Conference Proceedings of the Society for Experimental Mechanics Series Series Editor Kristin B. Zimmerman Society for Experimental Mechanics, Inc., Bethel, USA i
The Conference Proceedings of the Society for Experimental Mechanics Series presents early findings and case studies from a wide range of fundamental and applied work across the broad range of fields that comprise Experimental Mechanics. Series volumes follow the principle tracks or focus topics featured in each of the Society’s two annual conferences: IMAC, A Conference and Exposition on Structural Dynamics, and the Society’s Annual Conference & Exposition and will address critical areas of interest to researchers and design engineers working in all areas of Structural Dynamics, Solid Mechanics and Materials Research. ii
Emily Retzlaff · Piyush Thakre · Marco Rossi · Sharlotte Kramer Editors Mechanics of Additive & Advanced Manufacturing, Inverse Methods and Machine Learning, Vol. 5 Proceedings of the 2025 Annual Conference on Experimental and Applied Mechanics River Publishers
Published, sold and distributed by: River Publishers Broagervej 10 9260 Gistrup Denmark www.riverpublishers.com ISBN 97887-438-0831-2 (Hardback) ISBN 97887-438-0836-7 (eBook) https://doi.org/10.13052/97887-438-0831-2 Conference Proceedings of the Society for Experimental Mechanics An imprint of River Publishers © The Society for Experimental Mechanics, Inc. 2025 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, or reproduction in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Preface Mechanics of Additive and Advanced Manufacturing, Inverse Methods and Machine Learningrepresents one of five volumes of technical papers presented at the 2025 SEM Annual Conference & Exposition on Experimental and Applied Mechanics organized by the Society for Experimental Mechanics and held in Milwaukee, WI, June 2-5, 2025. The complete Proceedings also includes volumes on: Dynamic Behavior of Materials; Advancement of Optical Methods & Digital Image Correlation in Experimental Mechanics; and Mechanics of Biological Systems and Materials and the Mechanics of Composite, Hybrid & Multifunctional Materials; and Fracture, Fatigue, Failure, Damage Evolution and Thermomechanics & Infrared Imaging. Mechanics of Additive and Advanced Manufacturing is an emerging area due to the unprecedented design and manufacturing possibilities offered by new and evolving advanced manufacturing processes and the rich mechanics issues that emerge. Technical interest within the Society spans several other SEM Technical Divisions such as: Composites, Hybrids and Multifunctional Materials, Dynamic Behavior of Materials, Fracture and Fatigue, Residual Stress, Time-dependent Materials, and the Research Committee. The topic of mechanics of additive and advanced manufacturing included in this volume covers design, optimization, experiments, computations, and materials for advanced manufacturing processes (3D printing, micro- and nano-manufacturing, powder bed fusion, directed energy deposition, etc.) with particular focus on mechanics aspects (e.g. mechanical properties, residual stress, deformation, failure, rate-dependent mechanical behavior, etc.). The topic of Inverse Methods and Machine Learning included in this volume covers the Virtual Fields Method, inverse methods for plasticity, identification for anisotropic and heterogeneous materials, optimal experimental design for inverse methods, and machine learning for mechanics with an emphasis on inverse methods. Editors: Emily Retzlaff – United States Naval Academy, Annapolis, MD, USA; Piyush Thakre, Dow Inc., TX, USA; Marco Rossi, Universita Politecnica Delle Marche, Italy; Sharlotte Kramer, Sandia National Laboratories, NM, USA. v
Contents 1 Effect of Coatings and Additives on the Water Absorption and Mechanical Characteristics of Additive Manufactured Nylon Polymers 1 James LeBlanc, Irine Chenwi, Julianna Martinez, Olivia Dube, Dillon Fontaine, Eric Warner, Tyler Chu, Arun Shukla, and Lewis Shattuck 2 Electrically Aligned Epoxy-Based Nanocomposites: Processing and Characterization 5 Rahman Julkarnyne M. Habibur and Gururaja Suhasini 3 Interfacial Characterization of Metal Wire Inlays for 3D Printed FDM Parts 13 Vereesh Ayyagari, Hugh A. Bruck, Amir H. Ohadi, and Michael M. Ohadi 4 Accelerated Mechanical Behavior Characterization of Structural Materials 23 J. Rathore, A. Imeri, E. Al Amiri, V. Shah, C. O’Brien, B. Wisner, and A. Kontsos 5 Visco-plastic and Damage Characterization of Sheet Metals for Railway Crash Applications 35 Valentina Arra`, Mattia Utzeri, Gianluca Chiappini, Marco Sasso, and Alberto Perticone 6 Design and Optimization of Shock Absorbers Made of Grade Density Foams 43 C. Sabbatini, G. Zandri, and M. Sasso 7 Design, Fabrication and Characterization of Layered Jamming Bistable Composite Structures for Assistive Robotics 51 Hugh A. Bruck, Supriya Shastry, and Oliver J. Myers 8 Assessing the Effects of Strain Rate, Print Orientation and Post Print Annealing in FDM PLA Specimens in an Undergraduate Materials Science Course 59 Matt Schaefer vii
Chapter 1 Chapter 1 On the Detection and Quantification of Nonlinearity via Statistics of the Gradients of a Black-Box Model Georgios Tsialiamanis and Charles R. Farrar Abstrac t Detection and identification of nonlinearity is a task of high importance for structural dynamics. On the one hand, identifying nonlinearity in a structure would allow one to build more accurate models of the structure. On the other hand, detecting nonlinearity in a structure, which has been designed to operate in its linear region, might indicate the existence of damage within the structure. Common damage cases which cause nonlinear behaviour are breathing cracks and points where some material may have reached its plastic region. Therefore, it is important, even for safety reasons, to detect when a structure exhibits nonlinear behaviour. In the current work, a method to detect nonlinearity is proposed, based on the distribution of the gradients of a data-driven model, which is fitted on data acquired from the structure of interest. The data-driven model selected for the current application is a neural network. The selection of such a type of model was done in order to not allow the user to decide how linear or nonlinear the model shall be, but to let the training algorithm of the neural network shape the level of nonlinearity according to the training data. The neural network is trained to predict the accelerations of the structure for a time-instant using as input accelerations of previous time-instants, i.e. one-step-ahead predictions. Afterwards, the gradients of the output of the neural network with respect to its inputs are calculated. Given that the structure is