481

A publication of

CHEMICAL ENGINEERING TRANSACTIONS

VOL. 61, 2017 The Italian Association
of Chemical Engineering
Online at www.aidic.it/cet
Guest Editors: Petar S Varbanov, Rongxin Su, Hon Loong Lam, Xia Liu, Jiří J Klemeš
Copyright © 2017, AIDIC Servizi S.r.l.
ISBN 978-88-95608-51-8; ISSN 2283-9216 DOI: 10.3303/CET1761078

Theory Analysis and CFD Simulation of the Pressure Wave

Generator
Bing Yang, Jianqiang Deng*, Kai Liu, Fanghua Ye
School of Chemical Engineering and Technology, Xi’an Jiaotong University, Xi’an, Shaanxi 710049, China
[email protected]

In a pressure wave generator mainly consisting of a stator and a rotor with several valve blocks, pressure
waves are generated by the continuous shear perturbation of fluids. In this paper, theory analysis of pressure
wave generated in water hammer principle was conducted. Then a simplified CFD model of a pressure wave
generator connecting the upstream and downstream pipes was established, and the grid and the time step
independence tests were carried out. Generating process of controllable pressure waves in pipe flows was
simulated, and the frequency and amplitude characteristics were obtained. The results suggest that pressure
waves in upstream and downstream pipes have the similar shapes of approximate sinusoidal waveform and
opposite phase. Pressure is inversely proportional to the first order derivative of x-velocity at a certain position,
which is in good agreement with theoretical analysis; pressure wave amplitude increases with an increment of
the rotational frequency of the rotor at the beginning of the pressure wave generator operation, but reaches
the maximum amplitude in an optimal frequency when the generator is in stable operation, which means there
exists a transition process for the generator. The critical frequency related to the first characteristics time can
be applied to avoid the waveform distortion.

1. Introduction
Pressure waves in transient flow are generally induced by flow disturbances which may be caused by planned
or accidental changes in a system (Chaudhry, 2014). Pressure waves with high propagation speed and
energy concentration properties can be utilized in transmission signal in measurement while drilling system
during drilling process (Patton et al., 1977). Another significant application of the pressure waves is the early
detection of gas influxes because of encountering gas-bearing fractures in gas drilling process to prevent the
accidents and ensure security and continuity of gas drilling (Meng et al., 2015). Pipe leakage may cause
wastes for water and energy resources and can produce contaminants in urban water supply systems.
Pressure waves can also be applied in identifying and detecting potential defects in water supply pipe systems
(Duan, 2017).
Fang and Su (2004) summarized the fundamental types of the pressure wave generator in a measurement
while drilling system in petroleum drilling engineering, and classified pressure wave generators into two main
categories including the depressurization type and throttle type. The principle of the pressure wave generator
based on the thin-walled cutting edge flow characteristics was analyzed and the curve orifice generating the
continuous sinusoidal pressure wave was designed and optimized (Jia et al., 2010). Xue et al. (2011)
designed a new type pressure wave generator applied in vibration cementing based on water hammer theory.
In order to obtain an optimized valve orifice satisfying the requirements for the continuous sinusoidal pressure
output, an improved arc-fillet-line triangular valve orifice was designed based on a general line triangular valve
orifice (Yan et al., 2015). Hansen and Kjellander (2016) employed the CFD method to explore the
consequences of a sudden opening of a storage tank due to an external aggression and analyzed
characteristics and attributions of the pressure waves.
Although many studies have been published concerning the applications of the pressure wave and generator
and qualitative analysis of pressure wave characteristics, a little of research has been done on generating
controllable pressure wave and theory analysis of pressure wave generator are inconsistent in different
literatures. In the present paper, theoretical analysis is carried out by extending classic water hammer theory

Please cite this article as: Yang B., Deng J., Liu K., Ye F., 2017, Theory analysis and cfd simulation of the pressure wave generator, Chemical
Engineering Transactions, 61, 481-486 DOI:10.3303/CET1761078
482

in continuous transient process and CFD simulations of pressure wave generated in the generator are
conducted. The simulation results reveal the amplitude and frequency characteristics which are discussed
quantitatively.

