Jan Svoboda. Magnetic Techniques for the Treatment of Materials
Подождите немного. Документ загружается.
360 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS
Table 5.20: Comparison of some operating parameters for various applications
of a cyclic magnetic separator.
Application Mag. fraction
in the feed
[kg/m
3
]
Flow velocity
[m/s]
Feed time
[hours]
Filtration of
thermal power
plant water
0.0001 0.24 1000
Steel mill waste
water
0.1 0.12 1.5
Kaolin purifica-
tion
1to3 ? 0.01 0.1
Hematite ore 250 0.16 0.0003
Table 5.21: The comparison of cyclic and continuous high-gradient magnetic
separators.
Parameter Cyclic Continuous
Duty cycle Lowtomedium 100%
Throughput Lowtomedium High
Specific processing cost High Moderate
Versatility Low High
Mechanical complexity Low High
cycle of the separator would be very short. Table 5.20 illustrates the range of
feed times for a cyclic high-gradient magnetic separator in dierent applications.
It can be seen that the main application of cyclic separators is filtration
of cooling and waste waters. For the beneficiation of weakly magnetic ores,
the utilization factor of the separator is very low. Although the situation can
be improved somewhat by the application of reciprocating technology, the use
of continuous magnetic separators is inevitable. A continuous separator which
operates with a 100% duty cycle, can treat a wide range of minerals with varying
content of the magnetic particles. The comparison of cyclic and continuous high-
gradient magnetic separators as applied in mineral processing is given in Table
5.21.
5.3.3 Magnetic field in high-intensity magnetic separators
The magnitude of a magnetic field generated by the magnet of a magnetic
separator is one of the most important parameters. It signifies the potential of
a separator for the treatment of various classes of magnetizable materials.
Various manufacturers of magnetic separators use dierent approaches upon
how to ascertain and specify the magnetic field strength, and the reported values
are often contentious and occasionally misleading. The magnetic field in mag-
5.3. WET MAGNETIC SEPARATION 361
y
x
r
a
B
0
ȥ
Figure 5.26: The geometrical configuration for calculation of magnetic induction
in the vicinity of a matrix element of radius d.
netic separators is generally very non-uniform. Particularly in high-intensity
separators, such as rare-earth rolls or WHIMS or HGMS, the value of the mag-
netic field varies considerably over a scale of several hundred micrometers, as
has been discussed in Section 1.6. Commonly available measuring instruments
are able to measure only the mean magnetic field over a relatively large cross-
sectional area, defined by the size, for example, of the Hall sensing element,
and the reported values of the magnetic induction generated, for instance, by
a rare-earth roll separators, may range from 1.0 T to 2.0 T, depending upon
which measuring (or estimating) technique was employed. In matrix separators,
the field is measured either as a background magnetic field or in the vicinity of
a matrix element. The importance of the latter value of the magnetic field is
insignificant, as it will be very similar for low-field and high-field separators.
The total external magnetic induction in the vicinity of a matrix element is
given by the superposition of the background magnetic induction E
0
and the
induction due to the matrix element E
p
:
E =
E
0
+
E
p
(5.5)
The induction E
p
causes distortion of the essentially uniform background
induction E
0
in the vicinity of each matrix element. This matrix material is
magnetized to polarization M, which is proportional to the background induc-
tion E
0
until the material is saturated. Above its saturation value M
v
,the
polarization is constant and does not increase with further increase in E
0
.
Under the assumption that the matrix material is magnetically saturated,
the total magnetic induction at a point u outside the matrix can thus be calcu-
362 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS
Figure 5.27: Pattern of the magnetic field outside a horizontally magnetized
steel sphere.
lated from the expression:
E =
·
E
2
0
+
1
4
M
2
v
³
d
u
´
4
+ E
0
M
v
³
d
u
´
2
cos 2#
¸
1@2
(5.6)
where d is the radius of the matrix element and the angle # is defined in Fig.
5.26.
The pattern of the magnetic induction and the radial component of the mag-
netic induction, for # = 0, in the vicinity of a magnetized steel wire, obtained
using the electromagnetic VF modelling software, are shown in Figs. 5.27., 5.28
and 5.29, for two values of the saturation polarization. It can be seen that, in
order to generate, for example, the magnetic induction of 2 T on the surface of
the matrix, a background magnetic induction of 1 T is needed. If a steel with
rather poor magnetic properties (e.g. M
v
= 1 T) is used for the matrix, the
background induction of 1.5 T is required to create 2 T on the matrix surface.
