Chattering detection in cylindrical grinding processes using the wavelet transform

Abstract

This paper presents a technique for chattering detection on ground cylindrical parts. Chattering is a defect caused by vibration during the grinding process and is best described as waviness in the surface of a workpiece. In this work, such waviness is measured with a mechanical stylus profiler featuring a diamond tip, and then converted into height values with the help of a precision displacement transducer. The output signal is later converted to digital form and transformed to the time–frequency domain by means of the wavelet transform, allowing its coefficients to grow as a function of surface defects and highlighting chatter marks. The method was validated experimentally with actual production parts and it was found to adequately measure the amplitude and to give location information of the chatter marks on the analyzed surface. Results show that this method can be used objectively to identify and quantify surface quality and that it is feasible to effectively integrate it into production processes for manufacturing control.