A Study On Parametric Appraisal of Fused Deposition Modelling (FDM) Process

Abstract

The manufacturing industries are contemplating to develop new technologies for production of complex end use parts possessing high strength and low product development cycle in order to meet the global competition. Rapid prototyping (RP) is one of the proficient processes having the ability to build complex geometry parts in reasonably less time and material waste. Fused deposition modelling (FDM) is one of the RP processes that can manufacture 3D complex geometry accurately with good mechanical strength and durability. Normally, the FDM process is a parametric dependant process due to its layer-by-layer build mechanism. As FDM build parts are used as end use parts, it is prudent to study the effect of process parameters on the mechanical strength under both static and dynamic loading conditions and wear (sliding) behaviour. In order to investigate the behaviour of build parts in a systematic manner with less number of experimental runs, design of experiment (DOE) approach has been used to save cost and time of experimentation. As the selection of input process parameters influence on build mechanism, the mechanical properties and wear behaviour of FDM build parts change with process parameters. Notably, the raster fill pattern during part building causes FDM build parts to exhibit anisotropic behaviour when subject to loading (static or dynamic). In this research work, an attempt has been made to minimise the anisotropic behaviour through controlling the raster fill pattern during part building by adequate selection of process parameters. Statistical significance of the process parameters is analysed using analysis of variance (ANOVA). Influence of process parameters on performance characteristics like mechanical strength, fatigue life and wear of build part is analysed with the help of surface plots. Internal structure of rasters, failure of rasters, formation of pits and crack are evaluated using scanning electron machine (SEM) micro-graphs. Empirical models have been proposed to relate the performance characteristics with process parameters. Optimal parameter setting has been suggested using a nature inspired metaheuristic firefly algorithm to improve the mechanical strength. Finally, genetic programming (GP) and least square support vector machine (LS-SVM) are adopted to develop predictive models for various performance characteristics