Reduction of the fatigue risk presented by acoustically induced vibration in flare header systems using mitigations that either reduce dynamic stress concentration effects or the level of vibration are of considerable interest to designers and plant operators. Assessments of the relative performance of different types of pipe fittings in reducing dynamic stress levels are presented based on the evaluation of data from full-scale laboratory tests of a pressure-relief system. A modal-analysis based finite-element methodology is also developed so that predictions may be extended to other piping arrangements that vary in thickness, size or connection type. The pipe fittings considered in the test are Pipet®, fabricated tee (Stub-on arrangement), sockolet (small-bore branch connections only), full-wrap reinforced fabricated Tee and Sweepolet®. For the finite-element method reducing tee connection is considered in addition. The test system produced significant levels of both turbulent-induced vibration (FIV) and acoustically induced vibration (AIV), which required differentiation of stress evaluations for the low-frequency FIV region and the mid-to-high frequency AIV region. The relative performance of mitigations (through selection of the type of pipe fitting) was found to be particularly relevant in the low-frequency FIV region. The reductions in dynamic stress and vibration of small-bore branch connections from installation of clamped bracing are also presented. The results show that the use of reducing Tees and full-wrap reinforcements for Stub-on connections for tailpipe and sub-header branch connections provide significant mitigation of dynamic stress and improvement of fatigue life over the use of Pipet® and Stub-on fittings. However, for the Sweepolet® connection which was expected to provide similar improvement the benefits are not fully realized in the 10S configuration.

The calculation of acoustic induced vibration (AIV) for piping downstream of a valve is a critical step in predicting the damage from extreme levels of noise generated by pressure relief valves in flare systems. Three noise prediction schemes are considered for this purpose: International Electrotechnical Commission (IEC) 60534-8-3, the Carucci-Mueller (C-M) formulation for sound power, and an industry valve noise prediction methodology published in the 1980’s. The application of these prediction methods is reviewed utilizing data from a full-scale test system consisting of an NPS6x8 pressure relief valve flowing into a NPS12 tailpipe that is connected through a tee to an NPS20 header. The results show good correlation between the IEC-based predictions and measured internal sound pressure and pipe wall vibration in the AIV frequency region. The industry method provides useful predictions without requiring the level of detailed information needed for the IEC method, whilst the C-M sound power model has limitations when applied to discrete predictions of vibration and strain levels. Observations are also made regarding the relative importance of the FIV contribution to the overall dynamic stresses and associated fatigue life.

Tests were performed on a mockup of a typical relief/blowdown system. The system consisted of an NPS6x8 pressure relief valve, an NPS12 tailpipe, and an NPS20 header. Small bore connections were installed in the tailpipe and header. Simple structures representing a valve mass and stiffness were attached to the small bore connections. Various industry standard tee connections between the tailpipe and header and the small-bore connections to the tailpipe and the header were studied. The goal of the testing program was to provide data to quantify the tee connection style, mitigation methods, and provide data for improvements and validation of Acoustic Induced Vibration (AIV) and Flow Induced Vibration (FIV) prediction and assessment methods. This paper describes the challenges with measuring the very high strain and vibration levels and the data processing used to extract meaningful data. The learnings about the shifting of internal acoustic modes, separating the contributions of AIV and FIV, and separating the contributions of pipe bending modes, higher order pipe modes, higher order acoustic modes, and acoustic/structural coincidence are also described. This work is connected to other presentations which focus on using the data to predict AIV and FIV in realistic blowdown systems.