Search

Bipolar plates are among the most important components of polymer electrolyte membrane fuel cells that are responsible for a high percentage of their weights and costs. Metals are a suitable replacement for thick graphite plates that require high machining costs. The present study investigates the manufacturing process of metallic bipolar plates with a thickness of 0.1 mm using the stamping process. One of the problems with the formation of metallic bipolar plates is their rupture during plastic deformation. This study is aimed to investigate the effects of the geometrical parameters of bipolar plates, including channel width, rib width, channel depth, draft angle, and corner radius, on the formation of sheets. By designing some experiments, the effect of each parameter on the thinning percentage and filling depth of the bipolar plates were evaluated. The required data is obtained using the commercial finite element code. The results show that the channel depth, draft angle, and corner radius have the most effect on the plate thinning. In addition, the corner radius and the draft angle have the highest effect on the maximum channel depth. Finally, the thinning percentage and the filling depth for other geometric parameters can be predicted by a mathematical equation.

In this paper, the stamping process was employed to fabricate metallic bipolar plates (MBPs). An account of low formability of the commercially pure titanium (CP–Ti), the fracture is the most common defect during its plastic deformation. Consequently, prediction of the fracture onset during the stamping was studied using three ductile fracture criteria including Rice-Tracey, Brozzo, Ayada, and a developed forming limit criteria based on consideration of the material size effect. The damage value in the lateral and central channel was evaluated to determine the critical channel and element. According to the results, the most accurate fracture prediction during stamping of titanium bipolar plates could be obtained via Brozzo ductile fracture criteria with an error rate of 3.68% compared to experiments. Moreover, the strain-based criteria represent higher fracture prediction errors compared to damage criteria. The stress state analysis showed the variation of stress triaxiality during the process leading to less accuracy of the strain-based criteria. According to the results, the damage function of the ductile damage criteria was more reliable for the semi-proportional loading path during the stamping of the titanium bipolar plates which makes them more suitable for accurate fracture prediction during the process.

In this paper, the modified Mohr-Coulomb (MMC) ductile fracture criterion was adopted to analyze the fracture behavior of the AA6061-T6 aluminum alloy sheet during the U-bending process. Appropriate calibration procedures were employed using various tension tests. A finite element (FE) model was built using the commercial FE code Abaqus/Explicit incorporating the calibrated MMC fracture criterion. It was found that the MMC criterion calibrated by the uniaxial, plane strain, notched, and modified in-plane shear tension tests can predict the onset of fracture in the U-bending process with reasonable accuracy. The error of the predicted fracture displacement was about 2% as compared to experiments. Due to the dependency of the fracture prediction accuracy on the stress state, the effects of process parameters on stress triaxiality and normalized Lode angle parameter were investigated. The negligible effect of process parameters on the stress state confirmed that the calibrated fracture criterion (MMC) was able to predict fracture in different forming conditions with equal accuracy. Also, the proportional stress and strain state in the U-bending process indicate that, besides the high fracture prediction accuracy of the MMC’s damage accumulation function, the fracture forming limit diagram (under proportional loading assumption) can serve as a reliable tool for estimating the onset of fracture in the U-bending process. Therefore, a series of studies were also conducted on damage evolution, crack propagation, principal strains, and fracture displacement during the U-bending process under various forming conditions.

In this study, the dimensional accuracy of metallic bipolar plates with stainless steel 316 was investigated in diagonal, longitudinal, and transverse direction (directional dimensional accuracy). The results showed that directional dimensional accuracy (uniformity of channel depth) would increase by raising the applied force. In addition, increasing the rubber hardness and rubber thickness leads to higher directional dimensional accuracy. After that, the dimensional accuracy in constant direction (total dimensional accuracy) was investigated. According to the result, central channels have lower channel’s depth than lateral channels. Difference between central and lateral channel’s depth was decreased when the applied force was increased. Also, total dimensional accuracy increased by increasing hardness and thickness of rubber. Most accurate bipolar plate is fabricated by 450 kN applied force and rubbers with hardness of Shore A 90 and 30 mm thickness. However, dimensional accuracy in conventional rubber forming process was not satisfying. Thus, a new method was used in order to improve the dimensional accuracy of fabricated bipolar plates named semi-stamp rubber forming. The results indicated that using semi-stamp rubber forming instead of conventional rubber forming would lead to 8.35, 3.72, and 3.3% improvement in directional dimensional accuracy and also 1.075% improvement in total dimensional accuracy.

