AP1_01.vp


1 Introduction
In all areas of particulate technology where powders are

handled, structures in contact with the powder will exhibit
wear. In some applications this wear may be so severe as to
limit the life of a component or plant, while in others it may
be negligible. All particles will cause some wear, but in general
the harder they are, and the more angular in shape, the more
severe will be the wear [1].

Mixing processes including the suspension of solids in
liquids have shown that the rate at which the impeller wears is
a strong function of the impeller blade inclination angle
(pitch angle) as well as the impeller tip speed [2]. The impact
rate of the solid particles affecting the blade of the pitched
blade impellers was studied experimentally with the aim of
explaining the significant dependency of impact rate on
impeller tip speed and impeller diameter [3, 4]. The impact
mostly occurs in the outer and upper corners of the blades
for a pitched blade impeller pumping downwards. This
fact corresponds to the results of experimental studies [5, 6]
investigating the erosive wear of pitched blade impellers in a
highly erosive suspension, where the impeller blade leading
edge is worn, together with a reduction of the surface blade.
Then, e.g., the impeller pumping capacity decreases.

This study deals with an attempt to express mathemati-
cally the shape of the worn blade of a pitched blade impeller
in two sorts of industrial suspension, i.e., in a suspension of
flue gas desulphurization product (gypsum) and a suspension
of silicious sand, and to interpret the influence of the impeller
erosion wear on the surface or on the thickness of the worn
blade of the various pitch angles under its erosion process.

2 Experimental
Experiments were carried out on pilot plant mixing

equipment (see Fig. 1), with water as a working liquid and two
types of solid phases:

1. silicious sand (volumetric concentration CV = 10 % vol.,
mean diameter of particles d = 0.4 mm),

2. waste gypsum (CaSO4 . 2H2O) from a thermal power station
(CV = 18.3 % vol., d = 0.1 mm) after a desulphurization
process.
Pitched blade impellers with four adjustable inclined

plane blades (see Fig. 2) made of rolled brass or construction
steel were used for the erosion tests. As an independent
variable we took a pitch angle � of the impeller with inclined
plane blades, and in all cases the frequency of the impeller
revolution, n, was kept constant under conditions of complete
homogeneity of the suspension. For the experiments we used
a cylindrical metal baffled vessel made of stainless steel (see
Fig. 1).

©  Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 23

Acta Polytechnica Vol. 41  No.1/2001

Erosion Wear of Axial Flow Impellers
in a Solid-liquid Suspension

I. Fořt, J. Medek, F. Ambros

A study was made of the erosion wear of the blades of pitched blade impellers in a suspension of waste gypsum from a thermal power station
(vol. concentration CV = 18.3 %, particle mean diameter d = 0.1 mm, degree of hardness “2.5”) and silicious sand (CV = 10 %,
d = 0.4 mm, degree of hardness “7.5”) in water under a turbulent flow regime of agitated charge when complete homogeneity of the
suspension was achieved. Experiments were carried out on pilot plant mixing equipment made of stainless steel (diameter of cylindrical
vessel T = 390 mm, diameter of impeller D = 100 mm, impeller off-bottom clearance h = 100 mm) equipped with four wall radial baffles
(width b = 39 mm) and an impeller with four inclined plane blades (pitch angle � = 20°, 30°, 45°, relative blade width W/D = 0.2) made
either of rolled brass (Brinell hardness 40–50 BH) or of structural steel (Brinell hardness 100–120 BH) always pumping the liquid
downwards towards the flat vessel bottom. Two erosion process mechanisms appear, depending on the hardness of the solid particles in the
suspension: while the particles of gypsum (lower hardness) generate a uniform sheet erosion over the whole surface of the impeller blade, the
particles of silicious sand (higher hardness) generate wear of the leading edge of the impeller blades, together with a reduction of the surface
of the worn blade. The hardness of the impeller blade also affects the rate of the erosion process: the higher the hardness of the impeller blade,
the lower the wear rate of the blade. This study consists of a description of the kinetics of the erosion process of both mechanisms in dependence
on the pitch angle of the tested impellers. While the wear of the leading edge of the blade exhibits a monotonous dependence on the pitch
angle, the sheet erosion exhibits the maximum rate within the interval of the pitch angles tested � � � 20°; 45°�. However, generally the pitch
angle � = 45° seems to be the most convenient angle of blade inclination when both investigated mechanisms of the blade erosion process are
considered at their minimum rate.