linear, the distribution of the aforementioned gradients should be unimodal and quite peaked, while in the case of a structure with nonlinearities, the distribution of the gradients shall be more spread and, potentially, multimodal. To test the above assumption, data from an experimental structure are considered. The structure is tested under different scenarios, some of which are linear and some of which are nonlinear. More specifically, the nonlinearity is introduced as a column-bumper nonlinearity, aimed at simulating the effects of a breathing crack and at different levels, i.e. different values of the initial gap between the bumper and the column. Following the proposed method, the statistics of the distributions of the gradients for the different scenarios can indeed be used to identify cases where nonlinearity is present. Moreover, via the proposed method one is able to quantify the nonlinearity by observing higher values of standard deviation of the distribution of the gradients for lower values of the initial column-bumper gap, i.e. for “more nonlinear” scenarios. Keyword s Structural health monitoring (SHM) · Structural dynamics · Nonlinear dynamics · Machine learning · Neural networks 1.1 Introduction In the pursuit of making everyday life safer, humans have extensively tried to model the environment around them. Structures are an important part of the environment, in which humans live. They are man-made and should be safe throughout their lifetime. Structures are exposed to numerous environmental factors, which may cause them to fail. Moreover, during operation, structures are subjected to dynamic loads, which, in time, may cause failure. Such failures will most probably result in economic damage to society and may even result in loss of human lives. Therefore, for the purpose of maintaining structures safe, the field of structural health monitoring (SHM) [1] has emerged. G. Tsialiamanis ( ) Dynamics Research Group, Department of Mechanical Engineering, University of Sheffield, Sheffield, UK e-mail: g.tsialiamanis@sheffield.ac.uk C. R. Farrar Engineering Institute, MS T-001, Los Alamos National Laboratory, Los Alamos, NM, USA e-mail: farrar@lanl.gov © The Society for Experimental Mechanics, Inc. 2024 M. R. W. Brake et al. (eds.), Nonlinear Structures & Systems, Volume 1, Conference Proceedings of the Society for Experimental Mechanics Series, https://doi.org/10.1007/978-3-031-36999-5_1 1 Effect of Coatings and Additives on the Water Absorption and Mechanical Characteristics of Additive Manufactured Nylon Polymers James LeBlanc, Irine Chenwi, Julianna Martinez, Olivia Dube, Dillon Fontaine, Eric Warner, Tyler Chu, Arun Shukla, and Lewis Shattuck Abstract The use of baseline material data taken in controlled laboratory conditions on dry, unaged specimens is largely insufficient for the design of critical structures to be deployed in deep-water, wetted applications. Recent studies have shown that Nylon based materials absorb a substantial amount of water when exposed to marine environments. This effort conducted an experimental investigation of the effectiveness of surface coatings and material additives on the reduction of water absorption into nylon based AM polymers and the corresponding effects on the mechanical properties of the materials. The study is divided into two unique parts: (1) effects of post print surface coatings on 3D printed specimens, and (2) effects of nanoclay additives on nylon base material. The study’s findings highlight that the use of surface coatings and nanoclay additives can reduce the amount of salt water absorbed by the respective materials. However, the coatings and nanoclays also have a degrading effect on the mechanical properties of the materials themselves, even in the absence of salt water exposure. Therefore, significant consideration must be given to material selection when choosing surface coatings and material additives. Similarly, the use of nanoclay additives are shown to change the material behavior from highly ductile to highly brittle. These changes can potentially cause premature failures when utilized with AM polymers without detailed material characterization data obtained before and after salt water exposure. Keywords Additive Manufacturing · Mechanical characterization · Water absorption · Material Additives · Surface Coatings Introduction Additive Manufacturing (AM) is one of the largest growing areas of novel manufacturing approaches, with applications ranging from home hobbyists up to full scale commercial production of parts. The technology lends itself to ready access to parts on demand, ease of manufacture for complex designs not otherwise possible, shorter design to manufacturing cycle times (rapid prototyping), lightweight designs, and cost savings. These manufacturing techniques provide the ability to achieve part geometries not possible with “subtractive” manufacturing methods (e.g. CNC milling), and eliminate the need for complex and expensive tooling/molds when small part quantities are needed, for example ad-hoc spare parts in an ondemand setting. Despite these advantages, there are also associated disadvantages with these technologies such as potential part variability based on print parameters, print post-processing, and increased part porosity. The effects of moisture and water absorption on the performance of additively printed materials has become a recent topic of research interests. Research findings have shown moisture and water absorption can result in decreases in modulus, James LeBlanc · Irine Chenwi · Julianna Martinez · Olivia Dube · Dillon Fontaine · Eric Warner · Lewis Shattuck Naval Undersea Warfare Center (Division Newport), Newport, RI 02841 e-mail: james.m.leblanc.civ@us.navy.mil; irine.n.chenwi.civ@us.navy.mil; julianna.a.martinez3.civ@us.navy.mil; olivia.m.dube.civ@us .navy.mil; dillon.t.fontaine.civ@us.navy.mil; eric.a.warner8.civ@us.navy.mil; lewis.b.shattuck.civ@us.navy.mil TylerChu· Arun Shukla Dynamic Photo Mechanics Laboratory, Department of Mechanical, Industrial and Systems Engineering, University of Rhode Island, Kingston, RI 02881 e-mail: tylerchu@uri.edu; shuklaa@uri.edu © The Author(s), under exclusive license to River Publishers 2025 Emily Retzlaff et al. (eds.), Mechanics of Additive & Advanced Manufacturing, Inverse Methods and Machine Learning, Vol. 5, Conference Proceedings of the Society for Experimental Mechanics Series, https://doi.org/10.13052/97887-438-0831-2 1
