Моделювання росту втомної тріщини в сплаві д16т методом скінченних елементів

Abstract

Створено модель плоского зразка з центральною тріщиною в пружно-пластичній постановці за допомогою методу скінченних елементів у програмному комплексі ANSYS. Обґрунтовано довжину проростання тріщини для усталення пластичної зони у вістрі тріщини. Змодельовано ріст втомної тріщини за регулярного циклічного навантаження з урахуванням залишкових деформацій та напружень в околі вістря тріщини та контакту її берегів. Розраховано швидкість росту втомної тріщини з використанням ефективного коефіцієнта інтенсивності напружень, отриманого при моделюванні розкриття тріщини. Отримано задовільне узгодження результатів моделювання та експериментальних даних.

The aim of this study was to simulate fatigue crack growth using ANSYS software, determine crack growth rate and compare it with experimental data. In order to simulate the fatigue crack propagation along the direction of its growth, contact conditions were applied on two separate surfaces – cracks faces. This made it possible to simulate the processes that occur at the crack closure due to the formation of residual plastic deformations, contact and compressive stresses in the crack tip. In order to simulate crack propagation, the rate of crack growth was determined by the Paris formula. Crack growth rate was determined after application of each cycle of regular loading and summarized using of the «cycle by cycle» technique. When the overall crack growth reached finite element size, the limitation of displacements in Y axis were removed from a pair of nodes that were at one point and modeled crack tip at that time. The minimal crack growth length which is required for stabilization of plastic zone, stress-strain state and crack tip opening level was determined. For realistic simulation of fatigue crack growth the crack should be grown for 2 mm at least. This can be explained by the stabilization of stress-strain state in the crack tip and plastic deformations on crack faces (behind crack tip). Results of stress-strain state simulation for each load step, including plastic deformations, were considered during simulation of next load step (cycle). During crack growth on its faces residual plastic deformations (extensions) were formed. Therefore during unloading the crack faces closed up prematurely, before unload if finished. This leads to the crack closure effect. In addition, after premature crack closer, residual contact compressive stress appears on the crack faces. Thus, the created model made possible to take into account the effect of residual plastic deformation (extensions) on crack faces and consequently residual compressive stresses. The series of FEM simulation experiments of fatigue crack growth that corresponded to the experimental data at different values of maximum stress intensity factor (SIF) and SIF ratios were modeled. It was determined that values of effective SIF ranges are by 15–20% higher than obtained experimentally or by calculations using formulas. Therefore, it was proposed to introduce correction coefficient in order to eliminate these differences in simulation algorithm of effective SIF range calculation. Corrected values of calculated SIF are in satisfactory conforming to experimentally obtained results. The reasons of overestimation of simulated effective SIF ranges relatively experimentally obtained are – neglecting roughness and accumulation of corrosion products at crack faces. These factors cause increasing of the level of crack tip opening stress and decreasing of effective SIF range. Proposed finite element model takes into account only plastic deformations of material and residual compressive stresses ahead and behind the crack tip. Proposed correction factor integrally takes into account an increasing of roughness and accumulation of corrosion products on crack tip faces. To compare the simulated and experimental results, the V  Keff diagram of D16T alloy in effective Keff values was built. The diagram shows satisfactory agreement of experimental and simulated results for each experiment.