Search

This study outlines an experimental and computational endeavor aimed at developing a machine learning model to estimate spring-back values utilizing the decision tree methodology. A design of experiment approach was employed to collect a dataset, and based on the experimental results, a precise model was constructed to predict spring-back values. The model considered parameters such as thickness, diameter of circle hole, distance between the center hole and flange edge, and hole spacing. Various hyper parameters, including max depth and minimum samples for split, were explored, with configurations such as (30,5), (20,8), and (10,2) being evaluated to identify the optimal model for spring-back prediction. Analysis of the results demonstrated that the decision tree models accurately estimated spring-back values in cold roll forming of pre-punched sheets based on the input parameters. The coefficient of determination in the test section for decision tree models with parameters (30,5), (20,8), and (10,2) was found to be 0.90, 0.98, and 0.96, respectively. Additionally, the percentage of absolute error in the test section for the same decision tree models was calculated as 8.84%, 6.18%, and 7.6%, respectively. [ABSTRACT FROM AUTHOR]

The multi-criteria decision-making method, along with experimental and finite-element (FE) methods, was employed to determine the optimal process parameter values for rubber pad forming of metallic bipolar plates. To achieve this aim, the rubber pad forming process was modeled by FE software, and the results were validated by conducting experimental tests. Three significant rubber pad forming parameters—punch force, hardness, and the thickness of the rubber layer—were considered as input variables. A full-factorial design involving 27 experiments was used to assess the alternatives. Filling percentage, thinning ratio, and stress were considered as distinct evaluation criteria. To solve the multi-criteria decision-making problem, a hybrid approach was used to weigh the alternatives. According to the results, the applied method was capable of selecting the optimal values of process parameters while considering different criteria simultaneously. The results demonstrated that a punch force of 50 kN, rubber hardness of 55 Shore A, and rubber thickness of 3.5 mm were the best processing conditions. For these parameters, the filling percentage, thinning ratio, and stress were 97.6%, 28.5%, and 1005 MPa, respectively. [ABSTRACT FROM AUTHOR]

In this study, experimental investigation and numerical simulation U-bending process are employed to construct a damage model based on ductile damage criteria. Also, an adaptive neural-fuzzy inference system (ANFIS) is extracted for the prediction of damage behavior based on the Cockroft-Latham and modified Mohr-Coulomb criteria calibrated using the proposed method. Appropriate calibration tests including uniaxial tension up to the in-plane shear tension are designed based on the bending deformation mechanics. In addition, the optimal parameters of the ANFIS system are obtained by the gray wolf optimization algorithm. The accuracy of the proposed model is investigated using experimental-numerical results. Results indicate that both the proposed damage model and ANFIS network are successful in predicting the damage behavior in the U-bending process. In addition, the maximum bending angle to achieve a damage index of less than 1 in the ratio of bending radius to sheet thicknesses of 0.5, 1, 2, and 3 is presented using the model presented in this study. According to the results, the minimum applicable ratio of bending radius to thickness is equal to 3 for when the sheet metal bends up to 90 degrees. [ABSTRACT FROM AUTHOR]

The present study uses lubricated and non-lubricated stamping processes to fabricate metallic bipolar plates (MBPPs) with a thickness of 0.1 mm from SS3316 with a serpentine flow field. Then, experimental tests, together with the finite element (FE) model, were utilized to define the directional thickness distribution. According to the results, using a friction factor of 0.1 during the FE analysis can provide accurate results in terms of thickness prediction under lubricated conditions. Furthermore, the largest thinning percentages in the modified conditions were found to be 27.02%, 30.40%, and 26.00% in the longitudinal, diagonal, and transverse directions, respectively, indicating that the diagonal direction is the most critical one. Finally, the effect of the lubricating condition on the wrinkling of the sample was investigated. According to the results, using the lubricated condition during the stamping process augments the wrinkling of the bipolar plate, and the wrinkling of the bipolar plate's margin was not symmetrical. [ABSTRACT FROM AUTHOR]