1

ICPS 1999 22. March 1999

Cement grinding ToBi
Lecture 6.1

Operating the efficient way.

Table of contents

1. Introduction
2. Feed and product
2.1 Product fineness
2.2 Grindability
2.3 Feed properties
2.4 Feed granulometry.
3. Ball mill
3.1 Ball charge
3.1.1 Compartment 1
3.1.2 Compartment 2
3.2 Diaphragm
3.3 Mill linings.
4. Closed circuit grinding
4.1 Separator efficiency
4.2 Separator configurations
 2
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

Operating the efficient way.

1Introduction

From a total of approximately 100 kWh/t, the grinding of cement in a modern cement
plant accounts for roughly 40 %.
Being responsible for almost half of the electricity bill would be a strong incentive to
attempt to make the operation as efficient as possible, as even minor improvements
will correspond to measurable savings in the budget.

Many attempts have been made over the years to make the cement grinding process
more efficient by introducing new grinding machines. The ball mill , however is still
the predominant grinding machine, and despite the simple nature of the process a
number of specific factors have a large influence on the performance and the
efficiency of the process.

Cement grinding, which used to be a simple matter of mixing clinker and gypsum in
the right proportions and perform the grinding in an open circuit ball mill, has become
a rather complex matter, involving a vast number of cement qualities and a variety of
equipment in different configurations. The complexity makes it difficult to operate with
the highest degree of efficiency under all circumstances.

The present paper will address a number of the areas involved in the cement
grinding operation, and look at details, where special attention can make a
difference, when it comes to paying the bills.

2Feed and product

2.1Product fineness

The cement produced normally has to live up to a number of specifications regarding

the composition as well as properties relating to the strength. Apart from setting limits
on the amounts of the individual components in the cement, the composition of the
cement has in itself an influence on the strength level of the product. Apart from this,
the decisive factor relating to the strength is the fineness.

Determination of the cement strength normally is a task that takes from 1 - 28 days,
and for final strength up to a year. This time from sampling to test result is of course
impractical when looking at the daily production, and for the operation a relationship
between fineness and strength level is established.
Once the proper relations have been established, the fineness determination will
serve as a relevant measure for the product quality.
 3
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

The fineness of cement can be stated in a number of different ways, using residue
determination, Blaine values or full particle size analysis. Normally the best way to
keep track of the product quality is to use more than one of the mentioned methods,
as none of them give the full picture on its own.

Individual residue values are used for a quick determination on regular sampling,
made in order to see that a target is kept. The value can be used in the operation as
a guideline for setting the production level or the separator speed. For practical
reasons only sieve residues on the coarse end of the particle size distribution curve
can be determined. The residue values determined in this way are basically related
to the late strength properties.

Blaine values are a measure of the specific surface of the product. The value is
written as m²/kg, according to the SI standard. The specific surface area of a
powdery product is largely found in the very fine end of the particle spectrum,
typically made up from particles below 5 micron in size. The Blaine value tells very
little about the coarse end of the spectrum. For a given Blaine value the residue on a
specific sieve can vary a lot depending on the composition of the cement. Also the
equipment used in the grinding process has a significant influence on the relationship
between residue and Blaine value.
As the Blaine determination is dealing with the finer particles in the sample, the value
is primarily related to the early strength level.

The full particle size analysis is a modern way, involving laser equipment, by which
the product can be described in a rather comprehensive manner, in a short time. The
result will in principle state the residue values on all particle sizes ranging from 1 to
100 micron.
Experience has shown that different makes of laser analysis equipment can give
different results on the same sample, and for this reason a comparison between
products should be made using the same analysis equipment.

Residue values and particle size distribution curves are normally plotted in a Rosen-
Rammler diagram. The diagram is drawn with logarithmic scaling of the axes.

Regular and frequent sampling of the product is required to be able to operate close
to the target values. Only in this way the necessary adjustments can be made to
keep the operation on target and avoid producing cement with excessive properties.

The fineness determinations are also used to detect changes in the operation
caused by gradual wear or faulty equipment.

