| Popis: |
Continuing development of aluminium alloys highlights the problem of choosing adequate constitutive stress-strain relationships, especially as aluminium has no pronounced yield strength and a significant degree of nonlinearity before reaching the 0.2% proof strength. Thus, for the sake of a more realistic approximation of the stress-strain relation more complex models are required. These complex models are continuous, defined either as a function of stress or strain. Unfortunately, basic mechanical properties available to engineers are often not adequate for material behaviour modelling in nonlinear numerical analysis, which are required for intricate design problems. Moreover, studies regarding suitability of various models are scarce. Since any deviations from experimental stress-strain behaviour consequently introduces errors in simulation results, it is necessary to quantify errors, and to enable a reliable choice among available models, depending on type of modelled phenomena. To gain insight on these problems, own experimental stress- strain data of coupons from alloys EN AW-6082-T6 and EN AW-6060-T66 was utilized. Alloys were selected as being representative of a group of commonly used alloys in structural engineering (T5, T6, and T66), due to their similar shape of the stress-strain curve. Experimental curves were divided into two zones, the dominantly elastic and plastic, and relative errors between six constitutive laws and experimental data by zone and in total behaviour were established. Descriptive and dispersion statistics of determined relative errors allowed for quantitative indicators of suitability, which then allowed for conclusions regarding choice of tested constitutive relationships. |
