Tasuku Shimizu

Residual stresses in girth butt welded pipes and treatments to modify these

Abstract A number of Type 304 stainless steel pipes are used in the primary cooling systems of nuclear plants. Intergranular stress corrosion cracks (IGSCC) were found at some welded joints in these piping systems due to very high tensile residual stress, sensitization of the material due to welding, and corrosive environment, all occurring simultaneously. Investigations have shown that at least one of the above factors must be eliminated to prevent IGSCC. This report describes experimental results on the temperature variations during pipe welding by conventional techniques and by the heat sink welding (HSW) technique. The mechanism of residual stress generation due to welding is also discussed. The pipe used in these experiments was 4B Sch80 Type 304 stainless steel. It was found that the temperature distribution through the thickness of the pipes was almost uniform for the conventional welding technique, but had a very sharp gradient for HSW. In the pipe axial direction, the temperatures varied sharply for both welding techniques. This implies that the sensitization of metal due to HSW is lighter than that of conventional welding and that the residual stress on the inside surface of the heat sink welded pipe is compressive. The induction heating stress improvement (IHSI) method has been investigated analytically and experimentally. In the IHSI method, a pipe is heated with an induction coil while cold water is pumped through it. This causes a temperature gradient throughout the pipe wall which generates high thermal stresses. This, in turn, generates compressive stresses on the inner surface of the pipe. This method is designed to eliminate tensile residual stresses near the weld heat affected zone on the inner surface. Temperature analysis and subsequent thermoelastic-plastic analysis show that tensile weld residual stresses at a joint were changed into compressive stresses on the inner surface of a pipe. It was confirmed experimentally that these stresses suppressed fatigue crack propagation in the heat affected zone (HAZ) of a welded pipe. Therefore, the IHSI method is effective not only in preventing crack initiation but also in suppressing crack propagation. As for the relaxation of residual stresses, no significant relaxation was measured when external loads were applied at as much as 80% of the yield strength in the experiments.

A two-dimensional thermal oxidation simulator using visco-elastic stress analysis

A two-dimensional thermal oxidation process simulation program, OXSIM2D, has been developed, taking into account viscoelastic material properties. Novel models for oxidation, stress dependency, and the white ribbon effect are introduced. The proposed approach can be used to analyze SiO/sub x/ growth on Si surfaces and the change in stress in the total structure, including the Si substrate. Simulation results showed good agreement with experiments for both a LOCOS and a shallow trench structure.<<ETX>>

Leakage and breakage estimation based on a net-stress approach for stainless steel pipes with circumferential cracks

Abstract Coolant leakage and breakage external loads for stainless steel pipes with circumferentially part-through cracks are predicted by the mechanical properties of pipe materials. In addition, the predicted loads at leaks and breaks are compared with other net-stress approaches. The plastic limit load criterion, a traditional and convenient method, is found to give an overly conservative estimation for shallow cracks and to be nonconservative for 60–80% wall-thickness cracks.

Consideration of allowable flaw size for pressurized pipes

Abstract Failure strength for circumferentially and axially cracked Type 304 stainless steel pipes is evaluated based on the net-stress approach. Using this approach, allowable flaw sizes for pipes containing circumferential or axial cracks are considered from the leakage and breakage viewpoint.

Elasto-plastic analysis of pipe with internal circumferential crack in bending

Abstract An elasto-plastic analysis program using the finite element method and a J- integral post-processor are developed. Program accuracy is examined using test problems of a single-edge notched specimen and a plate with a part-through wall crack. This program solves the Battelle/NRC finite element analysis round-robin problem of elasto-plastic analysis of a pipe with an internal circumferential crack in bending. Also, through this problem, the effect of the finite element mesh on the J- integral values is investigated.

Fracture margin of pipe with detectable crack by leakage

Abstract This paper describes a theoretical method for calculating a detectable crack size by leak detection systems in BWR plants. Crack opening areas for carbon steel pipes of various diameters containing circumferential through-wall cracks are analyzed. It was shown that large diameter pipes have a much higher safety margin, and that the 0.1 A Criterion (10% of pipe cross-section) for postulated leak cross-sections gives a conservative estimate.

Elastic-plastic fracture analysis of carbon steel piping using the latest CEGB R6 approach

Abstract The elastic-plastic fracture of carbon steel piping having various pipe diameter and circumferential crack angle subjected to a bending moment is analyzed using the latest CEGB R6 approach. The elastic-plastic fracture criterion must be applied instead of the plastic collapse criterion with increase of the pipe diameter and the crack angle. A simplified elastic-plastic fracture analysis procedure based on the R6 approach is proposed.

Assessment of the fracture margin and jet force caused by high-pressurized fluid for cracked pipes.

Piping systems for power plants are designed to withstand hypothetical events such as leakage from a crack and jet impingement. The assumed event, in this paper, is that high-pressure water leaks from circumferential through-wall cracks in a carbon steel pipe. Sizes of the cracks detected by leak monitors under normal operations are calculated from the analyses of fracture mechanics and fluid mechanics. Crack-opening areas for jet force assessment are obtained for various diameter pipes under abnormal conditions. The areas are compared with the 0.1F (10% of pipe cross section) criterion. Large diameter pipes are shown to have large safety margins against fracture from the viewpoint of leak monitors and jet force assessment.

Elastic-plastic analysis of pipe with internal circumferential crack in bending.

An elastic-plastic analysis program using the finite element method and a J-integral post-processor using the virtual crack extension method are developed. The program accuracy is examined by two test problems: a plate with a hole in tension and a single -edge notched specimen in bending. This program solves the Battelle/NRC finite element analysis round-robin problem regarding an elastic-plastic analysis of a pipe with an internal circumferential crack in bending. An investigation is made into the effect of the finite element mesh on the results of the elastic-plastic analysis. It is found that a finer mesh gives greater J-integral and COD values, while the mesh does not affect the load-load point displacement curve.

The development of a high-speed structural-analysis program, SIMUS

Abstract By putting supercomputers to use, there is a possibility of solving large-scale structural problems by the finite-element method at high computing speeds. The authors have developed a high-speed structural-analysis program, SIMUS, by applying algorithms suitable for supercomputers that have vector processors. To obtain a high processing speed by using vector processors, it is necessary to make the ratio of operation time in DO-loops to total execution time larger. It is therefore necessary to enhance the vectorized ratio. SIMUS has three characteristics for high-speed processing: 1. (1) the vectorized ratio and DO-loop ratio are enhanced; 2. (2) the number of repetitions in DO-loops is enlarged; 3. (3) plural calculators are used in parallel. As a result of applying four examples of linear-elastic stress analysis using this program, a vectorized ratio of 99% and a processing speed ratio in CPU time (scalar CPU/vector CPU) of 29 times are obtained for very large-scale problems, for example, over 10 4 D.O.F.