This paper proposes a multi-objective optimization method that integrates deep neural networks (DNN) with grey relational analysis (GRA) to optimize lay-up configurations for carbon fiber-reinforced plastic (CFRP) automotive components. Specifically, a lab-scale CFRP B-pillar structure was investigated to simultaneously maximize structural strength and failure safety. A DNN surrogate model was trained using finite element simulations of 2,000 random stacking sequences to achieve high predictive accuracy. The trained model was then used to evaluate all possible lay-up combinations to derive Pareto optimal solutions. Grey relational analysis was subsequently employed to select the final optimal configurations based on designer preferences. The selected lay-up designs demonstrated improvements in both strength and failure safety. To validate the proposed framework, laboratory-scale CFRP B-pillar was fabricated using a prepreg compression molding process and subjected to bending tests. The experimental results confirmed an error below 5% and failure trends consistent with the simulation results, thereby validating the reliability of the proposed method. The proposed DNN-GRA approach enables efficient multi-objective optimization with reduced computational effort and flexibility in reflecting different engineering priorities.

Carbon-Fiber-Reinforced Plastic (CFRP) is a typical lightweight material used in the aerospace industry. However, the automotive industry has focused on the application of composite materials in vehicle components for weight reduction. In particular, hybrid parts consisting of CFRP reinforcement and a steel outer have been investigated in many studies as a solution to satisfy weight reduction and high strength. In this paper, a steel/CFRP hybrid part was evaluated by impact analysis using several material models, such as the Johnson–Cook model, progressive damage analysis (PDA), and cohesive zone model (CZM). First, the mechanical properties of the steel were determined under different strain rates to assess collision effects. Subsequently, the material properties of the CFRP were evaluated to predict the failure of composite material in the tensile and compressive directions. In addition, the cohesive properties of adhesive film were evaluated under normal and shear modes. Finally, impact analysis using the obtained material properties was conducted to predict the behavior and strength of the steel/CFRP hybrid part under collisions, and the results were compared with the experimental results for verification.

This paper proposes a multi-objective optimization method that integrates deep neural networks (DNN) with gray relational analysis (GRA) to optimize lay-up configurations for carbon fiber-reinforced plastic (CFRP) automotive components. Specifically, a lab-scale CFRP B-pillar structure was investigated to simultaneously maximize structural strength and failure safety. A DNN surrogate model was trained using finite element simulations of 2000 random stacking sequences to achieve high predictive accuracy. The trained model was then used to evaluate all possible lay-up combinations to derive Pareto optimal solutions. Gray relational analysis was subsequently employed to select the final optimal configurations based on designer preferences. The selected lay-up designs demonstrated improvements in both strength and failure safety. To validate the proposed framework, laboratory-scale CFRP B-pillar was fabricated using a prepreg compression molding process and subjected to bending tests. The experimental results confirmed an error below 5% and failure trends consistent with the simulation results, thereby validating the reliability of the proposed method. The proposed DNN-GRA approach enables efficient multi-objective optimization with reduced computational effort and flexibility in reflecting different engineering priorities.

The increasing use of vinyl-coated metal (VCM) sheets in home appliances requires robust forming processes to prevent defects such as delamination and wrinkling, especially under elevated temperatures and humidity. This study presents a deep neural network (DNN)-based multi-objective optimization framework to determine optimal stamping parameters for VCM sheets. A delamination limit diagram (DLD) is experimentally established by combining limit dome height tests with immersion tests, defining the critical strain boundary under environmentally conditions. A finite element (FE) based dataset of four process variables was then used to train a DNN surrogate model with high predictive accuracy. Using the trained DNN model, Pareto-based optimization identifies nondominated solutions balancing delamination and wrinkling. The optimal condition was validated by FE simulation, confirming simultaneous suppression of both defects within the DLD. The proposed DNN–Pareto framework provides and efficient and reliable tool for defect prediction and optimization in VCM stamping, ensuring high surface quality and environmental durability.