Sage Journals: Discover world-class research

Abstract

Additively manufactured parts are distinctly influenced by diverse 3D-printing parameters, which directly affect their mechanical properties and overall performance. Key factors such as print orientation, infill density, nozzle temperature, and layer height play an essential role in determining the strength, durability, and dimensional accuracy of the printed parts. These parameters must be carefully optimized to get the appropriate mechanical characteristics and ensure that the final product meets quality and performance standards. This research investigates the mechanical behaviour of boron Fiber-reinforced glass bead-filled polyamide 12, focusing on fracture mechanics across various printing orientations. To optimize the composite formulation, this mechanical impact predictive analysis integrates deep learning and advanced optimization techniques, specifically Enhancing Spiking Neural Networks with Hybrid Top-Down Attention and Clouded Leopard Optimization (ESNN-HTDA-CLO). Additionally, post-heat treatment improves tensile strength and surface roughness, achieving nearly a 10% improvement. Also, this research aims to address challenges in enhancing stiffness and temperature resistance while showcasing the composite advantages. The findings underscore the potential of the developed composite in aerospace, automotive, electronics, construction industries, medical devices, and advanced manufacturing.

Keywords

Additive manufacturing material extrusion 3D printing composite material optimization deep learning mechanical properties evaluation post-heat treatment surface roughness

Introduction

Additive Manufacturing (AM), particularly using the technology of material extrusion, is a popular approach for prototyping and creating finished items. This approach has various advantages, including flexibility, the capacity to generate complicated forms, low user knowledge requirements, the ability to fabricate geometries not possible with other technologies, and compatibility with a variety of materials.¹ Polymers are predicted to lead the 3D printing materials market since 2021 due to their outstanding mechanical qualities, strength, and durability, all of which contribute to their widespread use. “3D printing polymers” include photopolymers, Poly Lactic Acid (PLA), acrylic styrene, PolyAmide (PA), polycarbonates, and other polymer varieties.² The performance of AM parts with good mechanical properties are highly influenced by material selection and 3D printing parameters like print orientation, infill density, nozzle temperature, and layer height. Optimizing these factors is essential to ensure strength, durability, and overall quality of the product.³ Failure to control these factors may result in defects such as poor layer adhesion,⁴ voids,⁵ and reduced structural integrity.⁶ As such, understanding and managing 3D printing parameters are essential for enhancing the reliability of additively manufactured components in engineering applications.⁷

By considering one of the polymer variants called PA-12 material, which is the combined product of conventional Nylon and Carbon Fiber (CF), gaining interest in the 3D printing because of its high mechanical properties in convergence with other polymers and its composite combinations. Different fabrication methods are tailored to additively manufacture the end products using PA-12 material, such as material extrusion 3D printing, the processing strategy of depositing material layer by layer through a nozzle to build 3D objects, which improves the mechanical properties of PA and discovered that wear rate decreases as nozzle temperature increases.⁸ The samples manufactured using Selective Laser Sintering (SLS) had the most effective tensile characteristics while having a lower layer thickness (0.1 mm). Furthermore, parts made with 0.2 mm layer thickness showed advantageous mixed-mode breaking characteristics.⁹ The Machine Learning (ML) approach predicted the impacts of crystallinity behaviour of the manufactured PA-12 samples using the Multi-Jet Fusion (MJF) technique, and the results revealed the time of the initial cooling, end temperature of the 3D printing process had a substantial influence on different thermal properties, impacting mostly.¹⁰

Several research investigations on mechanical behaviour analysis are carried out experimentally with PA-12 composites in order to optimize critical process parameters. The most important parameters are print orientation, infill density, nozzle temperature, layer height, and speed. The optimal control and regulation of these process parameters is increasingly important. There are many applications, including aerospace, that require custom products with a high stiffness-to-weight ratio as industries are highly weight-conscious and want less product weight. Such products should also have mechanical and thermal stability with a long fatigue life, as emphasized by Saeed et al.¹¹ In conjunction with a fair experimental setup, Rath et al.¹² studied mechanical properties of 3D printing Nylon-12CF composite, focusing on the impact of critical processing parameters. The tensile, flexural, and impact strength studies are: with morphologies and wear analysis. The outcomes revealed that the highest tensile properties with maximum tensile stress of 53.74 MPa are recorded, and optimal parameters are identified, such as 0.03 mm layer height, speed of printing (10 mm/s) for a hexagonal infill pattern. Also, the highest flexural strength (98.97 MPa) and impact strength (21.35 kJ/m²) are achieved for 30 mm/s print speed, layer height of 0.03 mm, and linear infill pattern. These parameters are found as optimal one when compared with Beylergil et al.,¹³ which have an impact strength of 10.54 kJ/m² only.

A Fused Filament Fabrication (FFF) 3D printer was recently used in an investigated by¹⁴ Sandeep Varma et al.¹⁴ to look into the non-linear connections between mechanical qualities and input parameters. Relative Error (RE) analysis is employed in this study to determine the best parameter combinations. Hossain et al.¹⁵ examined PA12 composite materials and their machine parameters, using the current state-of-the-art ML incorporated Universal Testing Machine (UTM). The approximation using ML having excellent agreement with the extrusion temperature, non-conductivity, extruded filaments (2500 Vs). Energy dispersive X-ray spectrometry, Field Emission Scanning Electron Microscope (FESEM) and particle analysis all attest to the polymer’s excellent surface texture and microstructure.

Parameters like infill percentages and nozzle temperature are crucial in the strengthening of the mechanical characteristics of the PA12. The analysis is conducted by Gómez-Ortega et al.¹⁶ to investigate the consequences of wall thickness, nozzle temperature, and infill percentage on the porosity content and microstructure state of PA-CF specimens using a Taguchi-L9 design of experiments. The outcomes of the tensile and flexural tests, the printing temperature had a significant effect on microstructural alterations and encouraged debonding behaviour in the interlayer because of voids and holes at the skin-core and wall interfaces. Additionally, it was shown that all thicknesses had a substantial influence on tensile strength, with the ideal ranges for wall thickness (1.15 to 1.5 mm) and maximum flexural strength (0.8 to 1.35 mm) being determined. Ferdousi et al.¹⁷ created a lightweight hybrid composite with ML integration that is 3D printed and made of thin-shell particles, microfillers, and an elastomer matrix. Microstructures may be designed with a great deal of freedom thanks to it, improving mechanical characteristics and densities by up to 91% and 70%, respectively.

