<p>A finite element model has been developed to simulate the forming of a channel section profile with the roll forming method. The model has been optimized to experimental results with respect to strains at the edge of the sheet and spring back of the sides of the profile. Finite element models with a coarse mesh have been compared to models with a finer mesh. The models with to fine mesh become instable and a model with a rather coarse mesh was finally chosen.</p><p>Both the models with shell elements and the models with solid elements have been used in the simulations. The simulations with shell elements gave very good results both for the geometry shape and the strains at the edge of the sheet. The reaction forces at the tools found in the simulations was only half of the reaction forces fond in the experiments.</p><p>The simulations with the solid element model showed very good results for the reaction forces while the geometry shape of the sheet was really bad. The spring back was much larger in the simulations than in the experiments.</p><p>The shell element model was chosen because of the excessive spring back with the solid element model. The spring back of the sides of the sheet differs only a few percent between the simulation and the experiment results when using the shell element model. The reaction forces at the tools in the simulation are only half of the reaction forces measured in the experiments but the results from the simulations are linearly proportional to the results in the experiments. The model that finally was chosen describe both the spring back and the strains at the edge of the sheet very well. Like in the experiments there were no signs of wrinkles at the sheet in any of the simulations.</p>

A finite element model has been developed to simulate the forming of a channel section profile with the roll forming method. The model has been optimized to experimental results with respect to strains at the edge of the sheet and spring back of the sides of the profile. Finite element models with a coarse mesh have been compared to models with a finer mesh. The models with to fine mesh become instable and a model with a rather coarse mesh was finally chosen. Both the models with shell elements and the models with solid elements have been used in the simulations. The simulations with shell elements gave very good results both for the geometry shape and the strains at the edge of the sheet. The reaction forces at the tools found in the simulations was only half of the reaction forces fond in the experiments. The simulations with the solid element model showed very good results for the reaction forces while the geometry shape of the sheet was really bad. The spring back was much larger in the simulations than in the experiments. The shell element model was chosen because of the excessive spring back with the solid element model. The spring back of the sides of the sheet differs only a few percent between the simulation and the experiment results when using the shell element model. The reaction forces at the tools in the simulation are only half of the reaction forces measured in the experiments but the results from the simulations are linearly proportional to the results in the experiments. The model that finally was chosen describe both the spring back and the strains at the edge of the sheet very well. Like in the experiments there were no signs of wrinkles at the sheet in any of the simulations.

Ultra-high strength steels is known to be susceptible for hydrogen embrittlement at very low concentrations of hydrogen. In this thesis three methods to prevent or reduce the hydrogen embrittlement in martensitic steel, with tensile strength of 1500 MPa, were studied. First, a barrier layer of aluminium designed to prevent hydrogen to enter the steel, which were deposited by vacuum evaporation. Second, a decarburization process of the steels surface designed to mitigate the induced stresses from cutting. Last, a hydrogen relief treatment at 150°C for 11 days and 200°C for 4 days, to reduce the hydrogen concentration in the steel. The effect of the hydrogen embrittlement was analyzed by manual measurements of the elongations after a slow strain rate testing at 5*10-6 mm/s, and the time to fracture in an in-situ constant load test with a current density of 1.92 mA/cm2 in a 0.5 M Na2SO4 solution. The barrier layer showed an increase in time to fracture, but also a decrease in elongations. The decarburized steel had a small increase in the time to fracture, but not enough to make it a feasible process. The hydrogen relief treatment showed a general decrease in hydrogen concentrations, but the elongation measurements was irregular although with a tendency for improvement. The simplicity of the hydrogen relief treatment makes it an interesting process to reduce the influence of hydrogen embrittlement. However, more investigations are necessary.