CFD Analysis With Fluid-Structure Interaction of Opening High-Pressure Safety Valves
CFD Analysis With Fluid-Structure Interaction of Opening High-Pressure Safety Valves
CFD Analysis With Fluid-Structure Interaction of Opening High-Pressure Safety Valves
a r t i c l e i n f o a b s t r a c t
Article history: A multi-mesh numerical valve model has been developed to analyze the opening characteristic of high-
Received 11 March 2011 pressure safety valves. Newtons law and the CFD result for the ow force are used to model the move-
Received in revised form 26 October 2011 ment of the valve. In incompressible transient ow simulations a large force rise and collapse is caused by
Accepted 16 May 2012
a redirection of the bulk ow. This ow-history effect cannot be incorporated in a quasi-steady approach.
Available online 24 May 2012
For real-gases at a set pressure of 40 bar oscillations have been observed during closing of the valve. They
are caused by the interaction between the ow in the cavity of the valve disk and the ow towards the
Keywords:
valve outlet. At a higher set pressure the ow force continually decreases, which indicates that only a suf-
Computation
Fluid mechanics
ciently fast inlet pressure rise forces the valve to open. With this tool the operation characteristics of
Safety safety valves can be assessed to optimize the valve design.
Simulation 2012 Elsevier Ltd. All rights reserved.
0045-7930/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compuid.2012.05.010
A. Beune et al. / Computers & Fluids 64 (2012) 108116 109
safety valve
vessel
Fig. 1. Process diagram (left) and construction drawing (right) of a high-pressure safety valve.
2.2. Discretization
Fig. 4 shows an example of a grid at h = 0.89 mm. From the the mass, momentum and energy equation, while a rst-order
smallest disk lift of 0.01 mm each grid is predened with a factor backward Euler scheme is applied for the turbulence equations.
1.5 larger disk lift compared to the previous one resulting in the All inlet and outlet boundaries have subsonic conditions. Hence,
17 predened meshes in total. The smallest nodal distance of the at an inlet boundary the total pressure and total temperature are
mesh with disk lift h = 0.01 mm is 0.001 mm which increases up prescribed, while at an outlet boundary the static pressure is pre-
to 0.025 mm for the larger meshes. The meshes have approxi- scribed. The walls of the high-speed ow are considered adiabatic.
mately 42,000 nodes. For a small part of a simulation with only A simulation with multiple moving meshes starts with the def-
two subsequent predened meshes a grid convergence study has inition le with basic solver parameters and one predened non-
been performed, in which four sets of grids with approximately deformed mesh that is consistent with the valve disk displacement
8000, 16,000, 42,000 and 100,000 nodes. The results for ow force dened in the denition le. The rst simulation is initialized with
and valve displacement on the coarsest mesh differ considerably the solution of a steady ow simulation at the same initial disk lift
from the ner meshes, whereas the results on the two ner grids with velocity zero. When during a simulation a second mesh is
almost collapse. Therefore, the grid with approximately 42,000 loaded additional parameters such as the spindle velocity and ow
points has been selected for the full analysis of the valve force are updated from the previous simulation to the CFX Com-
movement. mand Language (CCL).
A series of simulations with different predened meshes, the
2.3. Solution strategy communication of the variables necessary for the calculation of
the valve displacement and the export of the solution variables
In the CFD software package ANSYS CFX the numerical discret- for postprocessing are controlled with Perl scripts.
ization is node based and it uses shape functions to evaluate the As part of each time step rst the mesh movement during the
derivatives for the pressure gradient term and the diffusion terms next time step is evaluated, where the acceleration of the disk h n
in the momentum, continuity and turbulence equations. The Na- is determined from the valve displacement hn and ow force F nflow
vierStokes equations are discretized in a collocated way and by
solved by an algebraic multigrid solver. To avoid pressurevelocity
decoupling, a robust interpolation scheme similar to RhieChow n n
n F flow kspring h h0 mspindle g :
h 3
interpolation [12] is used. CFX solves the conservation equations
mspindle
of mass and momentum in one system of equations, with all equa-
tions being fully coupled [1]. The turbulence equations are solved Then the displacement of the disk hn+1 at the next time step
coupled as well and the energy equation is solved separately. becomes
CFX uses advection schemes such as rst-order upwind differ-
ences and numerical advection with a specied blend factor. This n1 1 n 2
h h_ n Dt h
n
blend factor can be varied between 0 and 1 to vary between a rst-
h Dt ; 4
2
and second-order differencing scheme and to control numerical
diffusion. The high-resolution scheme option will be chosen, which where Dt is the time step and h_ n is the velocity of the disk.
maintains the blend factor as closely as possible to 1 without vio- Moreover,
lating the boundedness principle that could result in non-physical
oscillations in the solution. For the turbulence equations the rst-
h_ n1 h_ n h
n Dt: 5
order accurate scheme is sufcient. The grid convergence study
mentioned in the previous subsection indicated that the order of After the mesh movement the simulation time step size is eval-
accuracy of the discretization method is sufcient for the purpose uated. In order to minimize errors in the interpolation of the
of this research. In unsteady simulations a second-order accurate numerical solution to the next predened mesh, the disk lift on
linear multi step method is applied for the time integration of the previous mesh should be as close as possible to the predened
Fig. 4. Front surface of axially symmetric computational domain of safety valve for h = 0.89 mm.
