See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/330741064

DYNAMIC DESIGN FOR GRINDING MILL FOUNDATIONS

Article · January 2019

CITATIONS READS
0 688

1 author:

Yingcai Han
Beijing University of Technology
143 PUBLICATIONS   638 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Piling Dynamics View project

Dynamics of piles View project

All content following this page was uploaded by Yingcai Han on 30 January 2019.

The user has requested enhancement of the downloaded file.

 CSCE 2014 4th International Structural Specialty Conference - 4e Conférence internationale
spécialisée sur les structures 2014 de la SCGC

Halifax, NS
May 28 to 31, 2014 / 28 au 31 mai 2014

DYNAMIC DESIGN FOR GRINDING MILL FOUNDATIONS

Yingcai Han

Fluor, 1075 W. Georgia Street, Vancouver BC, V6E 4M7 Canada

Abstract: The dynamic analysis of grinding mill foundation is a typical problem of soil-structure interaction.
The sub-structure method is used to carry out the dynamic analysis and design in practice. The concrete
mat foundation and piers are modelled by FEM, and the stiffness and damping of soil (rock) are
generated by a computer program. Then the impedance parameters are input into to the FEM model as
the base boundary condition. Radiation damping is the dominant energy dissipation mechanism in most
dynamically loaded foundation systems, and the values of radiation damping have been modified in the
program based on dynamic tests. In this study a practical case of grinding mill foundation is investigated
to illustrate the approach and the dynamic behaviour of structure. The rock is weathered in the site and
capacity is not high enough to provide strong stiffness of supporting piers. The weakened rock is a
challenge to the design of grinding mill foundation in this case. Different design options are compared to
obtain the better solution, that is, the vibrating amplitudes calculated should be less than the allowable
vibrating limit, and the construction cost should be reduced.

1. INTRODUCTION

The foundation acts as a rigid body assumed normally in classical empirical methods for dynamic
analysis, (e.g., Barkan1962). However, the structure of a mill foundation and piers with large dimensions
is flexible rather than a rigid body. Numerical methods such as the general finite element method (FEM)
are also difficult to apply, as the direct simulation of radiation damping is not possible. Radiation damping
is the dominant energy dissipation mechanism in most dynamically loaded foundation systems. The
dynamic analysis is challenging for the flexible mill foundations using the standard analytical or numerical
methods, and it is a typical problem of soil-structure interaction. Another challenge is that the rock is
weathered at the elevation of foundation bottom for this project, and the shear wave velocity is 660 m/sec.
The bearing capacity of soil is not high enough, and classification is very weak rock. The diameter of
grinding mills may be much large in mining industry. A practical case of ball mill foundation is examined
herein. The diameter of mill is 8.0 m with length of 13.1 m, operating at 12 rpm. The height of mill shaft is
18.1 m above ground. The weight of mill and charge (ore and grinding media) is 3,452 tons.

The sub-structure method is used for dynamic analysis of the ball mill foundation, that is, the structure
and soil are considered as two parts separately. The structure (mat foundation and piers) are modelled by
FEM model. The impedance of soil (stiffness and damping) are generated by a computer program
DynaN, and then input to the FEM model as the base boundary condition. A series of dynamic
experiments had been done in the field to verify the values of radiation damping, and it can be generated
by the program.

CST11-1
2. FOUNDATION OF GRINDING MILLS

There are a number of grinding mills in operation around the world with diameter up to 8 m. Aspect ratio
L/D varies for ball mills, L/D >1, typically 1.5 to 2.5 factor, as shown in Figure 1. Where L is the length and
D is the diameter of mill. Mining operations continually invest in new technologies to improve their energy
efficiency and capacity in their grinding circuit. There is no doubt that mills size will continue increasing.

Grinding mills are designed to break mineral ore into smaller pieces by the action of attrition and impact
using grinding media. Ball mills are basically a horizontal rotating cylinder partially filled with steel balls as
grinding media. Bearing pads are located at the end of the mills. One bearing has no axial float, while
the second bearing has sufficient float to accommodate the thermal expansion of the mill. In the mining
industry, ball mills normally operate with an approximate ball charge of 30% with a rotational speed close
to 11 rpm. The mill is fed at one end of the cylinder and the discharge is at the other.