2.2Grindability

Despite the many different additive components used in the manufacture of cement,
one point of interest is still the grindability of the clinker.
 4
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

Even though the clinker in all cases is produced by burning in a kiln, the quality in
respect of grindability can vary greatly due to differences in the burning- or the
cooling process. Clinker with a high content of small pores will be easy to grind and
clinker with a very dense structure will be more difficult.

Differences in the chemistry will also play a part as this will have an influence on the
formation of the clinker minerals. Clinker with a low LSF will normally have a high
content of C2S and is generally considered hard to grind.

Other factors contributing to differences in grindability are ageing of the clinker and
the granulometry.

Variations in grindability are experienced in everyday operation as long term as well

as short term variations.
Over a period of time, days or weeks, a change in mill output can be experienced
even though all operational parameters for the mill have been maintained. In many
cases things will revert to normal by themselves, but if they do not, an investigation
should take place to determine whether the change is caused by conditions in the
grinding system or it is a reflection of a grindability change. In the latter case the
difference may require changed operational parameters or even changes relating to
the ball charges in the mill, to be able to obtain the maximum output from the
installation.

Short term variations are experienced during the operation as minor variations in
output. These changes are taken care of by the mill control system, which is an
essential part of any modern and well functioning installation. Operating without an
automatic control system normally results in a lower average output because the
manual operation will require wider safety margins to avoid problems in the process.

The grindability is expressed as kWh pr. ton product produced, to a certain fineness.
For calculations of the grindability only the power consumption for the grinding
machine is regarded. The grindability can be expressed both in relation to residues
and to Blaine fineness.

When looking at residues a measure for the grindability can be expressed as the
grinding constant, W(45), meaning the grinding constant relating to a 45 micron sieve
residue. The calculation is valid for ball mills, which by experience have been found
to have a logarithmic grinding sequence.
In the calculations any sieve size can be applied.

W is calculated from the expression:

R0
E  W  Log
Rf

where: E = the specific power consumption

 5
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

W = the grinding constant

R0 = residue of the feed on the relevant sieve
Rf = residue of the product on the relevant sieve

The grinding constant is a measure for the grindability of the materials and, when
calculated, can be used for the detection of changes irrespective of the fineness level
of the product.

The calculations can be used to assess the required specific energy consumption to
a different fineness, for a product with the same composition, when the energy
consumption is known at any fineness.

The relation is as follows:

Log ( R0 )  Log ( R2 )
E 2  E1 
Log ( R0 )  Log ( R1 )

where: E1 = the known specific energy consumption

E2 = the specific energy consumption for the new fineness
R0 = residue of the feed materials
R1 = residue of the known product
R2 = residue of the desired product

In case the desired fineness criteria is a Blaine value similar calculations can be
made.
The expression below describes the ball mill grinding process:

E  W (S)  S n

where: E = the specific energy consumption to the Blaine fineness

W(S) = the grinding constant relating to Blaine
S = the Blaine fineness expressed as km²/t
n = the Blaine development constant, which assumes the value
1.6, when only the ball mill is regarded.

If the mill is operated in closed circuit with a separator the

value of n will be 1.3 - 1.4, depending on the separator type.

The calculation of the specific energy consumption for a different fineness can be
carried out according to the following expression:

n
S 
E 2  E1   2 
 S1 

where: E1 = the known specific energy consumption

 6
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

E2 = the specific energy consumption of the new product

S1 = the Blaine value of the known product
S2 = the Blaine value of the new product
n = the Blaine development constant

By making the calculations the task of determining the output rate, when changing
the fineness level of the product, is made easier, and the time operating with low
efficiency or away from a given target is limited.
By having knowledge about the grindability of the materials the job of identifying
irregularities in the operation is made a simpler task.

2.3Feed properties

Apart from the grindability of the feed materials, other properties may have a decisive
influence on the operation.

Clinker being used in a cement factory can originate from a number of places. They
can come more or less directly from the kiln, and will have a temperature that
depends on the way the kiln and the cooler is operated. Quite often the clinker is a
mixture of hot and cold clinker in varying proportions.