It also transpires from eq. (5.6) and Figs. 5.28 and 5.29, that with increas-
ing distance from the matrix, the role of the saturation magnetization of the
matrix material diminishes and, for a distance u =3d, the dierence between
the induction generated by M
v
= 2 T or 1 T materials is negligible.
Similarly, it is possible to plot the gradient of the magnetic induction outside
a magnetized steel wire. Figures 5.30 and 5.31 depict such patterns of the field
gradient. These graphs reveal several interesting facts. It can be seen that the
gradient is essentially independent of the values of the background magnetic
induction E
0
, with the exception of a case where E
0
=0.5TandM
v
=2T.At
the same time, the gradient is independent of the magnetic polarization of the
wire material for distances from the wire surface greater than 2d=
5.3. WET MAGNETIC SEPARATION 363
Figure 5.28: Distribution of the radial component of the magnetic induction in
the vicinity of a magnetized wire (d = 2.5 mm), as a function of
the normalized distance u
d
from the wire. M
v
=1T.
Figure 5.29: Distribution of the radial component of the magnetic induction in
the vicinity of a magnetized wire (d = 2.5 mm), as a function of
the normalized distance from the wire. M
v
=2T.
364 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS
Figure 5.30: Distribution of the gradient of the radial component of the magnetic
induction in the vicinity of a magnetized wire (d =2.5mm). M
v
=
1T.
Figure 5.31: Distribution of the gradient of the radial component of the magnetic
induction in the vicinity of a magnetized wire (d =2.5mm). M
v
=
2T.
5.3. WET MAGNETIC SEPARATION 365
Figure 5.32: Distribution of the radial component of the force index in the vicin-
ity of a magnetized wire (d = 2.5 mm), as a function of the nor-
malized distance from the wire. M
v
=2T.
The total magnetic force outside a magnetized steel wire, expressed as the
force index, is illustrated in Fig. 5.32. It can be seen that the force index on
the surface of the wire decreases with the decreasing background magnetic field.
However, with increasing distance from the matrix surface, the influence of the
background magnetic field diminishes and at distances greater than 3d, the force
is independent of the external background magnetic field.
A force holding already captured particles on the matrix surface increases,
therefore, with increasing background magnetic field. This force decreases for
particles captured onto more distant layers of the particle build-up. However,
the traction force that is supposed to attract particles from a distance onto
the matrix surface is, for practical purposes, independent of the background
magnetic field. These observations contradict simple models of particle capture
based on single collector-single particle approach, discussed in Section 3.4.1.
The eectofthemagneticfieldstrengthontheperformanceofa
magnetic separator
The performance of a magnetic separator is, to a great extent, determined by
the magnetic field strength, as discussed in Chapters 1 and 3. Simple theoretical
models show that an increase in the magnetic induction increases the e!ciency
of magnetic separation. A more realistic model of the process [S31, S32, S33],
and numerous practical applications of the technique, paint a somewhat dierent
picture and show that an increase in the magnetic field strength can have even
366 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS
Figure 5.33: The attractive and repulsive regions for paramagnetic particles,
around a magnetized matrix collector, as a function of magnetic
induction.
a negative eect on the performance of a magnetic separator.
There are several reasons for the failure of the magnetic field to play the role
that follows from simple theoretical models. Most materials that are treated
by high-intensity magnetic separation are not paramagnetic. They usually pos-
sess some degree of magnetic ordering, which results in the field dependence of
their magnetic susceptibility, as given by the Owen-Honda equation (1.44): =
4
+P@K.
In the first phase, before the onset of magnetic saturation of the ferromag-
netic (v=o=) components, the magnetic susceptibility remains constant when the
field increases, since the magnetization also increases. Once the saturation is
reached, which, for most ferromagnetic materials, occurs at quite low fields of
the order of a few mT, the susceptibility decreases with increasing applied mag-
netic field and the drop in the value of magnetic susceptibility can often be
quite appreciable. At high magnetic fields, the second term in eq. (1.44) can be
neglected and the susceptibility becomes constant and equal to
4
. Therefore,
the reduction in the magnetic susceptibility that accompanies the increase in
the magnetic field strength can often result in stagnation, or even a decrease,
in the recovery.
Another factor that adversely aects the recovery into the magnetic fraction
is the decrease in the surface area of the matrix elements which is available
for the capture of paramagnetic particles, when the applied field is increased.