2 J. LeBlanc et al. ultimate tensile strength, decreased energy release rate (GIC), and stress intensity factor (KIC). These studies have investigated common AM polymers such as ABS, PLA, and PETG. In addition to the specific materials themselves, the effect of layer thickness and associated print parameters can affect water absorption levels and material swelling. Furthermore, the effectiveness of post print surface treatments and coatings to reduce water absorption for FDM printed materials have been the subject of recent publications. Recently, the effects of salt water exposure have been investigated and highlighted the base material composition is a key contributor to the level of material degradation due to saline fluid exposure. Materials There are several distinct materials utilized in the study, which are categorized into baseline filaments, filament additives, and post print coatings. The Markforged Onyx material is utilized to manufacture mechanical specimens that were subsequently coated with one of two surface treatments. The Nylon filament was used as a base material to which the nanoclay additive was added to through a secondary filament extrusion process. Markforged Onyx is a nylon-based material, which contains micro carbon fibers, with a nominal material density of 1.2 g/cm3. The first post-print surface treatment applied to the Markforged Onyx specimens was an epoxy hardpot coating, Henkel Loctite Stycast 2651 Epoxy Encapsulant which was mixed with Henkel Loctite Catalyst 9. The second post-print surface treatment applied to the Markforged Onyx specimens was an epoxy based paint system, Sherwin Williams E2A933 Epoxy Primer and Sherwin Williams Genesis GC 3.5 Automotive Top Coat. Markforged Nylon is a pure nylon-based material with a nominal material density of 1.1 g/cm3. The nanoclay powder that was used as an additive for the Markforged Nylon was Sigma-Aldrich Montmorillonite Clay (MMT) with surface modification, particle size of <= 20 microns, and a density of 200-500 kg/m3. The nylon material was pelletized and subsequently combined with the nanoclay additives at weight percentages of 2% and 5%. Two extrusion cycles were carried out for each nylon composite to ensure that the nanoclay was sufficiently dispersed in the nylon. Water Absorption and Mechanical Characterization The water immersion of the epoxy hardpot and painted Markforged Onyx specimens was carried out in a novel test facility specifically designed for high-pressure water immersion experiments. The material specimens were placed in a 3.5% NaCl solution, pressurized to 5000 psi, and were left untouched for 60 days. The water immersion study for the nylon specimens with nanoclay additives was performed at ambient pressure and at room temperature. Both the neat nylon and nylon composites were immersed in a 3.5% NaCl solution for 30 days. The coated Markforged Onyx specimens were characterized for the following properties before and after the 60-day immersion: Tension, Compression and Flexure per ASTM standards. The Nylon with nanoclay additives underwent the following characterization: thermogravimetric analysis (TGA), Differential Scanning Calorimetry (DSC), Dynamic Mechanical Analysis (DMA), and Tension. Results This effort conducted an experimental investigation of the effectiveness of surface coatings and material additives on the reduction of water absorption into nylon based AM polymers and the corresponding effects on the mechanical properties of the materials. The study is divided into two unique parts: (1) effects of post print surface coatings on 3D printed specimens, and (2) effects of nanoclay additives on nylon base material. The study’s findings highlight that the use of surface coatings and nanoclay additives can reduce the amount of salt water absorbed by the respective materials. The surface coatings that were applied to the AM printed specimens were found to be effective in reducing both the rate of water absorption and the overall mass change due to water absorption from 6% to 3%. The incorporation of nanoclay additive into the baseline nylon material was shown to reduce the amount of mass absorbed during salt water immersion from 1.7% to 1.6%. The thermogravimetric analysis was performed to quantify the thermal decomposition of the baseline nylon and nanoclay modified nylon. The baseline nylon experienced the onset of decomposition at a temperature of 416 ◦C and the nanoclay modified material showed the onset of decomposition at 406 ◦C and 397 ◦C for the 2% and 5% material respectively. DSC quantified the changes in the thermal properties of the nylon and modified nylon materials. The quantities of interest included glass transition temperature (Tg) and melting temperature (Tm). The glass transition temperature for the pure nylon, 2% nanoclay, and 5% nanoclay materials were 45◦C, 44◦C and 43 ◦C, respectively. The melting temperatures were found to be 200 ◦C for the pure nylon, 203 ◦C and 204 ◦C for the 2% and 5% nanoclay materials. Therefore, the inclusion of
Effect of Coatings and Additives on the Water Absorption and Mechanical Characteristics of Additive Manufactured Nylon Polymers 3 the nanoclays into the nylon lower the Tg and increase the Tmtemperatures. The tension, compression, and flexure testing of the surface coated material indicate that both the coatings and salt water exposure affect the mechanical behavior of the Markforged Onyx material. When coated with both the epoxy hardpot and paint there is a reduction in modulus and strength as compared to the dry uncoated material. For each coating configuration there is a further drop in both stiffness and tensile strength after 60-day salt water exposure. The incorporation of the nanoclay additive into the pure nylon baseline material had two adverse effects on the tensile behavior of the material. The first is that the tensile strength of the nylon was reduced by the nanoclays, with the 2% nanoclay specimens having a lower strength than the 5% specimens. More significantly, the inclusion of the nanoclays changed the behavior of the nylon from a highly ductile material to one of high brittleness. The tensile strain at failure was reduced from above 100% for the nylon itself to 2% when nanoclay was incorporated. DMA results show that adding nanoclay increases the stresses and storage and loss moduli of the composite materials. The presence of the clay also reduced the overall loss in stiffness associated to water ingression. There was a remarkable increase in stiffness when the wet samples were tested at -2◦C indicating that the water had changed to solid. The nanoclays led to an overall brittleness in the nylon composites for both dry and wet samples. Acknowledgments The research presented in this study has been generously supported through the Internal Investment Program at Naval Undersea Warfare Center (NUWC) Division Newport. The authors greatly acknowledge the assistance of Jennifer M. Szarkowicz (NUWC Division Newport) in producing the 3D-printed samples for this study.