Variations in the clinker temperature will be experienced as fluctuations in the cement

mill parameters. The temperature of the material inside the mill affects the flowability
of the material and hence the grinding process. From experience it is known that the
most efficient grinding takes place when the temperature is above a minimum but on
the other hand below a maximum level.

Exceeding the maximum level will enhance the tendency of the material forming
agglomerates, which will form a coating of material on the grinding media and the
lining plates. When this takes place the grinding efficiency will be reduced.

For a pure OPC cement the conditions are relatively simple as the temperature
profile of the mill depends only on the temperature of the feed and the heat
generated by the grinding process. In case the heat is insufficient to reach the
recommended level, it is to be expected that the grinding process is less efficient and
that the mill operates with a reduced output, when compared to the optimum
conditions.

For cements containing additional materials the matter can be rather different, as
many of the additives contain moisture, which will have to be evaporated during the
grinding process. When producing these types of cement it is essential that the feed
materials in combination with the grinding heat provide sufficient heat to maintain the
correct temperature in the mill.
 7
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

If the heat content is insufficient to evaporate the moisture in the first part of the mill,
the risk is that moist material will block the passage in the diaphragm, and reduce the
output to an unacceptable level.

Cement grinding installations, which have to operate under conditions where it is

obvious that the heat from the process itself is insufficient to provide the required
conditions, should be provided with an additional heat source in the form of a
separate auxiliary furnace or in case surplus heat is available from the kiln system,
heat could be supplied from there.

Normal temperature variations in a system with sufficient heat should be dealt with by
automatic temperature control equipment, where water is injected into the mill at a
varying rate according to the amount of surplus heat.

The normal operating range for cement mills is a temperature at the diaphragm
between the first and the second compartment around 100 deg. C, and the
temperature at the outlet of the mill in the range 110 - 120 deg. C.
The temperature in the mill outlet ensures that all free water in the material has been
evaporated before the cement is sent to the silo.

The temperature level in combination with the residence time in the mill ensures that
the gypsum in the feed mix reaches the required level of de-watering.

2.4Feed granulometry.

Most ball mills have been filled with a charge, which is intended for grinding of feed
material sizes in a specific range. When looking at the size of the feed materials,
obviously the point of focus is the ball charge in the first compartment.

If the feed material consist of sizes in the correct range, no problems relating to the
feed sizes are experienced and the grinding follows the normal rules for ball mill
grinding, and only the power available is limiting the output from the mill.

If on the other hand one or more of the feed materials contain lumps in excess of the
normal size range severe operational problems can be experienced. Whether the ball
charge is capable of breaking the lumps will depend on the hardness as well as the
size of the material.

Only a small fraction of oversize material can cause major upsets, as the pebbles will
accumulate in front of the diaphragm and block the passage. The un-ground particles
will eventually be worn down but they will continuously be replaced by new ones, and
in the process they will reach a stage where they will fit into the slots in the
diaphragm and get stuck.
When the situation has occurred the first compartment will operate with a high level
of material, causing a reduction in the grinding efficiency, while at the same time the
fine grinding compartment will operate with a reduced level, making the mill difficult to
 8
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

control in terms of constant fineness of the product. The overall result will be a drastic
reduction in mill output

One way of dealing with the problem is to increase the size of the media. This can be
done only up to a certain point, where the media, because of the impact and attrition
forces will start to become a hazard.

Once the problem has been identified the oversize particles should be eliminated
from the process either by screening or by additional crushing.

3Ball mill

3.1Ball charge

3.1.1Compartment 1

Any ball mill operating with a compartment 1 charge, which is not optimised for the
actual grinding conditions defined by the size, amount and quality of the feed
materials, will perform with a reduced efficiency.
The main objective of the first compartment is to prepare the relatively coarse feed
materials to a level, where the smaller media in the subsequent compartment will
have no problems continuing the size reduction.

The combination of size and hardness of the feed materials will decide which
maximum ball size is required in the compartment. The length of the compartment
will be depending on the level of fineness required at the outlet of the compartment.

