The Journal of Engineering

7th International Symposium on Test Automation and Instrumentation

(ISTAI 2018)

Modal simulation and experiment analysis of eISSN 2051-3305

Received on 15th October 2018
Accepted on 8th November 2018
coriolis mass flowmeter doi: 10.1049/joe.2018.9169
www.ietdl.org

Juan Zhao1, Zhongxiang Li1, Dezhi Zheng1 , Shangchun Fan1

1School of Instrumentation Science and Opto-electronics Engineering, Beihang University, Beijing, People's Republic of China
E-mail: [email protected]

Abstract: This article describes the simulation and experimental determination of the oscillation modes of a Coriolis Mass
Flowmeter (CMF). The authors use the SolidWorks software to create a suitable model which is a double U-type of CMF. Based
on SolidWorks software, the authors established the double-U tube model for CMF, adopted ANSYS-Workbench simulation
software to make the modal analysis, and the exciting mode and Coriolis force mode of CMF are obtained by simulation.
According to the results of modal analysis, the authors made the harmonic response analysis of the Flowmeter, got the
displacement amplitude of the U-shaped tube excited by different frequency loads under the exciting force, and the maximum
displacement amplitude at the resonant frequency was determined. Next, the authors use a laser vibrometer to perform an
experimental modal test on the flowmeter. The result of the simulation of the finite element model was verified by modal test of
the flowmeter. Provides an effective reference for CMF modelling and simulation in the future.

1 Introduction will have n physical coordinates to describe its parametric model

[8]. The dynamic equations are:
Coriolis Mass Flowmeter (CMF) [1] is used to measure the mass
flow of liquid and gas. The basic measurement principle is based [M]{ẍ} + [M]{ẋ} + [K]{x} = {F(t)} (1)
on a flow pipe excited to vibrate in a fundamental mode, usually
called excitation mode, at the according resonance frequency. The where the [M], [C], [K] are the mass matrix, damping matrix,
amplitude of the motion is thereby very small, typically a fraction stiffness matrix, {ẍ} ， ẋ ， x are the acceleration matrix,
of a millimetre. Mass flow through the pipe then represents a velocity matrix, displacement matrix, t is time, {F(t)} is exciting
relative movement to the pipe, on which the Coriolis force acts,
force.
resulting in a superimposed vibration of another mode, called a
In the case of un-damped free vibration, the solution of the
Coriolis mode [2]. The amplitude of the Coriolis mode is hereby
natural frequency and vibration mode of the structure can be solved
proportional to mass flow through the pipe and usually two or more
to the problem of the eigenvalues and eigenvectors in the above
magnitudes smaller as the excitation mode and can be measured
(1). The linear equation of motion for free, un-damped vibration is:
through position sensors as a phase shift of the excitation mode.
The sensitive unit is a U-shaped elastic measuring tube, which [M]{ẍ} + [K]{x} = 0 (2)
is a key component of the CMF and has a great influence on the
sensitivity and natural frequency of the sensor. Generally, in order
where the [M] and [K] matrices are constant. Assume harmonic
to improve the precision of the flowmeter from the sensitive
motion:
mechanism and facilitate the structural optimisation design of the
sensor, it is necessary to grasp the influence of various structural
{x} = {ϕ}isin(ωit + θi) (3)
parameters of the U-shaped tube on the performance of the mass
flowmeter.
The modal analysis technique is used to determine the vibration {ẍ} = − ωi2{ϕ}isin(ωit + θi) (4)
characteristics (i.e. natural frequencies and mode shapes) of linear
elastic structures. It is the most fundamental of all dynamic Substituting {x} and {ẍ} in the governing equation gives an
analysis types [3]. Harmonic response analysis are used to eigenvalue equation:
determine the steady-state response of a linear structure to loads
that vary sinusoidally (harmonically) with time, thus enabling you [K] − ωi2[M] ({ϕ}i) = 0 (5)
to verify whether or not your designs will successfully overcome
resonance, fatigue, and other harmful effects of forced vibrations. This equality is satisfied if:
To understand the dynamic of mechanical system is of great
importance for the creation and improvement of new design as well i. {ϕ}i = 0 (trivial, implies no vibration)
as for solving the problems associated with the mechanical ii. or if det [K] − ωi2[M] = 0
vibrations of existing structures [4, 5]. An effective tool for
studying the dynamic properties of the system is the modelling of
• This is an eigenvalue problem which may be solved for up
the dynamic behaviour of structures using the finite element
method [6, 7]. Verification by modal test results is important for to n roots ω12, ω22, …, ωn2
designing a finite element model. • These roots are the eigenvalues of the equation
• For each eigenvalue, there is a corresponding eigenvector
2 Theory and modelling analysis ({ϕ}1, ({ϕ}2, …{ϕ}n)