The tensile and infill density parameters of different PA composite materials like CF-reinforced polyamide-6 (PA6-CF) material, are analysed by Al Rashid and Koc.¹⁸ The numerical model is designed to precisely predict the infill patterns and impact deflections densities and distortions during the FFF process. Furthermore, improved dimensional control is found for rectangular infill designs and detected enhanced infill density with the optimal management of ML. In the same way, Udu et al.¹⁹ examined the mechanical properties of 3D printed Carbon Fiber-Reinforced Polyamide (CF-RPA) and Carbon Fiber-Reinforced Acrylonitrile Butadiene Styrene (CF-ABS) composites, including their flexural, tensile, compressive, porosity, and hardness. Using ML approaches, namely ensemble tree learners and K-Nearest Neighbour (K-NN) regressor algorithms, the mechanical properties were predicted with an accuracy ranging from 80 to 99%. The improved effectiveness of the composite material was observed through the results when temperature and porosity were selected optimally. It also predicted the hardness of the material and tensile strength, flexion behaviour, and compression. Likewise, a mixed glass fiber and ABS called ABS/Glass fiber polymer composite was prepared by Kumar et al.²⁰ and the mechanical properties were predicted using Classification and Regression Trees (CART) algorithm based on predictor variables. Results show the peak tensile stress of 39 MPa was achieved with the process parameters like 250°C, 80°C temperatures at the nozzle and bed, respectively and an infill density of 60% and the most influencing parameter was infill density followed by nozzle temperature and bed temperature as per the prediction.

The prediction and optimal control of the mechanical parameters during the material extrusion 3D printing process in the existing strategies with ML techniques like random forest regression, grey relation analysis showed fewer predicted values below 96%. The other mechanical parameters are yield strength and toughness, and they were valued the highest, i.e., 0.9965 and 0.96, respectively, according to Ali et al.²¹ The best parameters were flow rate = 100%, nozzle temperature = 185°C, and layer thickness = 0.15 mm. Additionally, it was found that, at 53% of the total, the layer thickness has a substantial impact on the mechanical performance.

From the conducted review on the previous literature with the Nylon and its composites like PA12 and other mixed composite materials, play a pivotal role. Since it exhibits high mechanical characteristics compared to other polymers like PLA, ABS, and conventional Nylon. Also, wear rate decreases as nozzle temperature increases, which impacts the product quality mostly, and maximum tensile stress of 53.74 MPa and 39 MPa achieved with optimal process parameters like nozzle temperature of 250°C, bed temperature of 80°C and infill density at 60%-layer height (0.03 mm), print speed (10 mm/s). Also, found that the infill density and temperature at the nozzle impact the mechanical properties. However, there is a lack of studies on the surface, mechanical characteristics and tribological properties of polymer matrix composites integrated with boron fibers and glass beads. Furthermore, Ghabezi et al.²² optimised the filament production pa rameters to fabricate a high-quality composite and streamline the MEX printing parameters to produce composite parts with favorable mechanical properties and printability from recycled materials. Zhang et al.²³ designed an impregnation device for resin-injected fiber tows to achieve good impregnation. Most recently, the ML integrated process optimization for the glass fiber mixed PA12 composite identified thickness of layer as 0.15 mm, a temperature at the nozzle of 185°C, and 100% flow rate as the optimal parameters. The addition of glass fibres can significantly reduce the thermal expansion coefficient, enhancing the composite material stability with varying dimensions during temperature variations and concluded that, glass beads have demonstrated improved thermal stability and a high material modulus of stiffness. Unlike prior parameter-based optimization studies, this work introduces a data-driven mechanical impact predictive framework by advancing conventional parameter optimization by combining ML driven predictive modeling, fracture-focused mechanical analysis, and post-processing enhancement, thereby establishing a robust pathway for high-performance composite components suitable for advanced manufacturing applications.

Research gap

Researchers are looking at using composite materials to improve AM’s mechanical qualities and numerous studies have been published on this field. These materials include reinforced continuous fiber, short fiber (reinforced), nanofiber (reinforced) and glass beads (reinforced). A limitation of this 3D printing technology as a novel manufacturing method for a novel class of materials is that mechanical characteristics are still predicted with a low degree of accuracy (96–99%), and constitutive and strength models specialised to material extrusion 3D printing materials are not yet accessible. As a result, actual material extrusion 3D printing structures cannot be accurately produced with high precision and examined. The development and use of material extrusion 3D printing have been hampered by a lack of knowledge, which has led to the necessity of extensive study into forecasting the mechanical properties of its materials, especially with regard to failure strengths and elastic qualities. Even though elastic characteristics have been extensively studied, failure strength assessments are still unusual and rarely performed.

Novelty of the proposed research

The primary goal of this mechanical property analysis is to examine the tribological features of prepared novel composites of PA-12 with Boron Fiber and Reinforced Glass Bead (PA12BF-RGB). These commercially available materials are often considered more durable than those commonly used in 3D printing. Furthermore, Boron fibers and glass beads enhance strength, wear resistance, and stability, making the composite suitable for high-performance applications. This combined mixed composite named PA12BF-RGB is prepared by utilizing FDM a 3D-printing method based on material extrusion, which has properties like improved tensile strength, stiffness, enhanced fracture toughness and thermal resistance. The specimens used for evaluation are oriented in three directions in the build chamber of the material extrusion 3D printing machine. The tribological characteristics of PA12BF-RGB composites are examined. This research examined the impact of fiber orientation and the mechanical qualities using a novel ML strategy of ESNN-HTDA-CLO. Also, the outcomes are compared with state-of-the-art reinforced Nylon 12 samples from the literature. This research experiment is illustrated in Figure 1.

Figure 1.

Organisation of the proposed research experimental analysis.

Main objective of the proposed experimental mechanical behaviour analysis of the PA-12 composite framework are listed as follows:

✓ To develop a PA12BF-RGB composite with superior tensile strength, stiffness, fracture toughness, and temperature resistance for advanced engineering applications.

✓ To investigate the influence of key material extrusion 3D printing parameters, such as layer height, printing speed, nozzle temperature, and infill density, on the mechanical performance of the composite.

✓ To use a hybrid framework to balance critical mechanical properties and identify the optimal composite formulation.

Proposed mechanical performance analysis description

The composite mixture of the proposed PA12 base material preparation and the test sample fabrication are presented in this section. Experimental procedures from the mixing of the composite materials to evaluation procedures are described briefly.

The suggested mechanical property analysis experiment is separated into three stages: input, processing, and output, as shown in Figure 2. In the input step, the basic material NYLON - PA12 is combined with composite materials such as boron fibers and glass beads following the conventional technique¹ and the mechanical characteristics are listed in the Table 1. The manufactured composite, designated as PA12BF-RGB test samples, enabled mechanical property assessment with state-of-the-art PA12 reinforced composite materials.²⁴ This analysis is conducted by using the ESNN-HTDA-CLO machine learning technique. With the advanced use of ML technology, an effective PA12 composite with effective mechanical and thermal characteristics are analysed, and optimal process parameters in the material extrusion 3D printing process such as print orientation, infill density, nozzle temperature, and layer height are determined and the results are compared to recent literatures. The processes used in this experimental investigation are described in the following sections.

Figure 2.

Proposed optimal mechanical property analysis architecture.

Table 1.

Mechanical characteristics of the materials used in this research.