In many cases a grinding installation will be used for more than one product. Often
products with a significant difference in fineness level and output rate, and as the first
compartment ball charge is almost entirely utilised on crushing the feed materials, the
difference in the amount of fresh feed will result in less than optimum conditions for
at least some of the products.

A coarse grinding compartment with an insufficient ball charge will cause a reduction
in the mill output, not only caused by the resulting coarseness of the material at the
end of the compartment, but in most cases the grinding in the following compartment
will be affected, as the media charge will not be capable of dealing with the size of
the particles.

A coarse grinding compartment with a surplus of power, and provided that the
grinding media are of the correct size, will have no problems feeding the following
compartment with adequately ground material, and the grinding in this compartment
takes place in a controlled and efficient manner. The surplus energy relating to the
excessive ball charge is wasted, and should be as small as possible.
 9
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

A superior compartment will normally operate with a low level of material, which
tends to increase the wear rate of balls and lining plates.

Where more than one product is produced in a ball mill, the length of the first
compartment is a compromise.
The important thing is, that the compartment is long enough to cater for the power
requirement needed for the product with the highest feed rate, which is often the
product with the lowest fineness level.

Based on experience it is found that in most cases the maximum ball size required in
a first compartment of a cement mill is 90 mm. Normal practice is to charge the
compartment with a grading of balls ranging from the biggest size and downwards in
certain proportions. By adhering to the figures given below, for a new charge, the
replacement for wear can be carried out by addition of only the largest ball size. Over
time an equilibrium will be established and the composition will remain constant.

Initial compartment 1 charge.

Ball size Percentage of charge

90 mm 23
80 mm 32
70 mm 21
60 mm 24

The above charge is suited for a ball mill operating in closed as well as open circuit,
with a normal feed of clinker and gypsum.

3.1.2Compartment 2

The charge in the second compartment, normally accounting for 65 - 75 % of the

total power consumption for the mill, is reducing the rather coarse material from the
first compartment to the desired fineness level at the outlet of the mill.
The size of the media should be suited for the initial grinding at the inlet of the
compartment, requiring larger media than at the outlet of the compartment, where the
material is finer. A general rule is that the finer the material has to be ground, - the
smaller the media for efficient operation.

The required fineness level at the outlet of a mill depends on the product fineness
but it also largely depends on the operational mode of the mill. The fineness level at
the outlet of an open circuit mill is at the level of the product fineness. In a closed
circuit mill system, operating with the same product, the fineness level will be lower
because of the intermediate separation process.

The fineness of the material entering the compartment can be ground down
effectively, in the first compartment, to a level requiring only moderate ball sizes for
 10
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

the further grinding. All practical experience has shown that the most efficient
grinding is obtained, when the media sizes are as small as possible, but at the same
time being of sufficient size to exert the required forces.

Fine grinding in open circuit to a high fineness requires very small media to be
effective. The use of very small media is in general not practical for a number of
reasons, one being that media in the required size range 5 - 8 mm have a short life
span as they will be worn down quickly, and subsequently find their way out of the
compartment.

Very small sizes of grinding media forms a rather dense charge, through which the
transport of material will be slowed down, limiting the amount of material that will
pass through the mill cross section. In connection with closed circuit operation this
could reduce the benefit of using the separator, as the circulation factor will be
limited.

A normally operating ball mill will have a natural fineness level at the outlet of the fine
grinding compartment in the range 180 - 200 m²/kg Blaine, when operating in closed
circuit and producing an OPC type cement. This fineness level can be maintained for
a different fineness of the product, by varying the circulation factor. For this reason a
closed circuit mill can operate efficiently with the same charge composition in the fine
grinding compartment, at different levels of fineness for the product.

Below is a suggestion for a practical initial charge for fine grinding in closed circuit.
The charge can be maintained by using entirely 25 mm balls as replacement.

Initial fine grinding charge, closed circuit.

Ball size Percent of charge

25 mm 40 %
20 mm 40 %
15 mm 20 %

For an open circuit mill the recommended charge is finer because of the increased
demand for fineness level. The ball sizes for wear compensation are 25 and 20 mm.

Initial fine grinding charge, open circuit.

