A Framework for Parametric Modeling of Additive Manufacturing Mechanical Properties

Thesis Abstract

Additive manufacturing (AM) has revolutionized the production landscape by enabling rapid prototyping and complex geometrical fabrication; however, the variability in mechanical properties resulting from diverse process parameters poses significant challenges to its broader industrial adoption. This study investigates the development of a comprehensive parametric modeling framework aimed at predicting and optimizing mechanical properties such as tensile strength, Young’s modulus, and fracture toughness, based on key AM process parameters including layer height, print speed, nozzle temperature, and material composition. The research aims to establish a systematic approach for correlating process variables with mechanical outcomes, thereby facilitating more reliable and tailored AM processes for engineering applications. The specific objectives are to (1) identify the relevant process parameters influencing mechanical properties, (2) develop a quantitative model connecting these parameters to the resulting mechanical characteristics, (3) validate the model through experimental data, and (4) propose a generalized framework adaptable to different additive manufacturing techniques and materials. The research questions focus on determining the extent of impact each process parameter has on mechanical strength, the accuracy of the proposed model in predicting properties, and its practical applicability in process optimization. Hypotheses are formulated to test the significance of process parameters within the model and the predictive capability of the framework. The methodology adopts a quantitative, experimental research design following a positivist paradigm, prioritizing empirical measurement and statistical analysis. The study population comprises 150 specimen sets fabricated using fused filament fabrication (FFF) with varying process parameters within a controlled laboratory setting. A stratified random sampling technique ensures representative coverage of the parameter space. Data collection involves mechanical testing through standardized tensile and impact tests, while process parameters are documented via digital control logs. The study employs the Design of Experiments (DOE) methodology to systematically vary process factors, ensuring efficient exploration of the parameter space. Instrument validity and reliability are established through calibration of testing equipment and repeated measurements, respectively. Data analysis centers on multiple regression analysis complemented by ANOVA to quantify the influence of individual parameters and their interactions on mechanical properties. The theoretical underpinning is grounded in the Theory of Material Property-Surface Relationship and the Process-Structure-Property framework, facilitating integration of process variables with material microstructure and resulting properties. Model specification involves the development of predictive equations, with the validation phase employing root mean square error (RMSE) and R-squared metrics to assess predictive accuracy. Expected findings include statistically significant relationships between process parameters and mechanical properties, with certain parameters—such as layer height and print speed—demonstrating a dominant influence. The model is anticipated to accurately predict mechanical characteristics within a confidence interval of 95%, and the developed framework aims to serve as a decision-support tool for process optimization. This study contributes to existing knowledge by formalizing a parametric modeling approach tailored to additive manufacturing, bridging the gap between process control and material performance prediction. It offers a validated, adaptable framework capable of informing practitioners and researchers on process-structure-property relationships, thereby advancing the reliability and customization of AM-produced components. The main conclusion underscores the importance of precise process parameter control to achieve desired mechanical properties, emphasizing the framework's potential to standardize process optimization practices across AM industries. Recommendations include incorporating real-time process monitoring for enhanced predictive accuracy and extending the framework to other AM techniques such as selective laser sintering (SLS) and stereolithography (SLA). Future research should focus on integrating microstructural analysis and machine learning algorithms to refine the model further, broadening its applicability and robustness for industrial adoption.