2.1 Modal analysis theory 2.2 Harmonic response analysis theory

In a linear system, the actual structural system can be discretised In a structural system, any sustained cyclic load will produce a
into vibrations with n degrees of freedom. The corresponding ones sustained cyclic or harmonic response. Harmonic analysis results

J. Eng. 1
This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
 are used to determine the steady-state response of a linear structure
to loads that vary sinusoidally (harmonically) with time, thus
enabling you to verify whether or not your designs will
successfully overcome resonance, fatigue, and other harmful
effects of forced vibrations. The purpose of the harmonic response
analysis is to determine the steady-state response of a known linear
structure subjected to a load that changes with harmonics over
time. This is a commonly used method of structural dynamics
analysis, also called frequency response analysis or sweep analysis.
All loads and displacement vary sinusoidally at the same known
frequency (although not necessarily in phase):
Fig. 1 Three-dimensional model of U-tube F(t) = F0sin(2π f t + φ0) (6)

where F0 is amplitude, φ0 is phase angle.

Assumptions and Restrictions:

• All loads and displacements vary sinusoidally at the same

known frequency
• All loads and displacements, both input and output, are assumed
to occur at the same frequency
• Calculated displacements are complex if:

i. damping is specified
ii. applied load is complex

2.3 Modelling analysis of CMF

Fig. 2 ANSYS-workbench mesh control
2.3.1 Modal simulation analysis: Here, SolidWorks software
was used to construct the model for the U-shaped elastic tube of
CMF. The three-dimensional model is shown in Fig. 1. It is
imported into the ANSYS-Workbench software for modal
simulation analysis.
In the simulation experiment, we adopted the double-U tube
model for CMF which was added with fixed distance plate, without
considering the influence of fluid, neglecting the additional mass
and external support etc. The tube material was set to 316L
stainless steel, the elastic modulus was 206 GPa, the Poisson ratio
was 0.3, and the density was 7800 kg/m3. We used a zero-
displacement constraint to simulate the boundary conditions of the
fixed end of the measuring tube. The grid division was generated
into hexahedron mesh automatically by automatic method, which
was shown as in Fig. 2. The modal frequencies of the first six
orders were obtained by modal analysis under boundary conditions,
as shown in Table 1.
CMFs work in two modes: the excitation mode and the Coriolis
mode. The basic measurement principle is based on a flow pipe
excited to vibrate in a fundamental mode, usually called excitation
mode, at the according resonance frequency. Mass flow through the
pipe then represents a relative movement to the pipe, on which the
Coriolis force acts, resulting in a superimposed vibration of another
mode, called a Coriolis mode. Corresponding to the second and
sixth modes of the Ansys simulation results, respectively, as shown
in Fig. 3a and b.
In addition, the excitation unit and the detection unit installed
on the U-shaped pipe are made of magnetic steel and coils, and
their equivalent mass has a great influence on the working
frequency of the measuring pipe. Therefore, the U-tube adopts an
additional mass point method to simulate the natural frequency
variation measured by the excitation unit and the detection unit.
Here, we only discuss the working frequency of the excitation state
of the measuring tube, i.e. the second-order natural frequency of
the simulation result, as shown in Table 2 and Fig. 4.
Fig. 3 The Ansys simulation results From Table 2 and Fig. 4, it can be concluded the natural
(a) The excitation mode, (b) The Coriolis mode frequency of the measuring tube decreases with the increase of the
additional mass. When the additional mass is 35 g, the natural
frequency is close to the exciting force of the flowmeter the natural