112 A. Beune et al. / Computers & Fluids 64 (2012) 108116
disk lift of the next mesh. In order to achieve equality of these two At time t = 0 s the mass ow rate equals m _ 0:013 kg=s, which
disk lifts within 1 nm, the time step size is changed with not more is comparable to a leaking safety valve. In the rst 3 ms the ow
than 1% between a minimum and maximum value. force remains constant although the disk lift increases to 0.2 mm.
The time step in the simulations has been optimized in order to In the rst part of the valve opening the acceleration forces domi-
have a good balance between calculation time and accuracy. It nate the valve movement. At t = 8 ms and disk lift h = 1.13 mm the
turned out that a time step size of 2 105 s for incompressible ow force drops relatively fast to half of its maximum value in only
ow and of 2 106 s for compressible ow lead to solutions in 0.3 ms. This results in decelerating of the valve spindle.
which the disk lift and ow force on the disk are practically inde- The oscillations of the ow force during the rst few iterations
pendent of the time step size. These values of the time step result of a new simulation with a predened mesh are small, as can be
in up to eight coefcient loops of the linear solver per time step seen in Fig. 6. These oscillations, if present, reduce within 10 time
after a few time steps after interpolation of the solution variables steps to zero and, therefore, do not affect the total behavior of the
from a deformed mesh onto a new predened mesh. The inner ow force. Especially for incompressible ows the ow force is
coefcient loop terminates when the maximum residuals of the sensitive to small interpolation errors and inaccuracies that occur
mass, momentum and energy transport equations become below during initialization of the simulation.
103 or after 50 iterations. The change in direction of the ow in this time period is visual-
All dynamic valve simulations have been performed on a single ized in four vector plots of the velocity eld around the lifting-aid
core of an Intel Xeon 5130 dual-core processor at a clock rate of in Fig. 7. It can be seen that at t = 7.2 ms and h = 0.89 mm the ow
2 GHz. For the incompressible ow simulations the computational remains attached to the valve cone and directly impinges on the
time is 8 days for a simulation time of 0.1 s. For the compressible
ow simulations a time span of 0.01 s could be simulated within
8 days.
3. Results
75
t = 0.0072 s t = 0.0076 s
50
25
t = 0.0089 s t = 0.0106 s
Fig. 5. Disk lift, ow force and mass ow rate versus time of an FSI simulation with Fig. 7. Vector plots of velocity eld of water ow at four different simulation times
liquid water at pinlet = 44 bar. at pinlet = 44 bar.
A. Beune et al. / Computers & Fluids 64 (2012) 108116 113
Fig. 8. Contributions of the signicant parts of the walls of the disk to the total ow force versus time of an FSI simulation with liquid water at pinlet = 44 bar.
114 A. Beune et al. / Computers & Fluids 64 (2012) 108116
u [m/s]
100
75
50
25
Fig. 10. Vector plots of quasi-steady (left) and transient (right) velocity eld for water ow at pinlet = 44 bar and h = 0.89 mm.
ow remains attached to the truncated cone and completely ows changes because the disk velocity is low. Nevertheless, the initial
into the cavity of the lifting-aid, inducing a larger total ow force. It increase of the ow force provokes rapid and stable opening of
is clear that in the transient approach the ow does not have time the valve. At t = 22 ms the valve is almost closed and the ow starts
to relax so that the redirection of the ow pattern is delayed. This to behave irregularly with high-frequent valve oscillations around
has a large impact on the ow force. This ow-history effect cannot 25 ms.
be incorporated in the quasi-steady approach. This oscillatory behavior is illustrated in Fig. 12. It shows two
vector plots generated at two opposite extrema of the ow force.
3.2. Gas valve ow In the left plot the ow in the cavity has a low velocity, resulting
in almost no contribution of the cavity to the ow force. In the right
Fig. 11 shows results of transient simulations with multiple plot the ow has changed its direction and the supersonic area with
meshes of the axially symmetric safety valve with nitrogen gas. bow shock is smaller, because the ow impinges on the cavity edge
The simulation started on a predened mesh with disk lift with less space to expand. Two strong vortices are present in the
h = 0.01 mm, initial spring force based on a set pressure pset = 40 - cavity leading to a signicant contribution to the ow force. When
bar and spring stiffness kspring = 25328 N/m, a constant 10% over- the valve opens again the ow oscillations damp out.
pressure at the inlet pinlet = 44 bar, inlet temperature Tinlet = 300 K, It is noted that this unstable ow behavior cannot always be
outlet pressure 1 bar, equivalent mass of the moving components avoided in valve design. It is important that the traveling time of
mdisk = 0.7662 kg and time step size Dt = 2 106 s. Due to higher pressure waves moving from the outlet to the valve should not be-
ow velocities the time step size is smaller by a factor of 10 than come close to a multiple value of the eigenfrequency of the valve.
in the simulations presented in the previous subsection. This interaction could induce oscillatory behavior of the valve disk
The force jump previously observed in the liquid simulations al- which is called valve chatter. In this simulation this cannot be
ready occurs at t = 1 ms and h = 0.035 mm. This detachment of the shown, because the computational domain is too small resulting
ow from the valve cone does not lead to such large ow force in too small traveling times. Finally, the damping in this simulation
is very low. This shows that ow-history effects are unimportant
and implies that this low-pressure safety valve can be well simu-
lated in a quasi-steady approach. The quasi-steady simulation re-
sults included in Fig. 11 are indeed in good agreement with the
transient results.
