Study of surface properties and residual stresses of aluminium matrix composites after electric discharge machining

Abstract

With the emergence of newer technologies, many advanced engineering applications require materials with enhanced properties and controlled coefficient of thermal expansion. One such class of materials are metal matrix composites (MMCs) that have reinforcements (such as fibers or particles) supported by binder (matrix) material. Particulate reinforced MMCs combine a conductive matrix which has been embedded with hard ceramic particles with an average size scale ranging from the molecular level to few microns. Such materials have vastly improved properties and are particularly difficult to machine with conventional machining methods. The use of traditional machinery to machine MMCs results in large tool wear due to the presence of abrasive nature of reinforcement. Electric Discharge Machining (EDM) provides an effective alternative to machine such materials especially when complex geometries are required. This research work has been conducted in three stages of experimentation. In Stage I, the experimental study was undertaken to identify the significant factors that affect the output responses while machining of 10vol%Al2O3/Al composite material. The material removal rate (MRR) and tool wear rate (TWR) have a direct relationship with current and an inverse relationship with pulse-on time. The results were optimized using Lexicographic Goal Programming (LGP) to predict the ideal parametric combinations for machining of MMCs. Optimal conditions for the significant parameters were listed depending upon the requirements of the machining process which may vary for rough machining (higher material removal rate) and finish machining (lower surface roughness). The mean thickness of the recast layer formed after machining was also studied. The results at this stage of experimentation show that all the responses (MRR, TWR, and SR) have a direct relationship with current but an inverse relationship with pulse-on time. This is because the increased pulse-on time decreases the frequency of spark occurrence. A non-linear mathematical model was developed to predict the optimum machining response parameters. The LGP proved to be a powerful tool in determining the setting of response parameters under given constraints. The recast layer thickness is significantly affected by current and pulse-on time. Also, the pulse-on time is highly significant in determining the thickness of recast layer. The formation of clusters or uneven distribution of reinforced particles deteriorates the properties of the machined surface; hence due to least presence of reinforced particles in recast layer, it is desirable to remove it. ii For the next stages of study, the kerosene dielectric medium was replaced with EDM oil for enhanced transfer of spark energy and least deposition of carbon particles on both the electrodes. The diameter of the selected tool electrode was also increased for a detailed study of responses. During Stage II, two variants of MMCs (65vol%SiC/A356.2; 10vol% SiC-5vol% quartz/Al) were machined and the material removal rate (MRR), surface roughness (SR) and residual stresses were measured. A method to obtain more reliable global weights of three different responses has been described for electric discharge machining (EDM) of different types of particulate reinforced metal matrix composites using the Analytic Hierarchy Process (AHP). Five different process parameters were varied to evaluate their effects on MRR, SR and residual stresses using a standard Taguchi’s orthogonal array L18. The machining process parameters such as electrode material (Copper, Graphite and Copper-Graphite composite), peak current, pulse-on/off time and dielectric medium were varied during the study. The residual stresses induced due to subsequent heating, and cooling shocks during the electric discharge process are of primary concern during machining. Artificial Neural Network (ANN) modeling technique was implemented to predict the residual stresses. The capability of ANN to predict residual stresses during EDM has been achieved with feed forward back propagation neural architecture with two hidden layers. The model accurately predicts the residual stresses and can be used as a reliable tool for study of residual stresses in case of complex problems that involve qualitative and quantitative factors. MMCs with low coefficient of thermal expansion and high reinforced particles exhibit lower residual stresses. Also, better conductive electrode materials used during machining cause lower residual stresses. Pulse-off time was identified as the most significant factor resulting in residual stresses in both the MMCs. The increase in pulse-off time causes a steep rise in residual stresses due to extended solidification period while pulse-on time has no effect. The addition of powder in the dielectric lowers the residual stresses. However, the conductivity of powder particles has no effect on residual stresses. The pulse-off time which had no significant effect on MRR or SR, had the largest effect on residual stresses followed by powder mixed dielectric, current and the type of electrode used. The XRD patterns clearly indicate the formation of new phases on the machined surface. The peak intensity after machining is reduced due to dislocation of atomic layer resulting in residual strains. iii The process conditions that affected the three responses were identified and optimized together using AHP and the most suitable process parameter settings for machining of MMCs. It reveals that machining of work piece with graphite electrode and higher setting of pulse-on time with lowest pulse-off time in the presence of suspended particles in the dielectric gives minimum residual stresses with desired MRR. Due to the presence of dense ceramic reinforced particles in Sample I as compared to Sample II, the desired results were obtained at intermediate level of current and choice of higher conductive powders in a dielectric medium. Also, copper powder in the dielectric medium resulted in an optimal solution for 10vol%SiC-5vol% quartz/Al MMC and graphite powder gave better results for 65vol%SiC/A356.2 metal matrix composite. Higher density of reinforced particles in the matrix results in lesser MRR, SR and residual stress as ceramic particle act as a shield of matrix material against sparks energy. MRR was observed to decrease when the pulse-on time was stepped up from 10 to 30 μs, further it increases drastically as prolonged pulse-on time causes intense melting and evaporation of matrix material and easy removal of reinforced particles by spalling mechanism. Stage III reports the optimal process conditions for machining of three different types of MMCs; namely 65vol%SiC/A356.2, 10vol% SiC-5vol% quartz/Al and 30vol%SiC/A359 using EDM process. MRR, TWR, SR, residual stresses, micro-hardness and recast layer were evaluated after each trial and contributing process parameters were identified using Taguchi L27 orthogonal array. Each work piece was examined by X-ray diffraction (XRD) followed by Scanning Electron Microscope (SEM) for surface integrity and material deposition. In this stage of study, the effect of material properties and machining parameters on the residual stresses of a machined surface during EDM was investigated by measuring the shift in selected peak at highest angle by Diffractometer method. It was observed that the residual stresses are tensile as well as compressive in nature due to conflicts in thermal properties of matrix and reinforced particles. The surface residual stresses were observed to be mainly dependent upon concentration/ particle size and the conductivity of the work piece. The residual stresses increased with an increase in pulse-off time due to higher resolidification time of the recast layer (as observed in Stage II experimental plan). The depositions on the work piece due to the presence of copper powder in dielectric resulted in higher residual stresses as compared to conditions where graphite powder is mixed. iv Due to weak bonding in composite electrode (Cu-Gr composite), significant quantities of disintegrated particles were deposited to form thick recast layer on the work piece. The study also reports the phenomenon of surface modification while machining with EDM process. The density of reinforced ceramic forming oxides at elevated temperature (above 17000C) was the most significant factor affecting micro-hardness on the machined surface. The XRD spectra revealed formation of copper oxide as the major transferred element from electrode/ dielectric medium. The deposition of carbon was observed when the machining parameters setting were at the highest level. Finally, the cross-sectional view of recast layer analysis reflected the profile of sparks generated during EDM. The SEM analysis revealed that the thickness of recast layer was maximum for dense reinforced particles MMCs due to the higher heat absorption tendency of doped particles. Hence, high density reinforcement also showed high extent cross-sectional residual stress distribution. It also depended upon the size of reinforced particles. The powder mixed EDM with copper as an additive showed the least recast layer thickness due to improved electric conductivity and consistent discharges.