔钻具进行了改进，使其具有更高的性能。最初的工具在光学系统中安装了一个He-Ne激光器来检测其姿态，并使用凸轮作为致动器。新的一台配备了氩气激光器和压电式转换器。用JIS A2017-T4硬质合金工件预钻孔108 mm进行的基本实验表明，该刀具具有较高的操纵性和导向性能。(2)通过工业内窥镜观察，利用CCD摄像机和图像处理，可以在离线信息处理系统上简单地表征表面粗糙度。利用图像处理技术可以在线观察切屑的排屑和尖端情况。(3)利用切削声作为输入，构建了一个人工神经网络来检测异常切削量。在车床上进行了实验，以检验人工神经网络能否对深孔加工中可能出现的颤振、多角孔形成和刀具磨损等切削条件进行分类。因此，可以脱机对切割条件进行分类。利用人工神经网络对整个深孔钻进系统进行了评价。

The research was conducted to bore high-quality ultra-deep-hole (L/D*1000) by monitoring the cutting conditions, i.e., by giving functions of such five senses as sight, hearing, touch, taste and smell to a deep-hole boring tool. The following three items were made clear. (1) The old-type laser-guided deep-hole boring tool was improved to be of high-performance. The original tool had a He-Ne laser in the optical system for detecting its attitude and used cams as actuators. The new one is equipped with an argon laser and piezoelectric translaters. Basic experiments using duralumin workpieces (JIS A2017-T4) with a prebored 108 mm diameter hole showed that the new tool has high performance in manipulation and guidance. (2) The surface roughness could be simply characterized on the off-line information processing system by observation through an industrial endoscope, using a CCD camera and image processing. The chips exhausting-and tip-conditions could be observed on-line using image processing. (3) To detect the abnormal cutting, an Artificial Neural Network was constructed using cutting sound as input. Experiments were conducted on a lathe to examine whether the ANN can classify cutting conditions of chattering vibration, formation of multi-cornered hole and tool wear, which are expected to occur in deep hole boring. As a result, the cutting conditions couled be clasified off-line. Evaluation of the total deep-hole boring system using ANN continues hereafter.