Bioprocess Biosyst Eng (2016) 39:169180

DOI 10.1007/s00449-015-1501-9

ORIGINAL PAPER

Investigation on novel raceway pond with inclined paddle wheels

through simulation and microalgae culture experiments
Fanxue Zeng1 Jianke Huang2 Chen Meng1 Fachao Zhu1 Jianpei Chen1
Yuanguang Li2

Received: 26 October 2015 / Accepted: 30 October 2015 / Published online: 13 November 2015
Springer-Verlag Berlin Heidelberg 2015

Abstract The open raceway ponds are nowadays the

most used large-scale reactors for microalgae culture. To
avoid the stacking of microalgae, the paddle wheels are the
most widely used to circulate and mix the culture medium.
In this paper, a numerical simulation using computational
fluid dynamics (CFD) was used to investigate the hydrodynamic characteristics of open raceway ponds with different types of paddle wheels (the traditional paddle wheels
and the novel paddle wheels with specially inclined angle
of the blades). The particle image velocimetry (PIV) was
used to validate the reliability of the CFD model. The CFD
simulation results showed that the novel raceway pond
with 15 inclined angle of the blades had the best mixing
efficiency under the same power consumption. Lastly, the
results of microalgae culture experiments showed that the
growth rates of Chlorella pyrenoidosa in the novel raceway
pond with 15 inclined angle of the blades were higher than
those in the traditional reactor. The results of the culture
experiments and CFD simulations were identical with each
other. Therefore, a novel paddle wheel with 15 inclined

angle of the blades was obtained for better microalgae

cultivation.
Keywords Open raceway pond  Computational fluid
dynamics  Particle image velocimetry  Optimal mixing
system  Microalgae cultivation
List of symbols
L
Length of open raceway ponds, m
W
Width of open raceway ponds, m
H
Height of open raceway ponds, m
a
Inclination angle to agitation axis for paddle wheel
blades,
TKE Turbulent kinetic energy, J kg-1
U
Average velocity, m s-1
Uz
Average velocity magnitude along the light
direction, m s-1
Ux
Average velocity magnitude along the horizontal
direction, m s-1
Dz
The volume percentage of dead zone, %
h0
Depth of deionized water, mm
H
Depth of water, m
Vtip
Impeller tip velocity, m s-1

F. Zeng and J. Huang contributed equally to this study and share first
authorship.
& Jianpei Chen
chenjianpei@ecust.edu.cn
Yuanguang Li
ygli@ecust.edu.cn
1

State Key Laboratory of Chemical Engineering, East China

University of Science and Technology, Meilong Road 130,
Mail box 373, Shanghai 200237, Peoples Republic of China

State Key Laboratory of Bioreactor Engineering, East China

University of Science and Technology, Meilong Road 130,
Mail box 301, Shanghai 200237, Peoples Republic of China

Introduction
Due to the extensive application in food, drug, energy
sources, and CO2 mitigation [13], microalgae production
has received increasing attention over the last few years.
Microalgae culture systems can be broadly classified into
open systems where the culture medium is directly exposed
to the environment and closed photobioreactor (PBR)

123

170

systems where the culture is enclosed in a transparent

vessel. However, open raceway ponds have been the most
common choice for outdoor algal production because of
their less cost of building, operation and low mixing energy
requirement [4, 5].
As a low cost and high efficiency method, CFD (computational fluid dynamics) technology has been widely
used to design various types of chemical reactors and
bioreactors [6, 7]. CFD is gradually becoming the most
promising method in the future PBRs design and optimization, but the published studies are basically for closed
PBRs [813]. For example, Pruvost et al. [14] investigated
hydrodynamic conditions in a torus PBRs, and validated
the numerical model with experimental measurements
using PIV. Wu et al. [15, 16] studied the mixing condition
and light intensity histories of cells in a series of spiral tube
PBRs. However, there are few reports in literature on open
raceway ponds with the application of CFD. To choose a
proper turbulence model is very important for the turbulent
kinetic energy distribution prediction during the CFD
simulation process. The three main strategies used for
turbulent flow simulation are direct numerical simulation
(DNS), Reynolds averaged NavierStokes (RANS) and
LES models. The DNS method is not practical due to the
massive computational resources required to resolve all the
length and time scales in the turbulent flow spectrum.
Therefore, the ke turbulence model (RANS) and LES
model were chosen for the validation experiment in the
following PIV test.
With the development of laser technology, image
acquisition technology and computer technique, particle
image velocimetry (PIV) has been widely used in fluid
mechanics measurement. Compared with other traditional
measuring methods, the biggest advantage of PIV is to
achieve plane measurement [17]. In recent years, more and
more researchers use PIV to verify the validity of the CFD
analyses [1821]. However, few researchers have applied
this technique to check the rationality of CFD model in the
design of PRBs, let alone use for the open raceway ponds
[22, 23].
To improve the efficiency of microalgae cultivation in
open raceway pond, optimization of open raceway pond
structure is necessary for a better mixing. Although various
measures have been proposed, paddle wheels remain by far
the most practical means to mix and circulate the culture
medium [24, 25]. However, the study of the paddle wheel
structure is of very little utility now. Rainier Hreiz et al.
[26] put forward a new paddle wheel configuration which
has non-aligned blades. It was the only reported
research on the design of paddle wheel in raceway pond.
In this study, numerical simulation validated by PIV was
adopted to investigate hydrodynamic characteristics in
open raceway pond. In order to improve the microalgae

