Moisture Content Prediction in Polymer Composites Using Machine Learning Techniques

Das, Partha Pratim; Rabby, Monjur Morshed; Vadlamudi, Vamsee; Raihan, Rassel

doi:10.3390/polym14204403

Cited by 8 publications

(10 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The input vector is processed through the activation function Equation (15), and the resulting output is passed to subsequent nodes in the network. 42,44…”

Section: Artificial Neural Network Regressionmentioning

confidence: 99%

“…The F1 score, sensitivity, and recall values for each class were computed using the equations documented in the literature. 42 Table 4 shows the model's F1 score, precision, and recall value. It is important to calculate precision and recall because it evaluates how well the model deals with identifying and predicting true positives.…”

Section: Degree Of Cure Inspection With Machine Learningmentioning

confidence: 99%

“…C is a constant term, often referred to as the “bias” and C ≥ 0. Finally, SVM uses the decision function stated in Equation () to predict the class 42

f (x, α) = \sum_{i = 1 ∊ SV}^{n} y_{i} α_{i} K (x_{i}, x) + β .

…”

Section: Supervised Machine Learning Algorithmmentioning

confidence: 99%

“…Every input vector has been added a zeroth dimension, which is equal to 1, known as bias input. The input vector is processed through the activation function Equation (), and the resulting output is passed to subsequent nodes in the network 42,44

a = w^{T} x = \sum_{i = 0}^{n} w_{i} x_{i}

Z (bold-italicx) = h (bold-italica) = h (\sum_{i = 0}^{n} w_{i} x_{i}) .

…”

Section: Supervised Machine Learning Algorithmmentioning

confidence: 99%

See 3 more Smart Citations

Fast and accurate prediction of cure quality and mechanical performance in fiber‐reinforced polymer composite using dielectric variables and machine learning

Rabby,

Das,

Rahman

et al. 2023

Polymer Composites

Self Cite

View full text Add to dashboard Cite

Fiber‐reinforced polymer (FRP) composites are widely used in the aerospace and automotive industries due to their high strength‐to‐weight ratio. However, the manufacturing process of FRP composites is intricate, requiring precise control over multiple parameters, which can be challenging. To ensure top‐notch product quality and bolster confidence in the durability of composite materials, it requires a uniform curing evaluation technique and a predictive strength model. This study presents a rapid nondestructive quality inspection technique for composites, involving three distinct studies that employ machine learning techniques in combination with the frequency‐dependent dielectric response of materials. The first study introduces a nondestructive method for swiftly inspecting the cure state (degree of cure) of composite samples. This technique combines broadband dielectric spectroscopy with supervised machine learning algorithms, particularly support vector machines. The second study employs advanced artificial neural networks like multi‐layer perceptron to predict the tensile strength, a measure of mechanical performance, of composite materials. The results demonstrate an accurate classification of the curing state with 96.7% accuracy and a prediction of tensile strength with 87.5% accuracy. The final study explores the application of machine learning in quality monitoring of prepreg (raw materials for FRP) aging at room temperature.Highlights Integration of machine learning with frequency‐dependent dielectric measurement offers efficient method for inspecting and monitoring the curing state and mechanical performance of composite materials. Machine learning and frequency‐dependent dielectric response used to accurately assess the curing state of fiber‐reinforced polymer composites with 96.7% accuracy, which saves the time and effort and also reduces the need for destructive testing. Advanced artificial neural networks, specifically multi‐layer perceptron, employed to predict mechanical performance of composite materials with 87.5% accuracy, enhancing confidence in product reliability. Novel dielectric measurement technique integrated with machine learning algorithms enables prepreg age monitoring at room temperature.

show abstract

“…The input vector is processed through the activation function Equation (15), and the resulting output is passed to subsequent nodes in the network. 42,44…”

Section: Artificial Neural Network Regressionmentioning

confidence: 99%

Section: Degree Of Cure Inspection With Machine Learningmentioning

confidence: 99%

“…C is a constant term, often referred to as the “bias” and C ≥ 0. Finally, SVM uses the decision function stated in Equation () to predict the class 42

f (x, α) = \sum_{i = 1 ∊ SV}^{n} y_{i} α_{i} K (x_{i}, x) + β .

