Fuzzy logic cluster analysis of surface roughness waveforms generated using a laser-based system

The automated measurement and analysis of surface roughness (finish) is an important part of computer integrated manufacturing (CIM) systems. In this study a laser based system was used to scan the surfaces of 6 steel sheets. The surface roughness of the sheets, quantified by the average roughness (Ra) parameter, were 0.05, 0.1, 0.2, 0.4, 0.8 and 1.6. Each sheet was scanned 20 times using a sampling rate of 4000 data points per millimetre. The resulting waveforms were pre-processed and then they were represented by a set of feature vectors. The relationships between the actual Ra values and the features obtained from the recorded waveforms were investigated. The performance of the fuzzy c-mean (clustering) algorithm for differentiating the recorded surface roughness waveforms was analysed. The study demonstrated that it was possible to accurately differentiate the surfaces by applying the devised methods to the recorded waveforms.