123

Bioprocess Biosyst Eng (2016) 39:169180

cultivation efficiency in open raceway pond, a novel paddle

wheel was proposed and further optimized for the first
time. Flow characteristics, including velocity field, volume
percentages of dead zone, and average velocity magnitude
along the light direction were compared in detail. Lastly,
the optimum paddle wheel was determined according to
numerical analysis and further verified by photoautotrophic
cultivation of microalgae.

Materials and methods

Materials for experiment
Materials for PIV experiment
A two-dimensional PIV system (Fig. 1) of ILA Company
(Germany) was used in the experiment; the system was
mainly composed of a pulsed laser light source, lens set,
high frame rate CCD camera, synchronizer and a computer.
PIV can precisely freeze the flow in a certain time interval.
By cross-correlation algorithm, the displacement of a large
number of tracer particles from the flow field was obtained
in the time interval, and then the velocity plane can be
obtained.
The open raceway pond for PIV test was made of perspex and its dimension is shown in Fig. 2a. In order to bate
the refraction of laser caused by the curved wall, the same
fabric square groove was installed in the bend of the open
raceway ponds and its liquid level was the same as open
raceway ponds. The paddle wheel was installed on one side
of the raceway pond at a height of 0.35 m. It was driven by
the motor the speed of which can be adjusted. The real
picture of raceway pond used for PIV test is shown in
Fig. 2b.

Fig. 1 The schematic view of PIV system

Bioprocess Biosyst Eng (2016) 39:169180

Fig. 2 a Structure and
dimension diagram of small
raceway pond for PIV
experiment; b Picture of real
raceway pond for PIV
experiment

(a)

171

monitoring points

(b)

The open raceway pond used for the optimization

of the paddle wheels

Numerical methods and computational models

Numerical methods for PIV experiment

The size of open raceway pond (Fig. 3a) for the optimization of the paddle wheel is of 4.5 m length (L), 1.9 m
width (W) and 0.35 m height (h). The paddle wheel was
composed of six blades. As shown in Fig. 3b, a is the
inclined angle of the blades to the axis. Because of the
crossed arrangement of the adjacent blades, the maximum
angle is 15. The novel paddle wheel was optimized by
changing the inclined angle of the blades according to
numerical analysis and further verified by photoautotrophic
cultivation of microalgae.

Fig. 3 a Structure and

dimension diagram of open
raceway pond used for the
optimization of the paddle
wheel; b The structure diagram
of paddle wheel

Two-dimensional instantaneous velocity of the flow field

can be obtained by the PIV, and then the average speed can
be obtained by averaging these instantaneous velocities of
different times.
The equation of turbulence kinetic energy (TKE) is as
follows:
k ~
u2x u~2y u~2z =2

where u~x ; u~y ; u~z is the fluctuating velocity in the three

directions.

(a)

(b)
monitoring points

123

172

Bioprocess Biosyst Eng (2016) 39:169180

As a result of the two-dimensional PIV system, the

fluctuation velocity in another direction can be obtained
using isotropic hypothesis [27, 28]:
u~2y ~
u2x u~2z =2