Metal is considered to be a good material for a bipolar plate due to its good electrical conductivity, excellent mechanical properties, and low cost. In this study, fabrication of metallic bipolar plates with serpentine flow field microchannels was investigated by a novel rubber pad forming process. Stainless steel 316 with thickness of 0.1 mm was used. Polyurethane rubbers with the hardness of shore A40, A55, A65, and A90 and the thickness of 10, 20, and 30 mm were used in this study to manufacture bipolar plates. At first, convectional rubber pad forming process was used to produce bipolar plates by some experimental tests. In this step, the effect of applied force, rubber hardness, and rubber thickness on filling percentage, thickness distribution, and dimensional accuracy was investigated. The results showed that filling percentage would increase by rising applied force, but channel depth was not equal in different directions. Increases in rubber hardness led to improving uniformity in channel depth. In addition, decreases in rubber hardness and increases in rubber thickness caused higher filling volumes. However, maximum filling percentage, dimensional accuracy, and thickness distribution in convectional rubber forming process were not satisfying. Thus, a novel method was developed in this study in order to improve the quality of fabricated bipolar plates named semi-stamp rubber forming. The results indicated that using semi-stamp rubber forming instead of convectional rubber forming would lead to 11.7, 9, and 1.075% improvement in filling percentage, thinning percentage, and dimensional accuracy, respectively. According to the results, the developed model (semi-stamp rubber forming) could be a feasible technique in fabricating metallic bipolar plates.

Bipolar plates are some of the most important parts of a fuel cell that constitute a high percentage of weight and cost in a fuel cells collection. Furthermore, metals due to lesser weight and cost are suitable replacements for thick graphite plates which need high cost of machining. In recent years, among of sheet forming productions the applications of stamping process have become increased due to reduction of cost and increase production speed. The present study investigates the manufacturing process of metallic bipolar plates made of SS316L with a thickness of 0.1 mm using stamping process. In this respect, the effect of lubricant on improvement of fabricated plates has been examined. Also, two cases of forming process with and without lubricant on depth of filling flow channels, thickness distribution, thickness reduction and uniformity of channels have been evaluated. The results obtained from the experiments showed that effect of lubricant is so important. Moreover, by using of lubricant the thickness reduction, uniformity of channels, depth of filling flow channels and thickness distribution could be improved.

The present study evaluates the fracture behavior of the AA6061-T6 aluminum alloy during the roll forming process using ductile fracture criteria. Three criteria including Ayada, Rice-Tracey, and normalized Cockroft-Latham were utilized to develop an accurate fracture model for predicting fracture onset during the roll forming process. To this end, the fracture criteria were calibrated using uniaxial, plane-strain, and notched tension tests. The model was developed by integrating the calibrated fracture criteria into the numerical analysis (commercial finite element code Abaqus/Explicit) through an appropriate user subroutine. Based on the results, the damage weighting function and calibration procedures had a significant effect on the accuracy of the fracture prediction during the process. When the uniaxial tension test was used as the calibration test, the normalized Cockroft-Latham and Rice-Tracey ductile fracture criteria were not able to predict the fracture onset during the roll forming process and the Ayada fracture criterion could only predict the bending angle of the forming station where the fracture initiated. On the other hand, calibrating ductile fracture criteria using the plane-strain tension test led to higher accuracy in predicting fracture. The most accurate prediction was obtained using the Ayada fracture criterion calibrated by the plane-strain tension test with an error of 7.85%. Overall, the Ayada fracture criterion and plane-strain tension test can be considered as the most appropriate ductile fracture criterion and calibration method for predicting fracture in the roll forming process, respectively.