Table 1 First six frequencies of the elastic tube

Order 1 2 3 4 5 6
frequency, Hz 141.57 163.26 204.16 259.87 313.42 365.23

2 J. Eng.
This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
Table 2 Additional mass impact analysis
Additional mass 0 10, g 20, g 30, g 35, g
frequency, Hz 163.26 153.35 144.92 137.72 134.5

Fig. 4 Relationship curve between natural frequency and additional mass Fig. 6 Test Point distribution

Fig. 7 Photo experimental stand

light in an interferometer to give the Doppler-shifted wavelength.

This shifted wavelength provides information of surface velocity in
the direction of the incident laser beam [9].
LVs significantly extend measurement capabilities with respect
to traditional vibration sensors [10]. LV techniques offer the
possibility of performing quick and flexible non-intrusive vibration
Fig. 5 Z-axis frequency function curve measurements at many points over a structure; the data quality and
quantity allows to use them for modal analysis and model updating
frequency is close to the working frequency under the excitation [11]. It has the accuracy and sensitivity that the general sensor
force of the flowmeter. cannot reach and plays an important role in the measurement of
weak vibrations or minor changes in the target.
2.3.2 Harmonic response analysis: Assuming that the two ends
of the U-shaped tube are rigidly fixed, and a simple harmonic 3.2 Laser vibration test
exciting force is applied to the inner side of the tube at the middle
position of the two tubes. The exciting force is: There is a Germany Polytec laser vibrometer in our laboratory. The
physical quantity measured by this instrument is displacement, and
F(t) = F0sin(2πft + φ0) (7) the vibration with an amplitude of 0.1 nm can be measured, and the
vibration response signal of the mass flow meter U-shaped tube
The amplitude of excitation force F0 = 450N, the initial phase φ0 = can be accurately collected.
0 rad, and the excitation frequency is the frequency of the The flowmeter was placed on the isolation table and the
corresponding order obtained by modal analysis. We only discuss boundary conditions are elastic support. Place 17 test points on one
the excitation mode of the flowmeter here. According to the results U-tube, as shown in Figs. 6 and 7. Based on the fundamental
of modal analysis and actual working conditions, the frequency principle of modal analysis, applying continuous excitation on the
range was set to 125∼150 Hz and the solution intervals was 25. U-tube and collecting the responding signals from a single point.
The frequency response function curve obtained by solving in the Since the excitation coil of the flow meter was in the middle of the
Z-axis direction is shown in Fig. 5. two tubes, the exciting point should be at this position. The LV was
Fig. 5 shows that with the increase of the excitation frequency, set to sample at the rate of 1 kHz. The LV measures the response
the U-shaped tube reaches the maximum amplitude at 134 Hz, and signal of each test point three times, and transmitted it to the
its maximum value is 0.1343 mm. controller by the laser head, collected it by the data acquisition
card, connects the PC, and recorded the response signal with
Labview.
3 Modal test analysis When the laser beam was continuously scanned and vibrated on
3.1 Laser vibration measurement technology the vibrating U-shaped tube, the vibration signal of the
corresponding test point was converted into a corresponding digital
The Laser Vibrometer (LV) basically uses the Doppler principle to voltage signal. The voltage signal was a vibration signal of the test
measure velocity at a point where its coherent laser beam is point at different positions and different moments. In order to
directed to. The reflected laser light is compared with the incident obtain the corresponding mode shape under the natural frequency

J. Eng. 3
This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
 Table 3 Comparison of the simulation and experimental
frequency and amplitude of the tube
Simulation Experiment Error, %
frequency, Hz 134.5 133.1 1.1
amplitude, mm 0.1343 0.1335 0.6

frequency of 134 Hz were obtained. Next, we used the LV to do a

modal experiment, comparing test mode result with simulation
mode analysis data, the finite element model is verified. The results
obtained from the verified finite element model allow a sufficiently
Fig. 8 Frequency response of test point high reliability in assessing the changes of flowmeter frequencies.