Keywords: pitched blade impeller, erosion wear, solid-liquid suspension.

Fig. 1: Geometry of pilot plant mixing equipment
(H = T = 290 mm, h = D = 100 mm, four baffles, b/T = 0.1)



During the experiments with silicious sand the shape of
the blade profile was determined visually from magnified

copies of the worn impeller blades on photosensitive paper
(magnification ratio: approx. 2:1). The average course of the
blade profile for the given length of the erosion process was
determined as an average curve from four individual worn
impeller blades. The surface of the worn blades was deter-
mined from their magnified copies using a planimeter.

During the experiments with waste gypsum the local
thickness of the blade si,j ( j = 1, 2, 3, 4) along the blade length
(see Fig. 3) was determined by means of a contact microme-
ter with an accuracy �0.01 mm. The erosion process when
a suspension of waste gypsum was used exhibits no change in

the blade profile, e.g., at its leading edge. Only its thickness
changes.

3 Results and Discussion
Two series of experiments were evaluated: studies of the

erosion wear of the impeller blades under the influence of
particles of silicious sand (higher hardness) and under the
influence of waste gypsum (lower hardness).

Axial impeller in a suspension of silicious sand
During the erosion process of the pitched blade impeller

caused by solid particles of silicious sand, its metal plane blade
changes shape mainly around its leading edge [6]. The radial
profile of the worn plane blade can be considered in
dimensionless exponential form (see Fig. 4):

� � � �� �H X C k X� � �1 1exp , (1)
where the dimensionless axial (vertical) coordinate along the
width of the blade is

� �H x W� z (2)
and the dimensionless longitudinal (radial) coordinate along
the radius of the blade is

X x D� 2 . (3)

The values of the parameters of Eq. (1) – the wear rate
constant k and the geometric parameter of the worn blade
C – were calculated by the least squares method from the
experimentally found profile of the worn blade; each curve
was calculated from 14–16 points (X, H) with a regression
coefficient better than R = 0.970.

Figs. 5 and 6 show the dependence (all the dependences
mentioned here were calculated from experimental data by
the least squares method) of the wear rate constant k on the
duration of erosion process t for the given impeller pitch
angle � both for rolled brass and for construction steel. The
independence of the wear rate constant from the duration
of the erosion process proves the reliability of the chosen
analytical function [1] describing the radial profile of the
worn blade. Nevertheless, parameter k depends significantly
on the pitch angle: the lower the pitch angle, the higher the
rate of the erosion process of the pitched blade impeller.
This result is in excellent accordance with the fact (1) that
metals tend to suffer the most severe erosion wear at impact
angles of 20° to 30° (measured from the plane of the blade
surface), i.e., the angles at which the particles strike the blade
surface.

Figs. 7 and 8 show the dependence of the worn blade
geometric parameter C on quantity T. For both tested metals
this parameter is linearly dependent on the duration of the

24 ©  Czech Technical University Publishing House http://ctn.cvut.cz/ap/

Acta Polytechnica Vol. 41  No.1/2001

W

�

�
D

� D0

4
x

90
°

Fig. 2: Design of a pitched blade impeller with four inclined
plane adjustable blades ( D = 100 mm, D0 = 25 mm,
W = 20 mm, s = 1.2 + 0.04 mm (experiments with silicious
sand), s = 1.0 + 0.04 mm (experiments with waste gyp-
sum), � = 15°, 20°, 30°, 45°, n1 = 479 RPM, n2 = 640 RPM)

Fig. 3: Distribution of measurement points for determining of the
local blade thickness along the blade length

Fig. 4: Radial profile of the leading edge of the worn blade of
a pitched blade impeller



©  Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 25

Acta Polytechnica Vol. 41  No.1/2001

process. This course confirms fairly well the strong dependen-
ce of the studied erosion wear on its duration. Moreover, this
is in accordance with the experimentally determined
distribution of the rate of particle impact along the blade
of the pitched blade impeller indicated on the artificially
soft surface of the impeller blades (4).

