Review Article

Proc IMechE Part A:

J Power and Energy
Regenerative flow pumps, blowers 0(0) 1–22
! IMechE 2021
and compressors – A review Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/09576509211018118
journals.sagepub.com/home/pia

M Sreekanth, R Sivakumar , M Sai Santosh Pavan Kumar,

K Karunamurthy, MB Shyam Kumar and R Harish

Abstract
This paper presents a detailed and objective review of regenerative flow turbomachines, namely pumps, blowers and
compressors. Several aspects of turbomachines like design and operating parameters, working principle, flow behaviour,
performance parameters and analytical and Computational Fluid Dynamics (CFD) related details have been reviewed
and summarized. Experimental work has been put in perspective and the most useful results for optimized performance
have been presented. Consolidated plots of specific speed-specific diameter have been plotted which can be helpful in
the early stages of design. Industrial outlook involving details of suppliers from various parts of the world, their product
description and applications too are included. Finally, future research work to be carried out to make these machines
widespread is suggested. This review is targeted at designer engineers who would need quantitative data to work with.

Keywords
Blowers, compressors, pumps/pumping systems, turbomachinery aerodynamics
Date received: 2 May 2020; accepted: 15 April 2021

Introduction
other aspects which are not covered in a recent review
Pumps, along with blowers and compressors form an work by Davies and Aggidis.1
integral part of almost all industrial and domestic
applications. They consume a huge chunk of energy
produced.1 Any improvement in their efficiency Historical background
would result in a large amount of energy savings. Even though it is not known when exactly the first
However, the design and manufacture of an efficient ever work on regenerative turbomachines was pub-
pump are time-consuming as well as expensive. lished, some of the earliest research can be traced to
Hence, for low end applications, cheap and inefficient the 1920s and 1950s2,3 with not much work reported
pumps are used for economic benefits, even though it in the mean period. Lehmann4 gives a detailed
is known that their operation is far from optimum. account of the historical development and traces
Regenerative Pumps, Blowers and Compressors back the history to 1884 when a water ring pump
are a class of turbomachines which operate similar was patented by F. Windhausen (Patent DRP
to any rotodynamic machine (like a centrifugal 28,893 Feb. 1884). This machine is claimed to have
pump or compressor) but share a few characteristics the self-priming and centrifugal pumping effect, the
of a positive displacement machine. Characteristics functions commonly attributed to regenerative
like low flow rate, high pressure rise per stage, pumps. Lehmann considers this to be the forerunner
linear variation of head with flow rate are some of to regenerative pump (which is also called the
the similarities that exist between a regenerative tur- Side Channel Pump). A power machine patented by
bomachine and a positive displacement machine.
These make the regenerative turbomachines very
attractive and employable in confined spaces where School of Mechanical Engineering, Vellore Institute of Technology
high delivery pressure is required. This review paper Chennai, Chennai, India
attempts to provide a consolidated database and Corresponding author:
R Sivakumar, School of Mechanical Engineering, Vellore Institute of
understanding of the existing work done till date Technology Chennai, Melakkottaiyur, Vandalur-Kelambakkam Road,
and also touch upon aspects of CFD modelling and Chennai 600 127, India.
Stress build-up in regenerative flow machines and few Email: sivakumar.r@vit.ac.in
2 Proc IMechE Part A: J Power and Energy 0(0)

B. Blank in 1890 (Patent DRP 56,529 May 1890) to ease of manufacturing is also attached to them.7 Some
handle dripping and aerated liquids has a character- researchers have successfully used rapid prototyping
istic feature of an eccentric impeller and represents methods like 3D printing to manufacture the impel-
the basic liquid ring pump, which is a regenerative lers.8 However, 3D printing is affordable when limit-
pump in today’s perspective.4 According to Davies ed to non-metals and Aluminium only. In the case of
and Aggidis,1 the earliest application of the underly- symmetrical blades, the direction of rotation can be
ing principle was used in 1874 during the design of a interchanged, thereby interchanging the inlet and exit
helical pump. It consisted of paddle shaped rotor ports. Also, some impeller designs have blades on
vanes and the casing containing stator vanes that ini- both sides. This results in minimal or zero axial
tiated the helical flow pattern to the fluid. This thrust, thereby reducing the load on the bearings.
designs were modified to develop self-priming Also, they are self-priming and reliable.
pumps in the 1950s and 1960s. The earliest patent
(German Patent Number 413435) is filed by Siemen
and Hinsch in 19205 and was intended for domestic
Applications of regenerative flow
pumping purpose. This pump was also popular by the turbomachines
name SIHI pump, a name derived from the first two Regenerative flow turbomachines are commonly used in
letters of Siemen and Hinsch.4 A subsequent patent applications where ease of manufacture, low cost,
(US2113116A) titled “Regenerative Turbine Pump” smaller size, stable operating conditions, high tempera-
by James O McMillan3 can be traced back to 1938 ture capability, reliability and self-priming abilities are
in which the inventor claimed that his invention can needed. Most applications of regenerative flow turbo-
develop the head of a multistage centrifugal machines are in the field of pumping water or liquids.
pump but by using a single impeller. Lehmann4 As mentioned by Davies and Aggidis,1 these machines
describes some of the noteworthy developments in can be used in the fields of biomedical pumping, con-
these pumps. They are due to inclusion of water and densate return, micro-turbines, shipping and mining,
air delivery channels by Siemen and Hinsch in 1940 sump service, car washers, chemical and food process
(Patent D. P. 8,42,448 Dec. 1940), noise reducing industries, refineries and many more. Regenerative flow
design of passages and channels by Luhmann in 1962 turbomachines are also used in handling gases, most
(Patent D. P. 11,82,960 Oct. 1962), the combination of commonly air. Also, one can find studies on the appli-
side channel stages to centrifugal stages by Fandrey cations of regenerative flow turbomachines in fuel cells
and Muller in 1977 (Patent DE 27,57,952 while handling hydrogen, natural gas, helium, R-22,
C2, Dec. 1977) and integration of monitoring cham- and ISOPAR G. Few applications involving gasoline
bers into the pump unit by Lehmann and Fandrey in injection into an IC engine also have been found.9
1981 (Patent specification EP 0,045,483 July 1981).
Even with over 150 years of history, the research and
development together with publications in this field Working principle
have not been as prolific as it should have been.6 Understanding of the working principle of any device
The name “Regenerative” has been used most often will be helpful in reducing the losses and improve the
because it was believed that the working fluid enters performance. Also, optimization can be carried out
and exits the space between each blade on the rotor depending on whether one wants better efficiency or
and as it does, the energy transfer takes place resulting higher delivery pressure or improved flow pattern.
in multi-staging or Regeneration. Other names used to For a centrifugal pump/compressor, the working
refer to this type of machines are peripheral, side chan- principle has been understood quite well and hence
nel, water-ring, liquid-ring, drag, turbine, traction, tan- those devices have been standardised as on date.
gential and vortex machines. Application wise, there is Unfortunately, the same cannot be applied with
no widespread use of this technology even though it respect to regenerative flow turbomachines. This is sur-
was available for a long time. The reasons for this prising since these machines have been in operation
could be due to (i) No clear understanding of the even before the 1950s. Also, the fact that there have
working principles, (ii) No established design proce- not been many publications in open literature reflects
dures and (iii) Low efficiency (<50%). Many existing on the limited understanding of these machines.
designs are based on trial and error, personal experi- However, some attempts have been made to understand
ence and empirical relations. Due to this, the regener- the working principle of these machines and they have
ative machine has remained a work of art rather than a been presented in the name of (i) Turbulence Theory
machine of science and technology. and (ii) Momentum Exchange Theory. These theories
are discussed to some extend below.
Advantages of regenerative machines
A major advantage of these machines is that their
Turbulence theory
design remained very simple compared to a rotody- This theory was the most earliest one proposed by
namic machine. Due to this, an extended advantage in Senoo10 and supported by Iversen11 and Balje12 to
Sreekanth et al. 3

explain and understand the working of regenerative

flow turbomachines. According to this theory, the
turbulence induced by the drag effect of the impeller
in the fluid results in the turbulent mixing (also tur-
bulence stresses in some variations) and this becomes
the driving force for the flow and is responsible for
the rise in head. According to them, the motion of a
rough surface (the impeller) in a confined space like
the channel causes the fluid to be dragged along and
the pressure rise takes place in the flow direction. This
theory gave reasonable shapes to the performance
curves.2

Momentum exchange theory

The turbulence theory was countered by Wilson
et. al.13 and highlighted its shortcomings. For exam-
ple, the helical motion of the fluid in the meridional
plane along the flow path,14 Figure 1(a), which is
observed in flow visualization techniques, is not
explained by the Turbulence theory. They proposed
a new theory which explains the circulatory motion as
well as the performance curves predicted by the tur-
bulence theory. This new model is popularly known
as momentum exchange theory or circulatory
momentum exchange theory. According to this, the
fluid receives energy (momentum) as it flows through
the impeller and this momentum is transferred to the
slower fluid which is in the space between the impeller
and casing (known as the annular channel) and there-
by increasing its pressure. The fluid again enters the
impeller (into the subsequent space between the
blades) and again absorbs energy from the impeller
and then transfers to the fluid in the annular channel.
This happens repeatedly until this chunk of the fluid
reaches the exit, after an angular rotation of almost
340
(discounting the angle subtended by the strip-
per). Thus the fluid undergoes internal multi staging
and hence the name Regenerative Flow.
This theory, although neglects the drag and shear
stress effects, satisfactorily explains the performance
of several machines and hence became popular and Figure 1. (a) Helical flow in the flow direction.14 (b)
was cited in several publications.2,6,7,9,14 Circulatory flow from root to tip of the blade..14
A later study conducted by Senoo15 compared the
two theories and concluded that both are similar in sufficient to explain the working principle and perfor-
principle. There is also another theory proposed by mance curves.
Shimosaka and Yamazaki16 which suggest that the
turbulent flow is actually caused by the difference in
the angular momentum at the tip and root of the Geometry and design parameters
impeller. Fluid within the space between the blades Like any turbomachine, regenerative flow machines
flows into the channel from the blade tip to the root also depend on energy transferred between the impeller
and returns to the tip of the blade, as shown in and the fluid. In case of pressure building machines like
Figure 1(b). This motion occurs at each blade result- pumps and compressors, the energy transfer takes
ing in the transfer of momentum from the fluid in the place from the impeller to the fluid and the increase
blade spacing to the fluid in the channel, resulting in in energy is manifested in the form of a rise in pressure
turbulence and hence rise in the pressure. The above energy (also known as flow energy). The amount of
discussion shows that both the theories (Turbulence energy transfer determines the performance which is
and Momentum Exchange) may be different in prin- measured in several ways. The amount of energy trans-
ciple but have few common aspects which are fer strongly depends on the geometry and design
4 Proc IMechE Part A: J Power and Energy 0(0)

parameters of the impeller and casing, where all the blades) result in higher pressures.19 However, the choice
pressure build-up takes place. Geometry parameters of blade type is also determined by the design and
include the shape of the vanes, location of the vanes, manufacturing complications as well as stress build
location of inlet and outlet ports, shape of the annular up. It is easy to design and manufacture straight
spacing, etc. Design parameters include the size of the radial blades compared to others. Also, the design of
impeller, number of vanes, rotational speed, number of symmetric blades without complex twisted angles are
stages etc. Most of the research work carried out to favoured compared to other shapes. One of the major
improve the performance of the regenerative flow tur- advantages of a regenerative flow turbomachine is its
bomachines is centred around the geometric and design simplicity and complex blade shape could nullify that
parameters.1 advantage.20 Also, from the stress point of view, straight
blades are capable of withstanding centrifugal stresses
Blade type compared to curved blades.21 Figure 2(a) shows straight
radial, backward swept, forward swept, chevron type of
The type of blade determines the amount of pressure
blades and Figure 2(b) shows the S-shaped blades.
rise. A variety of blades have been investigated and the
shapes vary from straight, curved, radial, symmetric,
aerofoil, chevron, T-shape, etc.1,17 In one of a recent Number of blades
study, unique S-shaped blades have been used.18 It The number of blades plays an important role in deter-
has been found that radial vanes give rise to low pres- mining the pressure rise. Higher the number of blades,
sure rise while chevron blades (also known as V-shaped lower will be the slip factor and hence the performance