The grinding mills are manufactured using steel plate and some casting parts. Mill design can be trunnion
supported or shell supported. In shell supported design, the mill shell supports the weight at the
circumference through T-shaped fabricated riding rings and slipper pad bearings. The load of the mill
body, lining, and charge is transferred directly from the sliding ring to the bearing shoes and then to the
foundations.

The motion of charge, rocks and balls, in grinding mills is performed by metal liners installed in the mills
shells. The purpose of installing liners in grinding mills is to protect the mill shell from wear and efficiently
transfer the energy to the grinding media. Liners lift the charge producing a cascade motion of the charge
inside the mill. The frequency of the cascade motion, and then the charge impact frequency, is a function
of the number of lifters and the rotational speed of the mill.

The motor design has neither a shaft or bearings. The mill is used as a direct rotor, moved by poles which
are divided into a number of segments. The poles are mounted directly on the mill shell through a flange
motor carrier ring. One option is to mount this flange by using bolting connection on to the mill shell into
the fixed mill bearing side. This is the location with least axial movement of the mill due to thermal
expansion. Also the mill deflection due to bending after shut down is smallest at this location.

Figure 1 Grinding mills sketch

CST11-2
The motor air gap, distance between poles and stator frame, is defined together with the mill
manufacturer and needs to be kept in the range of 16 mm. The stator frame is designed as a self-
supporting ring construction. Usually, the stator is split in four sections to allow easy transport. It is
mounted on a motor foundation with integral stator bedplates. During installation and overhaul the stator
can be moved on these bedplates.

3. SOIL-STRUCTURE INTERACTION AND RADIATION DAMPING

Many authors have made contributions to the subject of soil-structure interaction (e.g., Dobry and
Gazetas 1988, Gazetas and Makris 1991, Benerjee and Sen 1987, and Wolf 1988). Different approaches
are available to account for dynamic soil-structure interaction but they are usually based on the
assumptions that the soil behaviour is governed by the law of linear elasticity or visco-elasticity, and that
the soil is perfectly bonded to the footing. In practice, however, the bonding between the soil and the
foundation is rarely perfect, and slippage or even separation often occurs in the contact area.
Furthermore, the soil region immediately adjacent to the foundation can undergo a large degree of
straining, which would cause the soil-structure system to behave in a nonlinear manner. A lot of efforts
have been made to model the soil-structure interaction using the 3D Finite Element Method (FEM).
However, it is too complex and costly.

Several problems of soil-structure interaction are of concern for dynamic analysis in practice. It is a
consideration how to account for the nonlinear properties of soil. As an approximate analysis, a procedure
is developed using a combination of the analytical solution and the numerical solution, rather than using
the general FEM applied to the entire system composed by soil, foundation, mill and motor. The
relationship between the foundation vibration and the resistance of the side soil layers is derived using
elastic theory (e.g., Baranov 1967). A model for the boundary zone with a non-reflective interface was
proposed for nonlinear properties of soil (e.g., Han and Sabin 1995). The effects of soil-pile-structure
interaction on dynamic response were discussed (e.g., Han 2008).

The radiation damping is important to the dynamic response of grinding mill foundations. The elastic-
wave energy is dissipated from foundation vibration in three dimensions to form the radiation damping.
The radiation damping is the dominant energy dissipation mechanism in most dynamically loaded
foundation systems. The formula of radiation damping is derived based on elastic half-space theory in
which the soil is assumed to be a homogeneous isotropic medium. As a matter of fact, however, the soil
is not a perfect linear elastic medium as assumed. A series of dynamic experiments have been done and
indicated that the damping is overestimated in the elastic half-space theory (e.g., Han and Novak 1988).
The values of radiation damping are modified and reduced in the program based on the measurements
carried out in the field.

