Electrolysis promoted reductive amination of electron-deficient aldehydes/ketones: a green route to the racemic clopidogrel

Article abstract of DOI:10.1039/C8OB02500B

An electrocatalytic reductive amination of electron-deficient aldehydes/ketones was developed, which could be used in the synthesis of functionalized tertiary amines and large scale preparation of racemic clopidogrel. A plausible mechanism involving an iminium cation intermediate was proposed.

Full text of DOI:10.1039/C8OB02500B

IOP Conference Series: Earth and Environmental Science
PAPER • OPEN ACCESS
Related content
- On high efficiency operation of contra-
rotating axial flow pump with rotational
speed control toward effective energy
saving
Backflow vortex behaviours in contra-rotating axial
flow pump at low flow rates
L L Cao, S Watanabe, T Imanishi et al.
- Numerical simulation of internal flow in a
contra-rotating axial flow pump
To cite this article: D Zhang et al 2018 IOP Conf. Ser.: Earth Environ. Sci. 163 012036
S Momosaki, S Usami, S Watanabe et al.
- Blade rows interaction in contra-rotating
axial flow pump designed with different
rotational speed concept
View the article online for updates and enhancements.
L L Cao, S Watanabe, T Imanishi et al.
This content was downloaded from IP address 5.101.220.198 on 21/01/2019 at 08:45

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
Backflow vortex behaviours in contra-rotating axial flow
pump at low flow rates
D Zhang1, T Tsuneda1, Y Katayama2, S Watanabe2, S I Tsuda2 and A Furukawa3
1Graduate School of Engineering, Kyushu University, Fukuoka, Japan
2Department of Mechanical Engineering, Kyushu University, Fukuoka, Japan
3National Institute of Technology, Oita College, Oita, Japan
fmnabe@mech.kyushu-u.ac.jp
Abstract. Backflow usually exists at the inlet of rotors of many turbomachines at low flow
rates. In counter-rotating rotors applied for the axial flow pump, such vortical structures are
also able to form. In our previous researches, some broad-banded pressure fluctuations in low
frequency range have been observed between front and rear rotors, but we have not yet been
able to explain what causes such phenomenon. In this study, in order to find out the causes of
low frequency components at low flow rates, unsteady numerical simulations for the whole
front and rear rotors are conducted, and casing pressure is experimentally measured at the inlet
and outlet of front rotor and the inlet of rear rotor. It is found that vortical structures exist
between front and rear rotors at below 40% of design flow rate. These vortices seem to be the
result of shear layer instability at the impingement location of the exiting flow from front rotor
to the backflow of rear rotor. The behaviours of these backflow vortices and their interaction
with front rotor contribute the low frequency components in pressure fluctuations observed at
low flow rates.
1. Introduction
It is well known that backflow occurs at the inlet of pump/fan/compressor rotors at low flow rates, and
the backflow region increases with the decrease of flow rate. Yokota et al [1] have observed vortex
structures at the inlet of an inducer in experiment, and these vortices have been confirmed to be caused
by the strong circumferential shear flow between the swirling backflow and the oncoming main flow.
A numerical research based on Large Eddy Simulation (LES) is conducted by Yamanishi et al [2] to
investigate the backflow vortex structure at the inlet of an inducer. Depending on the number of
vortices and their moving velocities in circumferential direction, the backflow vortices induce pressure
fluctuations with high frequency, which are afraid to cause some structural troubles and mechanical
vibrations.
In contra-rotating axial flow pumps, such vortical flow structures may be also able to form.
Actually, in our previous study [3] aiming at understanding the pressure fluctuations due to rotor-rotor
interactions, some broad-banded pressure fluctuations in low frequency range are observed in the gap
between the front and rear rotors. Since the flow is too complex, it is very difficult to understand the
flow mechanism causing such pressure fluctuations only through experiment. Therefore, in this study,
unsteady numerical simulations considering the whole front and rear rotors are conducted at different
low flow rates. Special emphasis is placed on the vortical structures between the front and rear rotors,
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
1

