scope:

SUMMARY

This report documents the analysis work done at Stress Engineering Services to establish the load capacity of all flanges given in the April, 1986 Editions of API 6A (Fifteenth Edition) and 6AB. A total of 69 different geometries were analyzed. This is less than the total number of flanges in 6A and 6AB due to some duplication of flange geometries between sizes and styles. The various loads considered were: Bolt Makeup (Preload), Internal Pressure, Tension, and Bending Moment. All 69 flanges were analyzed with an axisymmetric finite element model for each of the four load cases. A post-processor program was written to calculate the maximum moment capacity for various levels of pressure and tension, based on linear superposition of results. Three different criteria were used to establish the maximum moment: 1) ASME Section VIII, Division 2 allowable stress categories for the flange with the basic membrane stress allowable established by API, 2) Allowable bolt stresses as established by API, and 3) Loss of preload on the ring joint. The results of this post-processing are presented in plots of pressure vs allowable moment for various tension levels in Appendix B. See Section 2.0 for details of the axisymmetric analysis and Section 3.0 for details of the load capacity calculations.

A side study was done to determine the effect of a reduced bore diameter on a flange's load capability. This study was done since the Taylor-Waters calculations show that the stresses may increase with a reduced bore. The stresses in the reduced bore flange were determined using revised axisymmetric finite element models. The load capacity remained the same or increased for the five geometries with reduced bores that were considered. This work is summarized in plots that show a comparison of results for a flange with the maximum allowed bore diameter and a flange with a reduced bore diameter. See Section 4.0 for details and results of this analysis.

A second side study was done to determine the effects of thermal gradients and the resulting thermal stresses on a flange's load capability at 250°F. This study was done since there has been some evidence that temperature gradients may lead to flange failures. Five of the axisymmetric finite element models used in the first task were modified to perform steady-state thermal analyses for the temperature distibution of 250°F on the inside and 32°F on the outside. These temperatures were used to calculate the resulting thermal stresses. The allowable stress criteria used in the first task was expanded to include thermal stresses. The flange load capacity was unaffected by thermal stresses for four of the five geometries that were considered. The allowable external loads would have to be reduced approximately 25 percent for the affected flanges (7-1/16 5000 6B) to meet the stress criteria. See Section 5.0 for details and results of this analysis.

Three-dimensional finite element analysis in addition to full scale testing of a 7-1/16″ 10,000 psi 6BX flange was done to establish the ultimate load capacity of the flange with a combination of applied loads (pressure, tension and moment). The analysis and test provided a good baseline to compare results with the axisymmetric results. The two and three dimensional finite element analysis showed a good comparison for prediction of loss of raised face contact on the tension side of bending. Stresses in the bolt also compared well for load conditions where the raised face remained in contact. The results of the test showed reasonable comparison to three dimensional finite element results for the ultimate load capacity of the flange. The allowable bending moment from the rating chart was 73 ft-kips, the ultimate load capacity from the 3-d analysis was around 600 ft-kips, and the test was carried to a moment of 198 ft-kips with no leakage. The 3-d analysis ultimate load capacity was based on bolt yielding and not gasket leakage.

As a side study to the ultimate load analysis, a second three-dimensional model was made for the flange with radial lockdown screw holes between each bolt hole. This side study was done to find the amount the stresses increase in the flange due to this loss of material. The finite element model was run for a linear, elastic solution with a single tension/moment load in addition to the working pressure and bolt makeup loads. The results of this analysis show the stress intensity at the flange O.D. to increase for makeup only, with the pressure added, and with pressure, tension and moment added when compared to a flange without any radial holes. However, the controlling stresses did not dramatically increase. See Section 6.0 for additional details.

Based on the work reported here, additional work may be considered in the following areas.

- Additional analysis of flanges with radial holes to establish the effect of the holes for other sizes.

- Additional thermal analysis of flanges with large makeup stresses that are most likely to have reduced load capacity when large temperature gradients are applied.

- Additional testing should be concentrated on defining the sealing limits of API gaskets when subjected to pressure, tension, and moments. Cyclic loading would be particularly instructive.