Base material	Young’s modulus ( $M P a$ )	Density ( $g / c m^{3}$ )	Elongation (%)	Ultimate tensile strength ( $M P a$ )	Ultimate flexural strength ( $M P a$ )	Supplier
NYLON–PA12^25,26	5160	1.01	26	60	122	Zortrax S. A., Olsztyn, Poland
BR5746-boron fibers²⁷	4500	2.34	16.5	41.73	43.29	Markforged, U.S.A
Glass beads²⁸	72	1.2	25	44	75	3DXTECH, Markforged ONYX

Materials and processing

This study investigates the mechanical enhancement of PA12 composites reinforced with boron fibers and glass beads. These reinforcements have been chosen because of their capacity to strongly influence the strength, stiffening character, and thermal behavior of PA12. Boron fibers aid to increased stiffness, strength, and heat conductivity of the composite, although they generally reduce ductility because of their brittle nature. Reinforcement of boron fibers and glass beads accounts for the improvement in stress response, with studies showing up to an almost 50% increase in the strength of reinforced PA12 when compared to non-reinforced PA12; however, this is still accompanied by good ductility. These fillers enhance the PA12 composite in load-bearing and mechanical wear resistance performances. All materials are supplied as 1.75-mm commercial filaments; therefore, the exact filler composition is not disclosed by the manufacturer. Due to this limitation, the study focuses on analyzing filler size and distribution in both raw filaments and 3D-printed specimens. Rather than quantifying each component individually, the objective of this work is to evaluate the overall consequence of reinforcing fillers and strain ratio on the mechanical behavior of PA12-based composites.

Fabrication of test samples

It is to establish the mechanical characteristics of the PA12 composites selected for the study that the material extrusion printer has produced a 3D printing test sample for modeling in computer programs. In fact, material extrusion 3D printing is a very convenient method to print polyamide and composite filaments because it has a fast and flexible printing speed, it is cheap, offers a broad range of materials, and extremely high strength and durability.

A digital 3D model is created in Computer-Aided Design (CAD) software and saved in a Stereolithography (STL) file. Slicing software which manages the material extrusion 3D printer and allows the user to modify parameters such as speed, layer thickness, and temperature is used to cut the object into horizontal layers. The fibres are dispersed into the matrix through a melt-based filament fabrication process. The chopped boron fibre, glass beads and PA12 pellets are put into a Zortrax M200 Plus 3D printer. The filament diameter is controlled through continuous monitoring during extrusion. The effective filament manufacturing parameters including nozzle temperature of 260°C, nozzle diameter of 0.6 mm and bed temperature of 70°C are modified to achieve favourable quality filaments in terms of diameter. Support structures are removed manually after printing. It is because of this that the material extrusion printer has manufactured for itself a 3D printing test model that is meant to be used to model its characteristics under consideration in computer modeling programs, and so it is patterned in such a way as to be seen as mechanical with the consideration of being associated with the PA12 composites. In fact, the reason material extrusion 3D printing is very convenient compared to other printing techniques for polyamide and even composite filaments is that while really fast and flexible, it’s also cheap with a wide variety of materials having incredibly strong and durable printing.

A digital 3D mold is created in CAD software and saved as a STL file. Slicing software which manages the material extrusion 3D printer and allows the user to modify parameters such as speed, layer thickness, and temperature is used to cut the object into horizontal layers. Until the model is finished, printing filament is extruded layer by layer. The structures supporting the object are removed manually.

For the assessment of mechanical properties, three test samples of composite materials like PA12BG, PA12CF-AM, and the suggested PA12Bf-RGB are developed. When considering the manufacturing process, an material extrusion 3D printing item is simplified to behave as a transversely isotropic material, with the material layer plane representing the plane of transverse isotropy. Three different samples with orientations like $0^{o}$ , $\pm 45^{o}$ and $90^{o}$ angles are fabricated for the conducted experimental performance evaluation

Test sample printing description

To prepare for printing, the prepared filaments are dried at 50℃ for 1 day. Markforged Onyx One™ 3D printer with the dimension $(width \times depth \times height)$ of $584 \times 330 \times 355 mm$ with the build volume of $320 \times 132 \times 154 mm$ is utilized to create specimens with a mixture of Z-NYLON™ -PA12 composite filament.²⁹ The GCode is generated using the specialised slicer program, Z-Suite. The default production settings are used, like 0.4 $m m$ nozzle diameter, 0.19 $m m$ of layer thickness. The 3D printing procedure used nozzle temperatures in the range of 210°C to 280°C, making it safe for material processing, according to the seller. These temperatures are chosen to optimize the materials processability during material extrusion 3D printing, and 70℃ is maintained at the print bed, and the printing speed is averagely has 40 $m m / s$ . A typical perforated printing bed with rafts is used to create the specimens. Because the reinforced filaments are quite abrasive, the setup is changed to incorporate a dual-drive extruder with a ruby nozzle. This change guarantees constant printing conditions and reduces nozzle wear. To improve the adherence of the first layer, a PVP-based adhesive and a perforated glass printing bed are used. All specimens had 100% infill density and are printed in a variety of orientations, including, $0^{o}$ , $\pm 45^{o}$ and $90^{o}$ to its axis. The following layers are printed using reversed route angles. Figure 3 shows printing process of printed specimens from the base materials with different orientations.

Figure 3.

Experimental evaluation of mechanical property evaluation procedure.

For fiber-reinforced specimens, an isotropic infill pattern is used and altering the infill type changes strength, stiffness, and failure modes. Fibre volume fraction estimate is also crucial, and is stated to be 35% utilizing characterisation methods. The fabrication process of boron fibers and glass beads reinforced NYLON - PA12 composites is shown in Figure 4. The fiber evenly distributed in the material are extruding from the nozzle and the printer melts filament and extrudes it onto a forming table to create a three-dimensional object. The extrusion and attachment of the printing filament layer by layer progresses until the model is finished.

Figure 4.

Fabrication process of boron fibers and glass beads reinforced NYLON - PA12 composite.

Testing parameters

Tensile testing is used to evaluate the performance of boron fiber and reinforced glass bead specimens, with a strain load (2 N) in the clamps before the test begins. Tensile tests are performed utilizing Materials Test Systems (MTS) with a load (100 KN) to impart a load test. Strain is measured both longitudinally and transversely using two extensometers. Utilising a 25.4 mm gauge length extensometer (Model number 634.11 F-5x), the test sample’s strain is measured longitudinally. The strain rate at which specimens are examined is 2 mm/min. To optimize the exactness of the results, 5 test specimens of each kind are analysed for tensile strength analysis.

The strain ( $ε$ ) is measured both longitudinally and transversely in order to calculate Poison’s ratio with the range from $ε = 1 e^{- 02}$ to $ε = 1 e^{- 06}$ are analysed according to the standards as described in the following sections.

Measurement standards description

For the experimental mechanical evaluation of the composite filaments, the Type 1 dog bone form geometry as shown in Figure 5 is taken into consideration in accordance with the, Standard ASTM-D638 for tensile strength. The samples are manufactured using the conventional dimensions of 165 mm × 19 mm × 3.2 mm.¹ A tensile machine is used to test the samples, and the fracture site is found to be outside of the gauge length. The nylon layer is removed by abrading the endpoints of the test specimens because the bonding is not flawless, and slippage happens without the nylon layer being removed.

