Neural networks models for wear patterns recognition of single-point dresser

Abstract

Grinding is a workpiece finishing process for advanced products and surfaces. However, the constant friction between workpiece and grinding wheel causes the latter to lose its sharpness, thereby impairing the result of the grinding process. When this occurs, the dressing process is essential to sharpen the worn grains of the grinding wheel. The dressing conditions strongly influence the performance of the grinding operation; hence, monitoring them throughout the process can increase its efficiency. The purpose of this study was to classify the wear condition of a single-point dresser using intelligent systems whose inputs were obtained by digitally processing acoustic emission signals. Two multilayer perceptron (MLP) neural networks were compared for their classification ability, one using the root mean square (RMS) statistics and another the ratio of power (ROP) statistics as input. In this study, it was found that the harmonic content of the acoustic emission signal is influenced by the condition of the dresser, and that the condition of the tool under study can be classified by using the aforementioned statistics to feed a neural network. © IFAC.