Prediction of gas and liquid leak rates in compression packing at room temperature

Abstract

Packing rings are the most sensitive element in pressurized mechanical assemblies such as valves and compressors. An incorrect choice or improper procedure of packing ring installation could affect the sealing performance and lead the system to leak. Moreover, leakage from valves can be potentially harmful to humans and the environment. Among the equipment requiring sealing compliance in the petrochemical industry, valves are reported as being the number one component that has leakage problems. Very often, the compression packing is the main element to blame. Unfortunately, the situation is made worse with the lack of a standard procedure for the design of packed stuffing boxes and standard test procedures for compression packing rings. In practice, there are only few experimental, theoretical and numerical studies that are conducted to characterise the flow through packing rings to be able to predict gas and liquid leak rates in valves at room and elevated temperatures. From this standpoint, the need to design a new test rig to test packing rings is essential to be able to predict leakage and reduce fugitive emissions to a minimum level. Therefore, the objective of this work is to carry out experimental tests on different types of packing rings to be able to characterise their sealing properties for the purpose of developing suitable theoretical models to predict their gaseous and liquid leakage. The work focuses specifically on experimental leak measurements of few packing ring materials to characterise their porosity parameters in order to be used in the developed theoretical fluid flow models to predict the leak rates. Leaks through the porous packing material are also simulated through numerical approaches using Ansys CFX. A comparison between the developed analytical models and the experimental and numerical approaches was conducted to validate their accuracy. The methodology of this work is to measure initially the leaks through different packing ring materials using the pressure rise and the mass spectrometry leak detection techniques under different conditions of gland stress and fluid pressure. A referenced gas such as helium is first used to characterise the pores size and their number. The developed analytical models based on Navier–Stokes equations are then used to predict the gaseous and liquid leaks in the wide range of 10-1 to 10-6 mg/s. Different molecule size of gas such as Argon, Nitrogen and air and liquids such as water and Kerosene are tested in the experiment to verify the developed analytical models. Finally, a special numerical model developed in Ansys CFX software is used to support the analytical models. A comparison of these approaches has allowed us to state that the prediction of leakage through packing rings is possible and can be conducted with reasonable accuracy. Finally, the control of leakage in packed stuffing boxes, using accurate prediction models, can improve the reliability of valves, increase the level of production, reduce maintenance costs and minimize fugitive emissions.

Date	10 Sept 2019
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Hakim A. Bouzid (Supervisor)