Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 Rudy i Metale Nieżelazne

1 2008

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: współczynnik sprężynowania

Sortuj według:

Ogranicz wyniki do:

Zastosowanie sztucznych sieci neuronowych do wyznaczania podstawowych parametrów procesu gięcia rur o przekroju kołowym

Litwin P., Stachowicz F.

Rudy i Metale Nieżelazne

2008

R. 53, nr 11

722-726

Sprężynowanie jest istotnym zjawiskiem towarzyszącym procesowi gięcia. Wpływa ono na końcowy kształt giętych elementów, a wielkość powrotnych odkształceń sprężystych jest trudna do oszacowania w przypadku znacznej deformacji przekroju poprzecznego giętych elementów. W artykule dokonano analizy podstawowych parametrów gięcia rur o przekroju kołowym - momentu gnącego i współczynnika sprężynowania. Przeprowadzono pomiary doświadczalne oraz obliczenia analityczne wymienionych parametrów. Pokazano zastosowanie sieci neuronowych do wyznaczania momentu gnącego i współczynnika sprężynowania.

Under bending relatively thin profiles will undergo cross-sectional distortion. It is not observed in case of beams with solid cross-section (more precisely, the distortion of a solid cross-sectional beam can be completely ignored). This characteristic in conjunction with the properties of the profiles material leads to a typical hardening-softening characteristic in the relation between the centerline curvature of the profile and applied bending moment. There is a critical value of curvature K/er. When K < K/er, the strain hardening of the material dominates the behaviour of profile and the bending moment increases with increasing curvature. When K > K/er, however, the reduction in rigidity caused by distortion of the profile cross-section overwhelms the effect of strain hardening and results in decreasing bending moment as the curvature increases. In the case of the bending of pipes with circular cross-section, the parameter describing the degree of cross-sectional distortion is the degree of ovalization. Analytical determination of the main bending parameters i.e. bending moment and springback coefficient is rather complicated because of above mentioned cross-sectional distortion and could be performed under several assumptions. The evaluation of elastic springback effect is a fundamental aspect in practice of profile forming operations. Springback takes place in a forming operation after removing the forming tools and introduces deviations from the desired final shape and consequently, the stamped profile does not conform the design specifications and could result unsuitable for the application. Since almost all forming processes are characterized by a significant amount of deformation introduced by a bending mechanics, the distribution of strain along profile cross-section is strongly inhomogeneous. Such a distribution, together with elastic-plastic behaviour of the workpiece determinates the occurrence of springback after the removal of the forming tools. It is well known from the tensile test that the elastic part of the total strain, which is recovered if the load is released, is equal to the ratio of the stress before unloading to the Young modulus. Thus the tendency to elastic springback increases at increasing the strain hardening coefficient and decreasing the elastic stiffness. This means that the cross-sectional distortion of a profile affected the tendency of elastic springback. The main goal of the work presented in this paper was to determine the relation between both the bending moment and springback coefficient as a function of bending curvature. For this reason another procedure was applied - the artificial neural network (ANN) method. Multi-layer Perceptron (MLP) neural networks were trained using measured process data of profile bending. The MLP had profile parameters as input and bending moment as well as spring-back coefficient as output. It was confirmed that this system is a valid alternative for the quick responsible method of main bending parameters determination.