2. Mechanism of a pressure wave generator

2.1 Theory analysis
Water hammer with spatial and temporal changes in velocity and pressure in pipe systems is caused by rapid
flow disturbances; such transient flows are assumed to be unidirectional because the axial fluxes of mass,
momentum and energy are far greater than their radial counterparts (Ghidaoui et al., 2005).
With the unidirectional assumption, the one-dimensional water hammer governing equations in transient pipe
flows can be derived by applying the principles of mass and momentum conservation to a control volume in an
oblique tapered pipe. The continuity equation for a control volume is expressed as follows:
P P V
V   c2  V sin   0 (1)
x t x

where v is the x-direction velocity, P is the pressure, α is the angle between pipe centreline and the horizontal
line, g is the gravitational acceleration, c is the wave speed, ρ is the density of fluid. Considering gravitational,
wall shear and pressure gradient forces as externally imposed, the momentum equation can be given by:

1 P V V fV V
V   0 (2)
 x x t 2D

where D is the inner diameters of the pipe, f is the friction coefficient. The wave speed can be calculated by:
K/
c2  (3)
1   K / E  D / e   c1

where K is the bulk modulus of elasticity of the fluid, E is the Young’s modulus of elasticity of pipe materials, e
is the thickness of the pipe, c1 is the parameters related to the constraint conditions of the pipes. Eq(1-2) can
be used in 1D compressible pipe flow.
Compared to other terms in Eq(1-2), the convective terms are negligible in magnitude because the wave
speed is considerably larger than x-direction velocity of fluid in most of the engineering applications
(Chaudhry, 2014). To analyze the pressure wave propagation characteristics, the fluid is supposed to be
inviscid. Based on the above assumptions and simplifying the equations, the pressure wave equation can be
written as:
2P 2P
 c2 2 (4)
t 2
x

which is a one-dimensional second-order wave equation.

For the case of a rapid closure of the downstream valve in a piping system with a constant level reservoir at
the upstream end and a valve at the downstream end, the general solution of Eq(4) is represented as:
P  cv (5)

It can be noticed from Eq(5) that the pressure wave amplitude is proportional to change in x-direction velocity.
The disturbance of the case for the rapid closure of the valve is instantaneous and discontinuous. For a
continuous disturbance case, the process can be viewed as a linear superposition of instantaneous closure
process at each moment. Assuming that ρ and c are constant, the partial derivative of time on right side of
Eq(5) can be calculated to obtain the relationship between the pressure and the x-direction velocity which can
be written as:

P t   k c  v t  (6)

where k is the correction coefficient. The Eq(6) can describe the quantitative relation between the pressure
and the velocity accurately, which may validate the simulation results in Part 4.
2.2 Construction and working principle
A pressure wave generator shown in Figure 1(a) mainly consists of a stator, a rotor, a shaft, an upstream
pipeline and a downstream pipeline. Both the stator and rotor are fan-shaped and have four valve blocks.
When the generator works, the stator remains stable and the rotor driven by a variable frequency motor
rotates at a certain speed to disturb the fluid flow in both upstream and downstream pipes continuously. As
mentioned for continuous disturbances, variation in pressure is closely related to change rate of velocity.
 483

Therefore, the periodic variation in velocity field which is induced by the continuous shear disturbance will
produce the pressure waves in the generator and waves will propagate along the pipelines.

