NEURAL NETWORK MODELING OF GRINDING PARAMETERS OF DUCTILE CAST IRON USING MINIMUM QUANTITY LUBRICATION

Abstract

This paper presents the optimization of the grinding parameters of ductile cast iron in wet conditions and with the minimum quantity lubrication (MQL) technique. The objective of this project is to investigate the performance of ductile cast iron during the grinding process using the MQL technique and to develop artificial neural network modeling. In this project we used the DOE method to perform the experiments. Analysis of variance with the artificial neural network method is used to investigate significant effects on the performance characteristics and the optimal cutting parameters of the grinding process. Ductile cast iron was used in this experiment and the ethanol glycol was applied in the conventional method and compared with the MQL method. During conventional grinding, a dense and hard slurry layer was formed on the wheel surface and the performance of the ductile cast iron was very low, threatening the ecology and health of the workers. In order to combat the negative effects of conventional cutting fluids, the MQL method was used in the process to formulate modern cutting fluids endowed with user- and eco-friendly properties. Aluminum oxide was used as the grinding wheel (PSA-60JBV). This model has been validated by the experimental results of ductile cast iron grinding. Each method uses two passes - single-pass and multiple-pass. The prediction model shows that depth of cut and table speed have the greatest effect on the surface roughness and material removal rate for the MQL technique with multiple-passes by showing improved surface roughness, preventing workpiece burning and enabling a more friendly environment. Thus, various other parameters need to be added for further experiments, such as the wheel speed, distance from the wheel to the workpiece zone contact, and the geometry of the nozzle.