It is also an important subject for the coupled horizontal and rocking vibration of an embedded foundation.
As for the approximate analysis, the plain strain method is considered as an efficient technique for solving
this problem (e.g., Luco 1982). The relationship between the foundation vibration and the resistance of
soil layers was derived using the elastic theory. Then, the solutions of coupled horizontal and rocking
vibration of embedded footings were formulated. Six vibration parameters horizontal stiffness Kx, damping
Cx, rocking stiffness Kand damping C, and cross coupled stiffness Kx and damping Cx are included in
the displacement expression. However, the foundation embedment conditions are very complex
practically. An inverse problem is often required in experimental research: all the parameters of the
embedded foundation need to be determined, while the dynamic response is given from measurements.
It is not convenient to back-calculate for all of the six parameters in the displacement expressions. A
simplified mathematical model of the coupled horizontal and rocking vibration of an embedded foundation
is proposed (e.g., Han 1989). Vibration tests of the foundation with different embedment were conducted
and compared with different methods. Four parameters are required in the displacement expression
based on this method, but six parameters are required in the traditional method. The four parameters Kx,
Cx, K and C can be back-calculated from the dynamic response of the foundation.

CST11-3
4. STIFFNESS AND DAMPING OF SOIL

The geotechnical investigation of the site was performed and the foundation recommendations were
provided for the mine site. A unit value for Modulus of Subgrade Reaction of 500 lb/in3 should be used for
the majority of the mill foundation in this case. In the grinding area, the bearing capacity is 6,000 psf.

The laboratory sample tests showed weathering grade decreases to slightly weathered. The worse case
is BH-18 at depth of 75 ft (close to the elevation of foundation bottom), completely weathered. The natural
density is 127 to 152 pcf (2,030 to 2,450 kg/m3 ).

The modulus of subgrade reaction is 500 lb/in3 = 72 kp/inch, per ft2, that is, the vertical stiffness, Kz = 136
kN / mm, per m2. This value is close to that output by the program for the shear wave velocity, Vs = 660
m/sec. Settlement of 6 mm assumed, P = 136 x 6 = 816 kN/ m2. Bear capacity, Pr = P/SF = 816 /2.5 =
326 kN/ m2 = 6,800 psf, here SF = 2.5. That value is closed to 6,000 psf provided from the geotechnical
report. The bearing capacity is not high and its classification is very weak rock.

The shear wave velocity, Vs = 660 m/sec, is used by the program. The vertical stiffness Kz = 7.16 x 10 7
kN/m at operating speed of 0.2 Hz. Foundation base area A = 24.7 x 20.12 = 497 m2. So, Kz = 144
kN/mm, per m2. This value is close to that by geotechnical report.

The stiffness and damping of soil output by the program are shown in Table 1 and 2, and the embedment
is considered. Where Kx, Kz and K are stiffness in horizontal, vertical and rocking direction, and Cx,, Cz and
C are damping constants in horizontal, vertical and rocking direction. The node is the base point of model
in the tables. The total damping distribute to six nodes by damper link element in FEM model by SAP
2000 program.

Table 1, Stiffness of foundation

Stiffness Total Per m2 Per node (0.836 m2)

Horizontal Kx (kN/m) 6.51 x 107 131,000 110,000
Horizontal Ky (kN/m) 6.51 x 107 131,000 110,000
Vertical Kz (kN/m) 7.16 x 107 144,000 120,000
Rocking Kx (kN-m/rad) 1.907 x 10 10 3.84 x 10 7 3.21 x 10 7
Rocking Ky (kN-m/rad) 2.34 x 10 10 4.71 x 10 7 3.94 x 10 7
Torsional Kz (kN-m/rad) 3.10 x 10 10 6.25 x 10 7 5.23 x 10 7

Table 2, Damping of foundation

Damping Constants Total Per Node

Horizontal Cx (kN/m/sec ) 1.75 x 10 6 292,000
Horizontal Cy (kN/m/sec ) 1.75 x 10 6 292,000
Vertical Cz (kN/m/sec ) 2.03 x 10 6 338,000
Rocking Cx (kN-m/rad/sec) 3.87 x 10 8 64,500,000
Rocking Cy (kN-m/rad/sec ) 4.65 x 10 8 77,500,000
Torsional Cz (kN-m/rad/sec) 5.74 x 10 8 95,700,000

5. VIBRATION OF GRANDING MILL

Two methods can be used to carry out the dynamic analysis for ball mill foundations, namely the free
vibration analysis and the forced vibration analysis.