Inlet Outlet Upstream pipe Gap Downstream pipe A-A

Stator Gap
Rotor Stator A Rator X Flow

Flow
Flow
o Valve
Shaft
blocks Flow
Upstream pipeline Flow
A
(a) Downstream pipeline
(b) Inlet Wall Interface Wall Outlet (c)

Figure 1: (a) The 3D model of a pressure wave generator, (b) The Main parts and boundary conditions of the
CFD model, (c) Left view of the rotor

3. CFD method
3.1 CFD modelling
In order to investigate the generating process and characteristics of pressure waves, the effect of the bend
pipe on pressure wave is neglected and the pressure wave generator system can be reduced to five parts,
including an upstream pipe (length: Lu = 5,000 mm, diameter: Du = 80 mm), a stator (inner diameter: Dsi = 40
mm, outer diameter: Dso = 80 mm), a rotor (Dri = 40 mm, Dro = 80 mm), a downstream pipe (Ld1 = 5,000 mm, Dd
= 80 mm) and a gap (thickness: δg = 3 mm, Dg = 80 mm) between the stator and the rotor which are shown in
Figure 1(b). Figure 1(c) shows the left view of the rotor to illustrate the valve block shape. The origin of
coordinate system is at the centre of the left surface of the gap. Grids of each part is generated by using the
ICEM CFD 14.5 and merged together. The junction surfaces of each part are defined as interface boundaries
to accomplish the data communication. Attributed to the complexity of each part, the grids of the gap, the
upstream and downstream pipes are structured; the stator and the rotor are meshed by tetrahedral
unstructured grids. ANSYS Fluent 14.5 with 3D mode was applied to perform CFD calculations. Simulations in
this paper mainly focused on the pressure wave characteristics caused by velocity fluctuations in a single
phase flow and involved strong unsteadiness occurring in the flow. The realizable k–ε turbulence model was
chosen to simulate the pressure wave better (Yang et al., 2015). The motion of the rotor was achieved by
setting in the cell zone motion dialog box. All types of boundary conditions and corresponding values are
presented in Table 1. Considering slight compressibility of water as the fluid in cases, the density of water was
determined by a user defined function(UDF) of pressure (ANSYS., 2012). The density of water was calculated
by the following equation:

   ref 1   P  P  K 
op (7)

Where ρref is the reference density, ρref = 998.2 kg/m3; Pop is the operating pressure, Pop = 101,325 Pa; K is
the bulk modulus of elasticity of the fluid, K = 2,180 MPa. All the details regarding the chosen solver settings
are given in Table 2. In all cases, the steady results remaining the rotor stationary were obtained to be the
initial conditions for unsteady calculations. Monitoring surfaces along the axis positioned by the different x-
coordinates were created to achieve the pressure and velocity at certain positions, which were named by the
x-coordinate and the measurement units in m.

Table 1: Boundary conditions

Part of geometry Inlet All walls All interfaces Outlet
Boundary condition type Pressure inlet Stationary wall, no slip Interface Pressure outlet
Value 1.005 MPa - - 1 MPa

Table 2: ANSYS Fluent solver settings

Pressure-velocity
Solve type Density Wave speed Turbulence models Convergence criteria
coupling
Pressure- 10-6 for energy
UDF 1,000 m/s Realizable k–ε SIMPLE
based 10-3 for the rest
484

3.2 Independence test

Grid independence test was performed on three meshes, consisting of the following number of nodes on per
meter in axial direction of the pipe: 200 (coarse mesh), 300 (medium mesh) and 320 (fine mesh). Due to the
periodic unsteady flow, time step independence test was conducted on several time steps determined by:

t  1/  N  n  fr  (8)

where fr is the rotational frequency of the rotor, N is the number of valve blocks, n is a specific integer. fr was
set to 30 Hz. Variations in x-direction velocity of the monitoring surface at x = -1 m in different mesh types and
time steps are illustrated in Figure 2(a) and (b). It can be observed from Figure 2(a) that velocity variation with
time between medium mesh and fine mesh is closely related. Figure 2(b) shows the velocity changes with
time in different time steps cases. The medium mesh and n = 24 were chosen to conduct the CFD calculation.
0.44 0.62
n=70 n=40
coarse mesh 0.60 n=24 n=21
0.42 medium mesh
fine mesh 0.58 n=17
0.40
0.56
Vx(m/s)

