Simulation and development of dry gas seal for supercritical CO2 carbon dioxide

Abstract

The supercritical CO2 closed loop Brayton cycle operates at high pressure to achieve higher energy conversion efficiency. One of the important components in turbomachinery of the power cycle plant is the dry gas seals. Dry gas seals are gas-lubricated, mechanical, non-contacting, end-face seals, consisting of a mating (rotating) ring and a primary (stationary) ring. Low leakage dry gas seals are considered as a key enabling technology for achieving the improved thermodynamic cycle efficiency in the supercritical CO2 power cycle. Alternate seal types, for example, labyrinth seals, suffers from high leakage. Even so there is a growing interest and importance of applying the small length scale dry gas seal in the small to medium scale supercritical closed loop Brayton cycle (1-20 MWe), there are still uncertainties for their operation at supercritical CO2 conditions. These include the real gas effects near the critical points and the methods of minimizing the deformation of the supercritical CO2 dry gas seal. In the supercritical region in the vicinity of the critical point (304 K, 7.4 MPa), CO2 behaves as a real-gas, exhibiting significant and abrupt nonlinear changes in fluid and transport properties and high densities.Comprehensive analysis is performed to simulate the supercritical CO2 dry gas seal. First, an isothermal simulation assuming rigid sealing ring walls in the gas film is performed using ANSYS Fluent to study the influences of real gas effect on performances of the dry gas seal. Then, conjugate heat transfer simulation is used to optimize the face geometry for a small to median scale supercritical closed loop Brayton cycle (1-20 MWe) using ANSYS Fluent. Finally, the pressure and the thermal outputs from the conjugate heat transfer analysis are used as boundary conditions for one way coupling fluid-structure-thermal simulations using ANSYS Static Structural to study the effect of deformation of the sealing rings under applied pressure-loads, thermal-loads, and centrifugal effect.Finding from the simulation results shows, close to critical point the real gas effect is significant, whereas far from the critical point the supercritical fluid resembles an ideal gas. The centrifugal effect is enhanced by the higher density due to the real gas effects, causing a reduction of average pressure in the dam region hence reduces the opening force, and seal leakage.Increasing the groove radius decreases the opening force whereas increasing the spiral angle, increases the opening force. However, the variation is small for all tested cases studies with variation up to 6%. Increasing the groove radius decreases the leakage rate whereas increasing the spiral angle, increases the leakage rate. The variation in changing groove radius is more significant than the spiral angles. The variation in leakage is up to 29.2% for all tested cases. In term of the film stiffness, there is no clear pattern when changing the groove radius but the impact of the spiral angle is significant, up to 46.6% for all tested cases studies. For a small diameter seal, groove radius, rg = 17 mm and a spiral angle α = 30o are recommended as it gives the optimum seal performances, owing to its lowest leakage rate and high film stiffness respectively. Results from one-way coupled analysis to explore sealing ring deformation, show that the thermal-induced deformation has more pronounced effects than the pressure-induced or centrifugal-induced deformations on the net deformation of the dry gas seal. It is shown that the larger thermal deformation is caused by the presence of a radial temperature gradient, which arises due to the much better heat transfer experienced at the seal outer diameter (to the sealed fluid), compared to heat transfer at the inner diameter of the rings (to fluid at ambient conditions). Reducing convection area, the portion of sealing ring exposed to high heat transfer is explored as a method to minimise this. It is shown that this has some benefit. However, overall deformation and coning remain large. Consequently finding a way to control heat transfer and temperature profiles of the rings is critical.The numerical models are validated with the previous computed data in term of the pressure, opening profiles and friction heat and a reasonable agreement have been achieved. The velocity profiles near the wall region are verified with the empirical formulas and reasonable agreement has been achieved.This project provides some new insights on how to design a seal for supercritical CO2 that reveal new flow physics and seal distortions management.