In this study, the modified Mohr-Coulomb (MMC) ductile fracture criterion was employed to investigate the ductile fracture during the roll forming process of the AA6061-T6 aluminum alloy. To this aim, the MMC fracture criterion was calibrated based on the tension tests and was integrated into the commercial finite element (FE) Abaqus/Explicit using an appropriate user subroutine. The experimental procedures were designed based on two distinct forming strategies, namely single-stage and multi-stage roll forming processes. The results indicate that the calibrated fracture model based on tension tests is capable to predict the fracture initiation in the single-stage roll forming process at an error rate of 4.47%. Meanwhile, the numerically predicted fracture initiation during the multi-stage roll forming process is inconsistent with the experiments. Hence, a nonlinear damage accumulation rule together with the additional test (multi-stage L-bending) was utilized to predict the onset of fracture throughout the multi-stage roll forming process, considering the effect of deformation mechanics and nonlinear loading path during the process. The results reveal that the fracture onset during the multi-stage roll forming process can be accurately predicted when the MMC fracture criterion with the nonlinear damage accumulation rule is used in the numerical fracture model (with an error of about 9%).

In this study, the forming of metallic bipolar plates micro-channel was investigated. Rubber pad forming process was employed to fabricate the parallel micro-channels. According to the experimental observation, the onset of failure was the main limiting factor during the deformation of micro-channels. Therefore, the fracture behavior was studied during rubber forming processes. Experimental analysis with the finite element model (FEM) simulation was adopted to investigate the fracture mechanism. Three different fracture criteria were used to predict the fracture. Critical damage values were determined through conducting the uniaxial tension test. The fracture surface was constructed using ductile fracture criteria, and the most accurate fracture prediction was obtained via normalized Cockroft-Latham fracture criterion (9.14% fracture prediction error). The effect of process parameters on stress triaxiality and normalized Lode angle parameter was analyzed by FEM simulation results. According to the results, process parameter variation did not significantly affect the stress state. Thus, the normalized Cockroft-Latham fracture criterion was employed to investigate the onset of fracture in different forming conditions and the effect of process parameters on damage evolution. Based on the findings, draft angle, corner radius, and the ratio of w/h should be set to 15°, 0.2 mm, and 1.6 in order to achieve 100% filling percentage.

In this study, three ductile fracture criteria were used to investigate the fracture in U-bending of AA6061-T6 aluminum alloy sheets. Considering the importance of the calibration procedure, three different tension tests including the uniaxial, plane strain, and notched tensions were conducted based on the stress state analysis. The Ayada, Rice-Tracey, and normalized Cockroft-Latham fracture criteria were calibrated through a hybrid experimental-numerical method to determine their critical damage values. Fracture envelopes (i.e., the locus of fracture strains) were constructed for the investigated fracture criteria based on the results of various tension tests in their calibration procedure. The calibrated fracture criteria were implemented in the commercial finite element (FE) code Abaqus/Explicit using an appropriate user subroutine. The numerical simulation of the bending process was performed and validated by relevant experiments. The accuracy of the fracture criteria was investigated by comparing fracture displacements predicted using the FE model with those measured by experiments. According to the obtained results, variations of damage function (fracture criteria) and the type of the calibration test utilized for a certain fracture criterion have significant effects on the critical damage value, the fracture envelope, and the accuracy of the fracture prediction. For all investigated criteria, the prediction is reasonable when the critical damage value was calibrated under the plane-strain tension. Thus, the plane strain tension is the best calibration method for predicting fracture in the U-bending process. The most accurate fracture prediction is obtained by normalized Cockroft-Latham fracture criterion, which was calibrated by the plane strain tension test with 2.52% error in fracture prediction. However, the prediction of fracture displacement in U-bending process is very imprecise and strongly overestimated, when the uniaxial tension was used as the calibration test. The Ayada criterion has the lowest error of 11.83% for predicting the fracture displacement when the uniaxial tension test was used as a calibration method. Also, the maximum prediction error (the lowest accuracy) was found to be 51.97% by the normalized Cockroft-Latham criterion calibrated under the uniaxial tension test. Globally, for the U-bending process, it is recommended to calibrate the critical damage values under the plane strain condition. But, when only the uniaxial tension test is available, the Ayada fracture criterion gives the best prediction.