It is also important that the frequency of the oscillatory behav-
ior observed in Fig. 12 is not close to the frequency of longitudinal
waves in the spindle. The eigen frequency of these waves can be in-
ferred from the stiffness of the spindle and its mass, resulting in a
frequency of 45,000 Hz. This is larger than the frequency observed
in Fig. 12 by a factor of 4. This also shows that spindle deformation
can be ignored in the present simulations.
Fig. 13 shows the results of FSI simulations of nitrogen gas ow
at a 10 times higher set pressure of pset = 400 bar, 10 times higher
spring stiffness of kspring = 253280 N/m, constant 10% overpressure
at the inlet of pinlet = 440 bar and the same temperature, outlet
pressure, equivalent mass of the moving components and time
step size as the previously presented gas ow simulation.
In this simulation the force always decreases as the valve opens,
which results in a permanent detachment of the ow from the
Fig. 11. Disk lift, ow force and mass ow rate versus time of an FSI simulation
valve cone. Furthermore, due to a high disk velocity h_ the damping
with nitrogen gas at pset = 40 bar. Solid line: transient simulation, dashed line: is larger and the ow history effects have a larger contribution to
quasi-steady simulation. the ow force. Indeed, the transient simulation results differ more
A. Beune et al. / Computers & Fluids 64 (2012) 108116 115
u [m/s]
700
600
500
400
300
200
100
0
Fig. 12. Vector plots at the time of minimum (left) and maximum (right) of oscillatory force of an FSI simulation with nitrogen gas at pset = 40 bar.
4. Conclusion
multimesh method to investigate the sensitivity to valve dynamics [4] Brk E. Numerische Strmungsberechnung ergnzt durch experimentelle
Methoden zur Untersuchung pneumatischer Strmungen. Ph.D. thesis, TU
and the steady ow method at small disk lifts to determine the
Dresden, Germany; 2006.
opening characteristics. The outcome of such investigation can be [5] Domagaa M. Cfd analysis of pilot operating relief valve. Czasopismo
used for the optimization of valve design. Moreover, Beune et al. Techniczne, wydawnictwo politechniki Krakowshiej z. 3-M; 2008.
[3] have shown that dynamic simulations provide insight into [6] ISO. EN ISO 4126-1:2004. Safety devices for protection against excessive
pressure. Part 1: safety valves. Berlin: Beuth Verlag GmbH; 2004.
the cause of deviations in the ow force between experimental re- [7] Li L, Zang H, Davis G, Kalb F, Hamid S. Improving the closing characteristics of
sults and numerical results based on steady simulations. subsurface safety valve with combined FEA and CFD modeling/numerical
analysis. In: SPE western regional meeting; 2005.
[8] Mahkamov K. An axisymmetric computational uid dynamics approach to the
Acknowledgements analysis of the working process of a solar stirling engine. J Sol Energy Eng
2006;128:4553.
The departments Safety & Fluid Flow Technology and Valves [9] Menter F. Two-equation eddy-viscosity turbulence models for engineering
applications. AIAA J 1994;32:1598605.
and Fittings (Metal Processing) of company BASF SE in Ludwigsha- [10] Reich A, Parthasarathy V, DiMeo A. Coupled uid-structure interaction
fen (Germany) are acknowledged for funding this research. simulation of the opening of the target rock vacuum relief valve. ASME Publ
PVP 2001;424:238.
[11] Reid RC, Sherwood TK. The properties of gases and liquids. 2nd ed. New
References York: McGraw-Hill; 1966.
[12] Rhie C, Chow W. Numerical study of the turbulent ow past an airfoil with
[1] ANSYS. Manual ANSYS CFX, Release 11.0. Canonsburg; 2006. trailing edge separation. AIAA 1983;21:152532.
[2] Beune A. Analysis of high-pressure safety valves. Ph.D. thesis, TU Eindhoven, [13] Soave G. Equilibrium constants from a modied RedklichKwong equation of
the Netherlands; 2009. state. Chem Eng Sci 1972;27:1197203.
[3] Beune A, Kuerten J, Schmidt J. Numerical calculation and experimental [14] Srikanth C, Bashker C. Flow analysis in valve using moving grids through CFD
validation of safety valve ows at pressures up to 600 bar. AIChE J 2011;22. techniques. Adv Eng Softw 2009;40:193201.
http://dx.doi.org/10.1002/aic.12534.