Figure 5.

Test sample geometrics.

The test duration time for the PA12BF-RGB specimen at $\dot{ε} = 1 e^{- 02}$ is 3.3 s. This helps define the strain rate range under investigation. PA12 composite specimens examined at $\dot{ε} = 1 e^{- 06}$ lasted 119.4 hours. Flexural properties evaluation is in accordance with ASTM D790 and the dimensions of composite materials with dimensions (125 ×12.7 × 3.2 mm) respectively. Creep examination is carried out under stress control at room temperature (around 23°C). The loading is being done in 0.1 seconds and maintained for a whole day. The specimen is emptied, after which it is kept unloaded for an entire day. Creep stress = 20 MPa is selected as the base level, whereas 40 MPa and 80 MPa are further tested. Creep test results are shown in Outcomes and Discussions.

To evaluate the fracture toughness, single-edge notched bending samples are printed based on ASTM D5045 standard. The height (W), width (B), depth notch (a), and total length (L) of the materials are selected as 12, 6, 5, and 60 mm, respectively. Loading is executed in three-point bending mode at ambient temperature and a displacement rate of 1 mm/min. The surface roughness measurement of the materials are performed befor and after polishing treatment using ARSURF PS10 surface profilometer.

Effect of annealing on PA12BF-RGB composites

In this research, annealing was used as a post-processing method for the improvement of the mechanical behavior of additive-manufactured composites. Specimens are prepared using material extrusion 3D printing through the reinforcement of Polyamide 12 (PA12) with boron filaments and glass beads. Post printing, the specimens are annealed at controlled temperatures to improve tensile strength, surface smoothness, and relieve internal stresses. For Annealing, the specimens are heated for a time at 90°C, 110°C, and 130°C to understand annealing affecting the mechanical properties of the specimens. Mechanical testing is completed with bending tests, hardness measurements, and Thermo Gravimetric Analysis (TGA) to understand the mechanical properties of the specimen. The radius of curvature is calculated for assessing the effects of different annealing cycles on the final geometry of the part.

Bending strength calculation:

σ = \frac{3 P_{a p p . l d} L_{l e n g h t}}{2 b T^{2}}

(1)

where

P_{a p p . l d}

is the applied load,

L_{l e n g h t}

is the span length,

b

is the specimen’s width,

T

is the specimen’s thickness

Radius of curvature calculation:

R = \frac{C^{2} + 4 D^{2}}{8 d}

(2)

where

C

and

d

are the measured variables to determine curvature.

Formulation of optimal mechanical property prediction

The mechanical stability prediction of the produced composite filament with proposed ESNN-HTDA-CLO is developed in this section. For every mechanical reaction ( $σ_{s}$ , ‘ $\dot{ε}$ ’, and ‘ $t$ ’), three ESNNs are created. The results of earlier, comparable studies and trial-and-error simulations are used to find the count of neurons in layers of hidden as depicted in Figure 6.³⁰ Performance is optimized by 12 neurons for $σ_{s}$ and ‘ $t$ ’, and by 4 neurons for ‘ $\dot{ε}$ ’. The models for ESNN are developed using MATLAB R2024b. Like artificial neurons, spike neurons use a weighted sum of inputs. However, the weighted sum increases the neuron membrane potential $(V_{m})$ rather than employing nonlinear activation like sigmoid or ReLU. When a neuron’s membrane potential hits a certain threshold $(V_{t h})$ , it transmits a spike to following connections. Also, the activity of spiking neurons mimics a resistor $(R_{r})$ -capacitor $(O)$ circuit, subsequently referred to as the Leaky-Integrated and Fire (LIF) neuron. In the training phase, proposed hybrid ESNN-HTDA-CLO algorithm is used as presented in Table 2.

Figure 6.

Enhancing spiking neural networks architecture.

Table 2.

Training phase with hybrid ESNN-HTDA-CLO³¹.

A hybrid neural network that uses LIF, Recurrent Leaky-Integrated and Fire (RLIF), and dense layers to change the pretraining input sequence.

L^{[q]} = ([{({\hat{A}}^{'})}^{1}, 0, h, H, 0], . . . . . . . ., [{({\hat{A}}^{'})}^{n}, {(\hat{l})}^{'}, h, H, S^{n - 1}])

(3)

To predict the strain and stress development and the respective time step-based represented in equation (2) as follows,

M^{[q]} = ([Δ ε_{e}^{- 1}], . . . . . . . ., [Δ ε_{e}^{- n}]) .

(4)

Update mechanism of CLO

The CLO algorithm updates each solution using two update rules:

1. Exploration update (Phase 1)

2. Exploitation update (Phase 2)

These rules decide how each clouded leopard’s position (solution) changes during the optimisation process.

Exploration update – phase 1

The exploration update allows each solution to make a large movement based on the position of another randomly selected solution (called prey). If the prey has a better objective value than the current solution, the movement is directed toward the prey. Otherwise, the movement is directed away from it.

The exploration updation equation is

X_{i, j}^{P 1} = {\begin{cases} X_{i, j} + r_{i, j} (p_{i, j} - L_{i, j} X_{i, j}), i f F_{p} ≺ F_{i} \\ X_{i, j} + r_{i, j} (X_{i, j} - L_{i, j} P_{i, j}), o t h e r w i s e \end{cases}

(5)

Here, $X_{i, j}^{P 1}$ is the updated position, $p_{i, j}$ is the position of the selected prey, $r_{i, j}$ is a random number in [0,1], and $L_{i, j}$ is randomly chosen as 1 or 2. This equation either moves the solution closer to or farther from the prey, depending on which one has a better objective value.

The new position is accepted only when it improves fitness, which is handled by:

X_{i} = {\begin{cases} X_{i}^{P 1}, i f F_{i}^{P 1} ≺ F_{i} \\ X_{i,} o t h e r w i s e, \end{cases}

(6)

Exploitation update – phase 2

The exploitation update performs a small adjustment around the current position to refine the solution. The step size becomes smaller over iterations since the term is raised to the power $t$ , allowing gradual convergence.

The exploitation updation equation is:

X_{i, j}^{P 2} = X_{i, j} + {(L_{j} + r_{i, j} (U_{j} - L_{j}))}^{t} (2 r_{i, j} - 1)

(7)

Here, the expression $(2 r_{i, j} - 1)$ creates either a positive or negative movement, and the term ${(L_{j} + r_{i, j} (U_{j} - L_{j}))}^{t}$ controls the magnitude of the local step, which decreases as the iteration number $t$ increases. This creates a fine-tuned local search around the current solution.