Figures 9 and 10 illustrate the dependences of the experi-
mental parameters of function (1) on the impeller pitch an-
gle. The linear form for the two tested metals

� �� � �k 8 79 10 48. sin cos .� � ,
rolled brass, � � �15°, 45°� (4a)

� �� � �k 4 73 6 31. sin cos .� � ,
construction steel, � � �20°, 45°� (4b)

characterizes the relation between the erosion wear of the
impeller blade and the corresponding (axial) component

of the mean velocity in the discharge stream leaving a pitched
blade impeller [5] expressed by means of the product
sin� cos�. At the same time our results confirm the general
view of the relation between the hardness of the worn metal
and its wear rate: the higher the hardness of the worn metal,
the lower its wear rate.

The linear forms

C t � � � 	 �� �214 10 121 104 2. .� ,
rolled brass, � � �20°, 45°� (5a)

C t � � � 	 �� �0 41 10 0 27 104 2. .� ,
construction steel, � � �20°, 45°� (5b)

characterize the fact that the geometric parameter of the
profile of a worn blade C depending on the duration of
the erosion period corresponds to the development of the
exponential profile of the worn blade with increasing time
T. Similarly as for parameter k the geometric parameter
of the profile of the worn blade exhibits lower values for
construction steel (Brinell hardness 100–120 HB) than for
rolled brass (Brinell hardness 40–50 HB).

Figs. 11 and 12 illustrate the dependence of the surface of
the impeller worn blade Aworn (related to the unworn blade
surface A0) on the duration of the erosion process for both
rolled brass and construction steel impeller blades as well as
being compared with predicted values [6]. The rate of the
blade surface reduction can be expressed in linear form of the
found dependence, i.e., for rolled brass:

A A tworn 0 1 0 0014� � . , � = 20° , (6a)
A A tworn 0 1 0 0008� � . , � = 45° , (6b)

5

7

9

0 100 200 300
t [h]

�
k

[-
]

k = 7.612�

k = 6.77�

k = 6.075�

– 20°,� 	 – 30°,� 	 – 45°� 	

Fig. 5: Dependence of the wear rate constant on the duration of
the erosion process (impeller blades made of rolled brass)

– 20°,� 	 – 35°� 	

3 5.

4 5.

5.5

0 200 400 600
t [h]

�
k

[-
]

k = - 4.30

k = - 4.78

Fig. 6: Dependence of the wear rate constant on the duration of
the erosion process (impeller blades made of construction
steel)

– 20°,� 	 – 30°,� 	 – 45°� 	

0 00.

0 40.

0 80.

1 20.

1.60

0 50 100 150 200 250 300
t [h]

C
[-

]

C t= 0.0053

C t= 0.0079

C t= 0.0025

Fig. 7: Dependence of the worn blade geometric parameter on
the duration of the erosion process (impeller blades made
of rolled brass)

0

0 2.

0.4

0.6

0.8

0 100 200 300 400 500 600
t [h]

C
[-

]

C t= 0.0018

C t= 0.0011

– 20°,� 	 – 35°� 	

Fig. 8: Dependence of the worn blade geometric parameter on
the duration of the erosion process (impeller blades made
of construction steel)

0

0 005.

0.01

10 20 30 40 50

[ ° ]

C
t/
[h

]

Fig. 10: Dependence of the worn blade geometric parameter on
the impeller pitch angle

3

5

7

9

0 2. 0 3. 0 4. 0 5. 0.6

�
k

[-
]

Fig. 9: Dependence of the wear rate constant on the impeller
pitch angle



and for construction steel:

A A tworn 0 1 0 0004� � . , � = 20°, (7a)

A A tworn 0 1 0 0003� � . , � = 35°. (7b)

All the above mentioned relations express the complex
influence of the duration of the erosion process and the hard-

ness of the impeller blade on the rate of its surface reduction,
i.e., the positive effect of both the increase of the impeller
pitch angle and the increase of the metal hardness on
the rate of the blade erosion wear. As follows from Fig. 11,
the proposed model of the erosion wear of pitched blade
impellers (see Eq. 1) enables us to make a sufficiently accu-

rate theoretical calculation of the rate of the surface blade
reduction [6].