(a)

(b)

Figure 2. (a) Types of blades as summarized by Choi et al.19 (b) S-shaped blades as used by Jeon et al.18
Sreekanth et al. 5

Table 1. Number of blades investigated.

Number of blades (Z) References

Z < 15 16

15 < Z < 50 6,8,18–20,22–33,36–45

50 < Z < 75 29,33,35,38

75 < Z < 100 9

100 < Z < 125 6,46,47

Z > 125 16,21,46

will be closer to design. However, higher blade number

will also result in a large contact surface area between
the fluid and wall resulting in excessive frictional losses.
Hence, a balance must be struck on the number of
blades. Also, due to the shear nature of regenerative
flow machines, the fluid should repeatedly enter and
leave the blade passage and hence a large number of
blades will reduce this space, offering more resistance.
Researchers have used blade numbers varying from as
low as 1216 to as high as 15016,21 These are extreme
numbers but most of the other researchers used blade
numbers in the range of and around 30–50.6,8,9,19,20,22–45
No particular trend is observed indicating the depen-
dency of the number of blades on the working fluid.
Table 1 shows the number of blades investigated by
various researchers. It can be seen that most
of the researchers have used a blade number between
15 and 50.

Impeller type and size

Impellers can be of two types, depending on where the
blades are located, (i) on the periphery or (ii) on the
sides close to the outer circumference. The second
type can be further classified depending on whether
the blades are on one side or on both sides. From the
pressure build-up or leakage point of view, there
seems to be no consensus about the choice of the
type of impeller. However, the second type with Figure 3. (a) Impeller having blades on one side only.6 (b)
blades on both sides has an advantage of having Impeller having blades on both sides.39
equal and opposite axial thrust on both sides of the
impeller resulting in low load on the bearings and in the broad range of 50 to 150 mm. Table 3 gives sev-
hence better than that with blades on only one side. eral geometric details. Table 4 shows the ratio of
Figure 2 shows impellers having blades on the periph- number of blades and the impeller outer diameter for
ery. Figure 3(a) and (b) show the impeller with single different fluids. This ratio ranges from 0.11 to 0.77,
side and double side blades respectively. Among the except for the case of Badami20 where it is having a
impellers with blades on the sides, the type having high value of 1.28. Most of the ratios fall in the
blades on both sides is more common. narrow range of 0.3 to 0.55. Based on the available
Impeller size (outer diameter including the blade tip) data, a power based correlation between the diameter
determines the pressure rise in the machine as it indi- and number of blades is obtained as given in equation
rectly fixes the blade tip speed, a parameter which plays (1). Here Z is the number of blades, d is impeller outer
an important role in energy transfer.21 Different sizes diameter in mm and the correlation coefficient is 0.6
ranging from 20 mm to 320 mm have been found in the
literature. The size is determined by the space availabil- Z ¼ 3:4  d 0:54 (1)
ity and the rotational speed that the prime mover
(motor) can provide. If the motor speed is limited, the
impeller size must be increased to meet a certain pres-
Multi-staging
sure ratio. Table 2 summarizes the various impeller sizes A major advantage of regenerative flow turboma-
used by different researchers. Most of the diameters fall chines is their ability to generate a high pressure
6 Proc IMechE Part A: J Power and Energy 0(0)

Table 2. Impeller sizes. ratio compared to centrifugal pumps/compressors.

This ability avoids regenerative machines to go for
Impeller
size (mm) References multi-staging and hence there are very few work
reported, which investigated multi-stage machines.
20–50 Badami,20 Mason,22 Song et al. 23 Sixsmith and Altmann46 have developed a 2-stage
50–100 Quail et al.,8 Shimosaka and Yamazaki,16 Jeon regenerative flow compressor with the two stages
et al.,18 Quail et al.,25 Quail et al.,27 Quail
located side by side. They could produce a pressure
et al.,28 Quail et al.,30 Annadurai et al.,39
ratio of 2 with water as the working fluid. Badami20
Nejadrajabali et al.,40 Nejad et al.,6 Nejad
et al.,48 Nejadrajabali et al.49 developed a two stage pump with different diameters
100–150 Raheel,9 Shimosaka and Yamazaki,16 Badami on the same impeller for handling ISOPAR G fluid. A
and Mura,31 Badami and Mura,32Lee and channel to transfer the fluid from 1st to 2nd stage is
Kil,33 Lee et al.,38 Zhang et al.,41 Nejad constructed in this pump. There are 39 blades in each
et al.,6 Zhang et al.,42 Pei et al.,44 Zhang stage. Raheel et al.54 constructed and investigated the
et al.,45 Raheel and Engeda.50 performance of a 4-stage regenerative flow compres-
150–200 Meakhail and Park,24 Choi et al.,19 Yoo et al.51 sor to handle natural gas and air. Their impeller
200–250 Badami and Mura,29 Yoo et al.51
arrangement is similar to that of Sixsmith and
250–300 Sixsmith and Altmann,46 Song et al.,23 Jang and
Jeon,35 Song et al.,47 Mekhail et al.52
Altmann46 except that in this case, there are 4
>300 Hollenberg and Potter53 stages located side by side. With this arrangement,
they could achieve a pressure ratio of 4 for the

Table 3. Impeller types, sizes and blade geometry details.

Blade location Impeller diameter Blade height Blade width Blade thickness
References on impeller (mm) (mm) (mm) (mm)

Mekhail et al.52 Peripheral 300 50 23 3

Jeon et al.18 S-shaped peripheral 145.2 0.72–0.88
Pei et al.44 Peripheral 150 35 15 2
Badami and Mura31,32 Peripheral 145 30 1.3
Jang35 Peripheral 256.3 1-3
Badami and Mura26 Peripheral 138 24 1.75
Badami and Mura29 Peripheral 138, 248, 220 24, 40, 20 16, 20 1.75, 3.2, 2
Choi et al.19 Peripheral 170 21.2 11 2
Quail8,27 Double side 74.5 1.3
Nejad et al.48 Double side 65 9 3.1
Nejadrajabali et al.49 Double side 65 9 3.1

Table 4. Ratio of number of blades and impeller outer diameter.

Number of blades/impeller
References Fluid outer dia (mm1)

Badami20 Air 39/30.4 ¼ 1.28

Raheel9 Natural gas 81/105 ¼ 0.77
Song et al.47 Air 120/280 ¼ 0.43
Badami and Mura26 Hydrogen 41/138 ¼ 0.297
Badami and Mura29,31 Hydrogen 36/145 ¼ 0.25
Jang and Jeon35 Air 54/256 ¼ 0.21
Meakhail and Park24 Water 20/170 ¼ 0.11
Quail et al.8,25,27,28,30 Water 30/74.5 ¼ 0.4
Choi et al.19 Water 45/170 ¼ 0.26
Annadurai et al.39 Water 36/65 ¼ 0.55
Nejadrajabali et al.40,49 Water 36/65 ¼ 0.55
Zhang et al.41,42,45 Water 24/150 ¼ 0.16
Nejad et al.6,48 Water 36/65 ¼ 0.55
Jeon et al.18 Water 48/72.6 ¼ 0.66
Pei et al.44 Water 24/150 ¼ 0.16
Sreekanth et al. 7

4-stage compressor. On the whole, it can be said that channel is not needed. Table 4 provides the side chan-
multi-stage compression has not been investigated to nel dimensions employed in various publications.
a larger extent.
Stripper
Clearances Stripper is a solid component often an integral part of
The clearances in a regenerative flow machine are the casing and its function is to prevent the pressurized
provided for the fluid to pass as well as to avoid fluid at the outlet zone from entering the low pressure
any surface to surface contact while in operation. inlet zone. It stops the pressurized fluid by being a phys-
Surface to surface contacts are possible due to rota- ical barrier. However, a small amount of fluid leaks into
tional imbalances, vibrations and material displace- the low pressure zone through the clearance between the
ment (elongation) due to high revolutionary speeds stripper and rotor. The amount of this leakage reduces
and in some cases, thermal expansion too. Also, the with a reduction in the clearance gap and with the
stripper, which is a common component in all regen- increase in stripper angle.53 Clearance gap cannot be
erative machines, separates the high pressure side too low as it may result in physical contact between
from low pressure side. However, it should have the rotor and stripper. Increase in stripper angle (the
some clearance so as to avoid contact with the rotat- angle by which the fluid rotates while traversing from
ing impeller. This clearance is critical as it allows the inlet to outlet) will decrease the amount of regener-
undesirable flow of pressurized fluid into the low ation and hence detrimental to pressure build-up.
pressure side. Table 5 provides the values of clearan- Stripper design involves the determination of stripper
ces adopted by investigators. The radial and axial angle (also known as extension angle or effective angle)
clearances are extremely small, a fraction of a milli- and the clearance space between itself and the rotor.
metre, thus posing manufacturing, installation and The literature is silent on the design aspects of the strip-
operational challenges. per and very few researchers provided limited data.
However, no mention has been made about the design
Side channel procedure. Most of the researchers used empirical
values for the design parameters. Quail et al.8 suggested
The side channel, as the name suggests, is a narrow
the use of 2 to 3 blade pitches as a measure of stripper
channel which receives fluid exiting from the blade
angle. Cates55 suggested that the clearance should be
grooves and returns the fluid into the blade grooves
“few thousandths of an inch”. Song et al.47 mentioned
along the flow path. The high velocity fluid exiting
that they used a clearance space of 0.305 mm in their
the blade grooves mixes with the almost stationary
compressor and mentioned that a 3% leakage is
fluid in the side channel and thus the fluid velocity
observed. Intuition says, depending on the fluid viscos-
reduces and pressure rises. The mixing zone is a
ity, stripper clearance can be larger for more viscous
highly turbulent zone and is not easily amenable to
fluids and vice versa. Table 6 lists the stripper angles
analytical analysis. It embraces an angle of around
and clearances used in the open literature.
340
and carries the fluid in it towards the outlet. It
performs the job of volute casing in a centrifugal com-
pressor. Side channel is a prominent part of machines Suction and discharge ports
with impeller blades on the sides. For machines An advantage of regenerative flow turbomachines is
having peripheral blades, like chevron blades, a side their flexibility in locating and orienting the suction

Table 5. Clearances and side channel dimensions.