The updated position is accepted only if it leads to an improved objective function, given by:

X_{i} = {\begin{cases} X_{i}^{P 2}, i f F_{i}^{P 2} ≺ F_{i} \\ X_{i,} o t h e r w i s e, \end{cases}

(8)

This optimization step is essential in this concept because it calibrates the parameters of the Enhanced Spiking Neural Network (ESNN) to accurately determine the mechanical properties of the composite material. By applying the Clouded Leopard Optimization (CLO) algorithm, the optimization methodology manipulates the parameters of the neural architecture to find the optimal solution through a balance between exploration (wide search) and exploitation (fine-tuning). Hence, predictions become more precise for tensile strength, wear rate, or any of the properties. This process would help design superior materials by providing optimal printing and annealing conditions to achieve better performance.

In the subsequent section, this simplified optimal prediction model for the results of mechanical property analysis is further elaborated, including the comparison of the effectiveness of the proposed novel composite filament with that from the existing literature, and the enhanced mechanical properties of the proposed composite.

Outcomes and discussions

The outcomes of the proposed mechanical property analysis with optimal prediction of parameters are presented in this section. By conducting tensile tests, the mechanical effectiveness of the proposed material is analysed with the parameters as listed in Table 3 based test conditions. The printing direction (orientation angle), temperature, infill density and speed properties are studied using a statistical approach. The considered samples with the composite filaments including conventional PA12, BubbleGlass™ (PA12BG), NanoCarbon AM™ (PA12CF-AM) and the proposed PA12BF-RGB.

Table 3.

Parameters for testing of tensile strength for the proposed composite specimens.

Testing parameters	Specifications
Fill density	100%
Layers of fibre	25
Layer-external	1
Top layer count	1
Bottom layer count	1
Orientation angle	$0^{o}$ , $\pm 45^{o}$ and $90^{o}$
Type of fibre fill	Isotropic
Speed of printing	25 $m m / s$

Tensile strength analysis

The outcomes of the tensile strength analysis with the relating the stress and strain responses for the proposed composite filaments and the composite filaments from the existing literature, are presented.

Figure 7 illustrates the tensile stress $σ (M P a)$ with the tensile strain $ε (%)$ with the test samples for different time durations from $\dot{ε} = 1 e^{- 02}$ to $\dot{ε} = 1 e^{- 06}$ at $90^{o}$ . Figure 7(a) shows the relation with the composite of PA12BG sample having tensile stress value of 40 $M P a$ maximum and its ultimate strength lies between 0.01% and 0.035 % of the tensile strain variations and sample is breaks at the level of 0.065% strain for the time of $\dot{ε} = 1 e^{- 03}$ . Figure 7(b) shows the relation with the composite of PA12CF-AM sample having tensile stress value of 62 $M P a$ maximum and its ultimate strength lies between 0.03% and 0.035 % of the tensile strain variations and sample is breaks at the level of 0.06% strain for the duration of $\dot{ε} = 1 e^{- 06}$ . Figure 7(c) shows the relation for the composite of proposed PA12BF-RGB sample having the tensile stress of 80 $M P a$ maximum and its ultimate strength lies between 0.03% and 0.063% of the tensile strain variations and sample breaks at the level of 0.063% strain for the time of $\dot{ε} = 1 e^{- 06}$ .

Figure 7.

Tensile stress strain performance curve for different time duration at $90^{o}$ (a) PA12BG (b)PA12CF-AM (c) PA12BF-RGB.

Failure at the interfaces between the layers in 3D printed materials would indicate weak bonding between the layers. Since the PA12BF-RGB composite shows a higher tensile strength and better performance, it suggests that the layer-to-layer bonding in our method is stronger, indicating better overall material integrity compared to the other composites.

The stress-strain curve shown in Figure 8 has some irregularities and the load is then transferred among the fibers till complete failure.

Figure 8.

Stress-strain curve for fabrication orientation with 45°. (a) PA12BG (b)PA12CF-AM (c) PA12BF-RGB.

Similar to the problem Saeed et al.¹¹ observed, failure results from stress concentration at the radiused corners, which are between the neck and the flange area. When fibers are orientated from $0^{o}$ to ±45°, the composite material softens and deforms more easily, resulting in increased elongation. In contrast, specimens with a $90^{o}$ fibre orientation have lower stress concentrations due to elongation. Fibre fracture, fiber pull-out, and delamination all contributed to failure in most of the test samples.

Infill density and nozzle temperature

During the process of material extrusion 3D printing, the infill density is maintained at 100% and the orientation angle of the printing process have an effect in the nozzle temperature and this performance relation is presented in Figure 9. Three orientation angles are compared with the proposed PA12BF-RGB composite filament.

Figure 9.

Relation curve for infill density and nozzle temperature.

It is proven that, the proposed composite material with the $90^{o}$ angle impacts greatly to the density of the composite filament. The filament density is decreased from 2.1 ${g / c m}^{2}$ to 1.78 ${g / c m}^{2}$ for temperature at the extrusion nozzle from 210℃ to 223℃. After that the density values are averagely in the range of 1.75 ${g / c m}^{2}$ until the maximum limit of the suggested nozzle temperature. The orientations like $0^{o}, \pm 45^{o}$ has less impact when compared with the orientation angle $90^{o}$ .

Layer height and nozzle temperature

During the process of material extrusion 3D printing, the with varying layer height levels impact with different nozzle temperature for the orientation angles considered for this mechanical property analysis with pure PA12 and PA12BF-RGB composite filament performance relation is illustrated in Figure 10. Figure 10(a) depicts the relation curve with $0^{o}$ angle, and the performance curves with the angle like $90^{o}$ , $\pm 45^{o}$ are respectively depicted in Figure 10(b) and (c).

Figure 10.

Relation curve for layer height and nozzle temperature (a) $0^{o}$ (b) $90^{o}$ (c) $\pm 45^{o}$

It is proven that, the proposed composite material with the $90^{o}$ angle impacts greatly with change in layer height of the composite filament. The orientations like $0^{o}, \pm 45^{o}$ has less impact when compared with the orientation angle $90^{o}$ . Overall, the proposed composite filament of PA12BF-RGB performance with varying layer height and nozzle temperature outperforms pure PA12 filament.

Post-processing heat treatment description

After the material extrusion 3D printing process, geometric defects occur that reduce the robustness of the end product. Post-processing is required to boost its tensile strength, despite the fact that errors are inevitable owing to the use of modern production procedures.³² By using post-heat treatment processes, it is evident that considerable improvements, like the capacity of load carrying and internal stresses, are enhanced.

During the post-heat treatment, the test specimens are heated in a hot air oven with a capacity of heating up to 400°C. For the safety of test specimens, find sand is covered around the oven trays before placing them inside the heating oven. Three test specimens of PA12BF-RGB composite filament are heated for half an hour with a constant temperature of 165°C. After this, the specimens are rotated and heated again for an additional half an hour to ensure the heat exposure is fully on the surface of the test specimens. For this time, the temperature is constantly regulated at 183°C. Finally, the test specimens are taken out from the oven and kept in the surrounding environment for cooling. Mechanical testing and microscopic inspection are used to evaluate material extrusion 3D printing specimens after heat treatment. The outcomes with enhanced mechanical properties like tensile strength, surface roughness, specific wear rate, elongation at break before (Samples Before (SB)) and after heat treatment (Samples After (SA)) is listed in Table 4 and illustrated in Figure 14 (a)–(d).

