Data-Driven Optimization of Dynamic Properties in Additively Manufactured Parts Using Linear Regression Algorithms

Structural components produced through Selective Laser Melting (SLM) are often subjected to vibrational tests to evaluate and optimize their dynamic behavior under random vibrations and loads. Modal Analysis serves as a powerful tool for determining mode shapes and resonant frequencies, which are in...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Proceedings of ... International Conference on Mechanical and Intelligent Manufacturing Technologies (Online) S. 226 - 232
Hauptverfasser: Toprak, Can Baris, Dogruer, Can Ulas
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 16.05.2025
Schlagworte:
Additive Manufacturing
Aerodynamics
Heuristic algorithms
Laser modes
Linear regression
Modal analysis
Optimization
Prediction algorithms
Resonant frequency
Selective Laser Melting
Shape
ISSN:2694-3182
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Structural components produced through Selective Laser Melting (SLM) are often subjected to vibrational tests to evaluate and optimize their dynamic behavior under random vibrations and loads. Modal Analysis serves as a powerful tool for determining mode shapes and resonant frequencies, which are influenced by variations in material properties such as elasticity modulus and mass density. However, the computational expense of Modal Analysis, exacerbated by the large-scale meshing requirements, presents a significant bottleneck, particularly during complex optimization cycles. This study introduces a novel approach to streamline the modal analysis process by leveraging a linear regression algorithm to approximate the dynamic behavior of SLM parts. The algorithm was trained using an experimental database, correlating material properties (elasticity and density) with resonant frequency predictions. The resulting predictive model was integrated into a Sequential Quadratic Programming (SQP) optimization framework to efficiently identify desired dynamic properties in the produced parts while minimizing computational time and cost. The proposed method demonstrated a substantial reduction in processing time during the optimization cycle, approximately 300 times faster than default algorithm, enabling rapid determination of optimal dynamic characteristics. Additionally, this approach highlights the potential for linking SLM processing parameters (e.g., laser power, scanning speed) with critical dynamic features such as resonant frequencies and mode shapes. By significantly accelerating the design and optimization processes, this study offers a valuable tool for enhancing the dynamic performance of SLM-produced parts with minimal computational overhead.
ISSN:2694-3182
DOI:10.1109/ICMIMT65123.2025.11092164