…”

Section: Supervised Machine Learning Algorithmmentioning

confidence: 99%

a = w^{T} x = \sum_{i = 0}^{n} w_{i} x_{i}

Z (bold-italicx) = h (bold-italica) = h (\sum_{i = 0}^{n} w_{i} x_{i}) .

…”

Section: Supervised Machine Learning Algorithmmentioning

confidence: 99%

See 2 more Smart Citations

Fast and accurate prediction of cure quality and mechanical performance in fiber‐reinforced polymer composite using dielectric variables and machine learning

Rabby,

Das,

Rahman

et al. 2023

Polymer Composites

Self Cite

View full text Add to dashboard Cite

show abstract

“…Machine learning methods have been widely utilized in various fields over the last two decades [8][9][10], including the processing and analysis of thermograms in IRT-based defect detection. Typical thermographic data analysis methods include principal component thermography (PCT) [11], sparse PCT [12], thermographic sequence reconstruction (TSR) [13], blind source separation [14], autoencoder [15], and convolutional neural networks [16].…”

Section: Introductionmentioning

confidence: 99%

Data-Augmented Manifold Learning Thermography for Defect Detection and Evaluation of Polymer Composites

Liu

Wang

et al. 2022

Polymers

View full text Add to dashboard Cite

Infrared thermography techniques with thermographic data analysis have been widely applied to non-destructive tests and evaluations of subsurface defects in practical composite materials. However, the performance of these methods is still restricted by limited informative images and difficulties in feature extraction caused by inhomogeneous backgrounds and noise. In this work, a novel generative manifold learning thermography (GMLT) is proposed for defect detection and the evaluation of composites. Specifically, the spectral normalized generative adversarial networks serve as an image augmentation strategy to learn the thermal image distribution, thereby generating virtual images to enrich the dataset. Subsequently, the manifold learning method is employed for the unsupervised dimensionality reduction in all images. Finally, the partial least squares regression is presented to extract the explicit mapping of manifold learning for defect visualization. Moreover, probability density maps and quantitative metrics are proposed to evaluate and explain the obtained defect detection performance. Experimental results on carbon fiber-reinforced polymers demonstrate the superiorities of GMLT, compared with other methods.

show abstract

Dielectric characterization of fiber‐ and nanofiller‐reinforced polymeric materials

Quader,

Narayanan,

Caldona

2024

J of Applied Polymer Sci

View full text Add to dashboard Cite

This review provides an in‐depth analysis of the electrical responses and relaxation behaviors of various polymeric materials including fiber‐reinforced polymer composites, polymer nanocomposites (PNC), and nanofiller‐reinforced polymeric films and coatings using dielectric spectroscopy (DS). The underlying theory of DS, corresponding instrumentation, various dielectric parameters, relevant equations, interpretations, and polarization mechanisms are thoroughly explored. The impact of synthetic and natural fibers and a diverse range of nanofillers on the dielectric responses and relaxation mechanisms of fiber‐reinforced composites, and PNC and films, respectively, are also examined. For each summarized article, we provide an overview of the composite or film preparation, followed by concise explanations of the dielectric behaviors under diverse conditions, including variations in temperature, frequency, moisture levels, fiber/filler ratios, and chemical treatment. In essence, this review contributes to a nuanced understanding of the intricate relationship between material structure, properties, and performance via DS data interpretation across a spectrum of scenarios.

show abstract

Moisture Content Prediction in Polymer Composites Using Machine Learning Techniques

Cited by 8 publications

References 43 publications

Fast and accurate prediction of cure quality and mechanical performance in fiber‐reinforced polymer composite using dielectric variables and machine learning

Fast and accurate prediction of cure quality and mechanical performance in fiber‐reinforced polymer composite using dielectric variables and machine learning

Data-Augmented Manifold Learning Thermography for Defect Detection and Evaluation of Polymer Composites

Dielectric characterization of fiber‐ and nanofiller‐reinforced polymeric materials

Contact Info

Product

Resources

About