Table 4.

Percentage improvement of the mechanical properties with post heat treatment.

	Before heat treatment			After heat treatment			Overall improvement
Test specimens	SB-1	SB-2	SB-3	SA-1	SA-2	SA-3	(%)
Tensile strength	80	81.5	82	87	88.5	90	9–10
Surface roughness	2.78	2.18	3	2.07	1.98	2.5	19
Specific wear rate	20.5	25.3	20.5	15.5	15.5	15.5	35
Elongation at break (SB-SA)	40 (SB1-SA1)	45 (SB2-SA2)	42 (SB3-SA3)	—	—	—	42

The tensile strength value of the proposed PA12BF-RGB composite is 80 $M P a$ . Following the process of heat treatment, the value is increased to 87 $M P a$ averagely with nearly 9-10 % improvement as illustrated in Figure 11(a). The surface roughness value of the proposed composite is decreased from 2.78 $μ m$ to 2.07 $μ m$ averagely as depicted in Figure 11(b). From the experimental specific wear rate analysis test similar to the setup in Ref. 8, it is evident that there is a reduction in the range from $25 \times 10^{- 5} k g / N m$ (maximum) to $15.5 \times 10^{- 5} k g / N m$ (minimum) averagely as illustrated in Figure 11(c), This reduction in wear rate is ascribed to the enhanced layer-to-layer bonding after the heat treatment process, improving the composite’s resistance to surface degradation. Additionally, Figure 11(d) shows the results of the repeated creep tests as well as the variation in material elongation before and after the heat treatment. In the load range of 1–10 $K N$ , three material extrusion 3D -printed tensile specimens demonstrated remarkable elongation, with a greater rate of 154.55% prior to and 112.56% following heat treatment.

Figure 11.

Mechanical Performance improvement analysis curves (a) tensile strength (b) surface roughness (c) specific wear rate (d) elongation at break.

It has been discovered that post-heating had a substantial effect on each of the material extrusion 3D printed -produced test samples. It is presented that the suggested composite filament of PA12BF-RGB had a significant impact on material extrusion 3D printed components and that heat treatment significantly enhanced its mechanical characteristics. Among the processing parameters for PA12BF-RGB, the 0° orientation around 270°C nozzle temperature, 85°C bed temperature and infill density 100% at a layer height of 3.287 $m m$ , exhibits superior performance in terms of layer bonding, optimal filament density, and stable extrusion, making it the most favorable for the composites under material extrusion 3D printing conditions. Henceforth, the TGA, tribological, and mechanical property evaluations are conducted using specimen fabricated under this particular optimized condition.

Thermogravimetric analysis

The outcomes of the thermogravimetric analysis of composites are presented in Figure 12. When compared to PA12BG and PA12CF-AM, the presence of BF-RGB offers thermal enhancement with proposed PA12BF-RGB. The weight loss (%) versus temperature (℃) of all evaluated composite materials is compared. TGA results confirmed the filler ratio in the nanocomposite material. The remaining material after the sharp weight loss is approximately equal to the filler’s weight ratio in each composite material. This is a qualitative estimate demonstrating a well-dispersed BF-RGB in the PA12. For the temperature 500℃, the weight loss percentage is increased to 40% when compared to PA12CF-AM and 60% with PA12BG. It shows that the material degradation starts from 423℃, but the injection molding temperature is 280℃ as suggested by datasheet of the material. This indicated the increased thermal stability rate with the proposed composite filament of PA12BF-RGB.

Figure 12.

TGA curve with composite filaments.

The performance comparison of the composite filament materials like PA12BG, PA12CF-AM and proposed PA12BF-RGB with its mechanical properties like tensile strength, flexural strength, tensile modulus of elasticity, fracture toughness, micro–Hardness Vickers (HV), flexural modulus of elasticity values are related and presented with spider curve is illustrated in Figure 13.

Figure 13.

Overall mechanical property analysis.

The mechanical property parameters like tensile strength of the proposed PA12BF-RGB when compared to PA12CF-AM composite filament has increased up to 80 $M P a$ with 20% improvement, fracture toughness to 25 $M P a$ with 15% improvement and flexural modulus of elasticity to 1115 $M P a$ with 12% improvement.

Table 5 summarises the mechanical property comparison between PA12BG, PA12CF-AM and the proposed PA12BF-RGB composite. The results clearly indicate that the proposed material exhibits superior performance across key parameters, particularly in flexural modulus, tensile modulus, tensile strength and fracture toughness.

Table 5.

Comparative mechanical properties of PA12BG, PA12CF-AM, and the proposed PA12BF-RGB composite.

Mechanical properties	PA12BG	PA12CF-AM	PA12BF-RGB (Proposed)
Flexural modulus of elasticity (MPa)	920	1020	1120
Flexural strength (MPa)	9	12	14
Tensile modulus of elasticity (MPa)	35	40	60
Tensile strength (MPa)	85	80	95
Micro hardness (HV)	15	17	18
Fracture toughness (kJ/m²)	18	19	24

Tribological properties

The friction coefficient of PA12BG, PA12CF-AM and the proposed PA12BF-RGB composite is displayed in Figure 14. The friction process is divided into two phases. In the intial phase, the friction coefficient of the materials increases considerably for about 120 s. The friction coefficient of the proposed PA12BF-RGB composite is lower than the others. Then, the friction coefficients enter into the second phase and display a mild increase to a steady state. It is monitored that the friction coefficient of PA12BF-RGB is less than that of PA12BG and PA12CF-AM around 10%.

Figure 14.

Evolution of the friction coefficient.

The wear losses and wear rates of the PA12BG, PA12CF-AM and the proposed PA12BF-RGB composites are displayed in Figure 15. The wear loss of the PA12BG and PA12CF-AM are 1.92 mg and 1.73 respectively. It is shown that the wear loss of PA12BF-RGB composite is 0.78 which is about half that of PA12BG. The corresponding wear rate of the materials are 0.23 $\times 10^{- 4} m m^{3} / (N m)$ , 0.20 $\times 10^{- 4} m m^{3} / (N m)$ and 0.17 $\times 10^{- 4} m m^{3} / (N m)$ respectively. PA12BF-RGB composites revealed a minimum wear loss and specific wear rate and the wear loss is decreased by 53%, and the specific wear rate is decreased about 20% than the others. The wear resistant of PA12BF-RGB is enhanced with the addition of boron fibers and glass beads.

Figure 15.

Wear loss and specific wear rate.