Through Eqs. (1) and (2), the TKE can be expressed as

follows:
k 3 ~
u2x u~2z =4

Computational models for CFD simulation

Mesh generation and CFD model setting The raceway
pond was divided into dynamic area and static area, the
area containing the paddle wheels was considered as
dynamic area, the rest was considered as static area. It was
meshed with ANSYS ICEM 12.0. The hexahedral mesh
and tetrahedral mesh were created for static and dynamic
areas, respectively. In the simulation of raceway pond used
for PIV test, the total count of the grids was different
because of the different turbulence models. LES simulation
needs more grids to achieve accurate prediction. Therefore,
its total count of the grids was 600,000 while it was only
200,000 based on ke turbulence model. The total grids
count of the raceway pond used for the paddle wheel
optimization was 300,000. The average mesh cell size of k
e turbulence model and LES was 14 and 10 mm, respectively. Quality check of grids is required after the mesh
generation. All the quality of grids can meet the requirement of the simulation.
The mixing performance was numerically simulated by
ANSYS CFX 12.0. Transient rotorstator method was used
for coupling the dynamic and static areas. Scalable wall
function was used for the laminar flow near the wall. The
top of raceway pond was considered as opening
boundary which allows the gas; liquid goes in and out. Free
surface model was chosen to simulate the surface fluctuation. The heat transfer option was set as isothermal and
liquid temperature was 30 C. Temperature of air was set
as 30 C and reference pressure was atmospheric pressure.
The initial air velocity was set as 0 m s-1. The viscosity of
the algal medium was 0.91.2 mPa s, and the density was
10001030 kg m-3 [29], which was similar to the water.
Therefore, the liquid water was chosen as the fluid material. The unsteady model was selected. The residual should
be less than 10-4. The torque of paddle wheel and velocity
of some points were monitored. The location of the monitoring points is shown in Figs. 2 and 3 when the monitoring target reached a stable value; it indicated that the
simulation can be finished.

123

Governing equations
The governing equations based on ke turbulence model
can be written by the following equations [30]:
Continuity equation
op
r  qU 0
ot

Momentum equation
opU
r  qUU  r  leff rU
ot
r  q r  leff rUT B

Turbulent kinetic energy equation and turbulent dissipation rate equation


 
opk
lt
r  qUk r  l
rk Pk  qe
ot
rk
6

 
ope
l
e
r  qUe r  l t re Ce1 Pk
ot
k
re
 Ce2 qe
7
where Ce1 1:44, Ce2 1:92, rk 1:0, re 1:3.
The governing equations based on LES turbulence
model can be written by the following equations [31]:
o
o
op
pui
pui uj 
ot
oxj
oxi  

o
oui ouj 2 out

 dij
oxj
oxj oxi 3 oxt
osij

oxj
8
We can get the sij by sub-grid stress model. In this
study, the standard SamagorinskyLily model was chosen
to solve it.
Calculation of mixing characteristic parameters
Average velocity magnitude along the light direction (Uz),
average velocity (U), turbulent kinetic energy (TKE) and the
volume percentage of dead zone (Dz) were regarded as main
parameters to evaluate PBRs mixing characteristics in the
present study. For microalgae culture, light is the most
important factor affecting biomass composition, growth rate
and product formation [9, 32]. In the research of Huang et al.
[33], the growth rate of Chlorella pyrenoidosa has a direct
correlation with fluid velocity along the light attenuation
direction. Thus, average velocity magnitude along the light

Bioprocess Biosyst Eng (2016) 39:169180

direction is a critical parameter to evaluate the performance of

the raceway ponds. With the increasing of average velocity
(U) and turbulent kinetic energy (TKE), the mixing efficiency
of raceway pond becomes better [34], however, power consumption also increase accordingly. Therefore, we should try
to increase the values of the U and TKE with limited power
consumption. During microalgae cultivation process, the
location where the liquid velocity is low (we call it dead zone)
can easily cause algae to stacking and death. To avoid this
phenomenon of microalgae cultivation, the proportion of dead
zone should be minimized. The region where the liquid
velocity is lower than 0.05 m s-1 is characterized as dead
zone [29, 35]. All parameters were gained from volume-average values of raceway pond, such as U value could be
expressed as follows:
Z
1 v
U
9
jU j dv
v 0

173

Plane 1 was located in the outflow of paddle wheel rotation

direction while plane 2 was situated on the flow channel
without paddle wheel. The light sheet which was formed by
pulse laser through the lens group and vertically shoot into
the top of raceway pond, then the particle image of flow
field was caught with CCD camera. The collected PIV data
was post-processed by Tecplot software.
Culture experiments conditions
In outdoor conditions, Photoautotrophic cultivation of
Chlorella pyrenoidosa (obtained from Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China) was
used to evaluate the performance of raceway ponds with
different paddle wheels. Initial biomass concentration was
0.29 g L-1, and the depth of culture was 0.15 m. Microalgae
was cultivated in the modified FSi medium [36] under

where V was the volume of PBR and jU j was the absolute

value of average velocity within the limited control area.
Experiment conditions
PIV experiments conditions
The medium was deionized water with a depth (h0) of
0.2 m. The Al2O3 particles with diameters of 10 lm were
selected as tracers in PIV experiment. Paddle wheel was
stirred at 10 rpm. As shown in Fig. 4, Cartesian coordinate
system was used in the test and its origin was in the right
side, the middle position of the raceway pond bottom.
Plane 1 and plane 2 were chosen as the detection planes.