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
which may induce the above-mentioned low frequency pressure fluctuations. By conducting
experiments below 30% of design flow rates, fluctuating pressure signals are measured at the inlet of
front rotor, outlet of front rotor and inlet of rear rotor.
In the present paper, by using numerical results, we will explain that these vortices between front
and rear rotors originate from the boundary between main flow and backflow regions. The vortex
behaviors will be used to understand the frequency-domain pressure signals measured in experiments.
2. Methods
The test contra-rotating axial flow pump has been designed with different speed of front and rear
rotors [4]. By considering higher cavitation performance of rear rotor, as well as the design flow rate
Qd =70 L/s and total design head Hd,t =4 m, the optimized rotational speed combination was
determined: rotational speed of front rotor Nf =1311 rpm, rotational speed of rear rotor Nr =1123 rpm.
The diameter of hub is 100 mm, while the diameter of casing Dc =200 mm, and the blade tip clearance
is 1mm. The blade numbers of front and rear rotors are 4 and 5 respectively. For more detail of the
design were shown in the reference [4].
2.1. Numerical simulation
Figure 1. Numerical model
Numerical simulations are conducted by using a commercial CFD code, ANSYS CFX 16.2. As shown
in figure 1, the whole front and rear rotors are modelled. The inlet boundary is located at 4Dc upstream
of the leading edge of front rotor, while the outlet boundary is located at 1.3Dc downstream of the
trailing edge of rear rotor. Furthermore, the inlet boundary type is chosen for the inlet boundary with
specifying the mass flow rate, and the opening boundary type is chosen for the outlet boundary with
constant static pressure. In the numerical model, two domains have been built up: Front Rotor and
Rear Rotor. The total number of nodes in both of domains is over 1.5 million, and at the same time, 10
nodes are radially located in the blade tip clearance to reproduce the tip leakage flow.
Even though the Large Eddy Simulation (LES) should be more appropriate to predict the unsteady
flow field. The pump performance and internal flow have been well predicted by using the unsteady
Reynolds Averaged Navier Stokes (RANS) based simulation as shown in our previous study.
Therefore, the unsteady RANS based simulation is again used in the present study. As will be shown
later, the RANS based numerical results can provide us with sufficient information to understand the
2

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
frequency distribution of pressure fluctuation measured in experiments. Furthermore, the k-ω based
SST (Shear Stress Transfer) turbulence model is selected to well predict the flow separation.
The time step in this study is determined with the following equation:
(1)
This time step represents the time duration of 0.5 degree relative rotation of front and rear rotors. All
the simulations have been conducted over 5 full relative revolutions, where stable flow variable
fluctuations are achieved. The quasi-steady results are used to analyse the vortical structures.
2.2. Experimental setup
In order to measure the pressure signals at the inlet of front rotor (Ch.1), outlet of front rotor (Ch.2)
and inlet of rear rotor (Ch.3), three pressure transducers have been set in the three positions displayed
in figure 2. These pressure taps will collect 220 (1048576) samples with 10kHz sampling frequency.
All the pressures are processed for the Fast-Fourier Transformation (FFT). In order to achieve fine
frequency resolution, the whole samples of pressures are divided into 63 parts, and each part has 50%
overlap with the adjacent part, then average spectral of every part’s FFT results is obtained.
Figure 2. Pressure taps position
3. Results and discussions
3.1. The origin of vortices
As shown in figure 3, vortex structures can be seen in the gap region between front and rear rotors.
These vortex structures are observed only at the flow rates below 27 L/s (about 40% of design flow
rate). The circumferential velocity normalized by front rotor tip velocity are used to display the main
flow and backflow regions in the figure. The positive value represents the flow swirling in front rotor
rotating direction, which shows the flow is exiting from the front rotor. On the other hand, the negative
value represents the flow swirling in rear rotor rotating direction, which indicates this region is the
backflow region from the rear rotor. As we can see, the vortex cores just exist on the boundary of main
flow region (red colour) and backflow region of rear rotor (blue colour). This indicates these vortices
may be the result of shear layer instability between the main flow and the rear rotor backflow. In order
to confirm the origin of these vortex structures, we compare the axial positions of backflow region
boundary and vortex cores. At the same time, the observed circumferential moving velocity of vortex
cores will be compared with the calculated velocity by using circumferential velocity of main flow and
backflow of rear rotor.
3

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
Figure 3. Vortex structures and normalized circumferential velocity distribution on 0.975 span at
30% of design flow rate
Figure 4. Contours of averaged normalized circumferential (a) and axial (b) velocities
4