Chapter 2 Chapter 1 On the Detection and Quantification of Nonlinearity via Statistics of the Gradients of a Black-Box Model Georgios Tsialiamanis and Charles R. Farrar Abstrac t Detection and identification of nonlinearity is a task of high importance for structural dynamics. On the one hand, identifying nonlinearity in a structure would allow one to build more accurate models of the structure. On the other hand, detecting nonlinearity in a structure, which has been designed to operate in its linear region, might indicate the existence of damage within the structure. Common damage cases which cause nonlinear behaviour are breathing cracks and points where some material may have reached its plastic region. Therefore, it is important, even for safety reasons, to detect when a structure exhibits nonlinear behaviour. In the current work, a method to detect nonlinearity is proposed, based on the distribution of the gradients of a data-driven model, which is fitted on data acquired from the structure of interest. The data-driven model selected for the current application is a neural network. The selection of such a type of model was done in order to not allow the user to decide how linear or nonlinear the model shall be, but to let the training algorithm of the neural network shape the level of nonlinearity according to the training data. The neural network is trained to predict the accelerations of the structure for a time-instant using as input accelerations of previous time-instants, i.e. one-step-ahead predictions. Afterwards, the gradients of the output of the neural network with respect to its inputs are calculated. Given that the structure is linear, the distribution of the aforementioned gradients should be unimodal and quite peaked, while in the case of a structure with nonlinearities, the distribution of the gradients shall be more spread and, potentially, multimodal. To test the above assumption, data from an experimental structure are considered. The structure is tested under different scenarios, some of which are linear and some of which are nonlinear. More specifically, the nonlinearity is introduced as a column-bumper nonlinearity, aimed at simulating the effects of a breathing crack and at different levels, i.e. different values of the initial gap between the bumper and the column. Following the proposed method, the statistics of the distributions of the gradients for the different scenarios can indeed be used to identify cases where nonlinearity is present. Moreover, via the proposed method one is able to quantify the nonlinearity by observing higher values of standard deviation of the distribution of the gradients for lower values of the initial column-bumper gap, i.e. for “more nonlinear” scenarios. Keyword s Structural health monitoring (SHM) · Structural dynamics · Nonlinear dynamics · Machine learning · Neural networks 1.1 Introduction In the pursuit of making everyday life safer, humans have extensively tried to model the environment around them. Structures are an important part of the environment, in which humans live. They are man-made and should be safe throughout their lifetime. Structures are exposed to numerous environmental factors, which may cause them to fail. Moreover, during operation, structures are subjected to dynamic loads, which, in time, may cause failure. Such failures will most probably result in economic damage to society and may even result in loss of human lives. Therefore, for the purpose of maintaining structures safe, the field of structural health monitoring (SHM) [1] has emerged. G. Tsialiamanis ( ) Dynamics Research Group, Department of Mechanical Engineering, University of Sheffield, Sheffield, UK e-mail: g.tsialiamanis@sheffield.ac.uk C. R. Farrar Engineering Institute, MS T-001, Los Alamos National Laboratory, Los Alamos, NM, USA e-mail: farrar@lanl.gov © The Society for Experimental Mechanics, Inc. 2024 M. R. W. Brake et al. (eds.), Nonlinear Structures & Systems, Volume 1, Conference Proceedings of the Society for Experimental Mechanics Series, https://doi.org/10.1007/978-3-031-36999-5_1 1 Electrically Aligned Epoxy-Based Nanocomposites: Processing and Characterization Rahman Julkarnyne M. Habibur and Gururaja Suhasini Abstract This study examines the electric-field alignment of amine-functionalized multi-walled carbon nanotubes (MWCNTs) within an epoxy matrix, aiming to enhance their mechanical, electrical, and thermal performance. Amine-functionalized MWCNTs were dispersed within the epoxy resin using high-shear mixing to achieve a homogenous nanocomposite. 0.1, 0.2, 0.3 wt% of nanomaterials are used to prepare the samples. A key feature of the research involved the application of an external AC electric field before curing to align MWCNTs in the desired orientation. The voltage and the distance of electrodes were varied to examine the effect of the AC electric field on the alignment time. Response surface methodology (RSM), along with the central composite design (CCD), has been used to analyze the influence of those process parameters. Analysis of variance (ANOVA) was performed to identify the significant parameters. A voltage of 1000V and a distance of 50 mm between the electrodes was the optimal solution to minimize the alignment time while aligning MWCNTs in the epoxy matrix. Keywords Nanocomposite · Electric field alignment · ANOVA· RSM· Optimization Introduction Polymer matrix composites (PMCs) are the combinations of different organic polymers with reinforced fibers. Manufacturers can produce PMCs from polymers such as thermoplastics and thermosets. Because of their excellent mechanical properties, thermal stability, and chemical resistance, people widely employ epoxy resins—the most commonly used thermoset—in the aerospace, automotive, and electronics industries [1]. Nanofillers, such as carbon nanotubes, can further enhance the performance of epoxy resins. The efficacy of these epoxy-nanocomposites hinges on various factors such as the dispersion quality [2], [3], weight fraction [2], and alignment [4], [5], [6], [7] of nanofillers in the epoxy matrix. Alignment is generally done by applying external forces such as an electric field [8], magnetic means [6], high-shear mixing, etc. Electric field-assisted alignment has emerged as a versatile technique for orienting conductive MWCNTs, achievable via alternating (AC) or direct (DC) potential fields. Pothnis et al. [9] used the non-uniform AC electric field to align the CNTs, and Khan et al. [10] and Oliva et al. [11] used the DC electric field to align the CNTs in the matrix. The alignment time mostly depends on the rotation and migration of the MWCNTs in the epoxy matrix [12]. Viscosity plays a major role in the alignment as the lower viscosity facilitates the alignment [13]. Factors such as voltage and electrode distance significantly influence this alignment by determining the electric field strength. During the alignment process, the applied electric field generates a torque needed to overcome the epoxy matrix’s drag force, enabling CNT rotation and subsequent migration towards an electrode [12] Y. Minimizing the alignment time is an important factor to facilitate the large-scale fabrication of these aligned nanocomposites [14]. However, literature reports a wide disparity in these parameters—ranging from 10 V to 1000 V and distances of 1 mm to 50 mm—owing to variations in MWCNT type, epoxy viscosity, and system geometry [10], [11], [15]. While recent studies have leveraged RSM with ANOVA to optimize shear rate effects in MWCNT nanofluids along elongated surfaces [16] and employed Taguchi design for nitrogen-doped CNT synthesis [17], few have targeted the optimization of electric field parameters for minimizing CNT alignment time in epoxy matrix. This oversight limits the scalability and reproducibility of CNT/epoxy nanocomposite fabrication. Rahman Julkarnyne M. Habibur · Gururaja Suhasini Department of Aerospace Engineering, Auburn University, Auburn, AL 36849 e-mail: jzr0114@auburn.edu; suhasini.gururaja@auburn.edu © The Author(s), under exclusive license to River Publishers 2025 5 Emily Retzlaff et al. (eds.), Mechanics of Additive & Advanced Manufacturing, Inverse Methods and Machine Learning, Vol. 5, Conference Proceedings of the Society for Experimental Mechanics Series, https://doi.org/10.13052/97887-438-0831-2 2
