Difficulties in Predicting Cycles of Failure Using Finite Element Analysis of Acoustically Induced Vibration (AIV) Problems in Piping Systems

Abstract

While finite element modeling analysis is becoming more frequent for analyzing AIV problems, in the absence of experimental data in large diameter pipe, there is no industry wide accepted methodology for representing the pressure excitation for the pipe so accurate cycles to failure may be predicted. The assumptions involved in determining the actual amplitude of the acoustic excitation, which modes may develop and how they couple with the structure all contribute to the overall uncertainty of the problem. Depending on the degree of correlation assumed between the structural and acoustical mode shapes the results vary dramatically. There are also variations based on the number of participating modes assumed. Relative strengths of a Weldolet®, an Insert Weldolet® that is a variation of Sweepolet® and a Reducing Tee connection were analyzed for a 24×6 inch Sch. 10S and STD connection assuming various degrees of correlation and mode participation. Wide fluctuations in the cycles to failure were observed based on the assumptions; however, the stress ratios between the connections are relatively stable. This suggests the use of an acoustic Stress Intensification Factor (SIF) in conjunction with Fatigue Strength Reduction Factors (FSRF) to determine suitability of connections in AIV service rather than an absolute value of cycles to failure. Further investigation of the trends in the value of SIF as the D/d (branch to header diameter) and D/t (diameter to thickness) ratios over a range of pipe diameter are required before these SIF’s could be put into use. Experimental data for a few controlled failure cases are required to ground the finite element prediction in reality. As the experiment is more likely to be conducted with air the possible pitfalls in extending the results from air to commonly used process fluid are also discussed.