5 Acknowledgments
This research was supported by the National Youth Talent Support
Program held by Beihang University.
When the paper was completed, I would like to thank Professor
Fan Shangchun and Prof. Zheng Dezhi. The successful completion
of this paper cannot be separated from the careful guidance of the
two teachers and the help of the students in the lab. At the same
time, I also thank my advanced sensing and intelligent instrument
laboratory of Beihang University of Ministry of the Education
Innovation Team for providing me with an excellent learning
platform and experimental environment.
Fig. 9 Displacement amplitude of test point
6 References
vibration, a series of continuous voltage signals were collected and [1] Fan, S., Zhang, X., Zhang, Q., et al.: ‘The influence mechnism of nonlinear
imported into MATLAB. Fig. 8 is the frequency response diagram vibration and its reduction for coriolis mass flowmeter’, Sci. Technol. Eng.,
of the sampled signal after FFT, and Fig. 9 is the displacement 2005, 5, (9), pp. 569–571
[2] Ruoff, J., Hodapp, M., Kück, H.: ‘Finite element modelling of coriolis mass
amplitude of each test point obtained by fitting. flowmeters with arbitrary pipe geometry and unsteady flow conditions’, Flow
It can be concluded that the first-order natural frequency Meas. Instrum., 2014, 37, pp. 119–126
measured by the LV is 133.1 Hz, and the maximum displacement [3] He, J.: ‘Modal analysis theory and application’ (Shanghai Jiao Tong
amplitude is 0.1335 mm from Figs. 8 and 9, which is basically University Press, Shanghai, China, 2000)
[4] Yaushev, A.A., Taranenko, P.A., Loginovskiy, V.A.: ‘Study of the oscillation
consistent with the results of the finite element simulation. Here modes of a coriolis flowmeter using a parametric finite element model’,
gives the comparison of the simulation and experimental frequency Verified by the Results of Modal Testing. Procedia Eng.., 2016, 150, pp. 336–
and amplitude of the tube in Table 3. 340
Due to the objective difference between the finite element [5] Zheng, D., Liu, B.: ‘Analysis of the sensitivity of DN1 coriolis mass flow
sensor’. IEEE, Int. Conf. on Electronic Measurement & Instruments, Harbin,
model and the actual structure in the material physical and China, 2013, pp. 379–384
mechanical properties, damping characteristics, size and shape of [6] Raszillier, H., Durst, F.: ‘Coriolis-effect in mass flow metering’, Arch. Appl.
the structure, constraints etc. the simulation results and measured Mech., 1991, 61, (3), pp. 192–214
results are slightly different in the modal frequency and mode [7] Segerlind, L.J.: ‘Applied finite element analysis’ (John Wiley & Sons,
Hoboken, NJ, USA, 1984), p. 357
shape. [8] HanMin, S., QiPo, H.: ‘Mechanical vibration systerm’ (Huazhong University
of Science & Technology Press, Wuhan, China, 2013), pp. 23–27
4 Conclusions [9] Hong, V.M., Abe, M., Fujino, Y., et al.: ‘The eigensystem realization
algorithm for ambient vibration measurement using laser Doppler
Here, the vibration characteristics of the mass flowmeter were vibrometers’. Proc. of the IEEE American Control Conf., Arlington, VA,
USA, 2001, vol. 1, pp. 435–440
analysed based on the finite element simulation software. The first [10] Kojima, N., Hirayu, T., Mikami, M., et al. ‘Vibration intensity measurement
six natural frequencies and mode shapes obtained; only considered on U-shaped shell using laser vibrometers’, Trans. the Japan Soc. Mech. Eng.
the vibration mode of the flowmeter, that is, the second mode of C, 1998, 64, pp. 1654–1661
the simulation result. The U-tube was analysed with a harmonic [11] Castellini, P., Martarelli, M., Tomasini, E.P.: ‘Laser Doppler vibrometry:
development of advanced solutions answering to technology's needs’, Mech.
response analysis in the frequency range of 125–150 Hz. The Syst. Signal Process., 2006, 20, (6), pp. 1265–1285
frequency response and the maximum displacement amplitude at a

4 J. Eng.
This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)