C' Type PTFE Gaskets Performances In High Pressure Sour Applications

Abstract

Sealing performance in general in centrifugal compressors and, in particular, in sour gas applications is extremely important to prevent safety issues and EHS concerns. As known, H S starts to become lethal at concentrations in excess of 100 ppm. In general, elastomeric O-ring gaskets (such as FKM-Fluor elastomer or FFKM-Perfluorelastomer) are the most appropriate to guarantee near-zero leakage towards the environment thanks to their very high viscosity, liquid-like behavior. Unfortunately, O-rings are generally quite permeable to gas and have poor mechanical properties in high pressure applications (above 200 bars). If exposed to explosive decompression or extruding forces through the downstream gap between the casing and the head-cover, they can deteriorate quickly. This finally leads to sealing performance degradation and compressor shutdown. To avoid such issues, a thermoplastic material such as PTFE-Politetrafluoroetilene (Teflon®) is normally used. PTFE has good mechanical properties and is not permeable to gases and liquids, and consequently it is not affected by explosive decompression. In addition, for high pressure applications a special profile, generally C shape, is used and a special spring is installed within the C. This configuration is known as an energized gasket. As plastic materials are only slightly elastic, they cannot rely on their mechanical properties to offer sealing capabilities as is the case with elastomers. Hence, PTFE requires a very low surface roughness of the metal parts in contact with it. (...) Download a pdf of this article s (...) During the leakage test of a BCL304/C (design pressure 425 barg), excessive leakage was measured on the inboard gasket ('C' type) with respect to the allowable value, i.e., ~700 NmLit/min @ 425 barg versus a requirement of 0.56 NmLit/min (equivalent to 0.01 cm /s). The casing tested and discussed in this paper is a barrel compressor architecture with an integral head-cover on the casing discharge side and a suction head-cover fixed to the casing by shear rings. Leakage from the discharge side is automatically avoided in such an architecture. A leakage recovery system is provided for the suction side. This leakage recovery system consists of 3 different chambers created by 4 gaskets set between the casing and its head-cover. The first two inboard gaskets (moving from the process side to the atmospheric side) are of the PTFE 'C' type followed by two elastomeric O-rings on the outboard side. During the leakage test, the leakage is measured at full pressure across the first C PTFE gasket using properly calibrated flow meters. For safety reasons, the leakage across the first inboard C gasket has to be guaranteed at very low level (equivalent to 10 bubbles of 1mm volume per second). The present work reports on the activities conducted to solve the issue of excessive leakage flow detected from the inboard seal and the design solutions are discussed. The impact of several parameters on the final leakage has been analyzed to better assess this phenomenon: a. Gasket groove geometry (circularity and roughness) and potential damages (scratches, dents, etc), b. Gasket surface roughness and potential damages (scratches, dents, etc), c. Gasket-groove assembly and gasket compression including the eccentricity between casing and head-cover causing non-uniform compression over the 360° arc. Since no major defects were found on either the grooves or gasket surfaces, the main focus of the activity was to assess the impact of reduced gasket compression on leakage flow. A special 1D gas dynamic model was developed to physically relate leakage flow with gas pressure and some other key parameters: i. Thermodynamic conditions, ii. Elasto-plastic material properties, iii. Contact surface tribological characteristics. The model was validated through tests on the actual compressor going from 0 to full pressure (425 barg) using nitrogen gas. Finally, the root cause of the excessive leakage was understood and the gasket design changed. This allowed fully meeting the very tight contractual requirements even with significant margins vs. the maximum allowable values. The robustness of the new design was confirmed with three repetitions of the tests and by using spare gaskets of the same design. Moreover, the impact of the groove surface roughness was tested using the new gasket design and comparing the results at two different roughness values (Ra=0.4m and Ra=0.8m). These results are of capital importance in assessing the sensitivity of the configuration to possible deterioration of the surfaces. next: Page 2/ ... Reproduced with permission of the Turbomachinery Laboratory (http://turbolab.tamu.edu). Proceedings of the Forty-First Turbomachinery Symposium, Turbomachinery Laboratory, Texas A&M University, College Station, Texas, Copyright 2012. 3