Machining Parameter Optimization of Hybrid Metal Matrix Composites using Artificial Intelligence

Abstract

The objective of this work is to study the various artificial intelligence methods used for optimizing machining parameters while machining of hybrid metal matrix composites, by conventional and unconventional methods. A hybrid composite is formed when two or more reinforcements are added to the matrix. Hybrid Composites are manufactured by stir casting method. Machining of composites is done to create holes, slots and other features that are not possible to obtain during manufacturing of the part. Various types of machining operations are done in hybrid composites with lathe, drilling, milling and EDM machine to get the desired surface roughness, tolerance. Cutting speed, feed rate and depth of cut are the machining parameters optimized while machining in lathe with desired target like surface roughness, MRR, cutting force and tool wear. Drilling of fiber-reinforced plastics (FRP’s) composites facilitates assembly of several components by means of mechanical fastening. Spindle speed, feed rate, drill type are optimized for performance characteristics thrust force, surface roughness, and tool wear in drilling operation. Spindle speed, feed rate, depth of cut is optimized for desired cutting force and surface roughness in milling operation. The influence of process parameters such as pulse on time, pulse off time, spark gap voltage, peak current, wire tension and wire feed rate on response variables such as cutting speed, surface roughness and spark gap are studied in EDM of hybrid composites. ANOVA, RSM, GRA, Taguchi method are used for optimization of various machining parameters. The AI techniques used for prediction include artificial neural network (ANNs), fuzzy logic (FL), adaptive neuro-fuzzy systems (ANFIS), decision tree, genetic algorithm (GA) and genetic programming. Fuzzy rules relate the relationship between input and output variables. Expert knowledge can be built into the system through the rule base. Fuzzy models are used for predicting the thrust force and torque for drilling hybrid composite. In adaptive neuro-fuzzy systems advantages of FL and ANNs are combined for adjusting the membership functions, rule base and related parameters for training the data set. It can continuously improve the initially obtained rough model based on the daily operating data. AI techniques are used for predicting the automatic selection of inputs and predicting surface roughness, other response variables.