Ball size Percent of charge

25 mm 20 %
20 mm 40 %
15 mm 40 %
 11
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

Any ball mill charge, even one that has been maintained fairly regularly, should at
intervals be sorted out and the smallest sizes removed. A worn charge containing a
large amount of under sized media is inefficient and a direct cause for reduced
output. The intervals at which the charge should be sorted depends on production
levels and media quality. It is recommended that the charge is monitored, by taking
representative samples of the charge, for determination of the sizes, at least a couple
of times a year. A decision to sort the charge can be based on such sampling and
the job can be planned ahead for a suitable period.

As the power consumption of the mill depends on the amount of charge, it will be
possible to decide when to add media as wear replacement, based on records of the
mill motor power consumption. The relationship between a loss of power and the
amount of media required to bring the power back up, should be established. In that
way the correct amounts of media can be made ready and available at the right time.

3.2Diaphragm

The main purpose of a diaphragm inside a mill is to separate the ball charges in the
two adjoining compartments.
A secondary function is a screening effect, caused by the rotation of the mill in
combination with the slots in the diaphragm plates.

Most modern diaphragms are provided with facilities for controlling the flow of
material from one compartment into the next. By properly adjusting the flow device, a
correct level of material can be assured in the compartment in front of the diaphragm.
A correct level of material in a compartment is essential for the grinding efficiency
and will ensure the lowest wear rate on the lining plates.

The screening effect is depending on the sizes of the slots in the diaphragm plates
and will to some extent reduce the possibility of letting oversize particles into the
subsequent grinding compartment. The importance of the screening will increase if
the coarse grinding compartment is on the short side and the finer grinding
compartment is charged with only small media.
In all cases, if the coarse grinding compartment has the correct length and is
charged with the correct media sizes, the screening effect is of less importance.

The size of the slots in a diaphragm mainly depends on the amount of material which
has to pass through. The diaphragm plates must have a sufficient area for passage
to cater for the required flow of material. For open circuit mills problems are seldom
encountered, but for a closed circuit mill the flow of material in the same cross
section is much greater and the corresponding free area must be available.

During operation a diaphragm is subjected to impact from the grinding media and
over time the slots in the plates may start to close up leaving a reduced passage
 12
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

area. It is often experienced that bits of grinding media and hard un-ground particles
get stuck in the slots. This as well will cause a reduction in the passage area.

One of the common reasons for blocked diaphragm plates is the widespread practice
of using the mill as a deposit for the material swept from the floors in the building and
as a recipient for the material from nib traps. The material from these locations will
contain a certain quantity of media bits and nibs of exactly the sizes which get stuck
in the slots.

The diaphragm should be regularly inspected and the slots cleared of foreign bodies,
in case the slots are closing up they should be cut open again, as otherwise the
output will suffer as a consequence of the restrictions of the material flow.

3.3Mill linings.

The main purpose of the mill lining is to protect the mill body from impact by the
grinding media and prevent it from having contact with the material that has to be
ground.

Obviously damaged lining plates should be replaced as soon as the damage is

observed as otherwise the mill shell will suffer from the exposure to the grinding
process.

Mill lining plates have different thickness and shapes depending on where they are
located. The coarse grinding compartment is fitted with plates of substantial
thickness enabling them to withstand the impact from the large balls in the
compartment. The plates are shaped in a way that forms steps, or a wavy surface,
on the circumference inside the grinding compartment. The steps are creating a
lifting effect for the charge. The shape of the plates have a decisive influence on the
power consumption for the compartment as the lifting effect is expressed in the
torque factor as used in the expression below:

N  k  F n Da

where: N = power consumption

k = constant for unit conversion
F = amount of charge
n = mill revolutions pr. min
m = torque factor
D = internal diameter of compartment
a = distance from mill centre to centre of gravity for the charge

The expression is valid for any grinding compartment in a mill, provided the correct
figures are inserted for each compartment.
 13
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

For a mill with more than one compartment, each compartment is calculated
separately and the power consumption for the mill is the sum of the individually
calculated values. The calculated value for a mill corresponds to the net shaft power,
excluding all losses in transmission and motor.