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
3.1.1. Comparison of axial positions of backflow boundary and vortex cores. In order to better
understand the development of backflow region, especially the backflow region of rear rotor, axial
velocity is also normalized by the front rotor tip velocity. Small range (-0.001~0.001) is defined to
emphasize the regions with flow moving in opposite directions (swirling in front and rear rotors
rotating directions, moving in upstream and downstream directions). As shown in figure 4, red-color
regions can be seen as regions with positive normalized velocities, while blue-color regions can
represent regions with negative normalized velocities. Both normalized circumferential and axial
velocities are not only arithmetically averaged in one relative revolution’s duration but also
circumferentially averaged. Figure 4 (a) displays the averaged normalized circumferential velocity
distributions on the r-z plane in whole domains. The backflow region of front rotor grows with the
reduction of flow rate, which is similar to the situation of an inducer [1]. However, the backflow
region of rear rotor seems to have no significant change until the flow rate is reduced to 7 L/s (10% of
the design flow rate). As shown in figure 4 (b) displaying the axial velocity distribution, a reverse flow
region exists at the front rotor hub area, and with the reduction of flow rate, the reverse flow region
moves upstream. We also notice that the size of the hub reverse flow grows rapidly at the flow rates
from 40% to 20% of design flow rate, while the size of the hub reverse flow has no obvious change
with the further decrease of flow rate. This reverse flow region grows rapidly from the flow rate of 27
L/s to 14 L/s, which seems to restrain the growth of the backflow region of rear rotor. When the flow
rate is decreased to 7 L/s, this hub reverse flow region may grow to its limitation, and as the hub
reverse flow region moves upstream, the suppression of hub reverse flow is weakened, then the rear
rotor backflow region develops significantly to fill the space with the further reduction of flow rate.
Table 1. Averaged backflow
region boundary axial position
Flow
rates [L/s]
Boudary axial
position z [cm]
7.264
27
21
14
7
7.257
7.336
5.887
0
4.366
Figure 5. Axial position of backflow vortex cores on casing
In order to specify the axial location of vortices, we investigate the vortex cores from the limiting
streamlines on casing walls for over 1000 steps (720 steps for one relative revolution of rotors), the
vortex cores are checked in every 100-step, then over 60 samples are obtained at each flow rate.
Figure 5 displays the axial position of backflow vortex cores. The y-axis represents the probability
density function of the location of vortex cores. We can find that the peak of axial location of
backflow vortex cores stay near the leading edge of the rear rotor at the flow rates over 14 L/s. They
stay at the nearest position to the rear rotor leading edge at 14 L/s. With the flow rate decreased from
7 L/s to 0 L/s, the axial position of vortex cores greatly moves toward the exit of front rotor.
We also compare the axial position of vortex cores with the location of rear rotor backflow
boundary near the casing. In order to determine the location of the backflow region boundary, we
define it by the region with small arithmetically-averaged normalized circumferential velocity between
-0.001 and 0.001. Table 1 shows the results at each flow rate. We can see that this averaged backflow
region boundary axial position agrees well with the axial position of vortex cores.
5

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
3.1.2. Comparison of velocities. Yokota et al [1] has found that the backflow vortex moving velocity is
about the average of the maximum circumferential velocity in the main flow and the backflow.
Therefore, in our simulations, the actual moving velocity of vortices, which call here “vortex observed
velocity”, is obtained by directly observing the vortex cores in every 100-step for total 1000 steps. At
the same time, the averaged maximum circumferential velocity just upstream and downstream of
vortex core is calculated, which is herein called “vortex calculated velocity”. In order to make
circumferential velocity available to be measured, both of observed and calculated velocities are taken
at cylindrical surface with 0.975 span.
Figure 6. Comparison of the velocity observed and calculated with circumferential velocities of main
flow and backflow
Figure 6 displays the comparison of probability density function of observed and calculated
velocities at the flow rates below 40% of design flow rate. The vortex moving velocity has been
normalized by front rotor tip velocity, therefore, the positive value denotes the vortex moving in the
front rotor rotating direction, while the negative value represents the vortex moving in rear rotor
rotating direction. As shown in the figure, the calculated velocity agrees well with the observed
velocity from 40% to 20% of design flow rate, while at 10% of design flow rate, the peak of observed
velocity distribution seems to have the opposite moving direction shown by calculated velocity
distribution. At 10% of design flow rate, the vortex moving velocity shows multi-peaks distribution in
observed moving velocity. As will be shown later, these observed moving velocities of vortices are
related to the unsteady and aperiodic behaviour of vortices.
Based on the comparison of axial position of vortex cores and backflow region boundary as well as
on the comparison of observed and calculated velocities, we believe these vortices occurring between
front and rear rotors are the result of shear layer instability between main flow from front rotor and
backflow of rear rotor.
6