CST11-4
5.1. Free Vibration Analysis

For a small ball mill, with a mill diameter less than 3.6 m and small dynamic loads, the method of free
vibration analysis (also call modal analysis) can be used.

The natural frequencies of foundation and piers can be calculated using the free vibration analysis to
avoid the resonance. The natural frequency should be less than 0.7 f n or larger than 1.4 f n, where f n is
the operation frequency of the machine.

5.2. Forced Vibration Analysis

For large mill, the method based on forced vibration analysis is recommended. The vibration amplitudes
should be calculated to meet the requirement of allowable vibration limit. Dynamics analysis is difficult for
the flexible mill foundations using standard analytical or numerical methods. Classical empirical methods
assume that the foundation acts as a rigid body. However, the structure of mill foundation and piers with
large dimension is flexible rather than a rigid body. Numerical methods such as the general finite element
method are also difficult to apply, as the direct simulation of radiation damping is not possible. Radiation
damping is the dominant energy dissipation mechanism in most dynamically loaded foundation systems.

The sub-structure method is used for dynamic analysis of grinding mill foundation, that is, the structure
and soil are considered separately. The structure part (mat foundation and piers) are modelled by FEM
model. The impedance of soil (stiffness and damping) are generated by the computer program, and then
input to the FEM model as the base boundary condition. So the reasonable values of radiation damping
can be used with the help of program. Different design options are compared to get the better solution.

The diameter of mill is 8.0 m with length of 13.1 m, operating at 12 rpm as shown in Fig. 2. The speed of
motor is 180 rpm. The height of mill shaft is 18.1 m above ground. The weight of mill is 2,280 tons and
charge (ores) is 1,172 tons. The mat foundation is 24.7 m x 20.1 m with thickness of 2.13 m, and the
height of piers is 15.8 m above ground. The concrete volume of bearing piers and mat foundation is 2,415
m3 for each unit. The ratio of foundation mass with mill and charge mass is 1.68. In general, the ratio of
mass of foundation with machine (including charge) should be 1.5 to 2.5, depending on the soil properties
and the foundation structure. Normally four to six mills are placed for a large plant, and the amount of
concrete is huge for the mill foundation construction. If the foundation design is over conservative, it may
lead to higher costs.

Figure 2 Grinding mill supported on foundation

CST11-5
The operating speed of the mill is relatively slow, usually in the range of 12 to 30 rpm (0.2 to 0.5 Hz), the
vibration amplitudes calculated are less than the allowable vibration limit in general. The motor driving the
mill operates at a different speed, synchronous low speed motor in the range of 90 to 200 rpm, and
induction motor in 1800 rpm. The motors are typically well balanced before they leave the factory. Any
residual imbalance normally does not give rise to significant excitation forces. It is desirable to tune the
supporting piers so that their lowest natural frequency is at least 33% above the operating speed of the
synchronous motor.

Unbalanced force from mill and charge rotation

[1] F = M e2

Where M is the eccentric mass of mill and charge, e is the eccentricity and  is the circular frequency.
M = As x L x = 0.5 x 3.14 x 4.0 2 x 13.1 x 2200 = 724,000 kg
As = 50% Ao is assumed. Here As is the charge section and Ao is the cross section of mill.
Mill diameter, D = 8.0 m and length, L = 13.1 m.
Considering cascading effect of charge in operation, the charge mass can be reduced. Assumed the
average charge unit weight  = 2,200 kg /m3
Eccentricity e = C = 0.425 r = 0.425 x 4.0 = 1.7 m
Mill rotational maximum speed 12 rpm
f = 12 / 60 = 0.2 Hz
= 2 x 3.14 x 0.2 = 1.243 1/ sec
Unbalanced force,
F = 724,000 x 1.7 x 1.243 2 = 1,902,000 N = 1,902 kN
As start up, horizontal force Fx = 1,902 kN.
In operation (steady state operation), cascading reduction considered, 30% off
The unbalanced force, F = 1,902 x 0.7 = 1,330 kN

