Modeling and Prediction Using Regression, ANN and Fuzzy Logic of Real Time Vibration Monitoring on Lathe Machine in Context of Machining Parameters

Abstract

Machine tool vibration plays a dominant rolein the surface finish, dimensional and geometrical tolerancesof the machined work piece. Condition of the machinesincludes collected data, such as vibration analysis, oil andwears debris analysis, ultrasound, temperature andperformance evaluation. Out of these the vibrations have beenmeasured and its effect has been studied. The present paperdeals with the measurement of acceleration during machiningof Cast Iron on lathe machine. 33full factorial design ofexperiments were selected, experiments are performed byvarying machining parameters such as spindle speed, feedrate and depth of cut. ANOVA and Regression analysis hasbeen carried out to know the significance of these parameters.Even Artificial Neural Network (ANN) and Fuzzy Logic basedmodels have been developed to predict Acceleration in thecontext of these input parameters. The predicted resultsobtained from the developed models are compared with theexperimental one. Results shows that the developed modelshaving more than 95% accuracy, which leads the use of it inpredicting the acceleration within the rang