Radial Axial
clearance clearance Side channel Side channel Side channel Side channel
Publication Type of device (mm) (mm) radius (mm) height (mm) width (mm) area (mm2)

Zhang et al.41,42,45 Pump 0.2 0.2 17.6

Badami20 ” Ac ¼ 2.6
Song et al.23 ” 0.01–0.127
Nejadrajabali et al.40 ” 4.3
Pei et. al.44 ” 0.2 0.2 17.6
Fathi et al.43 ” Ac ¼ 3.7–4.2
Lee and Kil33 Blower 0.2 23 9
Lee et al.38 ” 0.2 14–24 7–12
Badami and Mura26,31 ” 0.2 Ap ¼ 353
Badami and Mura32 ” 0.3–0.5 Ap ¼ 353.3
Badami and Mura29 Compressor 0.05–0.9 Ap ¼ 1250
Raheel and Engeda50 ” 2 0.1056 8.95 4.47
Ac: cross sectional area; Ap: peripheral area.
8 Proc IMechE Part A: J Power and Energy 0(0)

Table 6. Stripper angles and stripper clearances.

References Working fluid Stripper angle Clearance (mm)

Cates55 Gases, helium, air, R-22 Few thousandths of an inch

Badami20 ISOPAR G 290
, 311

Song et al.47 Air 0.305
Raheel9 Natural gas, air 330

Raheel et al.54 Natural gas, air 330

Yoo et al.51 Air 0.1
Quail et al.8,27 Water 330
(2 blade pitch)
Badami and Mura26,31,32 Hydrogen 340

Badami and Mura29 Air, helium 270
, 340

Choi et al.19 Water 330

Nejadrajabali et al.40 Water 344

Fathi et al.43 Water 316

Pei et al.44 Water 315
, 330
, 345

Table 7. Materials used for impellers and casing.

References Impeller material Casing material

Mason22 Steel Steel

Sixsmith and Altmann46 Duralumin Duralumin
Quail et al.8 Acrylonitrile butadiene, polycarbonates Acrylonitrile butadiene, polycarbonates
Karanth and Sharma34 Metal Metal
Meakhail and Park24 Plastic Plastic
Annadurai et al.39 Polyamide (nylon) Polyamide (nylon)
Zhang et al.41,42,45 Aluminium Plexiglas
Pei et al.44 Aluminium Plexiglas

and discharge ports. Axial entry and exit, as well as clearances would not be maintained during operation.
radial entry and exit ports, have been regularly used. Quail et al.8 reported material failure while 3-D print-
However, as it is needed for any turbomachine, these ing using plastic. Very few authors made a mention of
machines too must have smooth entry and exit in the material used to manufacture the impeller and
order to minimize losses. Karanth et al.36 varied the casing. Table 7 lists the materials used by various
inlet and exit port shapes. They used tapered inlet and authors.
exit ports as well as in one case, they used ports with
constant width (no tapering). In another case, they Types of losses
inserted a splitter vane in the outlet port. In this
Like in any energy consuming device, the input
case, they observed that the rotating stall found
energy is not converted completely to output energy
with no splitter vane has now reduced by a good
and the difference is known as loss. Knowledge of the
margin and a 4.6% rise in total pressure was
types of losses and their causes can help in minimizing
observed. No particular reasons have been imparted them and improve performance. As far as the regen-
to the port orientation. Radial ports were used erative flow turbomachines are concerned, losses will
by33,36,40,54,56 etc., while axial ports were used result in low output pressure as well as low flow rate
by8,26,31,32,35 etc. Perhaps, mounting and installation and hence excessive energy consumption. Davies and
requirements determine the port orientation. Aggidis1 have listed out the types of losses in a regen-
erative flow pump and the same can also be attributed
Impeller and casing materials to blowers and compressors.
The choice of impeller material is determined by its
ability to withstand the centrifugal stresses as well as Slip losses
wear and tear. In some cases, the choice is determined Slip losses occur due to the inability of the fluid to
by the manufacturing method. With the development adhere to the blade profile. This can happen when the
of 3-D printing technology, complex shapes can be fluid flows with high velocity and when it encounters
precisely manufactured but the materials are restrict- the sharp blade curvature, it is not able to adjust its
ed to plastic and aluminium. Stronger materials need direction resulting in flow separation and hence the
to be cast. Also, plastics yield under stress and so the formation of eddies.
Sreekanth et al. 9

Shock losses pressurized fluid into the low pressure entry zone
through the stripper clearance. It is generally accepted
Whenever the fluid entry or exit from the blade is not
that the latter is of greater magnitude. Badami and
aligned with the blade angle at inlet or outlet, it
Mura31 listed a third type of leakage that occurs
results in the rough operation and produces
through the clearance between external impeller
shock. The shock effect is accounted for by an inci-
annulus and the casing.
dence factor.
Apart from the types of losses discussed above, the
losses due to wear also exists. This results in the devel-
Circulation losses opment of rough surfaces over a period of time or due
These losses take place when the fluid enters and to handling particle-laden fluids.
leaves the blade grooves as it proceeds towards the The losses mentioned above were investigated by
exit. Entry and exit into the blade grooves take place several researchers either analytically or experimen-
from the side or annular flow channel. So circulation tally.9,19,20,46 The loss coefficients arrived at by these
losses can be due to (i) turning of the fluid in the researchers fall in a wide range and hence minimize
channel, (ii) mixing of the faster moving fluid exiting their utility for a wide range of operations. Table 8
the blades with slow moving fluid in the channel, and provides various estimates of the loss and leakage
(iii) change in flow area whenever the fluid enters or coefficients.
leaves the blades. Detailed equations to compute the loss and flow
coefficients are given by Lee et al.38
Peripheral losses
Peripheral flow is the flow component that deter- CFD Studies in regenerative flow
mines the flow rate of the machine. Since the periph- machines
eral flow takes place along the curved channel of the
machine, the loss can be estimated by standard prin- Computational Fluid Dynamics (CFD) can give insight
ciples of losses in pipe flow. Hence, the hydraulic into fluid flow aspects, which cannot be obtained ana-
diameter, flow velocity and surface roughness will lytically or experimentally. Improvements in perfor-
determine this type of loss. mance by decreasing losses are possible with CFD
studies. Also, costly fabrication of prototype can be
Inlet and outlet losses made after ascertaining convincing performance using
CFD models. With the advent of high speed comput-
The sizes of inlet and outlet ports can be greatly dif- ing, applications of CFD has become very popular in
ferent from the channel in which flow takes place. So, engineering design and analysis.
whenever the flow enters or leaves the ports, it results Since the working and flow pattern is not properly
in sudden contraction or sudden expansion losses. understood in regenerative flow machines, one has to
Also, if the ports are not aligned with the flow direc- resort to either experimental or CFD studies for flow
tion, it results in inlet and outlet losses. visualization. Experimental flow visualization has
been attempted by using transparent casing,45 the
Leakage losses simple and cheap technique which is not suitable for
The scooping action of the impeller and blades results colourless gases like air. Therefore, it becomes man-
in pressure build-up and flow. However, the scooped datory to opt for CFD techniques in order to visual-
volume is greater than the actual volume due to the ize the flow.
fluid that is trapped in the clearances and not deliv- One of the earliest works on the applications of
ered. These are called as leakage losses and they are CFD was done by Abdalla.57 He adopted a simple
mainly of two types: (i) Leakage through the radial case of unsteady flow outside the blade region and
and axial clearance between the impeller and casing, ignored any losses. This throws doubts over the spec-
and (ii) Leakage through the stripper clearance. The ification of boundary conditions. A later work by
second one is caused due to the unwanted flow of the Andrew58 adapted the streamline curvature technique

Table 8. Loss and leakage coefficients.

Circulating flow Peripheral flow loss Casing-impeller leakage Stripper leakage

Reference loss coefficient, Km coefficient, Kp flow coefficient (Kf1) flow coefficient (Kf2)

Badami20 1.5 4 3 1.5

Song et al.47 1
Badami and Mura26 1 4.7 75 1.5
Badami and Mura29 0.9–1 3.5–4.7 4–4800 26–2650
Badami and Mura31 1.6 5 23–28 5
Badami and Mura32 1.5 2.16 66.57 37.97
10 Proc IMechE Part A: J Power and Energy 0(0)