6 R. J. M. Habibur and G. Suhasini In this work, a design of experiments approach integrating RSM and ANOVA has been adopted to optimize voltage and electrode distance for MWCNT alignment in the epoxy matrix. A CCD was utilized within the RSM framework to construct a systematic test matrix, incorporating critical process parameters: applied voltage, electrode distance, and MWCNT weight percentage. The response variable, alignment time, was experimentally determined for each condition outlined in the test matrix. Linear regression models were subsequently developed to map the relationship between input parameters and alignment time, enabling the identification of optimal process conditions. Additionally, ANOVA was performed to assess each parameter’s statistical significance and interactions, elucidating their individual and combined effects on the alignment process. Materials Epon-862 (diglycidyl ether of bisphenol F) and EPIKURE curing agent W (bisphenol-F epoxy resin and an aromatic amine) purchased from Miller-Stephenson were used as the resin and hardener following manufacturers recommended ratio of 100:26.4. Amine-functionalized MWCNTs (length: 10∼30 µm, outside dia: 10∼20 nm) were sourced from MKnano. The MWCNT dimensions were confirmed via SEM. All the materials were used as received. Nanocomposite Fabrication with Electric Field Alignment The nanocomposite is fabricated following a systematic approach to ensure proper dispersion. Initially, CNTs were preheated at 105◦C for 10 minutes to eliminate moisture. The required amount of preheated nanofillers was then added to the epoxy resin, and dispersion was performed using an overhead stirrer (ONiLAB Electric Overhead Stirrer) equipped with a propeller-shaped blade. The stirring process was conducted at approximately 1000 rpm for 4.5–5 hours at 80◦C to achieve effective dispersion. Afterward, the curing agent (Epikure W) was introduced into the mixture, followed by additional stirring for 7 minutes. The mixture was then vacuum-degassed for 10 minutes to remove the entrapped air. Following the dispersion process, the nanocomposite mixture was maintained at 80◦C in an oven before being transferred into a preheated silicone mold. The mold was rectangular, with dimensions of 180 mm×40mm (L×W), and was initially preheated in the oven to ensure uniform thermal conditions. Subsequently, the mold was placed on a hotplate set to 125◦C, ensuring the mixture temperature remained around 80◦C to maintain low viscosity during the alignment process. Two copper electrodes were positioned at a fixed distance and immersed in the nanocomposite mixture. An alternating current (AC) voltage was applied via a function generator coupled with a power amplifier to induce the alignment of carbon nanotubes (CNTs). The current was monitored with a digital multimeter throughout the alignment process to assess the progression of CNT movement. The voltage supply was discontinued after a predetermined duration, which depended on the applied voltage and electrode spacing. The mixture was then allowed to cool for 25 minutes before being transferred to an oven for curing. The curing process was conducted at 121◦C for 2 hours, followed by post-curing at 177◦C for 2 hours. Using this methodology, CNT/epoxy nanocomposites were fabricated with CNT weight fractions of 0.1, 0.2, and 0.3 wt.%. Figure 1 depicts the process flowchart. Experimental Procedure Three voltage levels of 500 V, 870 V, and 1000 V were used for these experiments. Distances of 50 mm, 75 mm, and 100 mm were used and combined with the mentioned voltage level to find the minimum alignment time. The frequency was fixed at 10 kHz. An RSM approach employing CCD was utilized to assess alignment time across various parameter combinations. Key process parameters—voltage, electrode distance, and CNT weight percentage (wt%)—served as experimental inputs (Table 1). The objective was to identify the optimal parameters yielding the minimum alignment time. The DOE approach reduced the experimental trials required while enabling analysis of voltage and distance effects on alignment time. Table 1 Process parameters Parameters Code Level of parameters Voltage (V) A 500 870 1000 Distance (mm) B 50 75 100 Weight percentage (wt%) C 0.1 0.2 0.3
Electrically Aligned Epoxy-Based Nanocomposites: Processing and Characterization 7 Fig. 1 Nanocomposite fabrication process with electric field alignment process Results and Discussion Using Stat-Ease 360 software, DOE tables were generated based on these inputs. Two experimental cases were considered for this approach. The first case involved only 0.1 wt.% CNT/epoxy, while the second case included 0.1, 0.2, and 0.3 wt.% CNT/epoxy. Table 2 presents the design matrix and corresponding output responses for 11 experimental runs in the first case, where voltage and distance were varied. Similarly, Table 3 details the design matrix and output responses for 15 experiments conducted under the second case based on the selected process parameters. In all these experiments, alignment time was monitored by measuring current (A) data from the multimeter. An empirical model was developed for both cases utilizing RSM. Following RSM implementation and identification of significant input parameters, ANOVA was performed to ensure the responses’ statistical accuracy and validate the proposed model’s suitability for predicting alignment time. Table 2 Experiment matrix and obtained results for 0.1 CNT/Epoxy Factor 1 Factor 2 Response 1 Run A: voltage (V) B: Distance (mm) Alignment time (min) 1 500 75 15 2 870 50 6 3 870 100 16 4 870 75 10 5 500 100 18 6 1000 50 3 7 1000 100 15 8 870 75 10 9 870 75 10 10 500 50 8 11 1000 75 12
