Statistical Analysis of Effects on Weld Response Variables during Friction Hydro-Pillar Processing

Abstract

R & D Journal of the South African Institution of Mechanical Engineering 2019, 35, 9-20 http://dx.doi.org/10.17159/2309-8988/2019/v35a2 http://www.saimeche.org.za (open access) © SAIMechE All rights reserved. 9 Abstract: The article provides insight to the complex relationships that exist between the selected weld process parameters and the associated effects on weld responses by application of statistical analysis relating to response predictive models for Friction Hydro-Pillar Processing (FHPP). The literature review focused at obtaining the current assumptions made with regards to the relationship between FHPP and conventional Friction Welding (FW) variations. Experimental welds were produced at varying selected process parameters; motor speed, axial force, consumed length and forging time. These parameters were compared to five weld response variables through a 27-run full factorial Design of Experiment and multiple regressions. The analysis focussed on quantifying the main effects of process parameters on energy input, temperature profile, friction time, torque and consumed rate. Comparison with experimental results served to validate the effect of dominant process parameters. The process and statistical analysis are explained in detail to assist further understanding of the applied methodology and the effect of the various process parameters on weld responses presented. Results indicate that the mathematical equation based models predict the responses adequately within the limits of welding parameters used and that no single parameter solely control the weld responses during FHPP. This study provides a clearer understanding of FHPP showing that generalised conclusions, with regards to the influences of process parameters on weld responses during conventional FW, cannot be made as the effects of these inputs differ depending on the combination of levels included in a parameter set.