in the axisymmetric machine. His work too did not outlet etc. and found that the pressure developed
consider the blade-to-blade flow, which is a key to increases with the number of blades. Also, inserting
regenerative flow machines. Subsequently several splitter vanes in the outlet reduced the recirculation
studies were carried out to understand the influence zone and improved the performance. They also found
of design and operational parameters on the perfor- that pump configuration with radial inlet and outlet
mance of the regenerative flow machine, viz. the pres- demonstrated better performance. Lee et al.38 pro-
sure rise, flow rate and in some cases losses too. posed new methods to estimate pressure loss and
leakage inside a regenerative blower. They also com-
Types of models and studies pared the results with CFD analysis of the blowers
and found CFD to be quite useful. Jang and Jeon35
Since the flow is complex, all the models attempted studied the influence of blade angle and thickness at
were three-dimensional. However, wherever possible, the tip of the impeller of a blower meant for a fuel cell
researchers opted for the axisymmetric condition application using CFD. They investigated three dif-
since it reduces the computational burden. Several ferent angles (14
, 18
, and 22
) and thicknesses and
researchers used steady as well as unsteady flow. found that 1 mm tip thickness gave better pressure
Unsteady flow is useful to capture the flow in the rise compared to 2 mm and 3 mm. Mekhail et al.52
intricate paths. Quail et al.,25,27,28,30 Badami and conducted experiments and CFD analysis to study
Mura,32 Mekhail et al.,52 Griffini et al.,56 Zhang the influence of inlet blade angle on the performance.
et al.,41 and Fathi et al.43 have used a steady flow They kept the outlet angle constant at 90
. They
model. Raheel,9 Karanth and Sharma,34 Karanth noticed that the fluid enters the impeller from the
et al.,36 Maity et al.,37 Lee and Kil,33 Jang and side of the blade and leaves from the tip as well as
Jeon,35 Lee et al.,38 Nejad et al.,6 Jeon et al.,18 Pei partly from the blade side. They studied four different
et al.,44 and Zhang et al.45 used unsteady flow inlet angles viz. 90
, 115
, 125
, and 135
and found
models. Nejadrajabali et al.40 used a quasi-steady that best performance is obtained at an angle of 125
.
flow model. Griffini et al.56 used CFD as a tool to arrive at an
Raheel9 in his dissertation performed CFD analy- efficient design of a regenerative flow compressor.
sis on a compressor which was used and reported by The initial design was done using a 1-D analytical
Sixsmith and Altmann.46 He found that the tradition- model and that model was enhanced to a 3-D
al inlet design is not able to provide the necessary model with the aim of obtaining the expected mass
circulatory flow pattern and the pressure rise flow, pressure rise and efficiency. They noticed that
obtained during experiments. This prompted him to Mach numbers at different compressor sections are
modify the inlet by providing a constriction and hence around 0.5 and hence compressibility effects must
creating a nozzle effect. This change, though did not not be neglected. Maity et al.37 also performed experi-
predict the experimental results completely, it brought ments and tried to simulate using CFD. They found
about a significant improvement in the pressure rise. that providing curved walls at the outlet resulted in
Quail et al.25,27,30,59 in their various works, used CFD better performance of the pump due to reduced
to predict the fluid flow pattern in a regenerative flow vortex flow in the impeller. They also concluded
pump, especially the circulatory motion observed that the blade pitch and orientation affect the efficien-
during experiments. They also compared the 1-D cy by altering the toroidal motion. Nejadrajabali
(Analytical) and 3-D CFD models and concluded et al.40 numerically studied the influence of symmetric
that though CFD consumes more time and effort and non-symmetric blades on regenerative pump per-
than 1-D models, it is worthwhile considering the formance. They considered 10
, 30
and 50
angles for
accuracy of the results. Badami and Mura32 applied symmetric blades while 0
was considered as the inlet
CFD to compare the results of the 1D model for a angle for non-symmetric blades. They also varied the
hydrogen blower meant for use in an automotive fuel blade orientation and found that forward swept sym-
cell. They found that CFD provides a better idea of metric blades gave higher pressures at an angle of
the flow and also that for higher flow rates, the fluid 15.5
. Zhang et al.41 studied the influence of three
enters the rotor lesser number of times, resulting in a different blade angles, viz. 10
, 20
and 30
on a
lesser amount of regeneration and hence lower rise in pump. They found that a 30
blade angle to be the
pressure. This explains the decreasing trend of the best one for rise in head. They also noticed radial
head versus flow rate curve. Lee and Kil33 carried vortex on the impeller flow passage has detrimental
out experiments on four different regenerative blow- effect on the performance. Nejad et al.6 studied the
ers handling air and hydrogen and analysed the flow effect of blade and casing geometry on pump perfor-
field inside using CFD, with the intending to develop mance. They could obtain a maximum efficiency at a
a new blower design. Karanth and Sharma34 and pitch-chord ratio of 0.4. They also observed that effi-
Karanth et al.36 carried out a CFD analysis on a ciency increases with an increase in the arc of admis-
radial blade regenerative pump. They studied various sion. They suggest symmetrical aerofoil blades for
design configurations by varying the number of best performance. Jeon et al.18 conceived an innova-
vanes, inlet and outlet port design, inserts in the tive S-shaped peripheral blade on the impeller of a
Sreekanth et al. 11

pump and compared its performance with a conven- mass flow rate to a given pressure and temperature
tional radial and slightly curved blades. They con- conditions. Also, to capture the complex flow charac-
cluded that S-shaped blades perform better. They teristics, various turbulence models have been used.
conducted a parametric study by varying the The rotational effects have been captured by multiple
number of blades, blade thickness and radius of the reference frame model or the frozen rotor model
blade rib. With their optimized results, they found depending on the software package used. Table 9
that the pump efficiency increased by 3.7% compared below lists the various boundary conditions and tur-
to the conventional case of radial blades. Pei et al.44 bulence sub-models used in the literature.
studied the effect of stripper angle (which is called as Results of CFD studies would be useful to visualize
wrapping angle). They considered three pump cases the flow pattern and identify any flaws in the flow like
having three stripper angles viz. 15
, 30
, and 45
. As flow separation. In some applications, the hot spots
intuition suggests, they noticed that the 15
angle can also be identified, which are extremely expensive
resulted in higher pressure rise. Zhang et al.45 studied to detect using experiments. Figure 4 shows the stages
the influence of pressure and velocity fluctuations on of a CFD study from impeller geometry creation, sec-
noise, vibration and harshness. They employed CFD tional view of the flow domain and pressure profile of
techniques to identify the regions of high fluctuations a regenerative flow pump. The impeller shown is a
which occurred mostly near the stripper. double sided radial impeller. In the sectional view,
As a general practice, CFD studies are validated by the region occupied by the side channel and stripper
experimental results. However, Nejadali60 developed can be seen. In the domain view, the blue colour rep-
an analytical method to compute the flow character- resents water and the rest is the casing.
istics in a Regenerative Flow Pump (RFP). He treated
the fluid flow trajectory as helical and wrapped in a Operational and optimized parameters
Taurus. The flow components were resolved into
rotational and peripheral components. Head and Similar to centrifugal pumps and compressors, the
hydraulic efficiency as a function of flow rate were common operational parameters are the rotational
estimated analytically and compared with those of speed, discharge rate and delivery pressure. In few
CFD. The trend of performance curves predicted by cases, studies have been conducted using different
the analytical model were close to the CFD results. fluids in the same test facility. Cates55 studied different
However, the flow patterns predicted by analytical gases like Helium, air, R-22 etc. Hollenberg53 studied
air and water. Raheel9 and Raheel and Engeda50 ana-
method and CFD differed by a good amount at
lyzed natural gas and gasoline but in different experi-
large flow rates. This was attributed to the flow com-
mental setups. Badami and Mura29 and also Lee and
plexity and turbulence effects at those conditions.
Kil33 studied air and helium. Shirinov and Oberbeck17
considered nitrogen, helium and argon. Zhang et al.42
Initial, boundary conditions and sub-models used used a gas-liquid two phase fluid.
The inlet and outlet boundary conditions used in the Optimization is done by conducting parametric
models found in the open literature varied from given studies. The optimized function is usually the

Table 9. Boundary conditions and turbulence models.

References Fluid Inlet boundary condition Outlet boundary condition Other details

Badami and Mura32 Hydrogen Total pressure and Mass flow rate Steady flow, CFX, k-e,
temperature Frozen rotor
Karanth and Sharma34 Water Mass flow inlet Outflow RANS, k-e No slip at walls
Jang and Jeon35 Air Atmospheric pressure Mass flow RANS, SST, frozen rotor,
No slip at wall
Mekhail et al.52 Air Mass flow rate Total pressure and temperature CFX steady, k-e
obtained from experiment
Karanth et al.36 Water Mass flow rate Pressure condition RANS, k-e No slip at walls
Maity et al.37 ” Mass flow rate Outlet k-e Pressure based solver
Nejadrajabali et al.40 ” Total pressure Mass flow rate RANS, Frozen rotor
k-e, k-x SST
Zhang et al.41 ” Total pressure Velocity RANS, k-x, SST
Nejad et al.6 ” Total pressure Mass flow rate CFX, RANS frozen rotor,
k-e, k-x SST
Fathi et al.43 ” Mass flow rate Pressure outlet Reynolds Stress Model
Jeon et al.18 ” Atmospheric pressure Mass flow rate No slip CFX, RANS, SST
Pei et al.44 ” ” Velocity obtained from CFX RANS, SST, k-x
experiments
Zhang et al.45 ” ” ” 3D, unsteady, SST, k-x
12 Proc IMechE Part A: J Power and Energy 0(0)

certain operational or design parameter on the per-

formance. The parameters varied were the impeller
size and type, blade size, type and angles, casing
design (including stripper, inlet and outlet ports),
working fluid and rotational speed. The performance
evaluation is carried out by measuring the fluid flow
rate, discharge pressure, power input and computing
the efficiency. Frequently, comparisons are made with
centrifugal pump/compressor performance. The per-
formance parameters like flow rate, discharge pres-
sure and power are either represented in absolute
units or as dimensionless parameters, as defined by
Dixon and Hall.61 The definitions are given below for
convenience

Q
/¼ (2)
D22 u2

gH
w¼ (3)
u22

1
NQ2
Specific Speed ¼ 3 (4)
ðgHÞ4

mRTln PP21
gisothermal ¼ (5)
Shaft Power
Figure 4. Stages of a CFD study. (a) Impeller geometry 4; (b)
sectional view of computational domain. (c) Pressure build-up, qQgH
in bar. ghyd ¼ (6)
Shaft Power

efficiency or the discharge pressure. The parameters Power

varied are the impeller size,16,23,33,38,51 blade inlet and Power Coefficient ¼ (7)
qx3 D5
exit angles6,19,23,32,39–42,46,52,55 that ranged from 0
to
90
: blade shape6,9,17–19,31,39,47 varied among radial, Arc of inlet
forward swept, backward swept, twisted, symmetric, Arc of Admission ¼ (8)
Dm
asymmetric, chevron as well as S-Shaped blade thick-
ness,18,35 rotational speed6,16,20,26,31–33,38,43,47,55 varied
Stripper Angle
from 1450 to 25,000 rpm and stripper angle20,29,44 Regenerative Number ¼ Z (9)
varied from 270 to 344
. No optimization study 360
used any of the established engineering optimization
techniques. They have conducted a parametric study In one of the earliest experimental studies,
by varying in a range and chose the best parameter Mason22 attempted to study and establish a relation-
depending on the efficiency or pressure rise. ship between circulatory flow and pump geometry of
In general, there is a consensus that chevron blades a regenerative flow pump. In his study, he used two
gave higher pressure rise than any other type of different sized impellers (1.25 and 2 inches) each
peripheral blades. Among the blades on the sides, having 20 and 40 blades respectively. He concluded
forward swept blades gave higher discharge pressures. that there is no clear relation between the two. Later,
Shimosaka and Yamazaki16 identified a unique
parameter formed by the ratio of the sectional area
Experimental studies of the flow channel and circumferential length of the
Several experiments were conducted on regenerative vane, which they called a “characteristic dimension”.
flow machines since early 1950s, mostly on pumps Their study suggested that the influence of clearances
and the rest on blowers and compressors. and number of vanes on performance depends on the
Experiments are necessary to test out theoretical characteristic dimension and width of the vane
designs or empirical theories using which prototypes respectively. The pump efficiency is high in an appro-
have been fabricated. Several researchers have con- priate range of impeller diameter and characteristic
ducted experiments to investigate the influence of a dimension. In one of the earliest studies on
Sreekanth et al. 13

