Online monitoring of surface roughness and grinding wheel wear when grinding Ti-6Al-4V titanium alloy using ANFIS-GPR hybrid algorithm and Taguchi analysis

Abstract

Grinding wheel wear adversely affects the quality of machining surface, the working ability of grinding wheel as well as the grinding machine during the machining process. Challenges in the machining process, especially in the process of grinding Ti-6Al-4V alloy, which is a material with high adhesion, durability, and toughness in combination with poor thermal conductivity that leads to low economic - technical indicators in the grinding process, susceptibility to disabilities of the grinding details, and fast wear of the grinding wheel. Therefore, the precise prediction of the grinding wheel wear and surface roughness in the machining process is a prerequisite to minimize the damages caused by the grinding wheel wear when grinding Ti-6Al-4V alloy. This work presents a model for monitoring grinding wheel wear conditions using the grinding force signal obtained at the processing time in combination with the adaptive neural fuzzy inference system - Gaussian process regression and Taguchi analysis to predict the abrasive wear in the different stages of grinding process. Experimental results show the ability to accurately predict the amount of grinding wheel wear and surface roughness from the proposed model when grinding Ti-6Al-4V alloy. With the ability to accurately predict the indicators of surface roughness and abrasive wear with a corresponding average error of 0.31% and the reliability percentage of the measurement of 98%, the proposed model can be used in the industry for the online forecast of surface roughness and the time to repair the grinding wheel directly in the grinding process.