Thermomechanical processes such as hot rolling have been used in the industry for a long time to process and shape metals to a desired form with specific properties. However it can be difficult to make changes to the different process parameters. That's where it is beneficial to use a hot deformation simulator such as the Gleeble 3800-GTC. It can be used to test metals in a controlled environment where the deformation, temperature and other parameters are easily changed. When the machine uses a Hydrawedge module, it is able to simulate hot rolling using uniaxial compression at high temperatures. Swerim AB has one such machine and has requested to investigate what occurs inside a specimen during testing in the Gleeble, specifically inside two low-alloyed steels with a hardness between 400 and 500 HV. Such tests were replicated using LS-DYNA, an FE software. The goal was to acquire true stress-strain graphs that showed similar behaviour to the data from the Gleeble and plots of the effective plastic strain which could be correlated to the grain structure pattern inside the deformed cylinders. An FE-model was created which replicates the procedure. The model was verified through numerous steps. An initial mesh verification was done where the simulation time took at least 5 hours and at most 86 hours. Using a technique called mass scaling, the elements inside the model were manipulated using additional mass to increase their time step and reduce the computational time. A verification of the mass scaling was done where the computational time was weighed off against accuracy. Afterwards the friction had to be verified where it was found that the Gleeble test specimens were deformed more than necessary which was taken into account and the models were adjusted for friction verification. After all was said and done, the model had a reasonable friction coefficient with an optimal mesh and mass scaling configuration. The resulting model simulated a test of 0.5 seconds in 15 minutes and only costing at most 10 MPa in accuracy when experimental results have maximum values between 110 to 220 MPa depending on the scenario. This equals an approximate error of around 5-10%. When investigating the grain structure after 100 seconds of relaxation, the computational time amounted to 52 hours but could be reduced to 12 hours when simulating 30 seconds as there was no change in the effective plastic strain after that time. The final model has a high enough accuracy which, when combined with the Gleeble, is able to confirm material models and describe what occurs in the material during conditions akin to hot rolling.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:liu-204962 |
| Date | January 2024 |
| Creators | Tregulov, Farhad |
| Publisher | Linköpings universitet, Mekanik och hållfasthetslära |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0025 seconds