regenerative flow compressors, Cates55 conducted losses that take place in the compressors. Song
experiments on gases like Helium, air, R-22, etc. on et al.47 studied the flow in the developing region of
a compressor in the speed range of 5000 to a regenerative pump. They developed an analytical
14,000 rpm. He studied the influence of Mach model to evaluate the circulatory flow losses and
number, impeller, channel geometry and clearances. compared the predictions with the experimental
He reported stable operation except at high Mach results of Mason.22 Quail et al.25,28,59 conducted
number. Sixsmith and Altmann46 were one of the extensive experimental work on regenerative flow
first investigators to study a double-stage regenerative pumps. They estimated the pressure developed, flow
flow compressor with streamlined blades, instead of coefficient, efficiency for radial and optimized blades
radial blades which were common till that time. They and compared with CFD predictions. They have also
used a 30 cm diameter impeller at 4,000 rpm. discussed rapid manufacturing techniques like 3D
Uniquely, they evaluated the performance using iso- printing to produce the impellers in a shorter
thermal efficiency rather than hydraulic efficiency. time.8,27 The impellers manufactured in plastic were
They identified that slip losses are large in magnitude not strong enough and hence they proposed Acrylo
compared to leakage losses and suggested any Nitrile Butadiene and polycarbonates for pump
attempts of improvement should be aimed at slip impellers. Using these materials, they manufactured
losses. Hollenberg and Potter53 investigated the per- and conducted experiments successfully. Shirinov and
formance of a regenerative flow machine using air as Oberbeck17 conducted experiments using Nitrogen
well as water, in three different devices. They devel- Helium, and Argon as working fluids. They varied
oped a correlation relating pressure and flow driving blade shape (C, V, T and Y shapes) and the number
torque. They found that increasing the stripper clear- of blades as parameters. They found that ‘V’ shaped
ance resulted in a drop in head and efficiency. They blades gave highest pressure rise compared to others.
also suggested that higher efficiency may be possible Choi et al.19 studied the influence of blade angles (0
,
at higher specific speeds. Badami20 put forth a theory 10
, 15
, 45
) and angle (15
, 30
, 45
) of chevron
to evaluate the circulatory flow field in the side chan- shaped (V shape) blades on pump performance by
nel of a gasoline pump. He also designed and con- using 45 blades at 1000 rpm, a result that agrees
structed a 2-stage pump and also an innovative design with Shirinov and Oberbeck.17 They found that a
pump with vanes on the periphery and flow channel. chevron angle of 30
gave the highest pressure rise.
Badami and Mura,26,29,31,32 in a series of extensive Nejad et al.48 also conducted a similar study in which
experiments on air, hydrogen and helium, evaluated they varied the blade angle from 0–50
in a double
the performance in terms of head developed and flow sided impeller and measured the head and efficiency
rate obtained at a speed of 5,000 to 20,000 rpm. They at different flow rates. They observed that while the
also compared their experimental results with the head is a weak function of the blade angle, the effi-
model estimates. They also studied the influence of ciency varied with the angle. Maximum efficiency was
leakage on the performance of hydrogen blower.31 obtained at 15
forward curved blade angle. Mekhail
For this, they varied the clearance from 0 to et al.52 experimentally investigated the influence of
0.6 mm. They found that as the clearance increases, inlet blade angle on the performance of regenerative
the head developed and efficiency drops. Raheel and blowers. They used peripheral blade impeller with
Engeda50 and Engeda62 discussed a systematic design four inlet blade angles (90
, 115
, 125
, 135
) and a
procedure for a regenerative flow compressor. They common outlet angle of 90
and operated at
have used the experimental data available till date and 3000 rpm. They found that the best performance
derived correlations to estimate the impeller diameter, (pressure rise and efficiency) is at an inlet angle of
number of blades, rotational speed, axial and radial 125
. Maity et al.37 carried out experiments on a
clearance and other important design parameters, pump having radial blade impeller at a speed of
given mass flow rate and pressure ratio as input. 2880 rpm. They measured the pressure developed
This is the first of a kind work that provides a and flow rate to compute the efficiency. Further,
design procedure using simple equations. However, they validated their CFD model with the obtained
the estimated rotational speeds are extremely high experimental results, investigated different designs
and it is to be verified if such high speeds are actually and concluded that curvature in the outlet flow
necessary especially since several researchers have domain, as well as different number of blades on
used lower speeds to obtain higher pressure rise. either side of the impeller arranged with an offset,
Raheel,9 Raheel and Engeda50 and Raheel et al.,54 will give higher pressure rise. Annadurai et al.39 man-
have conducted experiments on air, natural gas and ufactured impeller made of nylon using rapid proto-
gasoline for micro turbines and automotive fuel typing and used the impeller to test a regenerative
pumping systems. Their work included a single flow pump for its ability to develop pressure. They
stage as well as multi-stage compressors. They could obtain the standard pressure versus flow rate
found that higher efficiencies are possible at higher curves for different conditions. They used a standard
mass flow rates but the efficiency drops as the flow impeller with radial blades and a modified impeller
rate increases further. They also identified several with twisted blades. They found that the modified
14 Proc IMechE Part A: J Power and Energy 0(0)

impeller had twice the discharge pressure compared angle in this article) on the pump performance. They
to the standard impeller. Nejadrajabali et al.40 and considered stripper angles of 315
, 330
, and 345
. The
Nejad et al.6 conducted experiments on double sided study revealed that larger stripper angle has a signifi-
impeller with symmetrical blades having angles 10
, cant influence on the pressure developed and very low
30
, and 50
. They also used bucket shaped blades influence of efficiency. In a unique experiment con-
and operated at speeds of 1,450 and 2,900 rpm. By ducted by Nejadrajabali et al.,49 the performance of
interpolation, it was established that the highest a regenerative flow pump is studied at different fluid
head is developed at an angle of 15.5
. For the case viscosities obtained by mixing Polyacrylamide in
of bucket shaped blades (aerofoil shape), 41
blade water. It found that the performance becomes poorer
angle was found to be optimum. In comparison, as the viscosity increases. They conclude that the RFP
bucket shaped blades fared better than radial is not suitable for high viscous fluids or those having
blades. Zhang et al.41,42,45 investigated the influence low Reynolds number.
of the base blade angle on head developed and found The test facilities are very much similar to those of
the best performance at an angle of 30
. They used commonly used machines like centrifugal pumps,
Plexiglas casing for flow visualization and hence had compressors and blowers. Figure 5 shows the general
to limit the speed to 1,500 rpm. In a different work,42 schematic and instrumentation commonly used in
they investigated the performance in the presence of testing regenerative flow machines. For pumps,
some gas as the working fluid. It was noticed that Figure 5(a), a sump which is open to the atmosphere
head and efficiency as well as the operating range is needed if the working fluid is water or any liquid
drop in the presence of gas. In a first of its kind, which does not evaporate or chemically react with the
they45 also studied noise, vibration and harshness of elements present in the atmosphere. For reactive
a regenerative pump and found that the fluctuations liquids, the sump must be a sealed one. The discharge
increase from inlet to outlet and the maximum was line is fitted with pressure and flow measurement
found at the stripper. Jeon et al.18 used novel S- devices. A flow regulator valve too is fixed on the
shaped peripheral vane assuming that the shape gives discharge side to operate the pump at various flow
more robustness compared to the cantilever type rates. The motor that drives the pump is preferred
blades used in every other impeller. They understood to be having variable speed capability, which would
that among blade thickness and radius of hub rib, the facilitate the variable speed operation and thereby
number of blades is an important parameter which generate data useful to plot constant efficiency
determines the head developed. After optimization, a curves. The discharge line is directed back into the
rise in efficiency by 3.7% was obtained. Pei et al.44 is sump so as to operate in a closed circuit mode. In
one of the most recent work and also the only study on the case or regenerative flow blower/compressor, a
the influence of the stripper angle (called as wrapping pressure gauge is needed on the suction side also.

Figure 5. Schematic and instrumentation diagram of test facilities. (a) Typical test facility of a regenerative flow pump. (b) Typical test
facility of a regenerative flow blower/compressor.
Sreekanth et al. 15

If the working fluid is air, the inlet is directly taken The lowest specific speeds are obtained for positive
from the atmosphere and the outlet is discharged into displacement machines and the highest for axial
the atmosphere. For any other gas, the inlet is drawn machines with radial and mixed flow machines
from storage cylinders and the outlet is discharged spread in the middle. After making a comparison of
into different pressurized storage cylinders. For various pressure building machines like pumps and
both pumps and blower/compressors, a tachometer compressors, Balje64 inferred that for low specific
is needed to measure the rotational speed of the speeds (below 30), the regenerative pumps and com-
impeller and a torque transducer is needed to measure pressors displayed higher efficiencies compared to
the torque. Using the speed and torque, the input the corresponding centrifugal machines. This is
power can be measured. Using the flow rate and pres- attributed to the multi-stage nature of regenerative
sure head developed, the output power can be esti- flow machines.
mated. The ratio of output power and input power Dubey et al.66 have consolidated the data of sever-
can be used to compute the overall efficiency. al types of turbines and plotted the Cordier diagram
as shown in Figure 7. The figure indicates that
Design and performance evaluation impulse machines need a high specific diameter and
low speed to operate at their best efficiency whereas
In one of the earliest studies, Balje12 predicted a low reaction turbines are at the other end of the spectrum
efficiency of about 35% for regenerative machines. with mixed flow machines lie in between.
Later, in his seminal work, Balje identified that spe- Balje63,64 has plotted constant efficiency contours
cific speed (Ns) and specific diameter (Ds) of a partic- for turbines, pumps and compressors in a wide range.
ular type of machine can be used to map the Figure 8 shows the Cordier line which is obtained by
performance of turbines63 as well as compressors
extracting points from the contours plotted by Balje
and pumps.64 He used the available experimental
in63 for single disk turbines Points from Balje64 are
data and together with analytical analysis developed
used to plot the Cordier line for single disk
plots containing contours of efficiencies for different
pumps and compressors. Region marked as point 2
types of turbines, compressors and pumps. A Cordier
diagram obtained from the contour plots of Ns-Ds
diagrams which can be used to design a turboma-
chine. A simplified Cordier line diagram is provided
by Dick65 and is shown in Figure 6 for a narrow range
of specific speed. The Cordier line for pumps and
compressors (P-C) lies over that of turbines (T).
Based on the specific speed and desired efficiency, a
particular specific diameter can be chosen to begin the
design process. The specific speed (Ns) and specific
diameter (Ds) are defined by equations (10) and
(11). The form shown in equation (10) is more com-
monly used compared to the one in equation (4)
pffiffiffiffi Figure 7. Cordier line for power producing turbomachines..65
n Q
Ns ¼ 3 (10) 1: impulse turbine, 2: radial machine, 3: axial machine, 4: pro-
H4 peller turbine.
1
DH4
Ds ¼ pffiffiffiffi (11)
Q