Motor weight is 62,200 kg, two motors in the fixed bearing end, operating at speed 180 rpm.
Rotor weight, Wr = 30,400 kg, the unbalanced force from motor,
F = Wr x rpm / 6,000 = 30,400 x 180 /6,000 = 912 kg = 9.0 kN

6. FEM MODEL FOR DYNAMIC ANALYSIS

The concrete mill foundation is modelled using solid element with dimension of 0.91 x 0.91 x 0.91 m as
shown in Fig. 3. At the fixed and free bearing ends, the thickness of piers supporting mill is 1.83 m. For
the fixed end, additional walls of 2.74 m thick are added at each side. For the free end, an additional wall
of 2.74 m thick and 4.57 m wide is added. A horizontal slab with thickness of 1.83 m and width of 4.57 m
is added to connect the fixed end and free end which increases the stiffness of the support. To examine
the dynamic behaviour between the soil and concrete foundation, the mill machines are assumed as a
sole rigid body and modelled by rigid link element.

The stiffness of soil generated from the program is distributed at each base node of mat foundation as the
values of spring in six directions, as shown in Fig. 4. The damping constants generated by the program
input into the model by six link elements of damper, as shown in Fig. 5. The time history analysis is
carried out to get the vibration amplitudes. The sine function is used for harmonic loads, and time step is
taken as 1/20 of period T. Thus, time step is 0.25 second for mill and charge rotation, and 0.0167 second
for motor rotation.

From vendor, the allowable movement of pier is 0.16 mm under lateral load of 1,000 kN, and the
maximum allowable movement is 0.5 mm at pier top. The deflection of piers is calculated as shown in Fig.
6. Lateral load of 1,000 kN applied at top of fixed bearing end and free end respectively. The
displacement calculated is 0.29 mm at fixed end, and 0.84 mm at free end. That is, the movements
calculated are larger than those allowable. As start up, the horizontal centrifuge force, Fx = 1,942 kN. The
displacement at free end is 0.84 x 1.942 = 1.6 mm > 0.5 mm.

CST11-6
Figure 3 FEM model of grinding mill foundation using solid elements

Figure 4 Springs distributed into each node at mat bottom

CST11-7
Figure 5 Damping distributed into 6 Figure 6 Stiffness of piers
damper link elements at base

Figure 7 Dynamic response under unbalanced Figure 8 Dynamic response under

forces by mill and charge unbalanced forces by motor

The dynamic response is calculated under the unbalanced forces of mill and charge, as shown in Fig. 7.
The displacement amplitude at free bearing pier, Ax = 18 m, Ay = 590 m, Az = 84 m. The
displacement amplitude at fixed bearing pier, Ax = 16.5 m, Ay = 212 m, Az = 53 m. Allowable
vibration limit = 175 m at frequency of 0.2 Hz. The maximum amplitude of 590 m calculated is larger
than the allowable limit.

The dynamic response is calculated under the unbalanced forces of motor, as shown in Fig. 8. The
displacement amplitude at free bearing pier, Ax = 0.68 m, Ay = 3.5 m, Az = 0.05 m. The
displacement amplitude at fixed bearing pier, Ax = 0.65 m, Ay = 6.0 m, Az = 1.4 m. Allowable
vibration limit = 135 m at frequency of 3.0 Hz. The maximum amplitude of 6.0 m calculated is less than
the allowable limit.

The stiffness of soil-foundation system is low, so the larger deflections are caused. If a strong rock with
higher values of shear wave velocity, such as higher than Vs = 900 m/s, located in this site, the
deflections of pier could be reduced significantly. To increase the stiffness of pier supporting and reduce
the vibration, the short walls are added to the two sides of free bearing end, as shown in Fig. 9. The wall
dimension is 2.74 x 6.4 m, with height of 7.3 m.

CST11-8
The deflection calculated is 0.25 mm at fixed end, and 0.37 mm at free end under lateral load of 1,000
kN. The maximum deflection calculated is 0.7 mm as start up. It is a little larger than the allowable
deflection, and the stiffness of piers is acceptable.