Fig. 4 The position of detective planes for PIV test

Fig. 5 Contours of average velocity with PIV and simulation based

on ke turbulence model. a the results in plane 1; b the results in
plane 2

123

174

Bioprocess Biosyst Eng (2016) 39:169180

natural illumination. Mixing was achieved by impeller stirring. Pure CO2 was injected into pond through a small
sparger located after paddle wheel zone to maintain the pH of
microalgae at approximately 7.0. And gas flow was just
11.5 L min-1. Thus, aeration influence could be negligible.
Biomass concentration was measured in the photoautotrophic culture using a method reported in our previous
study [36]. The experiments were carried out under the same
conditions (rotation speed at 10 rpm, pH, light intensity,
temperature and wind speed) to compare the culture efficiency of raceway ponds with different paddle wheels.

Results and discussion

Comparison of the PIV measurement and CFD
simulation
Comparison of the velocity field
To validate the numerical predictions, flow fields within the
raceway pond were experimentally measured by PIV. The
ke model was chosen as the turbulence model in the

1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0

PIV
Simulation
z/H=0.5
Uz / Vtip

Ux / Vtip

(a)

simulation process. As shown in Fig. 5, contours of average velocity with PIV and simulation were compared under
various x/L (dimensionless distance between investigated
points to the origin of the raceway pond (x) by the length
(L) of raceway ponds) and z/H (dimensionless height of
liquid by the height (H) of water) values. As shown in
Fig. 5a, the fluid was stirred by the paddle wheel, and
thrown back into the raceway pond. As a result, the highvelocity area was formed near the water surface while the
velocity was low in other region. The high-velocity area
(U/Vtip [1.2 dimensionless average velocity with tip
velocity) was located at the place where x/L range from
0.66 to 0.82 and z/H range from 0.85 to 1, respectively. The
fluid moved directly into the curved channel and then
turned downward rapidly when it reached the wall. It
formed the vortex which was located in the place where x/
L was 0.91 and z/H was 0.1. As shown in Fig. 5b, the fluid
turned downward rapidly because of the wall barrier in the
curved channel, and it rebounded at the bottom, then it
slowly restored the horizontal flow after it got out of the
curved area. The high-velocity region was formed at the
place where x/L range from 0.75 to 0.9 and z/H range from
0 to 0.25, respectively. Main characteristics of these

0.6

0.7

0.8
x/L

0.9

1.0

(b)

0.4
PIV
0.3
Simulation
0.2
z/H=0.5
0.1
0.0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-0.7
-0.8
0.6
0.7

1.0
0.8
0.6

0.8
x/L

0.9

1.0

PIV
Simulation
z/H=0.5

Ux / Vtip

Uz / Vtip

0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.8
x/L

0.2

0.4

0.6
x/L

0.8

1.0

Fig. 6 Compare the PIV values and simulation results of Ux and Uz. a the values of Ux/Vtip and Uz/Vtip in plane 1; b the values of Ux/Vtip and Uz/
Vtip in plane 2

123

Bioprocess Biosyst Eng (2016) 39:169180

175

velocity contours between PIV measured and CFD simulation were greatly consistent.
As shown in Fig. 6, a specific z/H value (0.5) was used
as a further work to compare the measured values and
simulation results of Ux and Uz. It showed that the
changes of Ux/Vtip (dimensionless average velocity

magnitude along the horizontal direction by tip velocity)

and Uz/Vtip (dimensionless average velocity magnitude
along the light direction by tip velocity) were the same
for PIV and CFD simulation with the increase x/L values.
Thus, the numerical predictions were preferably validated
by the PIV test. It was clear that the established CFD

Fig. 7 Contours of the turbulent kinetic energy (TKE) distribution with PIV and simulation which is based on ke turbulence model and LES
model, respectively. a the results in plane 1; b the results in plane 2

(a)

(b)