Figure 8. Cordier line for turbines. 1: multi stage turbine, 2:

single stage partial admission, 3: drag turbine with low pressure
Figure 6. Universal Cordier line diagram..65 ratio, and 4: single stage full admission.63
16 Proc IMechE Part A: J Power and Energy 0(0)

in Figure 9 is representative of drag flow machines or obtained from Thomas et al.68 Both Grabow67 and
regenerative flow machines. It is evident that drag Thomas et al.68 computed specific speed and diameter
flow machines have a low specific speed lying in the as shown below
range of 3–20, especially when compared to mixed
pffiffiffi 1
flow and axial flow machines. However, the specific 2 p Q2 n
diameters are higher for drag flow machines. Ns ¼ 3 (12)
ð2wÞ4
In recent literature, no such Cordier lines have
been reported by any researcher. Performance pffiffiffi 1
curves of Pressure rise Vs. discharge or pressure coef- p ð2wÞ4 D
Ds ¼ 1 (13)
ficient Vs. flow coefficient are often reported. Hence, 2 Q2
to make a comparison of different types of regenera-
tive flow machines, operating points have been These equations are different from equations (10)
extracted from reported experimental as well as com- and (11) and hence the specific work ‘w’ in equations
putational results and the specific speed and specific (12) and (13) has been re-written in terms of ‘head’.
diameter have been computed using equations (10) Accordingly, the points from67 and68 have been mod-
and (11). A consolidated plot is shown in Figure 10.
ified before plotting in Figure 10. Data for screw and
This figure also contains data of non-regenerative
piston compressors has been obtained from Bláha
flow machines and these points have been extracted
and Smid.69
from Grabow.67 Data related to scroll compressors is
The data represented in Figure 10 falls in several
bands. As a general observation, one can notice that
machines handling liquids are characterized by high
specific diameter and low specific speed while it is the
opposite for those handling gases. Further, positive
displacement pumps like gear pump and reciprocating
pump are on the extreme left while positive displace-
ment compressors like screw and reciprocating com-
pressors are to the right end. The one to the left
belongs to pumps having low specific speed (<10)
and high specific diameter while the one on the
lower right belongs to blowers showing very high spe-
cific speed (>100). The blowers correspond to the
region marked ‘4’ in Figure 9. The middle region is
Figure 9. Cordier line for single disk pumps and low pressure
ratio compressors. 1: positive displacement pumps, 2: drag occupied by compressors with specific speed around
pumps, 3: radial compressors, 4: mixed flow compressors, 5: 10 and specific diameter comparable to those of
axial flow compressors.64 pumps. This diagram is useful in the initial stages of

Figure 10. Cordier line for various regenerative and non-regenerative flow machines.
Sreekanth et al. 17

design where there is uncertainty around the diame- considers stresses, as well as displacement, must be
ters and speeds. A designer can pick a specific speed chosen to evaluate the safety of a design and the
corresponding to the pressure rise needed and arrive choice of impeller material.
at the specific diameter. These can be the initial design Surprisingly, there have been no studies involving
parameters that can be fine-tuned with other infor- stress build-up for regenerative flow turbomachines,
mation. Hydraulic efficiency is often reported as although several studies have been found for other
another performance parameter. For most of these roto-dynamic machines. Ray and Sinha70 developed
machines, the best efficiency point (BEP) is roughly a mathematical model and solved it analytically to
in the range of 40–50%. The BEP corresponds to a compute centrifugal stresses in a radial flow impeller.
specific speed of 14 for a blower and 1.5 for a pump. Their model predicts that stresses depend on the
Low hydraulic efficiency is a major drawback of material density, disk geometry, ratio of bore to tip
the regenerative flow machines. However, there are radii and the number of blades. They concluded that
several advantages claimed in favour of these more blades will bring in ribbing effect which is good
machines. Among these are simple design and hence for the integrity of the impeller but at the same time
ease of manufacture, ability to obtain high pressure in increase the stresses. Therefore, a balance must be
a single stage (i.e. obtain pressures of positive dis- struck in choosing the number of blades. Bert and
placement machines without surface contact), no Paul71 developed a model to predict the failure/frag-
vapour lock, self-priming,44 and the ability to develop mentation of rotating disks, which in general covers
high head at low flowrate.1 Applications where space pump and turbine impellers. They too predicted that
is a constraint and smooth operation is needed, regen- higher stresses are a result of high inner and outer
erative flow machines can be opted provided efficien- radii, just like Ray and Sinha.70 Satish and
cy is not of top priority. Kallurkar72 used Ansys package to predict the
stresses developed in a rotating centrifugal pump
impeller at various speeds ranging from 1,050 to
Stress build-up in impellers
1,800 rpm. They used the deformation and von
Since the regenerative flow machines operate at high Mises stresses to determine whether the impeller is
speeds and continuously churn a fluid, the impellers safe to operate. Raj et al.73 analysed the impeller of
will be subjected to stresses. The stress build-up is an 8-stage centrifugal pump using 3D modelling.
influenced by the impeller material properties, speed Using this analysis, they estimated the axial thrust
(RPM), blade design as well as the type of working on the bearings which was found to be in the range
fluid. Properties of the impeller material, like its den- of 100-480 kg. Then they proposed a modified design
sity, elasticity, Poisson’s ratio etc. will influence the to nullify the axial thrust. As can be seen, a variety of
stress formation. Also, sharp edges and corners on work has been carried out to study the stress and
the blades will contribute to stress concentration. failure probability of disks or centrifugal pump impel-
The method of manufacture will also determine the lers. However, there is no such work carried out for
amount of work hardening the impeller has been sub- regenerative flow machines. Therefore, a stress anal-
jected to. The viscosity of the working fluid will deter- ysis in these machines is the need of the hour and it
mine the amount of resistance offered to the churning can be done using analysis packages like Ansys or can
action of the impeller and thereby the shear stresses. also be done using analytical solutions to simplified
Intuition tells us that liquids offer more resistance cases (of rotating hollow disk), as provided by
than gases. Also the temperature of the fluid would Timoshenko and Goodier.74 These analytical solu-
contribute to the thermal stress development in the tions can give quick and approximate results which
impeller. However, most cases have none or very are useful to carry out an order of magnitude analy-
low temperature rise of the fluid as it passes through sis. Few assumptions can be made while attempting
the impeller and hence thermal stresses can safely be an analytical solution. They are (i) Considering two
neglected unless they have a profound influence in dimensional stresses only (ignoring the axial direc-
thermal expansion. Last but not the least is the rota- tion), (ii) No thermal stresses, (iii) No presence of a
tional speeds which determine the centrifugal stresses. fluid (ignoring shear stresses on the surface), and (iv)
All the above factors will contribute to a complex Uniform impeller thickness.
combination of stress build-up in the impeller. The
stresses can be in the radial, tangential and axial
directions and a representative stress at each location
Industrial outlook
along with a suitable failure theory will determine the Regenerative flow machines are capable of perform-
safety of the design. Also, as has been noticed in the ing functions of the regular pumps, blowers and com-
geometry and design parameters section, the clearan- pressors and hence they are capable of replacing
ces in the radial and axial directions are extremely them. They can also be employed in fields of biomed-
small. Hence the elongation (commonly called as dis- ical pumping, condensate return pumps, micro-
placement) of the impeller must also be within rea- turbines, shipping and mining, sump service, car
sonable limits. Therefore, a failure criterion that washers, chemical and food process industries and
18 Proc IMechE Part A: J Power and Energy 0(0)

refineries.1 Further, special applications in gasoline supplies pumps that can deliver 138 lpm at a pressure
injection,20 coolant pumping in helium cooled reac- of 24 bar. Besten Pumps85 is an India based company
tors, compression of natural gas, biogas and in water that supplied regenerative flow pumps having a head
treatment plants are some areas where they can be range of 3–42 m and discharge of 10–60 lpm. Their
used. Stanmech Technologies75 suggests that regener- applications are mostly domestic, gardening and for
ative flow blowers are suitable when the air needs to car washing purposes. Kirloskar Crompton86 is an
flow through a restrictive passage as they can gener- India based company which supplies 1 HP regenera-
ate enough pressure. However, their market penetra- tive pumps which can deliver a head of 39 m and flow
tion has been poor and one could not find many rate of 30 lpm.
applications in real life. This can be attributed to The above list is not an exhaustive one. There are
lack of publicity, poor understanding of the flow pat- few more manufacturers like Klaus Union, Teikoku,
tern and low efficiency. Despite these drawbacks, they CRI Group, Pentair, Xylem etc.87 The authors
could draw benefits from the fact that their geometry attempted to provide details of manufacturers from
is simple and size is compact for a given duty. various parts of the world. However, commercially
There are few manufacturers of regenerative flow available products are mostly meant for simple and
machines. Becker Pumps Corporation76 is a supplier common applications. Specialized applications may
of regenerative flow blowers. They supply single stage need application specific design and manufacture.
and two stage blowers and claim oil free, quiet and The regenerative flow pump market is expected to
contact less operation. Their two stage blower is con- grow from $263 million in the year 2021 to $362 mil-
vertible to single stage. Air Power Products Ltd.77 lion in the year 2026, with a cumulative annual
which is a compressed air and vacuum equipment growth rate of 5.42%.87 The market is split into
manufacturer. They supply blowers having a flow single-stage, multi-stage, agriculture, building and
rate of 1.7 m3/hr and an outlet pressure of 1.5 atm. construction, water and waste water, and industrial
They suggest that these blowers can be used in segments. Even though the market is influenced by
manufacturing, textile, chemical processing, printing, COVID 19 pandemic, just like similar markets, it is
food, aqua culture and electro-plating industries. expected to spring back and continue the growth.
Goorui is a Chinese company which claims to be
the world’s leading supplier of side channel blowers.78
They have products from single stage to three stage
Conclusions
blowers. They have capacities up to 200 m3/hr and This paper presents a review of experimental, analyt-
1.4 atm delivery pressure. ical as well as numerical work carried out over differ-
Speck is a German company specializing in both ent regenerative flow machines like pumps, blowers
regenerative as well as centrifugal pumps.79 They and compressors. A considerable amount of work has
claim their products to develop high pressure at low been done since the early 1950s but in comparison to
flow rates and capable of handling clear liquids. their counterparts in centrifugal pump, blower and
Applications are in coolant flow in laser devices and compressor, research on regenerative flow machines
fuel delivery in trucks. The pressure range of their is very less. This could be due to their low efficiency,
products is from 4 to 9 atm and power rating in the low flow rate and low specific speed. Engeda62 pro-
range of 0.12 to 4 kW. Blackmer80 is an American vided few guide lines for designing a regenerative flow
company specializing in regenerative flow pumps pump. In the first review on these machines, Raheel
that can develop a pressure of 14 bar at a flow rate and Engeda88 highlighted a few points, which are rel-
of 60 litres per minute. The impeller speed is close to evant even to this date. In another recent review
4000 rpm. Star Pump Alliance81 is a German compa- paper by Davies and Aggidis,1 a detailed review on
ny that supplies regenerative flow pumps which can pumps is given. The present work aims to give a direc-
develop a pressure of 10 atm at a flow rate of 15 m3/ tion for future work in all three types of machines,
hr. They claim that their pump is compact in design, namely pumps, blowers and compressors:
self-priming and high delivery head. It cannot handle
particle laden liquids due to small clearances. Roth (i) The two theories behind the working principle,
Pump Company82 is an American company which namely Turbulence theory and momentum
claims to supply regenerative pumps for volatile exchange theory explain several phenomena. If
liquids. This is typical of these pumps as there is no applications and fluids where these theories
vapour lock formation. Further, no cavitation is pos- hold good are identified, research work can be
sible. Their pump is capable of delivering water at narrowed down.
4 atm pressure. Inrag AG83 is a Switzerland based (ii) Even to this day, the hydraulic efficiency
company supply regenerative flow pumps which can seldom crosses 50%. This is a serious drawback
handle low viscosity fluids like propane, butane, and and needs to be addressed for all machines.
LPG. They claim that their pumps can handle (iii) Efficiency can be improved by a better under-
entrained vapours without loss and internal damage. standing of losses. As of now, a good amount of
Esbray Pumps84 is an Australian company which energy is expended in driving the circulatory
Sreekanth et al. 19