Microstructural analysis

Scanning Electron Microscope (SEM) observations of the fracture surface of the PA12BF-RGB material is displayed in Figure 16. The cross section of the composite is fully dense, and only a few pores exist. It indicated that the addition of boron filament and glass beads reduced the size and number of pores and increased the densification of the printed specimen. Few holes in the fracture surfaces are also be seen in the composite represented small amount of fibre pullouts. Meanwhile, the fracture surface morphology demonstrate the ductile-brittle transition and the smooth surfaces suggested relatively weak interfacial adhesion between the boron fibre and the polymer matrix.

Figure 16.

Fracture surface of the PA12BF-RGB material.

Validation experimenets

Inorder to verify the predicted output parameters as material extrusion 3D printed proposed sample (PA12BF-RGB) is again fabricated using these parameter settings and all mechanical and surface charecteristics are carried out to validate the obtained results, as depicted in Table 6. The outcomes valudated ESNN-HTDA-CLO module as the validations are closer to experimental values thereby achieving optimum values.

Table 6.

Obtained results for validation experiments.

Responses (Output parameter)	Experimental	ESNN-HTDA-CLO	Error (%)
Flexural modulus of elasticity (MPa)	1120	1149	2.59
Flexural strength (MPa)	14	15	7.14
Tensile modulus of elasticity (MPa)	60	63	5.00
Tensile strength (MPa)	95	97	2.11
Micro hardness (HV)	18	19	5.56
Fracture toughness (kJ/m²)	24	25	4.17

Performance comparison of hybrid optimisation algorithm

The performance of different hybrid optimization models was evaluated using three error indicators: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). Since lower values in these metrics represent higher prediction accuracy, the proposed ESNN-HTDA-CLO model achieved the best performance among all compared methods. It recorded the lowest RMSE, MAE, and MAPE, demonstrating superior error reduction capability. In contrast, conventional hybrid models such as ANN-GA and RSM-GA showed higher error levels, indicating comparatively weaker predictive efficiency. Overall, the results validate that the proposed model provides more accurate and reliable predictions than existing optimization-based neural network approaches as shown in Table 7.

Table 7.

Evaluation of hybrid optimisation models.

Method	Hybrid components	RMSE	MAE	MAPE (%)
ESNN-HTDA-CLO (Proposed)	Spiking neural network + hybrid top-down attention + clouded leopard optimization	0.821	0.795	0.89
ANN-PSO	Artificial neural network - particle swarm optimization.	0.921	0.885	0.95
DNN-HHO	Deep neural network + Harris Hawk optimization	0.945	0.898	0.97
RSM-GA	Response surface methodology-genetic algorithm	0.965	0.935	0.98
ANN-GA	Artificial neural network-genetic algorithm	0.98144	0.83984	0.96918

Conclusion

The proposed PA12BF-RGB composite material is compared to conventional PA12, BubbleGlass™ (PA12BG), NanoCarbon AM™ (PA12CF-AM), and conventional PA12 filament to upgrade mechanical properties such as mechanical strength and stability during the material extrusion printing process for customised 3D design production. Investigation into the mechanica1 process variables, including printing orientation, infill density, layer height, and nozzle temperature, is carried out to optimize boron fibre-glass-bead polymeric composites.

The research has shown that the orientation of a test specimen is greatly influence the tensile strength and Young modulus.

The 0° orientation exhibits superior performance in terms of density stability, layer bonding, and overall mechanical response

Compared with PA12CF-AM, the PA12BF-RGB loses 40% of its weight at 500°C after heat treatment and that of 42% elongation at break.

The tensile strength increased by around 9–10% and reached 87 $M P a$ after the heat treatment, whereas the average surface roughness increased by 19% and the value is 2.07 $μ m$ averagely.

After post-heat treatment, a 35% decrease in specific wear rate with the value $15.5 \times 10^{- 5} k g / N m$ is achieved.

The optimal predicted process parameters values of temperature in the nozzle and bed temperature, respectively 270°C, 85°C and infill density 100% at a layer height of 3.287 $m m$ .

The outcomes validated the ESNN-HTDA-CLO module, as the predicted results closely matched the experimental values, thereby achieving optimal validations.

Moreover, the printing challenges are successfully overcome by by ensuring homogeneous fibre and filler dispersion in the filament, which reduced nozzle clogging and inconsistent extrusion. Further investigation is suggested to optimize the 3D-printing process with respect to shape accuracy, quality, time, and cost parameters due to the composite material PA12BF-RGB being highlighted with strong mechanical and thermal stability and enhanced performance, which has large applications in several fields, including aerospace. Future research is expected to investigate advanced techniques of post-processing to improve the surface finish and dimensional accuracy of printed parts.

Footnotes

Author contributions

All the authors have contributed equally to the work.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Bollu Satyanarayana

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

References

Pejkowski

Seyda

Nowicki

, et al. Mechanical performance of non-reinforced, carbon fiber reinforced and glass bubbles reinforced 3D printed PA12 polyamide. Polym Test 2023; 118: 107891. https://doi.org/10.1016/j.polymertesting.2022.107891

Bayas

Kumar

Harne

. Impact of process parameters on mechanical properties of FDM 3D-printed parts: a comprehensive review. Eur Chem Bul 2023; 12: 708–725.

Bragaglia

Cecchini

Paleari

, et al. Modeling the fracture behavior of 3D-printed PLA as a laminate composite: influence of printing parameters on failure and mechanical properties. Compos Struct 2023; 322: 117379. https://doi.org/10.1016/j.compstruct.2023.117379

Sani

Zolfagharian

Kouzani

. Artificial Intelligence‐augmented additive manufacturing: insights on closed‐loop 3D printing. Adv Intell Syst 2024; 29(10): 6. https://doi.org/10.1002/aisy.202400102

Hosseini

Vaghefi

Mirkoohi

. The role of defect structure and residual stress on fatigue failure mechanisms of ti-6al-4v manufactured via laser powder bed fusion: effect of process parameters and geometrical factors. J Manuf Process 2023; 102: 549–563. https://doi.org/10.1016/j.jmapro.2023.07.014

Guo

, et al. Additive manufacturing of ni-based superalloys: residual stress, mechanisms of crack formation and strategies for crack inhibition. Nano Mater Sci 2023; 5(1): 53–77. https://doi.org/10.1016/j.nanoms.2022.08.001

Zhai

Wang

Chang

, et al. Robust optimization of 3D printing process parameters considering process stability and production efficiency. Addit Manuf 2023; 71: 103588. https://doi.org/10.1016/j.addma.2023.103588

Wang

Lin

, et al. Mechanical and tribological properties of FDM-printed polyamide. Tribol Int 2024; 191: 109198. https://doi.org/10.1016/j.triboint.2023.109198

Zehir

Seyedzavvar

Boğa

. Exploring mixed-mode fracture behavior and mechanical properties of selective laser sintered polyamide 12 components. Rapid Prototyp J 2024; 30(3): 529–546. https://doi.org/10.1108/rpj-08-2023-0270

10.

Chen

Teo

, et al. Predicting crystallinity of polyamide 12 in multi jet fusion process. J Manuf Process 2023; 99: 1–11. https://doi.org/10.1016/j.jmapro.2023.05.043

11.

