Local Failure Analysis of Bolted Flanges

Protection against local failure is one of the integral components in the design-by-analysis requirements in ASME BPVC Section VIII, Division 2. Of the methods offered by the ASME, the Local Strain Limit procedure outlined in 5.3.3.1 is the typical calculation method. However, it has been found that relying on this procedure alone can lead to untenable utilization results if used on certain analyses with varied load paths. The flange described in this study was calculated using “design by analysis” according to Part 5 of ASME BPVC Section VIII, Division 2. The elastic-plastic stress analysis method was used. The flange was loaded with an initial bolt pre-tension and then with internal pressure. During the local failure calculation, an abnormal condition was encountered in the form of a large spike in the history curve of the ratio between plastic strain and limiting triaxial strain. An investigation found that despite being in a stress state below yield stress, some nodes had a non-zero plastic strain and high triaxiality factor. This was caused by the load sequence: first, the bolt pre-tension and then internal pressure. The flange was first bent due to the pre-tension load, and later experienced bending in the opposite direction after the internal pressure load was applied. This resulted in a relatively low stress state with a high triaxiality factor and non-zero plastic strain in certain areas, which then showed high utilization under the local failure strain limit criterion. This paper will discuss how this issue can be avoided by using the strain limit damage calculation procedure 5.3.3.2 outlined in ASME BPVC Section VIII, Division 2.

Effect of Stress Triaxiality on the Strain Concentration Factor

Axisymmetric machine element with irregularities such as notches encountered with effects of stress triaxiality on the strain concentration factor (SNCF) at the reduced section. The effect of notch geometries on the triaxial stress state development in the critical section of a notched cylindrical bar is studied here using FEM. In addition, the effect of triaxial stress state (TSS) on the SNCF is evaluated. To this end, a notched cylindrical bars with notch depths from extremely deep notch (do/Do = 0.2) shallow notch (do/Do = 0.95) has been employed. The results show that the notches introduce a TSS at the critical section, which strongly affected by the notch depth as well as the notch radii. In this paper, a new concentration factor is introduced as the ratio of the stress triaxiality factor at the notch root (TFNR) to the average triaxiality on the critical section (), i.e. the triaxiality concentration factor KTF. The numerical results reveal that the variation of the average triaxiality factor with total strain shows the same trend as that of the SNCF. The variation of the elastic values of TFCN, , , and SNCF with do/Do and show that the minimum TFNR leads to the maximum elastic SNCF. It is prominent that elastic TFNR is less that elastic TFCN for 0.2 ≤ do/Do ≤ 0.85, while it is greater for shallow notches. The current results indicate a strong compatibility between the newly defined triaxiality concentration factor and the SNCF up to general yielding.

Influence of Lode Angle on the ASME Local Strain Failure Criterion

Failure strain at any point on a structure is not a constant but is a function of several factors, such as stress state, strain rate, and temperature. Failure strain predicted from the uniaxial tensile testing cannot be applied to the bi-axial or tri-axial stress state. ASME Sec VIII-Div-2, and −3 codes give methods to predict the failure strain for multi-axial stress state by considering the triaxiality factor, which is defined as the ratio of mean stress to the equivalent stress. Failure strain predicted by the ASME method (based on the Rice-Tracey ductile failure model) is an exponential curve that relates the failure strain to the triaxiality factor. The ASME VIII-3 method also gives procedures to calculate failure strain for various material types: ferritic, stainless steel, nickel alloy, aluminum, etc. Experimental results of failure strain at various stress states show that the failure strain is not only a function of the triaxiality factor, but also a function of the Lode angle. The Lode angle takes on the value of 1, 0, and −1 for tension, pure shear, and compression stress state, respectively. Experimental data shows that the failure strain is a 3D surface which has an exponential relation with triaxiality and a parabolic relation with the Lode angle. To validate the ASME failure strain prediction, this paper compares experimental failure strain test data from literature with the ASME predictions. The ASME predictions are non-conservative especially for moderately ductile materials such as aluminum and high strength carbon steel. A reduction factor on failure strain for low ductile material is presented using the relation between the R (yield/ultimate) and the stress ratio (shear/tensile stress). The ASME method does not account for the environmental effects while calculating the failure strain. High pressure, high temperature (HPHT) subsea components designed using ASME VIII-3 code are subjected to various environments in subsea, such as seawater, seawater with cathodic protection (CP) and production fluid (crude oil). Experimental data shows that the Elongation (EL) and/or Reduction in Area (RA) from tensile testing decrease in these environments. Therefore, to account for any environment effect on the failure strain, reduced EL and RA can be used to predict the failure strain.

A Study on Material Workability by Upsetting of Non-Axisymmetric Specimens by Flat Dies

Upsetting is a typical test for determining the workability diagram. In most cases axisymmetric samples are used for such tests. However, the shape of samples may have a significant effect on the ductile fracture initiation. Therefore, a greater variety of sample geometry should lead to a more accurate shape of the workability diagram. A difficulty here is that a theoretical treatment of samples in which three-dimensional flow occurs is more difficult and time consuming as compared to axisymmetric samples under axisymmetric loading. This difficulty can be overcome in the case of the ductile fracture criterion based on the workability diagram and the average value of the triaxiality factor. In particular, if fracture occurs at free surfaces then it is sufficient to determine experimentally in-surface strains after several stages of the upsetting process, up to the initiation of ductile fracture. After that, the corresponding point of the workability diagram can be found by means of a simple analytical procedure and numerical integration. This approach is used in the present paper to correct the workability diagram using non-axisymmetric upsetting of five different samples made of steel. Some previous results are combined with the new results to obtain the workability diagram over a wide range of the triaxiality factor.

Simplified Prediction of Creep Void Density in Heat-Affected Zone of Grade 122 Steel Considering Multiaxial Stress State

This paper deals with a simplified method for approximately predicting creep void growth in heat-affected zone (HAZ) of ASME grade 122 (11Cr-2W-0.4Mo-Cu-Nb-V) steel weldments. Authors have proposed a simplified prediction method based on the relationship between creep void density increasing rate and multiaxial stress state. This method has been applied to prediction of creep void growth behavior for grade 91 (9Cr-1Mo-Nb-V) tubular specimens with longitudinal weldments. In this study, the method has been also applied to grade 122 steel to clarify the applicability of the method. Internal pressure creep tests of grade 122 tubular specimens with longitudinal weldments subjected to several internal pressures have been conducted to reveal creep void growth behavior in HAZ. In addition, finite element creep analyses for the specimens at different creep strain rates in base metal, weld metal and HAZ have been carried out to investigate distribution of stresses and stress triaxiality factor in HAZ. A comparison between stress distributions and creep void distributions revealed that stress triaxiality factor affects growth behavior of creep voids. From the result, the relationship between creep void density increasing rate and the parameter as a function of principal stress and triaxiality factor was established. It was found that the slope of this relationship for 122 steel has a tendency to be slightly small compared with grade 91 steel. To demonstrate the applicability of the proposed simplified prediction method, the method was applied to the internal pressure creep test specimens at different experimental conditions. As a result, the predicted void distribution and void density increasing rates for grade 122 steel were in good agreement with the experimental results.