Vx(m/s)
0.38 0.54

0.36 0.52

0.50
0.34
0.48
0.32 0.46
0.00 0.02 0.04 0.06 0.00 0.02 0.04 0.06
(a) t(s) (b) t(s)

Figure 2: (a) Grid independence test, (b) Time step independence test

4. Results and discussion

Figure 3 shows the pressure variation with time in different positions at fr = 30 Hz. Time was
nondimensionalized by multiplying the frequency of the rotor and N. As shown in Figure 3, the pressure waves
in upstream and downstream pipes have the similar shapes of approximate sinusoidal wave and opposite
phase, which means that the compression wave in upstream and expansion wave in downstream are
generated simultaneously when the rotor rotates and disturbs the flow. In different observation positions, there
exists obvious phase difference because of the finite wave speed for compressible fluids.

1,040 1,060
-0.5m -1m -2m -3m -0.5m -1m -2m -3m
-4m 0.5m 1m 2m 1,050 -4m 0.5m 1m 2m
1,030 3m 4m 3m 4m
1,040
1,020 1,030
Pressure(kPa)

Pressure(kPa)

1,020
1,010
1,010
1,000 1,000
990
990
980
980 970
0.0 0.5 1.0 1.5 2.0 20.0 20.5 21.0 21.5 22.0
*
t t*

Figure 3: Pressure variation with time in different positions at fr = 30 Hz

Figure 4(a) shows changes in the pressure and x-direction velocity of the monitoring position at x = -3 m and
the opening of rotary valve with time. From Figure 4(a), it can be seen that the pressure reaches a maximum
when the valve was fully closed, and reaches a minimum when the valve were totally open. In order to validate
the simulation result, the first order derivative of velocity and the coefficient k in Eq(6) was calculated, and the
mean value of pressure were determined. Pressure of the monitoring position at x = -3 m can be computed by
using the following theoretical relation between pressure and the first order derivative of velocity:
 485

P  t   P  k c  v  t  (9)

where P is the mean pressure at a certain position in a time series. The pressure variations of the simulation
result and the theoretical calculation result with time are shown in Figure 4(b). It can be noted from Figure 4(b)
that CFD results are in good agreement with theoretical results.
1,025 0.240 1,025
-3m(P) -3m(vx) -3m(CFD) -3m(Th)
1,020
1,015 0.232 1,020

1,010
1,015
Pressure(kPa)

1,005 0.224

Pressure(kPa)
1,000 1,010
995 0.216
1,005

vx(m/s)
990
985 open 0.208 1,000
980
995
975 0.200
970 990
120 121 122 123 124 125 126 120 122 124 126 128 130
(a) t* (b) t*

Figure 4: (a) Pressure and velocity variations with time, (b) Pressure variation of both the simulation result and
the theoretical calculation result with time

Effects of the rotational frequency of the rotor on pressure waves in the pipes are shown in Figure 5(a). In
Figure 5(a), it can be found that the pressure wave amplitude increased with an increment of rotational
frequency at the early stage of the pressure wave generator operation. The reason for this could be that the
shear disturbance of rotor to flow might increase when frequency went up in the same initial conditions. Figure
5(b) shows the frequency characteristics of pressure wave more accurately by employing the fast Fourier
transform method. As shown in Figure 5(b), the amplitude of pressure increased when the rotational frequency
was from 0 Hz to 30 Hz; when rotational frequency was higher than 30 Hz, the amplitude decreased as the
rotational frequency went up. These phenomena may be attributed to the shear disturbance of rotor and the
flow response in different rotational frequency. The waveform distortion induced by the short pipes with low
rotational speed can be noticed from Figure 5(a). For a certain pipe system, there exists a critical frequency
which could be estimated by the first characteristics time Tic of pressure wave. Tic can be computed by:

Tic  2 Lp / c (10)

where Lp is the length of pipe. For a given number blocks of rotor, the critical frequency can be determined by:

fcr  1  N  Tic  (11)

where fcr is the critical frequency. To avoid the distortion, the rotational frequency should be higher than fcr ,
which is helpful to regulate the pipe system and rotational frequency to obtain a better waveform.