8 R. J. M. Habibur and G. Suhasini Table 3 Experiment matrix and obtained results for different weight percentages (0.1, 0.2, 0.3 wt%) Factor 1 Factor 2 Factor 3 Response 1 Run A: Voltage (V) B: Distance (mm) C: Weight percentage (wt%) Alignment time(min) 1 870 100 0.1 16 2 870 50 0.1 6 3 500 75 0.1 15 4 870 75 0.1 10 5 1000 75 0.3 9 6 500 50 0.2 10 7 500 75 0.3 10 8 500 100 0.2 18 9 1000 100 0.2 10 10 500 50 0.1 8 11 870 50 0.3 10 12 1000 75 0.1 12 13 870 75 0.2 11 14 1000 50 0.1 3 15 870 100 0.3 14 Analysis for Alignment time ANOVA for minimized alignment time has been represented in Tables 4 to 5, respectively. The F-value is a critical metric for identifying factors that significantly influence the physical quantity of interest. Specifically, a large F-value indicates that variations in a particular factor have a substantial impact on the observed output [17]. Table 4 shows that the model F-value of 54.93 indicates significant model performance, with only a 0.01% probability that such a high F-value is due to noise. P-values below 0.0500 denote significant model terms [18], specifically A and B, in this case. Table 4 ANOVA for alignment time for 0.1 CNT/Epoxy Source Sum of Squares (SS) df Mean Square F-value p-value Remarks & Percentage of Contribution (%) Model 209.64 2 104.82 54.93 <0.0001 A-Voltage 28.14 1 28.14 14.75 0.0049 12.51 (significant) B-Distance 181.50 1 181.50 95.12 <0.0001 80.70 Residual 15.27 8 1.91 CorTotal 224.91 10 Std. Dev. 1.38 R2 0.9321 Mean 11.09 Adjusted R2 0.9152 Coefficient of variance (%) 12.45 Predicted R2 0.8652 Adequate precision 20.9609 In Table 5, the Model F-value of 9.45 implies the model is significant, and there is only a 0.22% chance that an F-value this large could occur due to noise. Additionally, P-values less than 0.0500 indicate that model terms are significant. A (voltage) and B (distance) are significant model terms valid at 95% confidence in both cases. Additionally, the signal-tonoise ratio is evaluated using adequate precision measures, with a ratio exceeding four deemed desirable [16]. In Table 4, the adequate precision is 20.96, and in Table 5, it is 9.6, signifying both models are desirable and can be used to navigate the design space. Figure 2 represents the normal plot of the residual for the alignment time. The residuals align closely along the normal line, indicating that the errors follow a normal distribution [18]. In both cases, the empirical model obtained for the output
Electrically Aligned Epoxy-Based Nanocomposites: Processing and Characterization 9 Table 5 ANOVA for alignment time for different weight percentages of CNT/Epoxy Source Sumof Squares (SS) df Mean Square F-value p-value Remarks& Percentage of Contribution (%) Model 160.60 3 53.53 9.45 0.0022 A-Voltage 40.65 1 40.65 7.17 0.0215 18.23 B-Distance 131.81 1 131.81 23.26 0.0005 59.13 C-Weight percentage 0.1124 1 0.1124 0.0198 0.8906 0.05 Residual 62.33 11 5.67 CorTotal 222.93 14 Std. Dev. 2.38 R2 0.7204 Mean 10.73 Adjusted R2 0.6442 Coefficient of variance (%) 22.18 Predicted R2 0.4162 Adequate precision 9.6048 Fig. 2 Normal probabilities of residuals for (a) Alignment time using only 0.1 wt% CNT/Epoxy and (b) alignment time when different weight percentages of CNT used was deemed effective at a 95% confidence level. The regression model for both cases factoring voltage and distance for the alignment time is presented in equations(1) and (2). When the weight percentage is fixed at 0.1 wt% CNT, then the regression model of minimum alignment time is represented in equation (1) Alignment time =11.540−2.061∗A+5.5∗B (Equation 1) When the weight percentage is varied between 0.1, 0.2, and 0.3 wt% CNT, then the regression model of minimum alignment time is represented in equation (2) Alignment time =11.22−2.017∗A+3.939∗B−0.106∗C (Equation 2) Where A represents the voltage, B represents the distance, and C represents the weight percentage of the CNT. Multi-response optimization The desirability function approach was employed for multi-response optimization to determine the optimal input process parameters. This method is widely utilized for optimizing both single- and multi-response problems. In this approach, each output response is converted into a dimensionless desirability index (D), ranging from 0 (least desirable) to 1 (most desirable). The optimal parameter set is identified when the corresponding solution achieves the highest overall desirability, ensuring
10 R. J. M. Habibur and G. Suhasini the best possible outcome across multiple responses [18], [19]. In this study, the optimal values of voltage and distance are achieved after establishing the goal to minimize the alignment time. Figure 3 represents the optimized bar histogram for desirability. Tables 6 and 7 present the constraints for the input parameters and the corresponding optimal solutions for minimizing alignment time. In both cases, the predicted optimal parameters were determined to be 1000 V and 50 mm, achieving the highest desirability score of 0.8. Furthermore, Table 8 compares the predicted and observed values of the process parameters to validate the model’s accuracy. The average deviation between observed and predicted values is Fig. 3 Bar histogram demonstration for desirability (a) Alignment time using only 0.1 wt% CNT/Epoxy (b) alignment time when different weight percentages of CNT used Table 6 Constraints and optimal solutions when weight percentage is fixed at 0.1 wt% Name Goal Lower Limit Upper Limit Lower Weight Upper Weight Importance Optimal solutions A: Voltage is in range 500.0 1000.0 1.0 1.0 3 1000 B: Distance is in range 50.0 100.0 1.0 1.0 3 50 Alignment time minimize 2.0 18.0 1.0 1.0 3 3.980 Table 7 Constraints and optimal solutions when weight percentage is varied Name Goal Lower Limit Upper Limit Lower Weight Upper Weight Importance Optimal solutions A: Voltage is in range 500.0 1000.0 1.0 1.0 3 1000 B: Distance is in range 50.0 100.0 1.0 1.0 3 50 C:Weight percentage is in range 0.1 0.3 1.0 1.0 3 0.3 Alignment time minimize 2.0 18.0 1.0 1.0 3 5.165 Table 8 Observed and predicted results at optimum values of process parameters Voltage (V) Distance (mm) Weight percentage (wt%) Alignment time (Predicted bymodel) Alignment time (Experimental value) 870 75 0.1 10.35 10 500 50 0.1 9 8 1000 75 0.3 9.15 9 1000 100 0.2 13.15 15 500 100 0.2 17.18 18
Electrically Aligned Epoxy-Based Nanocomposites: Processing and Characterization 11 approximately 0.83, demonstrating the model’s reliability in predicting alignment time for different voltage and distance combinations. Effect of Process Parameters on Alignment Time From the ANOVA results (Table 4, Table 5), it can be observed that both voltage and distance play a significant role in minimizing alignment time. The percentage of contribution ratio (PCR) of each input parameter can quantify the effect and the interaction of those parameters. The PCR in both tables is calculated by dividing the sum of squares (SS) for each factor by the total SS [17]. As shown in Table 4, the electrode distance contributes 80.70% to minimizing the alignment time for 0.1 wt.