Axial impeller in a suspension of waste gypsum
During the erosion process when a suspension of the

waste gypsum was used, no change of the blade profile was
observed, and only its thickness was changed. Then for the
given duration of erosion period t the average thickness of the
blade was calculated (see Fig. 3)

s si i j
j

�

�

14

1

4

, (8)

as well as its standard deviation

� �
 s i i ji s s� ���
�
��


1
3

2 2
, (9)

and, finally, the relative change of the average blade thickness
for the given moment of the erosion process was determined

�s
s

s s
s

i i

0

0

0
�

�
(10)

where s0 is the initial average thickness of the unworn blade.

Table 1 and Fig. 13 illustrate the dependence of quantity
�si/s 0 on the duration of the erosion process. The values of
quantity 
si confirm that the uniform sheet erosion of the
impeller blade can be considered for both metals tested when
waste gypsum from the thermal power station is used. Its
degree of hardness exhibits quite a different effect on the

26 ©  Czech Technical University Publishing House http://ctn.cvut.cz/ap/

Acta Polytechnica Vol. 41  No.1/2001

0 6.

0 8.

1

1.2

0 50 100 150 200 250

t [h]

A
w

o
rm

/A
o

[-
]

� 	 20°: – theoretically predicted values, – experimental values
� 	 45°: – theoretically predicted values, – experimental values

Fig. 11: Dependence of the surface of the worn blade on the
duration of the erosion process – impeller blades made of
rolled brass

0 84.

0 88.

0 92.

0 96.

1

1.04

0 50 100 150 200 250 300 350 400

t [h]

A
w

or
m

/A
o
[-

]

– 20°,� 	 – 35°� 	
Fig. 12: Dependence of the surface of the worn blade on the

duration of the erosion process – impeller blades made of
construction steel

I. Impeller made of rolled brass

� = 20°

t [h] 0 93 141 183 241

si [mm] 1.020 0.9780 0.9400 0.9200 0.8900


si [mm] 0 0.02062 0.015 0.02062 0.03162

�si/ s0 �100 [%] 0 4.2 7.8 9.8 12.5

� = 30°

t [h] 0 91 118 256 329 444 521

si [mm] 1.040 0.870 0.850 0.780 0.740 0.670 0.640


si [mm] 0 0.01414 0.01732 0.03109 0.0216 0.02828 0.01915

�si/ s0 �100 [%] 0 16.3 18.2 25.4 28.8 35.6 37.0

Table 1. Dependence of blade thickness on duration of the erosion process

0 4.

0 6.

0 8.

1

1.2

0 100 200 300 400 500 600

t [h]

1
�

s/
so

– 20°,� 	 – 30°,� 	 – 45°� 	

Fig. 13: Dependence of relative blade thickness on the duration
of the erosion process – impeller blades made of rolled
grass



metal surface than previously used silicious sand with degree
of hardness approx. three times greater than the first one. It
can be proposed: the different mean particle diameters of
the two powders under study need not affect the erosion
mechanism [1], i.e., their differences are caused only via their
different hardnesses.

The relative change of the average blade thickness �s si 0
depends both on the type of metal and on the blade pitch
angle �. According with the above mentioned results, when
silicious sand was used the blade made of construction steel
exhibits a lower erosion rate than the blade made of rolled
brass. The dependence of the quantity �s si 0 on the duration
of the erosion process can be expressed in linear form

�s s ti 0 1 0 0005� � . , � = 20°, (11a)

�s s ti 0 1 0 0008� � . , � = 30°, (11b)

�s s ti 0 1 0 0007� � . , � = 45°. (11c)

The shape of relation � ��s s ti 0 � f does not exhibit
a monotonous dependence on the impeller pitch angle as in
the case of the harder powder used previously, and we can
consider the maximum rate of the erosion process when the
impeller blade pitch angle � = 30°. It follows from the above
results that more tests with various powders with different
particle hardnesses should be made, in order to gain better
and fuller evidence of the influence of solid particles on the
rate of erosion of pitched blade impellers.