1,060 16
15Hz 25Hz 15Hz
1,050 30Hz 40Hz 14 25Hz
25Hz 30Hz
1,040 12 30Hz
40Hz
Ampitude(kPa)
Pressure(kPa)

1,030 10

1,020 8

1,010 6 15Hz 40Hz

1,000 4

990 2

980 0
0.00 0.01 0.02 0.03 0.04 0 50 100 150 200 250 300 350 400 450 500
(a) t(s) (b) f(Hz)

Figure5: (a) Pressure variation with time in different rotational frequency, (b) Frequency characteristics of
pressure wave
486

5. Conclusions
In the present paper, theoretical analysis and numerical experiments on pressure wave generated in pipelines
have been conducted. Extending approximate solution of water hammer theory was obtained to reveal
correlation between pressure field and velocity field concisely, which may explain the generation of pressure
wave in continuous disturbance flow. The simulation results were validated by theory results and showed both
amplitude and frequency characteristics. Pressure wave amplitude increased with the rotational frequency at
the early stage of the pressure wave generator operation but existed an optimal frequency to reach the
maximum amplitude when the generator was in stable operation, indicating that there existed a transition
process in the generator. And the waveform distortion will appear when the rotational frequency is lower than
the critical frequency which can be avoided by regulating the pipe system and rotatory frequency. Further
research is needed to investigate practical operating conditions and the flow transition phenomenon in high
rotational frequency. The study on pressure wave characteristics can help to apply the pressure wave in
practice and achieve the velocity fluctuations in scientific research.

Acknowledgments
The authors gratefully acknowledge the financial support from the National Natural Science Foundation of
China under Grant No 21376187.

References
ANSYS., 2012, ANSYS FLUENT UDF Manual 14.5, ANSYS Inc, Canonsburg, PA 15317, USA.
Chaudhry M.H., 2014, Applied hydraulic transients, Springer New York, New York, USA.
Duan H.F., 2017, Transient frequency response based leak detection in water supply pipeline systems with
branched and looped junctions, Journal of Hydroinformatics, 19, 17-30.
Fang J., Su Y., 2004, Fundamental types and principles of hydraulic signal generator, Petroleum Drilling
Techniques, 32, 39-41.
Ghidaoui M.S., Zhao M., McInnis D.A., Axworthy D.H., 2005, A review of water hammer theory and practice,
Applied Mechanics Reviews, 58, 49-76.
Hansen O.R., Kjellander M.T., 2016, CFD modelling of blast waves from BLEVEs, Chemical Engineering
Transactions, 48, 199-204.
Jia P., Fang J., Li L., 2010, Design of the rotary valve of the drilling fluid continuous wave generator and
analysis of the signal characteristics, China Petroleum Machinery, 38, 9-12.
Meng Y., Li H., Li G., Zhu L., Wei N., Lin N., 2015, Investigation on propagation characteristics of the pressure
wave in gas flow through pipes and its application in gas drilling, Journal of Natural Gas Science and
Engineering, 22, 163-171.
Patton B.J., Gravley W., Godbey J.K., Sexton J.H., Hawk D.E., Slover V.R., Harrell J.W., 1977, Development
and successful testing of a continuous-wave, logging-while-drilling telemetry system, Journal of Petroleum
Technology, 29, 1215-1221.
Xue L., Li B., Liu A., Zou J., Zhang W., 2011, Study of compression wave caused by hydraulic cementing
tools, Petroleum Drilling Techniques, 39, 105-109.
Yan Z., Wei C., Geng Y., Shao J., Hu X., Li Y., 2015, Design of a rotary valve orifice for a continuous wave
mud pulse generator, Precision Engineering, 41, 111-118.
Yang S., Chen X., Wu D., Yan P., 2015, Dynamic analysis of the pump system based on MOC–CFD coupled
method, Annals of Nuclear Energy, 78, 60-69.