% CNT/Epoxy, whereas the applied voltage accounts for only 12.51%. Similarly, Table 5 presents the percentage contribution ratios (PCRs) for varying CNT weight fractions, indicating that distance and voltage contribute 59.13% and 18.23%, respectively, to reduce alignment time. These findings highlight the dominant influence of electrode distance over voltage in optimizing the alignment process. The ANOVA results indicate that electrode distance is the primary factor influencing alignment time. This is attributed to its significant impact on the electric field strength, which plays a crucial role in facilitating the formation of an aligned conductive network under the applied AC electric field [9], [10]. With the decrease in distance and the increase in voltage, the electric field strength increased which eventually increased the torque that facilitates the alignment of the CNT in the electric field direction[20]. In addition to the ANOVA results, the initial test matrix (Tables 1 and 2) highlights that a lower electrode distance and higher applied voltage enhance the electric field strength, thereby influencing alignment time. However, excessively high voltages can induce CNT aggregation, leading to the formation of rope-like structures, which may adversely impact the nanocomposite’s properties [13]. As shown in Table 5, the weight percentage of CNTs has a minimal effect on reducing alignment time, with a PCR of only 0.05%. However, in practice, CNT weight percentage does influence alignment time. A higher CNT concentration provides more conductive pathways, facilitating alignment under the applied voltage compared to lower concentrations [10]. Conclusions This study demonstrates the efficacy of integrating RSM and ANOVA to optimize voltage and electrode distance for minimizing the MWCNT alignment time in the epoxy matrix. • Higher voltage and shorter distance significantly enhance alignment efficiency. • Future work will involve COMSOL simulations to further investigate the effects of viscosity and epoxy curing dynamics, refining alignment kinetics, and paving the way for industrial scalability. • The study will be extended to optimize the voltage and distance for the GNP/Epoxy sample in the future. References 1. Kota, AK, “Commentary: Polymer/carbon-nanotube nanocomposites: from innovation to commercialization,” JNP, vol. 3, no. 1, p. 030307, Sep. 2009, 2. Zhang, D, Huang, Y, and Chia, L, “Effects of carbon nanotube (CNT) geometries on the dispersion characterizations and adhesion properties of CNT reinforced epoxy composites,” Composite Structures, vol. 296, p. 115942, Sep. 2022, 3. Shen, M-Y, Liao, W-Y, Wang, T-Q, and Lai, W-M, “Characteristics and Mechanical Properties of Graphene Nanoplatelets-Reinforced Epoxy Nanocomposites: Comparison of Different Dispersal Mechanisms,” Sustainability, vol. 13, no. 4, Art. no. 4, Jan. 2021, 4. Pothnis, JR, Kalyanasundaram, D, and Gururaja, S, “Mitigation of notch sensitivity by controlled alignment of carbon nanotubes in epoxy using electric field application,” Composites Part A: Applied Science and Manufacturing, vol. 149, no. 2021, p. 106544, Oct. 2021, 5. Moaseri, E, Fotouhi, M, Bazubandi, B, Karimi, M, Baniadam, M, and Maghrebi, M, “Two-dimensional reinforcement of epoxy composites: alignment of multi-walled carbon nanotubes in two directions,” Advanced Composite Materials, vol. 29, no. 6, pp. 547–557, Nov. 2020, 6. Moaseri, E, Karimi, M, Bazubandi, B, Baniadam, M, and Maghrebi, M, “Alignment of Carbon Nanotubes in Bulk Epoxy Matrix using a Magnetic-Assisted Method: Solenoid Magnetic Field 1,” Polymer Science, Series A, vol. 59, no. 5, pp. 726–733, 2017, 7. Wu, S, Ladani, RB, Zhang, J, Bafekrpour, E, Ghorbani, K, Mouritz, AP, Kinloch, AJ, and Wang, CH, “Aligning multilayer graphene flakes with an external electric field to improve multifunctional properties of epoxy nanocomposites,” Carbon, vol. 94, pp. 607–618, Nov. 2015, 8. Felisberto, M, Arias-Dura´n, A, Ramos, JA, Mondragon, I, Candal, R, Goyanes, S, and Rubiolo, GH, “Influence of filler alignment in the mechanical and electrical properties of carbon nanotubes/epoxy nanocomposites,” Physica B: Condensed Matter, vol. 407, no. 16, pp. 3181– 3183, Aug. 2012,
12 R. J. M. Habibur and G. Suhasini 9. Pothnis, JR, Kalyanasundaram, D, and Gururaja, S, “Controlled manipulation of CNTs in glass/epoxy composites with cut-outs using nonuniform electric field,” Advanced Composite Materials, vol. 30, no. 2021, pp. 205–221, 2020, 10. Khan, SU, Pothnis, JR, and Kim, JK, “Effects of carbon nanotube alignment on electrical and mechanical properties of epoxy nanocomposites,” Composites Part A: Applied Science and Manufacturing, vol. 49, pp. 26–34, Jun. 2013, 11. Oliva-Avile´s, AI, Avile´s, F, Sosa, V, Oliva, AI, and Gamboa, F, “Dynamics of carbon nanotube alignment by electric fields,” Nanotechnology, vol. 23, no. 46, p. 465710, Nov. 2012, 12. Monti, M, Natali, M, Torre, L, and Kenny, JM, “The alignment of single walled carbon nanotubes in an epoxy resin by applying a DC electric field,” Carbon, vol. 50, no. 7, pp. 2453–2464, Jun. 2012, 13. Amani, AM, Hashemi, SA, Mousavi, SM, Abrishamifar, SM, Vojood, A, Amani, AM, Hashemi, SA, Mousavi, SM, Abrishamifar, SM, and Vojood, A, “Electric Field Induced Alignment of Carbon Nanotubes: Methodology and Outcomes,” in Carbon Nanotubes - Recent Progress, IntechOpen, 2017. 14. Zhu, Y-F, Ma, C, and Zhang, W, “Alignment of multiwalled carbon nanotubes in bulk epoxy composites via electric field,” J. Appl. Phys, vol. 105, p. 54319, 2009, 15. Pothnis, JR, Gururaja, S, and Kalyanasundaram, D, “Mechanics of Advanced Materials and Structures Development and characterization of electric field directed preferentially aligned CNT nanocomposites,” 2019, 16. Baithalu, R and Mishra, SR, “On optimizing shear rate analysis for the water-based CNT micropolar nanofluids via an elongating surface: response surface methodology combined with ANOVA test,” J Therm Anal Calorim, vol. 148, no. 24, pp. 14275–14294, Dec. 2023, 17. Silva, RM, Fernandes, AJS, Ferro, MC, Pinna, N, and Silva, RF, “Vertically aligned N-doped CNTs growth using Taguchi experimental design,” Applied Surface Science, vol. 344, pp. 57–64, Jul. 2015, 18. Kumar, D and Gururaja, S, “Abrasive waterjet machining of Ti/CFRP/Ti laminate and multi-objective optimization of the process parameters using response surface methodology,” Journal of Composite Materials, vol. 54, no. 13, pp. 1741–1759, Jun. 2020, 19. Dhanawade, A and Kumar, S, “Experimental study of delamination and kerf geometry of carbon epoxy composite machined by abrasive water jet,” Journal of Composite Materials, vol. 51, no. 24, pp. 3373–3390, Oct. 2017, 20. Du, JR, Zheng, J, Ma, K, Du, C, and Zhou, D, “Electric Field-induced Orientation of Carbon Nanomaterials in Dispersions,” ChemistrySelect, vol. 9, no. 6, p. e202303547, 2024,
Chapter 3 Chapter 1 On the Detection and Quantification of Nonlinearity via Statistics of the Gradients of a Black-Box Model Georgios Tsialiamanis and Charles R. Farrar Abstrac t Detection and identification of nonlinearity is a task of high importance for structural dynamics. On the one hand, identifying nonlinearity in a structure would allow one to build more accurate models of the structure. On the other hand, detecting nonlinearity in a structure, which has been designed to operate in its linear region, might indicate the existence of damage within the structure. Common damage cases which cause nonlinear behaviour are breathing cracks and points where some material may have reached its plastic region. Therefore, it is important, even for safety reasons, to detect when a structure exhibits nonlinear behaviour. In the current work, a method to detect nonlinearity is proposed, based on the distribution of the gradients of a data-driven model, which is fitted on data acquired from the structure of interest. The data-driven model selected for the current application is a neural network. The