The fine grinding compartment in a mill should contain media of the smallest possible
and practical size. Apart from giving the best grinding efficiency they will also permit
the use of the thinnest lining plates, resulting in the largest internal diameter. This in
turn will give the mill the best power consumption corresponding to the largest output
rate.

In the following example the difference in power consumption for a given mill is
shown for two situations, using a thin lining and a lining suitable for a mixture of
larger balls.
 14
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

Mill speed Rpm 14.6 14.6

Diameter(nom) m 5.0 5.0
Lining thickness mm 42 80
Length(eff.) m 10 10
Filling degree % 28 28
Media bulk weight t/m³ 4.5 4.5
Amount of media t 239 232
Torque factor 0.7 0.7
Power consumption kW 3841 3665

The calculation demonstrates a significant difference in power potential for the two
situations, giving a better utilisation of the mill size and the corresponding higher
output rate.

The difference illustrated is often seen when comparing a mil with a thin lining and
only small media in the second compartment, with one that has a classifying lining
and the ball charge to match. A charge consisting of different proportions of balls
ranging from 60 mm and down to about 20 mm.

The lining plates should be designed to give a torque factor of the highest value,
which is obtained for a corrugated surface.

The normal range of values for torque factors are:

Coarse grinding compartment: 0.70 - 0.73

Fine grinding compartment: 0.65 - 0.73

The actual value of the torque factor will depend on a number of conditions apart
from the geometrical, the important one being the amount of material in the
compartment during operation. In case the mill is operating with an excessive level of
material the torque factor will decrease, resulting in a low power consumption and a
low output rate.

4Closed circuit grinding

The use of separators in closed circuit operation is intended to make the production
process more efficient by reducing the necessary fineness level in the mill.

By using the calculation method for Blaine fineness, it is observed that the cost in
terms of specific energy consumption is increasing for an increased level of fineness.

Below the graph is showing the specific energy consumption E as a function of the
Blaine surface S, using the expression E = W x S n.
 15
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

Specific energy consumption / Blaine

70.0
60.0
consumption, kWh/t
Specific energy

50.0
40.0
30.0
20.0
10.0
0.0
0 100 200 300 400 500
Blaine value

The graph is demonstrating that the production of 100 Blaine in the interval 200 - 300
Blaine requires 16.3 kWh/t, where 100 Blaine in the interval 300 - 400 Blaine requires
20.0 kWh/t.

A mill operating in open circuit with a production of 350 m²/kg Blaine, will have to
deliver the final product at the end of the mill. The flow of the material through the mill
will be identical to the output rate.

Assuming the following conditions:

Output rate: 100 t/h

Fineness, Blaine: 350 m²/kg
Grinding constant, W: 230 kWh/t
Blaine constant, n: 1.6

the specific energy consumption is calculated according to the expression:

E  230  0.351.6  42.9 kWh/t

If the same product were to be produced in closed circuit, with an efficient separator
the following assumptions would apply:

Output rate: 100 t/h

Fineness, Blaine: 350 m²/kg
Grinding constant, W: 230 kWh/t
Circulation factor: 2.5
Return material fineness: 85 m²/kg
 16
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

Based on the values for the circulation factor and the fineness levels the following
calculation will reveal the fineness level at the outlet of the mill:

C  Sm  1  Sf   C  1  Sg

where: C = circulation factor

Sm = material fineness at mill outlet
Sf = product fineness
Sg = fineness of return material

With the values inserted:

2.5  Sm  1  350   2.5  1  85

solving the expression Sm gets the value 191 m²/kg, which is the level of fineness at
the mill outlet.

By introducing the closed circuit operation, the mill feed now consists of the fresh
feed materials plus the material from the separator return. The fineness of the mill
feed, St can be calculated as:

  C  1  Sg  So 
St    ,
 C 

where: So is the fineness of the fresh feed, having the value 0 m²/kg.

Resulting in the value 51 m²/kg for the mill feed fineness St.

The mill, in closed circuit is now operating with a fineness at the inlet of St and at the
outlet of Sm. The flow of material through the mill is 2.5 times the output rate.