0.09
0.08

0.06

0.12
0.10

2
k / Vtip

k / Vtip2

0.07

0.05
0.04

PIV
model
LES model
z/H=0.5

0.08
0.06
0.04

0.03
0.02

0.14

PIV
model
LES model
z/H=0.5

0.02
0.6

0.7

0.8
x/L

0.9

1.0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
x/L

2
2
Fig. 8 Compare the PIV values and simulation results of turbulent kinetic energy (TKE). a the values of k=Vtip
in plane 1; b the values of k=Vtip
in plane 2

123

176

Bioprocess Biosyst Eng (2016) 39:169180

model for the turbulent kinetic energy (TKE) distribution

had a great difference with the PIV results, while the LES
results were almost the same.
As shown in Fig. 8, a specific z/H value (0.5) was used as a
further work to compare the measured values and simulation
2
results of TKE. It could be seen k=Vtip
(dimensionless TKE
by the square of tip velocity) gradually decayed with the
increasing distance from the paddle wheel in plane 1. The
2
value of k=Vtip
approximately reached the peak at x/L = 0.8
in plane 2. It showed that LES could predict the attenuation
and peak of turbulence kinetic energy accurately, which was
superior to ke model [37].
By the comparison of PIV test and CFD simulation, we
can get a conclusion that LES was more accurate than ke
model in the prediction of turbulence kinetic energy.
However, it caused a lot of computation cost, which is not
acceptable in practical applications [37]. CFD simulation
based on ke model could be adopted to precisely simulate
the fluid mixing characteristics in the raceway pond and
optimize the structure of the raceway pond.

simulation based on ke turbulence model could be

adopted to precisely simulate the fluid mixing characteristics in the raceway pond.
Comparison of the turbulent kinetic energy (TKE)
distribution
The turbulent kinetic energy (TKE) and average velocity
can reflect the mixing effect of raceway pond [34], and the
choice of turbulence model plays an important role in the
prediction of the turbulent kinetic energy distribution.
Therefore, the ke model and LES model were used to
simulate the distribution of TKE, respectively. The results
of PIV test and CFD simulation were shown as Fig. 7. By
PIV test, it can be seen that the high-TKE area was formed
in the middle-upper part of the plane 1 because of the stir
of paddle wheel. In the plane 2, the fluid changed the
direction and became disordered after it got out of the
curved channel, and then formed high turbulence kinetic
energy region in the middle. The simulation results of ke

0.80

-1

0.70

Dz (10

-1

Uz (10 m s-1)

0.75

0.65
0.60
0.55

10

15

20

(a)

25

30

35

40

0.60
0.55
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10

10

15

20

(b)

Pv

25

30

35

40

Pv

0.40

U (m s-1)

0.35
0.30
0.25
0.20

10

15

20

(c)
Fig. 9 Relationship between power per unit volume (Pv) and mixing
characteristic parameters for various raceway ponds with different
inclined angles of the blades. a relationship between Pv and average

123

25

30

35

40

Pv
velocity magnitude along the light direction (Uz); b relationship
between Pv and the proportion of dead zone (Dz); c relationship
between Pv and average velocity (U)

Bioprocess Biosyst Eng (2016) 39:169180

177

Optimization results of paddle wheels

Results of CFD simulation
A series of different situations were simulated under
specific power per unit volume (Pv) to quantitatively
compare the mixing effectiveness of raceway pond with
various inclined angles of the blades.

The relationships between Pv and three other mixing

characters, namely Uz, Dz and U, were analyzed, respectively in Fig. 9. It could be seen that the values of Uz and
U increased apparently with the increase of Pv. The novel
raceway pond with 15 inclined angle of the blades had the
highest values of the Uz and U while the traditional raceway pond had the worst results. As shown in Fig. 9b, the
proportion of dead zone (Dz) decreased gradually with the

Fig. 10 The comparison of mixing efficiency between traditional raceway pond (I) and novel raceway pond (II) with 15 inclined angle of the
blades. a Uz contours; b U contours; c Dz distribution sketch

Table 1 The comparison of CFD simulation results between the traditional raceway pond and the novel raceway pond with inclined angle of the
blades
The inclination angle of paddle wheel blades ()

N (rpm)

Uz (910-2 m s-1)

Dz (%)

U (m s-1)

TKE (910-2 J kg-1)

P (w)

0 (traditional raceway pond)

