Prediction of tensile strength in fused deposition modeling process using artificial neural network technique
Karthic Manoharan, K. Chockalingam, S. Shankar Ram
Abstract
Fused deposition modeling (FDM) has many different levels of parameters hence prediction method is required for FDM users. The main aim of this paper is to predict the mechanical strength of the components before fabrication of FDM process. In this paper, the most influenced FDM process parameters of Layer thickness, Infill density, print speed, temperature and build orientation were considered for prediction of tensile strength of the fabricated part. The tensile test specimen (ASTM-D638) was fabricated by FDM process with PLA (Polylactic Acid) material, based on the central composite design. The actual tensile strength values were obtained by conducting a test on Universal Testing Machine (UTM) for the PLA material. The results data were used to develop the mathematical models for the prediction of tensile strength of fabricated part from response surface methodology (RSM) and analysis of variance (ANOVA). Similarly, ANN tool was used to predict the tensile strength of the specimen with the aid of 33 experiments and result data. The predicted values from the results of RSM, ANN, ANOVA were compared with the actual tensile strength. The ANN technique gives the minimum value of percentage deviation in Error (1-2%).