The required specific energy consumption for the mill can then be calculated as:

E  W  C   Sm1.6  St 1.6   230  2.5   0.1911.6  0.0511.6   35.8 kWh/t

The calculations demonstrate that the closed circuit operation can reduce the
required specific energy consumption and that the reason for the reduction is the
lower level of fineness inside the mill.

Using the same power consumption for the closed circuit mill as for the open circuit
mill, the output would increase from the previous 100 t/h to:

 100  42.9 
Output     120 t/h
 35.8 
 17
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

The reduction in specific energy consumption and the corresponding increase in

output will only be achieved when the separator is operating efficiently. An in-efficient
separator will send too much of the finer material back to the mill, and by doing so
the level of fineness in the mill will be increased.

4.1Separator efficiency

Modern days separators are designed fairly strictly to operate with certain quantities
of material and rather fixed flows of air. In most cases, exceeding the limits will
reduce the efficiency of the separation, and the output will suffer.

The areas of the greatest importance are, that the separator has sufficient air, and
that the amount of material in the separator is in proportion to the airflow. A modern
separator will accept a loading of material between the values 2 - 2.5 kg material pr.
m³ of air in the separator. Another limit to observe is the material loading out of the
separator, where values in the range 750 - 850 g product pr. m³ of air are normally
recognised as being acceptable.

The efficiency of separation can be evaluated in a number of ways, some based on

calculations relating directly to the saving in power by using the separator, others
dealing with the ratios between the fineness levels of the different streams of material
going into and out from the separator.

The method which describes the separator function in the most complete way is the
calculation of the Tromp-curve. This method is based on complete size analysis on
all streams of material around the separator, and requires a laser instrument for the
determinations.

Below is an example of a Tromp-curve from an efficient separator:

90 Tromp
Percent to coarse

80

70
60

50
40

30
Delta = 6.06
20
Cut-size = 31.6
10
0 Kappa = 0.56
1 10 100
Sieve size, m icron
 18
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

The curve describes the percentage of the feed material, of any grain size, which is
found in the return material.

The calculation of the Tromp values are done by using the expression below:

  C  1  Rg 
T   100 %
 C  Rm 

where: C = circulation factor

DRg = percentage of return material in the interval representing the

particle size.

DRm= percentage of separator feed in the interval representing the

particle size.

The Tromp-curve has three characteristic values:

Delta = The lowest point on the curve, indicating the amount of by-pass. The
point for a well functioning separator is 10 - 15 %.
The delta value is particularly sensitive to the material loading in the
separator.

Cut-size = The sieve size where 50 % of the separator feed ends up in the returns
and 50 % goes to the product. The value depends on the speed of the
rotor.

Kappa = or sharpness, is the slope of the curve in the interval T = 25 - 75 %.

A good separator has values between 54 - 60
The sharpness depends on how well the material is dispersed in the air,
and is as well dependant on the geometric conditions in the separator.

The separator has a big influence on the overall performance of a closed circuit
grinding installation, and to make sure that the maximum benefit is obtained it is
recommended that the performance is monitored regularly. Differences in the
performance, when evaluated correctly, will often give good leads to where the
problem is located, and in many cases a mechanical fault is detected through the
evaluation.
Separators are exposed to a rather aggressive mixture of material and air, which will
cause gradual wear of some of the parts. The wear will eventually affect the
performance of the separator, if it is not detected in time. This type of gradual change
will show itself in the decline of the separator performance, and a problem can be
isolated to the separator, instead of calling for a complete investigation of the
grinding installation.
 19
ICPS 1999 22. March 1999
Cement grinding ToBi
Lecture 6.1

4.2Separator configurations

Separators can be fitted into a grinding circuit in a number of different ways. Quite
often the circuit was designed for different conditions than it is now operating under,
resulting in difficulties controlling and operating the installation.

In case the configuration is no longer suited for the actual conditions, considerations
should be made regarding possible modifications and changes in the lay-out.

In the following the most common configurations are described and their differences
are discussed.

Fig. 1

CM_1