4 Conclusions
Two mechanisms of the erosion process of pitched blade

impellers in a solid-liquid suspension were found in this
study:
a) particles of waste gypsum (lower hardness) generate uni-

form sheet erosion over the whole surface of the impeller
blade,

b) particles of silicious sand (higher hardness) generate pre-
dominantly erosion wear of the leading edges of the
impeller blade.

The higher the hardness of the impeller blade material,
the lower the wear rate of the blade.

While the wear rate of the blade leading edge exhibits
a monotonous dependence on the impeller pitch angle, sheet

erosion exhibits the maximum rate within the interval of the
pitch angles tested � � �20°, 45°�

The impeller pitch angle � = 45° seems to be the most
convenient angle of plane blade inclination when the two
investigated mechanisms of the blade erosion exhibit their
minimum rate.

All the results mentioned are valid for turbulent regime of
flow agitated charge.

5 List of Symbols
A dimensionless surface of impeller blade (A = A/W(D/2))
a surface of impeller blade, m2

b baffle width, m
C geometric parameter of the profile of a worn blade
CV volumetric concentration, m

3/m3

D impeller diameter, m
D0 hub diameter, m
d particle diameter, m
H dimensionless axial coordinate of the profile of worn

blade
HL height of liquid from bottom of vessel, m
h off bottom impeller clearance, m
k wear rate constant
n impeller frequency of revolution, s–1

R regression coefficient
s thickness of impeller blade, m
t time, h
W width of impeller blade, m
X longitudinal coordinate of impeller blade, m
z axial (vertical) coordinate along the width of the blade,

m
� pitch angle of blade, °

 standard deviation

Subscripts and superscripts
i summation index
j summation index
o initial value
worn worn
–– average value

©  Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 27

Acta Polytechnica Vol. 41  No.1/2001

� = 45°

t [h] 0 66 111 155 254

si [mm] 1.030 0.960 0.920 0.900 0.860


si [mm] 0 0.0238 0.0245 0.0294 0.02986

�si/ s0 �100 [%] 0 6.8 10.7 12.6 16.5

II. Impeller made of construction steel

� = 45°

t [h] 0 50 93 163 235

si [mm] 1.0400 1.0275 1.0275 1.0250 1.0250


si [mm] 0 0.00957 0.00957 0.00577 0.00577

�si/ s0 �100 [%] 0 1.2 1.2 1.44 1.44



This research has been supported by the Research Project
of the Ministry of Education of the Czech Republic
J04/98: 212200008.

REFERENCES

[1] Hutchings, I. M.: Wear by particulates. Chem. Eng. Sci.,
1987, 42, No. 4, pp. 869–878

[2] Kipke, K.: Erosiver Verschleiss von Rührorganen. Chem.-
Ing.- Tech., 1980, 52, pp. 658–659

[3] He, Y., Takahashi, K., Nomura, T.: Particle impeller im-
pact for a six bladed 45° pitched blade impeller in an agitated
vessel. J. Chem. Eng. Jap., 1995, 28, No. 6, pp. 786–789

[4] Kee, K. C., Reilly, C. D.: An experimental method for obtain-
ing particle impact frequencies and velocities on impeller
blades. Proceedings of the 10th European Conference on
Mixing, Delft (The Netherlands), Editors: H. E. A. van
den Akker, J. J. Derksen, Elsevier, Amsterdam 2000,
pp. 231–238

[5] Fořt, I., Ambros, F., Medek, J.: Study of wear and tear of
axial flow impellers. Proceedings of the Fluid Mixing VI
Conference (Editor: H. Benkreira), Bradford (England)
1999, pp. 59–68

[6] Fořt, I., Ambros, F., Medek, J.: Study of wear of pitched
blade impellers. Acta Polytechnica 2000, 40, No. 516,
pp. 5–10

Doc. Ing. Ivan Fořt, DrSc.
Ing. František Ambros, Csc.
Dept. of Process Engineering
Czech Technical University in Prague
Faculty of Mechanical Engineering
Technická 4, 166 07 Praha 6
Prof. Ing. Jaroslav Medek, Csc.
Dept. of Process Engineering
Technical University
Faculty of Mechanical Engineering
Technická 2, 616 69 Brno

Acta Polytechnica Vol. 41  No.1/2001

28 ©  Czech Technical University Publishing House http://ctn.cvut.cz/ap/