5

12
12

0.62
0.63

3.7
3.5

0.27
0.272

1.8
1.8

12.8
12

10

12

0.64

3.4

0.275

1.8

11.8

15

12

0.66

3.3

0.28

1.8

11.5

123

178

Bioprocess Biosyst Eng (2016) 39:169180

paddle wheels. Meanwhile, Table 2 showed that the average wind speed was within the scope of breeze which could
be negligible.

increase of Pv and the raceway pond with 15 inclined

angle of the blades had the lowest value of Dz.
As shown in Fig. 10, Uz contours, U contours and Dz
distribution sketch were used as further comparison of
mixing efficiency between different raceway ponds with a
certain, special inclined angle of the blades. The novel
raceway pond with 15 inclined angle of the blades and
traditional raceway pond were used as this comparison.
The string rate of paddle wheel was 12 rpm and the
investigated plane was 0.075 m from the bottom of raceway pond.
It can be seen from Fig. 10a that the Uz of raceway pond
with 15 inclined angle of the blades was apparently higher
than that of traditional raceway pond especially in the
stirring region. From the U contour, it can be seen that the
raceway pond with 15 inclined angle of the blades had an
apparently high-velocity region by comparing with the
traditional raceway pond. The distribution of dead zone is
shown in Fig. 10c. Compared with the traditional raceway
pond, the novel raceway pond had a lesser proportion of
dead zone especially in the stirring region.
Table 1 showed the CFD simulation results of traditional raceway pond and novel raceway pond with inclined
blades at the same stirring speed (12 rpm). With the
increase of inclined angle of the blades, the values of Uz
and U were gradually increased while the proportion of
dead zone (Dz) and power consumption were steadily
declined. It is shown that the raceway pond with 15
inclined angle of the blades can achieve a great mixing
effect with a low energy consumption.

Biomass productivity of Chlorella pyrenoidosa in raceway

ponds
Cultivations in the novel raceway pond with inclined
blades and the traditional raceway pond were conducted to
examine the effects of the novel raceway pond on
microalgae cultivation. Similar light conditions and cell
concentrations were initially provided to the cultures for
the two PBRs. The time courses of cell growth in various
PBRs are shown in Fig. 11. The growth rates of Chlorella
pyrenoidosa in the novel PBRs were higher than those in
the traditional reactor, indicating that the novel raceway
pond with inclined blades had a positive influence on
microalgal growth. Table 3 shows that the maximum biomass concentration in the novel raceway pond with
inclined blades was 11 % (0.92 g L-1) higher than the

1.2

novel raceway pond

traditional raceway pond

1.1

Dry cell weight (g L-1)

1.0

Culture experiments

0.9
0.8
0.7
0.6
0.5
0.4
0.3

The external culture parameters

0.2

Table 2 showed the external environment parameters and

measured power consumption during algal cultivation. It
was seen that the contrast experiment was carried out under
the same conditions and the traditional raceway pond had a
higher power consumption with the same rotation speed of

10

Time (days)

Fig. 11 Comparison of Chlorella pyrenoidosa growth between

traditional raceway pond and novel raceway pond with 15 inclined
angle of the blades

Table 2 External environment parameters and measured power of the culture experiments
Type of
raceway
pond

Average
atmospheric
pressure (K Pa)

Average wind
speed (m s-1)

Average pH of
algae medium

Mean
temperature of
air (C)

Mean temperature
of microalgae (C)

The rotation speed of

paddle wheels (rpm)

Measured
power (w)

Traditional
raceway
pond

102

0.3

7.2

29

30.8

10

13.7

Novel
raceway
pond

102

0.3

7.1

29

31

10

12.3

123

Bioprocess Biosyst Eng (2016) 39:169180

179

Table 3 Biomass production of photoautotrophic cultivation of Chlorella pyrenoidosa in the two types of raceway ponds
Type of raceway
pond

Maximum biomass
concentration (g L-1)

Areal productivity
(g m-2 d-1)

Productivity
(g m-3 d-1)

Measured power consumption

by biomass production (KJ g-1)

Traditional raceway pond

0.83

10.18

67.88

7.9

Novel raceway pond

0.92

11.89

79.25

6.1

0.83 g L-1 concentration in the traditional raceway pond

and areal productivity in novel PBRs (11.89 g m-2 d-1)
was increased by 17 % compared with the traditional
raceway pond (10.18 g m-2 d-1). Compared with the
productivity of Chlorella pyrenoidosa in other open ponds,
David et al. [38] achieved biomass productivity of
8.37 g m-2 d-1 in the conventional raceway pond. The
values of practical power consumption by algal biomass
yield were used to evaluate the utilization efficiency of
power energy. It was shown that the power consumption of
traditional raceway pond were 30 % (7.9 kJ g-1) higher
than that of novel raceway pond with inclined blades
(6.1 kJ g-1). The results indicated that the novel raceway
pond can significantly improve cell growth rate and biomass productivity with less energy consumption.