flow. This energy can be reduced by providing a Hence, anti-wear coatings must be identified
head start to the fluid at the inlet itself. so that particle-laden fluids too can be handled.
(iv) Leakage losses are also responsible for low dis- (xiv) No work has been found which studied the han-
charge and hence efficiency. It has been estab- dling of chemicals in the chemical process
lished that leakage losses across the stripper are industry. This is a vast field, which may find
higher compared to those in other regions. regenerative flow machines attractive.
Hence, a systematic study needs to be con- (xv) Since these machines deliver high pressure, they
ducted in the stripper region. may be used where a high speed jet of fluid is
(v) Stripper design has been very empirical and no needed, viz. automobile washing, fire extin-
established guidelines are available. More guishing nozzles, jet cooling etc.
research in the stripper region would lead to
proper guidelines in its design, mitigating the
Scope for further research
leakage losses.
(vi) Stripper angle is a crucial parameter and its Even though the concept of regenerative flow
choice in the literature has been quite arbitrary. machines is over a century old, the amount of
Also, pressure fluctuations and hence noise and research, quantified by the number of research
vibration are excessively high near the stripper. articles published in the open literature indicates
Hence, a wide range of studies near the stripper insufficient amount of study. The flow pattern is com-
is required. For example, there is no data to plex and hence there is no standard design procedure.
determine what should be the stripper clearance This calls for empirical correlations based on the
for a given liquid or a gas. available experimental and computational data.
(vii) Multi-staging is at its nascent stage and more Such correlations are scarcely found for regenerative
studies are needed. pumps and blowers of different types of impellers.
(viii) The rotational speeds required for some Even more useful would be developing nomograms
machines are over 20,000 rpm, which is difficult for the design process, which are currently not
to achieve. So, research must focus on reducing available.
the speed, without increasing the impeller size. Since it is known that these machines operate at
(ix) A major amount of work has been carried using low efficiency, any improvement in the efficiency is
water, a lesser amount of work can be found for welcome. Hence studies can be directed at reducing
gases like air, hydrogen, helium and natural the losses. Since the efficiency is low, the exit temper-
gas. Only one work has been found to be car- ature of the regenerative compressors could be high,
ried out on refrigerant. Hence, future work can even though actual data is not reported anywhere. If
be directed at replacing the existing compres- such compressors could be used in gas turbines, there
sors in air conditioners with regenerative could be saving in the fuel and reduction in CO2 emit-
compressors. ted even though the overall efficiency could still be
(x) Since regenerative flow machines are not prone low. Investigation into this can be carried out to iden-
to vapour locking, they may be used to pump tify if there could be any real benefit from this idea.
easily evaporating liquids like automobile fuels. For the same reason, compressors used in refrigera-
Also, the pumping of two-phase fluids needs to tion and air conditioning systems can operate at lower
be studied as there is very little work in that pressure ratio and still achieve temperature high
area. enough to transfer heat successfully in the condenser.
(xi) It has been established that double sided impel- Further, since the pressure is low, the condenser
ler is better than single sided impeller in terms tubes can be thinner and could result in cost savings.
of reducing axial thrust. Also, in the case of Optimization studies directed at understanding the
peripheral blades, chevron (V-shaped) blades cost savings due to thinner condenser tubes and
are better for head development. excess operation cost due to lower compressor effi-
(xii) No work has been found on stress analysis of ciency can be carried out.
regenerative flow machines. Since the clearan- No studies reporting the stress analysis carried out
ces are low and rotational speeds are high, high in the impeller of a regenerative flow machine was
stresses and elongation of material is possible, found. Such studies are even more important due to
which can result in the reduction of clearance the high rotational speeds and temperatures involved.
space, thereby altering the design and even This is an avenue open to research and could enable
worse, cause the metal to metal contact. the choice of lighter, stronger and cheaper materials.
Further, if suitable materials can be identified There are no studies carried out to understand the
which can withstand the stresses, we may also wear and tear in the impeller and compare it to reg-
be able to opt for rapid prototyping and man- ular turbomachines. This would help in life cycle
ufacture complex shaped blades and impellers. analysis.
(xiii) As of now, these machines can only handle Lastly, 2nd law analysis of systems using regenera-
clean fluids due to the possibility of wear. tive flow machines was never reported. If carried out,
20 Proc IMechE Part A: J Power and Energy 0(0)

it could throw some light on where the irreversibilities 15. Senoo Y. A comparison of regenerative pump theories
are occurring, thereby helping in improving the effi- supported by new performance data. Trans ASME
ciency. Moreover, such studies could help in compar- 1956; 78: 1091–1102.
ing systems having regular machines and those having 16. Shimosaka M and Yamazaki S. Research on the char-
acteristics of regenerative pump: 1st report, influences
regenerative machines and an overall system perfor-
of flow channel and impeller. Bull JSME 1960; 3:
mance evaluation could be compared.
185–190.
17. Shirinov A and Oberbeck S. Optimisation of the high
Declaration of Conflicting Interests vacuum side channel pump. In: Proceedings of the 7th
The author(s) declared no potential conflicts of interest with international conference on compressors and their sys-
respect to the research, authorship, and/or publication of tems, London, UK, September 2011, pp. 5–6.
this article. 18. Jeon SY, Yoon JY and Jang CM. Optimal design of a
novel ‘S-shape’ impeller blade for a microbubble pump.
Funding Energies 2019; 12: 1793.
The author(s) received no financial support for the research, 19. Choi WC, Yoo IS, Park MR, et al. Experimental study
authorship, and/or publication of this article. on the effect of blade angle on regenerative pump per-
formance. Proc IMechE, Part A: J Power and Energy
2013; 227: 585–592.
ORCID iD
20. Badami M. Theoretical and experimental analysis of
R Sivakumar https://orcid.org/0000-0003-1303-0519 traditional and new periphery pumps. SAE technical
paper, 24 February 1997.
References 21. Yahya SM. Turbines compressors and fans. New York,
NY: Tata McGraw-Hill Education, 1987.
1. Karlsen-Davies ND and Aggidis GA. Regenerative
22. Mason SC. Influence of internal geometry upon regener-
liquid ring pumps review and advances on design and
ative pump performance. Doctoral Dissertation,
performance. Appl Energy 2016; 164: 815–825.
Massachusetts Institute of Technology, USA, 1957.
2. Brown A. A comparison of regenerative and centrifugal
23. Song JW, Engeda A and Chung MK. A modified
compressors. West Lafayette, IN: Purdue University,
theory for the flow mechanism in a regenerative flow
Purdue e-Pubs, 1972, pp.229–233.
pump. Proc IMechE, Part A: J Power and Energy 2003;
3. McMillan JO. Regenerative turbine pump. Patent
2113116, USA, 5 April 1938. 217: 311–321.
4. Lehmann W. Seitenkanalpumpen und Kombisysteme. 24. Meakhail T and Park SO. An improved theory for
In: Faragallah WH (ed.) Seitenkanal- regenerative pump performance. Proc IMechE, Part
Str€omungsmaschinen, 1992, pp.98–122 (in German). A: J Power and Energy 2005; 219: 213–222.
5. Siemen O and Hinsch JA. Circulation pump which acts 25. Quail F, Scanlon T and Stickland M. Study of a regen-
partially with sealing circling auxiliary liquid. Patent erative pump using numerical and experimental
413435, Germany, 3 March 1920. techniques. In: 8th European turbomachinery conference,
6. Nejad J, Riasi A and Nourbakhsh A. Parametric study Graz, Austria, 23 March 2009.
and performance improvement of regenerative flow 26. Badami M and Mura M. Theoretical model with exper-
pump considering the modification in blade and imental validation of a regenerative blower for hydro-
casing geometry. HFF 2017; 27: 1887–1906. gen recirculation in a PEM fuel cell system. Energy
7. EQUIPMENT R. Consider regenerative pumps for Convers Manage 2010; 51: 553–560.
low-flow/low-NPSH applications. Hydrocarbon proc- 27. Quail FJ, Scanlon T and Strickland M. Development of
essing, August 2004. a regenerative pump impeller using rapid manufactur-
8. Quail FJ, Scanlon T and Strickland M. Rapid ing techniques. Rapid Prototyping J 2010; 16: 337–344.
manufacturing technique used in the development of a 28. Quail F, Scanlon T and Stickland M. Design optimisa-
regenerative pump impeller. Rapid Prototyping J 2010; tion of a regenerative pump using numerical and exper-
16: 337–344. imental techniques. HFF 2011; 21: 95–111.
9. Raheel M. A theoretical, experimental and CFD analysis 29. Badami M and Mura M. Setup and validation of a
of regenerative flow compressors and regenerative flow regenerative compressor model applied to different
pumps for microturbine and automotive fuel devices. Energy Convers Manage 2011; 52: 2157–2164.
applications. PhD Thesis, Michigan State University, 30. Quail FJ, Scanlon T and Baumgartner A. Design study
East Lansing, MI, 2003. of a regenerative pump using one-dimensional and
10. Senoo Y. Theoretical research on friction pump. Inst three-dimensional numerical techniques. Eur J Mech
Fluid Eng 1948; 5: 23–48. B/Fluids 2012; 31: 181–187.
11. Iversen HW. Performance of the periphery pump. 31. Badami M and Mura M. Leakage effects on the perfor-
Trans ASME 1955; 77: 19–22. mance characteristics of a regenerative blower for the
12. Balje OE. Drag-turbine performance. Trans ASME hydrogen recirculation of a PEM fuel cell. Energy
1957; 79: 1291–1302. Convers Manage 2012 ; 55: 20–25.
13. Wilson WA, Santalo MA and Oelrich JA. A theory of 32. Badami M and Mura M. Comparison between 3D and
the fluid dynamics mechanism of regenerative pumps. 1D simulations of a regenerative blower for fuel cell
Trans ASME 1955; 77: 1303–1316. applications. Energy Convers Manage 2012; 55: 93–100.
14. MTH Pumps. Regenerative turbine principles, http:// 33. Lee C and Kil HG. The design, performance and CFD
www.mthpumps.com/turbine.html (accessed 13th May analyses of regenerative blower used for fuel cell
2021). system. In: 2014 4th international conference on
Sreekanth et al. 21

