Stress Intensity Factor Distribution and Crack Growth in Pressure Vessel Problems by the Frozen Stress Method

摘要

This paper describes the application of a laboratory-based experimental method to three-dimensional (3D) cracked body problems in pressure vessels to determine the crack shape and stress intensity factor (SIF) distribution along the crack front when the crack shape is not known a-priori. Back in the mid-sixties, when an explosion of solutions to 3D fracture problems developed along with the digital computer, the authors decided that a laboratory-based experimental method should be developed to partially fill the gap produced by the lack of proof testing for many large structures (e.g., ships, aircraft, pressure vessels) because of the excessive costs involved. By effecting a marriage between the equations of linear elastic fracture mechanics (LEFM) with frozen stress photoelasticity, such a method has been developed. Over the years, it has been used to check many numerical solutions in a cost- effective manner and, in the process, has provided some unique results. One such result is that starter cracks may be grown above critical temperature by applying service loads before cooling the model. When the crack reaches its desired size and shape (controlled only by loads and body shape) the loads are reduced to stop crack growth and stress freezing is completed. If sufficient growth is permitted to occur, only Mode I will exist along the crack front. That is, the crack is a Class I crack, and by using data sufficiently close to the crack tip, a two-term expression for the local stresses becomes adequate for analysis. On the basis of experiments on nozzle corner cracks in reactor vessels and cracks in solid rocket grain, the authors conclude that stable crack growth in pressure vessels proceeds under pure Mode I. Moreover, cracks grown above critical temperature during the stress freezing process in photoelastic models appear to overlay those produced in steel structures by high cycle tension- tension fatigue when test scales are identical. (7 figures, 7 refs.).