Conclusions
By PIV experiments, it can be seen that LES was more
accurate than ke model in the prediction of turbulence
kinetic energy; however, it caused a lot of computation cost
which is not acceptable in practical applications. CFD
simulation based on ke model could be adopted to simulate the fluid mixing characteristics precisely in raceway
pond and optimize the structure of the raceway pond to
achieve a better mixing efficiency.
CFD was used to simulate the flow field of raceway
pond with different inclined angles of the blades. The
results showed that the novel raceway pond with 15
inclined angle of the blades had the best mixing effectivity
under the condition of same power consumption.
In microalgae cultivation, under the same environmental
condition, the growth rates of Chlorella pyrenoidosa in the
novel raceway pond with inclined blades were higher than
that in the traditional reactor, and areal productivity in the
novel raceway pond (11.89 g m-2 d-1) was increased by
17 % compared with the traditional raceway pond
(10.18 g m-2 d-1). The power consumption of traditional
raceway pond was 30 % (7.9 kJ g-1) higher than that of
new PBRs (6.1 kJ g-1). The results of the experiments and
CFD simulations were identical with each other. Therefore,
these studies showed a huge potential in high efficient
microalgae culture.

Acknowledgments This research was funded by National Basic

Research Program China (973 Program: 2011CB200903).

References
1. Pulz O, Gross W (2004) Valuable products from biotechnology of
microalgae. Appl Microbiol Biotechnol 65:635648
2. Huang JJ, Xia J, Jiang W, Li Y, Li J (2015) Biodiesel production
from microalgae oil catalyzed by a recombinant lipase. Bioresour
Technol 180:4753
3. Brune DE, Lundquist TJ, Benemann JR (2012) Microalgal biomass for greenhouse gas reductions: potential for replacement of
fossil fuels and animal feeds. J Environ Eng 135:11361144
4. Borowitzka MA, Moheimani NR (2013) Algae for biofuels and
energy. Springer, Netherlands
5. Mendoza JL, Granados MR, Godos ID, Acien FG, Molina E,
Banks C, Heaven S (2013) Fluid-dynamic characterization of
real-scale raceway reactors for microalgae production. Biomass
Bioenergy 54:267275
6. Fox RO (2006) CFD models for analysis and design of chemical
reactors. Adv Chem Eng 31:231305
7. Xia JY, Wang SJ, Zhang SL (2008) Computational investigation
of fluid dynamics in a recently developed centrifugal impeller
bioreactor. Biochem Eng J 38:406413
8. Perner-Nochta I, Posten C (2007) Simulation of light intensity
variation in photobioreactors. J Biotechnol 131:276285
9. Perner I, Posten C, Broneske J (2003) CFD Optimization of a
plate photobioreactor used for cultivation of microalgae. Eng Life
Sci 3:287291
10. Sato T, Yamada D, Hirabayashi S (2010) Development of virtual
photobioreactor for microalgae culture considering turbulent flow
and flashing light effect. Energy Convers Manag 51:11961201
11. Sato T, Usuib S, Tsuchiya Y (2006) Invention of outdoor closed
type photobioreactor for microalgae. Energy Convers Manag
47:791799
12. Su Z, Kang R, Shi S (2010) Study on the destabilization mixing
in the flat plate photobioreactor by means of CFD. Biomass
Bioenergy 34:18791884
13. Yu G, Li Y, Shen G (2009) A novel method using CFD to
optimize the inner structure parameters of flat photobioreactors.
J Appl Phycol 21:719727
14. Pruvost J, Pottier L, Legrand J (2006) Numerical investigation of
hydrodynamic and mixing conditions in a torus photobioreactor.
Chem Eng Sci 61:44764489
15. Wu LB, Li Z, Song YZ (2010) Hydrodynamic conditions in
designed spiral photobioreactors. Bioresour Technol 101:
298303
16. Wu LB, Li Z, Song YZ (2009) Numerical investigation of flow
characteristics and irradiance history in a novel photobioreactor.
Afr J Biotechnol 8:46724679
17. Wang HF, Liu ZH, Guo FS, Zhang CG (2004) Estimation of
turbulent kinetic energy dissipation rate In channel flow by PIV.
J Chem Ind Eng 55:10671071