The dynamic response under the unbalanced forces of mill and charge are reduced. The amplitude at
free bearing pier, Ax = 31 m, Ay = 254 m, Az = 78 m. The amplitude at fixed bearing pier, Ax = 28 m,
Ay = 186 m, Az = 50 m. The maximum amplitude of 254 m calculated is a little larger than the
allowable limit, and the dynamic response is acceptable.

The dynamic response under the unbalanced forces of motor is acceptable. The amplitude at free bearing
pier, Ax = 0.5 m, Ay = 1.2 m, Az = 0.05 m. The amplitude at fixed bearing pier, Ax = 0.5 m, Ay = 5.3
m, Az = 1.4 m.

Figure 9 FEM model of dynamic analysis for ball mill with walls strengthened

7. CONCLUSIONS

The soil-structure interaction is investigated based on the practical case of dynamic analysis for grinding
mill foundation. The dynamic response depends on both parts of soil and concrete foundation structure
(mat and piers), and the sub-structure method is efficient to solve the problem using the software. The
following results are concluded from this study.

The stiffness and damping of soil (rock) were generated by the computer program, and the values of
radiation damping were validated by many dynamic tests. Then the analysis of mat foundation and
supporting piers can be done using FEM models.

CST11-9
 In this case the deflections of pier are larger than the allowable movement under static loads on top, and
the vibrating amplitudes calculated are larger over the allowable vibrating limit, since the rock is
weakened in the site of project. If the values of shear wave velocity of rock increased, the vibration of
foundation would be reduced.

The stiffness of piers are increased with the short walls added to the two sides of free bearing end in the
final design, so that the deflections of pier and vibrating amplitudes become acceptable.

8. REFERENCES
Banerjee, P.K. and Sen, R. 1987. Dynamic Behavior of Axially and Laterally Loaded Piles and Pile
Groups. Chapter 3 in Dynamic Behavior of Foundations and Buried Structures, Elsevier App. Sc.,
London, 95-133.
Baranov, V.A., 1967. On the calculation of excited vibrations of an embedded foundation. Voprosy
Dynamiki Prochnocti, No.14, 195-209, (in Russian).
Barkan, D.D. 1962. Dynamics of bases and foundations. McGraw-Hill Book Co. New York.
Dobry, R. and Gazetas, G. 1988. Simple Method for Dynamic Stiffness and Damping of Floating Pile
Groups. Geotechnique, Vol.38, No.4, 557- 574.
DynaN 2.0 for Windows, 2003. Dynamic Analysis of Shallow and Deep Foundations, Ensoft.
www.ensoftinc.com.
Gazetas, G. and Makris, N. 1991. Dynamic Pile-Soil-Pile Interaction. I: Analysis of Axial Vibration. J.
Earthq. Eng. and Struct. Dyn. Vol. 20, No.2.
Han, Y.C. 1989. Coupled Vibration of Embedded Foundation, Journal of Geotechnical Engineering,
ASCE, 115(9), 1227-1238.
Han, Y.C. 2008. Study of Vibrating Foundation Considering Soil-Pile-Structure Interaction for Practical
Applications. J. of Earthquake Engineering and Engineering Vibration, Vol.7, No.3, 321-327.
Han, Y.C. and Sabin, G. 1995. Impedances for Radially Inhomogeneous Soil Media with a Non- reflective
Boundary. J. of Engineering Mechanics, ASCE, 121(9), 939-947.
Han, Y.C. and Novak, M. 1988. Dynamic Behavior of Single Piles under Strong Harmonic Excitation.
Canadian Geotechnical Journal, 25(3), 523-534.
Luco, J.E. 1982. Linear Soil – Structure Interaction: A Review. Applied Mech. Div., Vol.53, ASME, 41-57.

SAP 2000 v12.0.1Plus, 2009. Structure analysis program, Computers and Structures, Inc.
www.csiberkeley.com.

Wolf, J.P. 1988. Soil – structure interaction analysis in time domain. Englewood Cliffs, NJ: Printice - Hall,
446p.

CST11-10

View publication stats