5.3. WET MAGNETIC SEPARATION 367
The situation is depicted in Figs. 3.15, 3.16 and Fig. 5.33 and was discussed in
greater detail in Section 3.4.1. It is evident that by increasing the magnetic field,
the loading of material on the matrix increases as a consequence of a smaller
surface area of the matrix available for paramagnetic capture.
A high magnetic field often induces the formation of clusters of magnetic
particles through the process of magnetic flocculation. Since the formation
of such flocs is usually non-selective, the grade of the magnetic concentrate
deteriorates and the recovery can drop as a consequence of excessive matrix
loading and of increased interstitial velocity of the slurry in the matrix.
It can therefore be seen that a high magnetic field does not necessarily mean
better results. The optimum operating magnetic field must be determined in
such a manner that the recovery/grade relationship satisfies the objective of the
application.
In order to illustrate the role of the magnetic field strength in high-gradient
magnetic separation, typical examples, as observed in the laboratory or on pro-
duction scale, will be discussed.
Magnetic filtration The magnitude of the applied magnetic field is dictated
mainly by the magnetic susceptibility of the suspended solids. The usual pic-
ture in magnetic filtration is that the e!ciency of the removal of solids ini-
tially increases with increasing field strength up to a certain point, after which
the filtration e!ciency remains constant. In filtering ferromagnetic particles, a
magnetic induction of 0.3 T is usually su!cient. Typical applications are the
purification of steel mill waste water and cooling water in power plants, and the
removal of suspended solids from municipal waste water seeded with magnetite.
A typical example of such a straightforward "all-or-none" relationship between
the magnetic field and the separation e!ciency is shown in Fig. 5.34.
A similar observation was made in the application of a superconducting
magnetic separator to the purification of municipal sewage water seeded with
a ferromagnetic material. Isogami et al. [I5] reported that when a su!cient
concentration of the seeding material was used, the filtration e!ciency was
essentially independent of the magnetic field strength in the interval from 0.1 T
to 1.5 T. The experimental results are summarized in Fig. 5.35.
In more demanding applications, in which feebly magnetic solids are to be
removed, a much higher magnetic induction, sometimes as high as 2 T, is re-
quired. However, often little benefit is obtained with fields exceeding 1.5 T, as
is demonstrated in Fig. 5.36. It is evident that the e!ciency of filtration of very
fine, weakly magnetic kimberlite particles from diamond waste water does not
improve with increasing magnetic field for a magnetic field strength exceeding
1.5 T.
These two examples typify the situation usually encountered in magnetic
filtration, namely that the removal of solids is generally not improved by in-
creasing the field strength above a certain value.
368 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS
Figure 5.34: The e!ciency of the removal of solids from municipal waste water
seeded with magnetite (adapted from [A28]).
Figure 5.35: Removal of suspended solids from sewage water, seeded with a
ferromagnetic material, as a function of the magnetic induction
(adapted from Isogami et al. [I5]).
5.3. WET MAGNETIC SEPARATION 369
Figure 5.36: The e!ciency of magnetic filtration of kimberlite solids from three
types of diamond mine waste water (adapted from [S32]).
Removal of magnetic impurities from mixtures Although a high mag-
netic field strength is often desirable for an e!cient removal of magnetic im-
purities, care must be exercised in the selection of the magnetic induction. In
contrast to magnetic filtration, where excessive magnetic field strength usually
does not aect the performance of a magnetic filter, results of the removal of
magnetic impurities from mineral mixtures can often be impaired by undesirable
losses of the non-magnetic product.
As in magnetic filtration, the e!ciency of removal of weakly magnetic im-
purities does not increase indefinitely with increasing magnetic field strength
and a leveling-o of the removal e!ciency at higher magnetic fields is usually
observed.
This eect is illustrated in Figs. 5.37 and 5.38. Figure 5.37 shows the eect
of the magnetic field on the concentration of iron oxide in kaolin (South Africa).
It can be seen that, with a steel wool matrix, a magnetic induction greater than
0.8 T did not improve the grade of the kaolin product. With the ball matrix,
this threshold magnetic field was higher, namely 1.2 T. Moreover, the quality
of the product appears to deteriorate in the magnetic field of the order of 2 T.
An even more dramatic leveling-o of the grade of the non-magnetic glass sand
product, as a function of the magnetic field can be seen in Fig. 5.38.
It can be seen that very often there is a threshold field, which marks the
maximum field for which the grade of the non-magnetic product still improves.
Magnetic field smaller or equal to the threshold field should be used.
However, in some applications, particularly in kaolin beneficiation, very high
magnetic fields are reported to be beneficial. Figure 5.39 illustrates that with