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036
1234567890 ‘’“”
doi:10.1088/1755-1315/163/1/012036
3.2. Behaviours of backflow vortex
In order to be similar with the condition in experiment, these backflow vortices are also investigated
on the casing wall. Two main types of backflow vortex behaviours are observed: these backflow
vortices move circumferentially at the flow rates below 40% of design flow rate except 0 L/s, while
the backflow vortices at 0 L/s fluctuate in high frequency around the fixed position.
3.2.1. Circumferential propagation of backflow vortex. The backflow vortices are found to
circumferentially propagate on the casing at the flow rates from 40% to 10% of design flow rate.
Figure 7 displays their moving velocity distributions. The central moving velocity can be found in
27 L/s, 21 L/s and 14 L/s, which indicates the stable circumferential movement of backflow vortex.
We can find that there is an obvious moving direction change of central moving velocity with the
decrease of flow rate. All of the central moving velocity is below 10% of front rotor tip velocity. On
the other hand, the multi-peaks occur when flow rate is reduced to 7 L/s, which implies the aperiodic
backflow vortex behaviour.
Figure 7. Normalized moving velocity distribution of backflow vortices on casing
Yokota et al [1] have found that the number of backflow vortices in an inducer is decreased with
the reduction of flow rate, there is no obvious change of the number of backflow vortices in our
simulation, about 5~9 at each flow rate. In our case, shearing velocity, that is defined by the
circumferential velocity difference across the backflow boundary, is not significantly changed with the
reduction of flow rate, which may be the reason for almost constant number of vortex forming.
Otherwise, this may be caused by the limitation of RANS based simulation in unsteady flow
prediction.
3.2.2. Vortex vibration in high frequency. Figure 8 shows the typical example of trajectory of vortex
core on the casing at the flow rate of 0 L/s. The yellow crosses display the positions where vortex core
changes its moving direction. The vortex shown in figure 8 moves from R1 to F1, then R2, F2, R3, F3,
R4, F4… This moving trajectory seems like the backflow vortex vibrating at certain fixed position.
Table 2 shows the trajectory information of the vortices we observed at 0 L/s, and we find these
vortices take 280 ~360 steps’ time in one cycle, which results in high frequency fluctuation with
81Hz~104Hz.
7

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036
1234567890 ‘’“”
doi:10.1088/1755-1315/163/1/012036
3.3. Experimental results
Some low frequency components occur in pressure measured in experiments at low flow rates. Figure
9 displays the FFT processed results of pressure measured at Ch.1-Ch.3 on the casing shown in figure
2.
Table 2. Detailed trajectory information
Vortex No.1
1
2
3
4
Average steps
one cycle
280
R
8240
8320
1
8560
8680
2
8800
9000
3
F
340
Vortex No.2
4
R
8120
8320
1
8480
8640
2
8800 9120
8920
333
300
F
Vortex No.3
3
4
R
8320
8160
1
8680
8440
2
9000
340
320
F
9760 9120
Vortex No.4
3
4
R
8240
8120
1
8520
8360
2
8880
320
293
F
8680 9000
Vortex No.5
3
4
Figure 8. Trajectory of one vortex at flow rate
0L/s
R
8160
8280
1
8440
8560
2
8840 9120
9000
320
360
F
Vortex No.6
3
4
R
F
8080
8200
8360
8440
8720 9040
8920 9100
320
300
8

