Multi-Objective Optimization of Manufacturing Process in Carbon Fiber Industry Using Artificial Intelligence Techniques

Abstract

Seeking high profitability by improving energy efficiency and production quality is the prime goal of manufacturing industries. However, achieving this aim involves the realization of several conflicting objectives. In carbon fiber industry, the stabilization process is the most vital step with high energy consumption. The aim of this study is to use intelligent modeling methods in the stabilization process to maximize energy efficiency while considering better production quality, avoiding defects, and not scarifying the prediction accuracy. To this aim, a modified DOE method was used to reduce the number of required experiments. The mechanical and physical properties were then modeled based on input-output data derived from the experiments. In this way, the SVR method is used to develop a set of mathematical models for mechanical and physical properties of the fibers. The skin-core defect and energy consumption were considered as objective functions within the given range of physical and mechanical properties of fibers. The state-of-the-art NSGA-II algorithm used to find the optimum Pareto front, including non-dominated solutions among these conflicting objective functions. The results showed that by using the integrated NSGA-II and technique for order preference by similarity to ideal solution (TOPSIS), the energy efficiency of the system was improved. Moreover, the discussions showed how similar hybrid algorithms with high accuracy can be used by other industries to reduce the overall energy consumptions.

A Mini-Review and Perspective on Current Best Practice and Emerging Industry 4.0 Methods for Risk Reduction in Advanced Composites Manufacturing

“…As such, select systems have been successfully modeled using this strategy and few critical comparisons between studies can be made. Studies include modeling the production of carbon fibre [48] [49], optimization of the automated fibre placement process [50], modeling of the dynamic cure process [51], as well as the assessment of defects for quality control, namely fibre orientation [52] and delamination [53] [54].…”

A Novel Hybrid Machine Learning Algorithm for Limited and Big Data Modeling With Application in Industry 4.0

“…Support Vector Machine (SVM) is a novel supervised machine learning which can be used for both classification or regression challenges. SVM builds a separating hyperplane that maximizes the margin between the two classes [15]. SVM have high prediction accuracy, non-parametric, robust to outliers and low prediction time [5].…”

“…This model is then updated using limited data obtained from the online process. The online adaptation requires real-time algorithms while high runtime algorithms, such as evolutionary optimization methods can only be used for offline modeling [15]. To address the needs of both offline and online parts, we propose a new algorithm that builds a robust model based on the input-output data using an offline deterministic model and online updating parts.…”

“…In multi-objective optimization, both modeling and prediction errors are simultaneously considered as objectives. Using Nondominated Sorting Genetic Algorithm (NSGA)-II [15], Pareto optimum non-dominated models are obtained from the point of view of these two objective functions. An alphabetical chromosome is used to coding the structure and topology of general structure of GMDH (GS-GMDH).…”

Caracterização da etapa de estabilização do processo produtivo de fibra de carbono a partir de poliacrilonitrila utilizando algoritmos inteligentes

“…A distribuição dos erros para variável Densidade Volumétrica possui um segundo pico de menor intensidade entre 4 e 8%. Entretanto, apesar de haver um pequeno aumento na distorção dos valores simulados em comparação aos experimentais, ainda é possível considerarem-se estes valores aceitáveis para caracterizar a etapa de estabilização, visto que os erros apresentados estão na faixa de aceitação para processos similares, conforme descrito por Golkarnarenji et al (2019), Khayam et al (2017) e Golkarnarenji et al (2018).…”