123

180
18. Jakubowski M, Wyczalkowski W, Poreda A (2015) Flow in a
symmetrically filled whirlpool: CFD modelling and experimental
study based on particle image velocimetry (PIV). J Food Eng
145:6472
19. Sultana T, Morrison G, Taylor R, Rosengarten G (2015)
Numerical and experimental study of a solar micro concentrating
collector. Sol Energy 112:2029
20. Amokrane A, Charton S, Lamadie S, Paisant JF, Puel F (2014)
Single-phase flow in a pulsed column: particle image velocimetry
validation of a CFD based model. Chem Eng Sci 114:4050
21. Stogiannis IA, Passos AD, Mouza AA, Paras SV, Penkavova V,
Tihon J (2014) Flow investigation in a microchannel with a flow
disturbing rib. Chem Eng Sci 119:6576
22. Pruvost J, Pottier L, Legrand J (2006) Numerical investigation of
hydrodynamic and mixing conditions in a torus photobioreactor.
Chem Eng Sci 61:44764489
23. Lin C, Li YG, Wang WL, Shen GM, Chen JP, Wu HX, Huang JK
(2009)
Numerical and experimental investigation of a novel flat-photobioreactor with multistage-separator. J Chem Eng
Chin Univ 23:263269
24. Chiaramonti D, Prussi M, Casini D, Tredici MR, Rodolfi L, Bassi
N, Zittelli GC, Bondioli P (2013) Review of energy balance in
raceway ponds for microalgae cultivation: re-thinking a traditional system is possible. Appl Energy 102:101111
25. Ketheesan B, Nirmalakhandan N (2011) Development of a new
airlift-driven raceway reactor for algal cultivation. Appl Energy
88:33703376
26. Hreiz R, Sialve B, Morchain J (2014) Experimental and numerical investigation of hydrodynamics in raceway reactors used for
algaculture. Chem Eng J 250:230239
27. Chung K, Barigou M, Simmons M (2007) Reconstruction of 3-D
flow field inside miniature stirred vessels using a 2-D PIV technique. Chem Eng Res Des 85:560567
28. Sheng J, Meng H, Fox R (2000) A large eddy PIV method for
turbulence dissipation rate estimation. Chem Eng Sci
55:44234434

123

Bioprocess Biosyst Eng (2016) 39:169180

29. Zhu F, Huang J, Chen J, Li Y (2012) CFD simulation and optimization of an open raceway photo-bioreactor. Chem Ind Eng
Prog 31:11841192
30. Meng C, Huang J, Ye C (2015) Comparing the performances of
circular ponds with different impellers by CFD simulation and
microalgae culture experiments. Bioprocess Biosyst Eng
38:13471363
31. Ge C, Wang J, Gu X, Feng L (2014) Application of PIV and CFD
in impeller design and optimization. Chem Pharma Eng 35:2933
32. Su ZF, Kang RJ, Shi SY, Cong W, Cai ZL (2010) Study on the
destabilization mixing in the flat plate photobioreactor by means
of CFD. Biomass Bioenergy 34:18791884
33. Huang J, Li Y, Wan M, Yan Y, Feng F, Qu X, Wang J, Shen G,
Li W, Fan J (2014) Novel flat-plate photobioreactors for
microalgae cultivation with special mixers to promote mixing
along the light gradient. Bioresour Technol 159:816
34. Xu L, Liu R, Wang F, Liu CZ (2012) Development of a draft-tube
airlift bioreactor for Botryococcus braunii with an optimized
inner structure using computational fluid dynamics. Bioresour
Technol 119:300305
35. Sompech K, Chisti Y, Srinophakun T (2012) Design of raceway
ponds for producing microalgae. Rev Environ Sci Bio 3:387397
36. Han F, Huang J, Li Y, Wang W, Fan J, Shen G (2012)
Enhancement of microalgal biomass and lipid productivities by a
model of photoautotrophic culture with heterophic cells as seed.
Bioresour Technol 118:431437
37. Hou SD, Zhang Z, Wang YC, Shi LT (2000) Numerical simulation of turbulent flow in stirred tank agitated by axial impeller.
J Chem Ind Eng 51:7076
38. David C, Matteo P, David C, Mario R, Tredici Liliana R, Niccolo
B, Graziella CZ, Paolo B (2013) Review of energy balance in
raceway ponds for microalgae cultivation: re-thinking a traditional system is possible. Appl Energ 102:101111