selection of such a type of model was done in order to not allow the user to decide how linear or nonlinear the model shall be, but to let the training algorithm of the neural network shape the level of nonlinearity according to the training data. The neural network is trained to predict the accelerations of the structure for a time-instant using as input accelerations of previous time-instants, i.e. one-step-ahead predictions. Afterwards, the gradients of the output of the neural network with respect to its inputs are calculated. Given that the structure is linear, the distribution of the aforementioned gradients should be unimodal and quite peaked, while in the case of a structure with nonlinearities, the distribution of the gradients shall be more spread and, potentially, multimodal. To test the above assumption, data from an experimental structure are considered. The structure is tested under different scenarios, some of which are linear and some of which are nonlinear. More specifically, the nonlinearity is introduced as a column-bumper nonlinearity, aimed at simulating the effects of a breathing crack and at different levels, i.e. different values of the initial gap between the bumper and the column. Following the proposed method, the statistics of the distributions of the gradients for the different scenarios can indeed be used to identify cases where nonlinearity is present. Moreover, via the proposed method one is able to quantify the nonlinearity by observing higher values of standard deviation of the distribution of the gradients for lower values of the initial column-bumper gap, i.e. for “more nonlinear” scenarios. Keyword s Structural health monitoring (SHM) · Structural dynamics · Nonlinear dynamics · Machine learning · Neural networks 1.1 Introduction In the pursuit of making everyday life safer, humans have extensively tried to model the environment around them. Structures are an important part of the environment, in which humans live. They are man-made and should be safe throughout their lifetime. Structures are exposed to numerous environmental factors, which may cause them to fail. Moreover, during operation, structures are subjected to dynamic loads, which, in time, may cause failure. Such failures will most probably result in economic damage to society and may even result in loss of human lives. Therefore, for the purpose of maintaining structures safe, the field of structural health monitoring (SHM) [1] has emerged. G. Tsialiamanis ( ) Dynamics Research Group, Department of Mechanical Engineering, University of Sheffield, Sheffield, UK e-mail: g.tsialiamanis@sheffield.ac.uk C. R. Farrar Engineering Institute, MS T-001, Los Alamos National Laboratory, Los Alamos, NM, USA e-mail: farrar@lanl.gov © The Society for Experimental Mechanics, Inc. 2024 M. R. W. Brake et al. (eds.), Nonlinear Structures & Systems, Volume 1, Conference Proceedings of the Society for Experimental Mechanics Series, https://doi.org/10.1007/978-3-031-36999-5_1 1 Interfacial Characterization of Metal Wire Inlays for 3D Printed FDMParts Vereesh Ayyagari, Hugh A. Bruck, Amir H. Ohadi, and Michael M. Ohadi Abstract A novel process for creating 3D printed parts with metal wire inlay has been developed at the University of Maryland. By using metal wire inlay, parts can retain the properties of drawn metal wires that can be either embedded in polymer or fully exposed. As a result, the mechanical integrity of the metal wire interface with the FDM polymer can be controlled through the printing conditions and preparation of the wire surface. In this investigation, we develop a novel test specimen for characterizing the mechanical integrity of the metal wire/polymer interface. In particular, the test specimen is designed to quantify the ability of the interface to provide hermetic sealing when loaded by pressurized fluids. The effects of the processing and preparation conditions on the interfacial strength are also investigated via single fiber pullout tests. As a result, it is possible to design new 3D-printed metal-polymer composite structures with free standing metal wires that can be utilized in a a variety of different applications ranging from integrated flow sensing to enhanced thermal control. Keywords Metal-polymer composites · Polymer-metal interface · Failure characterization· Adhesion · Interfacial strength Introduction Polymer matrix-metal wire composite structures are increasingly being developed for strengthening polymer structures [1], heat exchangers [2], and conductive path tracing for electronic components [3]. For a mechanistic point of view, there could be two variants of the composites. In the first, the metal wire is wholly embedded in the polymer. In the second composite, the metal wires are partially exposed, resulting in a partially free-standing metal wire. The two variants of the composites are shown in Figure 1, where Figure 1a shows the polymer-metal composites where the metal wires are completely embedded into the polymer, and 1b shows the composites where metal wires are partially free-standing. While the fabrication of the composite can be achieved through conventional manufacturing techniques, additive manufacturing, commonly referred to as 3D Printing, has become popular choice of fabrication, primarily due to the complex nature of the composite. For 3D printing of the composites, Fused Filament Fabrication (FFF) has been widely used for depositing the polymer, and metal wires are thermally embedded. In FFF, thermoplastic materials, such as polycarbonates, acrylonitrile butadiene styrene (ABS), Polylactic Acid (PLA), and Poly Amides (PA), are used as feedstocks due to their favorable properties such as low melting temperature and appropriate melt viscosities [4]. Recently, we have developed a 3D printing process for metal wire-inlay to create cross-media heat exchangers [5–8] which contains an array of free-standing metal wires embedded into the polymer structure. The failure mechanisms of the polymer-metal composites are critical to understanding their limitations in the applications. In particular, to estimate how much pressure can be applied to the structure before the wires start pulling out i.e. interfacial strength between metal and polymer, wire push out tests are applied [1, 9–11]. While a few studies exist on the characterization of composites where metal fibers that are embedded continuously in a polymer matrix, very few have studied 3D-printed polymer-metal composites fundamentally, where the metal wires are partially exposed. The current study focuses on the failure characterization of the partially free-standing metal fibers in polymer-metal composites. Vereesh Ayyagari · Hugh A. Bruck· Amir H. Ohadi · Michael M. Ohadi Department of Mechanical Engineering, A. James Clark School of Engineering, University of Maryland, College Park, MD 20742 e-mail: veeresha@umd.edu; bruck@umd.edu; amir@umd.edu; ohadi@umd.edu © The Author(s), under exclusive license to River Publishers 2025 13 Emily Retzlaff et al. (eds.), Mechanics of Additive & Advanced Manufacturing, Inverse Methods and Machine Learning, Vol. 5, Conference Proceedings of the Society for Experimental Mechanics Series, https://doi.org/10.13052/97887-438-0831-2 3
RkJQdWJsaXNoZXIy MTMzNzEzMQ==