Comparative study on the performance of the MQL nanolubricant and conventional flood lubrication techniques during grinding of Si3N4 ceramic

• Main Authors: Dambatta, Yusuf S., Sayuti, Mohd, Sarhan, Ahmed Aly Diaa Mohammed, Abd Shukor, Mohd Hamdi
• Format: Article
• Published: Springer Verlag 2018
• Subjects: TJ Mechanical engineering and machinery
• Online Access: http://eprints.um.edu.my/22173/; https://doi.org/10.1007/s00170-018-1689-3

Silicon nitride (Si 3 N 4 ) ceramic is highly desired in various engineering applications due to its exceptional properties. However, machining the Si 3 N 4 ceramic suffers setbacks due to various degrees of damage inflicted on the ceramic during machining. Studies have shown that the minimum quantity lubrication (MQL) process is a better alternative to the flood cooling lubrication system during grinding of advanced engineering ceramics. The MQL technique is a highly efficient and eco-friendly lubrication method that can be used to reduce the different types of surface and subsurface damages, while significantly lowering the consumed lubricant. In this work, the performance of the MQL and flood cooling lubrication techniques during grinding of Si 3 N 4 ceramic was investigated. The MQL nanolubricant was formed by suspending silicon dioxide (SiO 2 ) nanoparticles in environmentally friendly vegetable oil (canola oil). Also, the effect of the input parameters, i.e., feed rate, depth of cut, type of diamond wheel, and lubrication type, were investigated on the output parameters, i.e., grinding forces, workpiece surface roughness, surface damages, and wheel wear. The Taguchi mixed-level parameter experimental design (L16) was used in the design of the experiment, and the signal-to-noise ratio was used to optimize the grinding process. Furthermore, the adaptive neuro-fuzzy inference system (ANFIS) prediction method was used to predict and analyze the variation of the input parameters with the grinding forces and surface roughness. Validation experiments also indicate that the ANFIS models for the normal grinding force, tangential grinding force, and surface roughness have accuracies of 98.45, 98.58, and 96.31%, respectively.