Saeed

McIlhagger

Harkin-Jones

, et al. Predication of the in-plane mechanical properties of continuous carbon fibre reinforced 3D printed polymer composites using classical laminated-plate theory. Compos Struct 2021; 259: 113226. https://doi.org/10.1016/j.compstruct.2020.113226

12.

Rath

Kumar

Duvedi

, et al. Characterization and optimization for mechanical and wear performance of reinforced nylon-12/carbon fiber matrix. J Reinforc Plast Compos 2024: 07316844241307079. https://doi.org/10.1177/07316844241307079

13.

Beylergil

Al‐Nadhari

Yildiz

. Optimization of charpy‐impact strength of 3D‐printed carbon fiber/polyamide composites by Taguchi method. Polym Compos 2023; 44(5): 2846–2859. https://doi.org/10.1002/pc.27285

14.

Sandeep Varma

Lal Meena

Chekuri

RBR

. Optimizing mechanical properties of 3D-printed aramid fiber-reinforced polyethylene terephthalate glycol composite: a systematic approach using BPNN and ANOVA. Eng Sci Technol Int J 2024; 56: 101785. https://doi.org/10.1016/j.jestch.2024.101785

15.

Hossain

Chowdhury

Mahamud

, et al. Electro-mechanical analysis of nanostructured polymer matrix composite materials for 3D printing using machine learning. Chem Eng J Adv 2024; 19: 100626. https://doi.org/10.1016/j.ceja.2024.100626

16.

Gómez-Ortega

Piedra

Mondragón-Rodríguez

, et al. Dependence of the mechanical properties of nylon-carbon fiber composite on the FDM printing parameters. Compos. Part A Appl. Sci. Manuf 2024; 186: 108419. https://doi.org/10.1016/j.compositesa.2024.108419

17.

Ferdousi

Advincula

Sokolov

, et al. Investigation of 3D printed lightweight hybrid composites via theoretical modeling and machine learning. Compos B Eng 2023; 265: 110958. https://doi.org/10.1016/j.compositesb.2023.110958

18.

Al Rashid

Koç

. Experimental validation of numerical model for thermomechanical performance of material extrusion additive manufacturing process: effect of infill design & density. Results Eng 2023; 17: 100860. https://doi.org/10.1016/j.rineng.2022.100860

19.

Udu

Osa-uwagboe

Adeniran

, et al. A machine learning approach to characterise fabrication porosity effects on the mechanical properties of additively manufactured thermoplastic composites. J Reinforc Plast Compos 2024; 4: 1060–1094. https://doi.org/10.1177/07316844241236696

20.

Kumar

. Material extrusion‐based additive manufacturing of sandwiched acrylonitrile butadiene styrene/glass fibers composites: machine learning approach to model tensile stress. Polym Compos 2024; 45(16): 15100–15112. https://doi.org/10.1002/pc.28823

21.

Ali

Nouzil

Mehra

, et al. Integrated optimization scheme for 3D printing of PLA-apha biodegradable blends. Progr Addit Manuf 2024; 10(1): 875–886. https://doi.org/10.1007/s40964-024-00684-z

22.

Ghabezi

Sam-Daliri

Flanagan

, et al. Mechanical and microstructural analysis of glass fibre-reinforced high density polyethylene thermoplastic waste composites manufactured by material extrusion 3D printing technology. Compos. Part A Appl. Sci. Manuf 2025; 194: 108930.

23.

Zhang

Wang

Zhou

, et al. In situ preparation and 3D printing of prepreg filaments for continuous carbon fiber reinforced nylon composites: principle, equipment, mechanical performance and defects. J Manuf Process 2025; 151: 1152–1168.

24.

Shanmugam

Babu

Kannan

, et al. The thermal properties of FDM printed polymeric materials: a review. Polym Degrad Stabil 2024; 228: 110902. https://doi.org/10.1016/j.polymdegradstab.2024.110902

25.

Vidakis

Petousis

Velidakis

, et al. Multi-functional polyamide 12 (PA12)/multiwall carbon nanotube 3D printed nanocomposites with enhanced mechanical and electrical properties. Adv Compos Mater 2022; 31(6): 630–654. https://doi.org/10.1080/09243046.2022.2076019

26.

Rodríguez-Reyna

Díaz-Aguilera

Acevedo-Parra

, et al. Design and optimization methodology for different 3D processed materials (PLA, ABS and carbon fiber reinforced nylon PA12) subjected to static and dynamic loads. J Mech Behav Biomed Mater 2024; 150: 106257. https://doi.org/10.1016/j.jmbbm.2023.106257

27.

Raja

Devarajan

Prakash

, et al. Sustainable innovations: mechanical and thermal stability in palm fiber-reinforced boron carbide epoxy composites. Results Eng 2024; 24: 103214. https://doi.org/10.1016/j.rineng.2024.103214

28.

Suresh Kumar

Shankar

Lalith Kumar

. Failure investigation on high velocity impact deformation of boron carbide (B4C) reinforced fiber metal laminates of titanium/glass fiber reinforced polymer. Def Technol 2022; 18(6): 963–975. https://doi.org/10.1016/j.dt.2021.04.006

29.

Tandale

Stoffel

. Physics-based self-learning spiking neural network enhanced time-integration scheme for computing viscoplastic structural finite element response. Comput Methods Appl Mech Eng 2024; 422: 116847. https://doi.org/10.1016/j.cma.2024.116847

30.

Liu

Zhao

. Enhancing spiking neural networks with hybrid top-down attention. Front Neurosci 2022; 16: 949142. https://doi.org/10.3389/fnins.2022.949142

31.

Trojovska

Dehghani

. Clouded leopard optimization: a new nature-inspired optimization algorithm. IEEE Access 2022; 10: 102876–102906. https://doi.org/10.1109/access.2022.3208700

32.

Balan

Raj

Adithya

. Effect of post-heat treatment on the mechanical and surface properties of nylon 12 produced via material extrusion and selective laser sintering processes. Polym Bull 2024; 81(11): 10149–10174. https://doi.org/10.1007/s00289-024-05197-x

Post-heat treatment with mechanical impact predictive analysis using fibre-reinforced polyamide 12 composites in 3D printing

Abstract

Keywords

Introduction

Research gap

Novelty of the proposed research

Proposed mechanical performance analysis description

Materials and processing

Fabrication of test samples

Test sample printing description

Testing parameters

Measurement standards description

Effect of annealing on PA12BF-RGB composites

Formulation of optimal mechanical property prediction

Update mechanism of CLO

Exploration update – phase 1

Exploitation update – phase 2

Outcomes and discussions

Tensile strength analysis

Infill density and nozzle temperature

Layer height and nozzle temperature

Post-processing heat treatment description

Thermogravimetric analysis

Tribological properties

Microstructural analysis

Validation experimenets

Performance comparison of hybrid optimisation algorithm

Conclusion

Footnotes

Author contributions

Declaration of conflicting interests

Funding

ORCID iD

Data Availability Statement

References