simulation and modeling methodologies, technologies and fluid viscosity enhancement. Mechanika 2018; 23:
applications (SIMULTECH), 28 August 2014, pp.751– 852–858.
755. Piscataway, NJ: IEEE. 50. Raheel MM and Engeda A. Systematic design
34. Karanth V and Sharma N. CFD analysis of a regener- approach for radial blade regenerative turbomachines.
ative pump for performance enhancement. In: World J Propul Power 2005; 21: 884–892.
conference on applied sciences, engineering & technology, 51. Yoo IS, Park MR and Chung MK. Improved momen-
Kathmandu, Nepal, 27–29 September 2014. tum exchange theory for incompressible regenerative
35. Jang CM and Jeon HJ. Performance enhancement of 20 turbomachines. Proc IMechE, Part A: J Power and
kW regenerative blower using design parameters. Int J Energy 2005; 219: 567–581.
Fluid Mach Syst 2014; 7: 86–93. 52. Mekhail TA, Dahab OM, Sadik MF, et al. Theoretical,
36. Karanth VK, Manjunath MS and Sharma YN. experimental and numerical investigations of the effect
Numerical study of a self priming regenerative pump of inlet blade angle on the performance of regenerative
for improved performance using geometric modifica- blowers. OJFD 2015; 05: 224–237.
tions. Int J Curr Eng Technol 2015; 5: 105–109. 53. Hollenberg JW and Potter JH. An investigation of
37. Maity A, Chandrashekharan V and Afzal MW. regenerative blowers and pumps. J Eng Ind 1979; 101:
Experimental and numerical investigation of regenera- 147–152.
tive centrifugal pump using CFD for performance 54. Raheel M, Engeda A, Hamrin D, et al. The perfor-
enhancement. Int J Curr Eng Technol 2015; 5: mance characteristics of single-stage and multistage
2898–2903. regenerative flow compressors for natural gas compres-
38. Lee C, Kil HG and Kim KY. The performance analysis sion application. Proc IMechE, Part C: J Mechanical
method with new pressure loss and leakage flow models Engineering Science 2003 ; 217: 1221–1239.
of regenerative blower. Int J Fluid Mach Syst 2015; 8: 55. Cates PS. Study of the peripheral compressor. Oak
221–229. Ridge, TN: Oak Ridge Gaseous Diffusion Plant, 1966.
39. Annadurai B, Arularasu M, Sudhakar T, et al. 56. Griffini D, Salvadori S, Carnevale M, et al. On the
Enhancement of regenerative pump performance development of an efficient regenerative compressor.
using rapid manufacturing techniques. Asia J Res Soc Energy Procedia 2015; 82: 252–257.
Sci Hum 2016; 6: 1824–1839. 57. Abdalla H. A theoretical and experimental investigation
40. Nejadrajabali J, Riasi A and Nourbakhsh SA. Flow of the regenerative pump with aerofoil blades. PhD
pattern analysis and performance improvement of Thesis, The Royal Military College of Science, UK.
regenerative flow pump using blade geometry modifica- 58. Andrew DN. The calculation of flow in regenerative
tion. Int J Rotating Mach 2016; 2016: 1–16. turbomachines by a streamline curvature method.
41. Zhang F, Fleder A, B€ ohle M, et al. Effect of suction Proc IMechE, Part A: J Power and Energy 1990; 204:
side blade profile on the performance of a side channel 121–129.
pump. Proc IMechE, Part A: J Power and Energy 2016; 59. Quail FJ, Stickland M and Baumgartner A. A one-
230: 586–597. dimensional numerical model for the momentum
42. Zhang F, B€ ohle M and Yuan S. Experimental investi- exchange in regenerative pumps. J Eng Gas Turbines
gation on the performance of a side channel pump Power 2011; 133: 093001, 1–7.
under gas–liquid two-phase flow operating condition. 60. Nejadali J. Calculation of flow in incompressible regen-
Proc IMechE, Part A: J Power and Energy 2017; 231: erative turbo-machines with bucket form blades based
645–653. on the geometry of flow path. HFF 2019; 29:
43. Fathi AR, Jalilvand R, Shirani E, et al. A new analyt- 2606–2621.
ical relation for performance of small regenerative tur- 61. Dixon SL and Hall C. Fluid mechanics and thermody-
bine pumps. In: Exergy for a better environment namics of turbomachinery. Oxford: Butterworth-
and improved sustainability. Cham: Springer, 2018, Heinemann, 2013.
pp.613–630. 62. Engeda A. Flow analysis and design suggestions for
44. Pei J, Zhang F, Appiah D, et al. Performance predic- regenerative flow pumps. ASME paper FEDSM2003-
tion based on effects of wrapping angle of a side chan- 45681, 2003.
nel pump. Energies 2019; 12: 139. 63. Balje OA. A study on design criteria and matching of
45. Zhang F, Chen K, Appiah D, et al. Numerical delinea- turbomachines: Part a – similarity relations and design
tion of 3D unsteady flow fields in side channel pumps criteria of turbines. Journal of Engineering for Power
for engineering processes. Energies 2019; 12: 1287. 1962; 84: 83–102.
46. Sixsmith H and Altmann H. A regenerative compres- 64. Balje OE. A study on design criteria and matching of
sor. J Eng Ind Trans ASME 1977; August: 637–647. turbomachines: Part B – Compressor and pump perfor-
47. Song JW, Raheel M and Engeda A. A compressible mance and matching of turbocomponents. Journal of
flow theory for regenerative compressors with aerofoil Engineering for Power 1962; 84: 103–114.
blades. Proc IMechE, Part C: J Mechanical Engineering 65. Dick E. Fundamentals of turbomachines. Berlin:
Science 2003; 217: 1241–1257. Springer, 2015.
48. Nejad J, Riasi A and Nourbakhsh A. Efficiency 66. Dubey M. Prasad BVSSS, Nema A. Turbomachinery,
improvement of regenerative pump using blade profile McGraw Hill Education (India), 1st Edition, 2019.
modification: experimental study. Proc IMechE, Part 67. Grabow GH. Das erweiterte “cordier”-diagramm für
E: J Process Mechanical Engineering 2019; 233: fluidenergiemaschinen und verbrennungsmotoren.
448–455. Forsch Ing-Wes 1993; 59: 42–50.
49. Nejadrajabali J, Riasi A and Nourbakhsh SA. 68. Thomas C, Margraf M, Stoeckel C, et al.
Performance evaluation of a regenerative pump under Implementation of scroll compressors into the cordier
22 Proc IMechE Part A: J Power and Energy 0(0)

diagram. IOP Conf Ser Mater Sci Eng 2019; 604: 87. https://www.marketwatch.com/press-release/regenera
012079. tive-turbine-pumps-market-revenue-was-263-million-
69. Bláha J and Smid V. Allgemeine klassifikation der usd-in-2021-and-will-reach-362-million-usd-in-2026-
anwendungsgebiete von pumpen und verdichtern. with-a-cagr-of-542-2021-03-10?tesla=y (accessed 26
Würzburg: Maschinenmarkt, 1981. March 2021).
70. Ray GS and Sinha BK. Computation of centrifugal 88. Raheel M and Engeda A. Current status, design and
stresses in a radial-flow impeller. Comp Struct 1991; performance trends for the regenerative flow compres-
40: 731–740. sors and pumps. In ASME international mechanical
71. Bert CW and Paul TK. Failure analysis of rotating engineering congress and exposition, New Orleans,
disks. Int J Solids Struct 1995; 32: 1307–1318. Louisiana, USA, 1 January 2002, Vol. 36444, pp.99–
72. Satish R and Kallurkar SP. Study of stresses developed 110.
on the impeller of centrifugal pump at different speed
using Ansys. Int J Eng Res Technol 2017; 5: 1–7. Appendix
73. Raj PP, Gupta PS, Kumar PP, et al. Model and analysis
of rotor and impeller of eight stage centrifugal pump. Notation
Mater Today 2020 ; 21: 175–183.
74. Timoshenko S and Goodier JN. Theory of elasticity.
3rd ed. New York, NY: McGraw Hill Education, 2017. d, D diameter (m)
75. Regenerative Vs. centrifugal blowers explained, https:// g acceleration due to gravity (m/s2)
www.stanmech.com/articles/regenerative-versus-centrif H pressure head (m)
ugal-blowers-what-are-they-and-when-do-i-use-what lpm litres per minute
(accessed 24 March 2021).
m mass (kg)
76. https://beckerpumps.com/regenerative-blower/
N rotational speed (s1)
(accessed 24 March 2021).
77. https://www.airpowerproducts.com/blog/about-regener Q volumetric flow rate (m3/s)
ative-blowers/ (accessed 24 March 2021). R specific gas constant (J/(kg K))
78. https://www.goorui.com/ (accessed 24 March 2021). T temperature (K)
79. https://www.speck-pumps.com/en/pc/speck/industrial- u blade tip speed (m/s)
technology/regenerative-turbine-pumpssmall-centrifu Z number of blades
gal-pumps/mit-gleitringdichtung/?v=3a52f3c22ed6 g efficiency
(accessed 24 March 2021). q density (kg/m3)
80. https://www.psgdover.com/blackmer/products/regener / flow coefficient
ative-turbine-pumps/rc20-regenerative-turbine-pump w pressure (head) coefficient
(accessed 24 March 2021).
x angular velocity (rad/s)
81. https://www.starpumpalliance.com/regenerative-tur
bine-pumps (accessed 24 March 2021).
82. https://www.rothpump.com/regenerative-turbine- Subscripts
pumps-volatile-fluids.html (accessed 24 March 2021).
83. https://www.inrag.ch/en_industrial-pumps-regenera
tive-turbine-pumps.html (accessed 24 March 2021).
84. http://www.pumpscout.com/supplier-directory/ebsray- Hyd hydraulic
pumps-cid312.html (accessed 24 March 2021). m mean
85. https://www.bestpumpsindia.com/centrifugal-regenera
s specific
tive.php (accessed 24 March 2021).
86. https://www.indiamart.com/proddetail/crompton-self-
1 inlet
priming-regenerative-pump-20431465673.html 2 exit
(accessed 26 March 2021).