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
Figure 9. Pressure measured in frequency domain
Table 3. Frequency components
1
2
3
4
5
6
87.4Hz, 93.6Hz
BPF of front and rear rotors
174.8Hz, 187.2Hz 2nd harmonics of BPF of front and rear rotors
181Hz
81.2Hz
Black curves
Interaction of front and rear rotors
Interaction of 2nd harmonics of rear rotor and BPF of front rotor
Backflow vortex behaviour
Interaction of BPF of front rotor and backflow vortex behaviour
The noticeable frequency components are summarized in table 3. In the previous numerical results,
backflow vortices move circumferentially in the velocity below 10% of front rotor tip velocity at flow
rates 21 L/s and 14 L/s. By considering the number of vortices we observed in simulation: 5~9, their
central frequency which can be calculated by (number of vortices) *(central moving
velocity)/(circumference of casing) is to be below 20 Hz. In figure 9, the central frequencies smaller
than 10 Hz can be seen only for Ch.3 at flow rates of 21 L/s and 14 L/s. This seems to be caused by
the vortices sweeping near Ch.3. The central frequency at 14 L/s (10 Hz) is higher than that (6 Hz) at
21 L/s, which seems to be the result of distance between pressure tap Ch. 3 and backflow vortex cores
position. In our simulation, compared with the axial position of vortex cores on casing at flow rate
21 L/s, vortex cores at 14 L/s exist closer to rear rotor. We also notice that the magnitude of Ch.3 at
flow rate 14 L/s is larger than that at 21 L/s, which also indicates the vortex at 14 L/s is closer to Ch.3.
With the flow rate decreased to 7 L/s, pressure fluctuation is becoming more pronounced in the
signal of Ch.2. In the numerical results, the backflow region of rear rotor grows rapidly when the flow
rate is reduced to 7 L/s, and the backflow vortices also move upstream approaching to the location of
Ch.2. Moreover, as the multi-peaks moving velocity distribution shows, these vortices slowly move
with aperiodic behaviour, resulting in the low frequency pressure fluctuations near 0 Hz and weak but
high frequency broad-banded fluctuations around 90 Hz.
In our simulation of flow rate 0 L/s, the backflow vortices fluctuate in a very high frequency at
fixed positions as has been shown in figure 8. These vortices exist near the front rotor blade passage,
indicating that these backflow vortices interact with the front rotor blades. As shown in figure 9, high
frequency component occurs at flow rate 0 L/s, which should be the result of backflow vortex
behaviours. As backflow vortices are too close to the front rotor blades, the nonlinear interaction
between front rotor and these backflow vortices should result in the low frequency components at 0L/s.
4. Conclusions
In this study, in order to find out the causes of low frequency fluctuations observed at low flow rates in
a contra-rotating axial flow pump, unsteady numerical simulations for whole front and rear rotors have
been carried out. Vortex structures have been analysed especially in the region between front and rear
rotors. Main findings are summarized as follows:

Vortical structures observed between front and rear rotors seem to be the result of shear layer
instability of rear rotor backflow.

Backflow vortices between front and rear rotors locate at the boundary between main flow
from front rotor and the backflow of rear rotor. They circumferentially move in a central
velocity below 10% of front rotor tip velocity. This velocity roughly agrees with the averaged
circumferential velocity across the backflow boundary.


Aperiodic behaviours of backflow vortex have been observed below 10% of design flow rate,
and backflow vortices at 0 L/s seem to fluctuate in a very high frequency around fixed
positions.
The low frequency components of pressure fluctuation measured in experiments seem to be
the result of the slow circumferential movements of multiple backflow vortices. However, at
9

Asian Working Group- IAHR’s Symposium on Hydraulic Machinery and Systems
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 163 (2018) 012036 doi:10.1088/1755-1315/163/1/012036
1234567890 ‘’“”
the flow rate of 0 L/s fluctuating motions of vortices with high frequency seem to produce the
pressure fluctuation in high frequency and the interaction of front rotor and backflow vortices
causes the low frequency components.
References
[1] Yokota K, Kurahara K, Kataoka D and Tsujimoto Y 1999 A study of swirling backflow and
vortex structure at the inlet of an inducer JSME International Journal B 42 3 pp451-459.
[2] Yamanishi N, Fukao S, Qiao X, Kato C and Tsujimoto Y 2007 LES simulation of backflow
vortex structure at the inlet of an inducer ASME Journal of Fluids Engineering 129 pp587-
594.
[3] Cao L, Honda H, Yoshimura H, Watanabe S and Furukawa A 2014 Experimental investigation
of blade rows interactions in contra-rotating axial flow pump Proc. ASME 2014 4th Joint
US-European Fluids Engineering Division Summer Meeting FEDSM2014-21901.
[4] Cao L, Watanabe S, Momosaki S, Imanishi T and Furukawa A 2013 Low speed design of rear
rotor in contra-rotating axial flow pump International Journal of Fluid Machinery and
Systems 6 2 pp105-112.
10