This action might not be possible to undo. Are you sure you want to continue?
ICESP X – June 2006
BackCorona Model
ICESP X – Australia 2006
Paper 3B3
BACKCORONA MODEL FOR PREDICTION OF ESP EFFICIENCY
AND VOLTAGECURRENT CHARACTERISTICS.
G. BACCHIEGA
1
, I. GALLIMBERTI
1
, V. ARRONDEL
2
, N. CARAMAN
2
AND M. HAMLIL
2
1: IRS srl  via Vigonovese 81  35124 Padova  ITALY (email: bacchiega@irsweb.it)
2: EDF R&D – 6, quai Watier – 78 401 Chatou  FRANCE (email :
veronique.arrondel@edf.fr)
ABSTRACT
The backcorona discharge is formed by a series of microdischarges in the air spaces
between the partcles of the dust layer deposited on the collecting plates. It starts when the
particle resistivity is quite high: it changes the voltagecurrent characteristics of the
electrostatic precipitator and significantly lowers its efficiency.
In the present work, the variation of the voltagecurrent characteristics, depending on
resistivity and particle layer thickness, has been successfully reproduced by a simple electrical
circuit model of the dust layer, including a backcorona current generator and the voltage drop
across the particle layer.
The model also includes an estimation of the variation of the captation efficiency: as back
corona lowers the ion space charge, the computation of the backcorona current allows to
evaluate the variation of the particle charging rate and therefore of their migration velocity.
2
ICESP X – June 2006
BackCorona Model
1. INTRODUCTION
The back corona discharge is a complex nonlinear phenomenon, which has been the subject
of a limited number of investigations [19]; its characteristics and effects are often described
only on qualitative basis. Its presence inside of an electrostatic precipitator is usually
associated to a threshold in the resistivity of the particle layer deposited on the plates. It is
experimentally evidenced by a strong bend in the voltagecurrent characteristics.
Previous mathematical models of backcorona inside the precipitator duct computed the
voltagecurrent characteristics without taking into account the particle layer resistivity and
voltage drop; they were based on a finite difference [10] or finite element [11] simulation of
the glowcorona discharge, including the backdischarge ion production associated to the
glow corona current density.
This paper describes a simple backcorona model describing the nonlinear effects of layer
resistivity and thickness. The model highlights the coupling between glowcorona and back
corona, and layer resistivity and thickness.
Back corona inception threshold is linked to the average electric field in the particle layer, and
it results one order of magnitude lower than the dielectric particle layer breakdown threshold
[1][6].
The value of the backcorona discharge current is defined on the basis of the overvoltage
with respect to the inception electric field. Using this value in an equivalent electrical circuit
of the dust layer, it has been possible to reproduce the change in the voltagecurrent
characteristics induced by backcorona.
The influence of the positively charged ions produced by backcorona on precipitation
efficiency has been estimated through the reduction of the particle charging rate. Finally,
using the modified voltagecurrent characteristics into an iterative method that takes into
account the reciprocal influence between efficiency, particle layer thickness and electrical
conditions, it has been possible to estimate the efficiency loss created by backcorona. The
model results successfully reproduce the efficiency losses described in literature [1][8].
2. BACKCORONA DISCHARGE MODEL
In a typical electrostatic precipitator configuration a negative applied voltage U
a
between an
emitter electrode and a collecting plate produce a negative glowcorona discharge at the
emitter electrode (Fig.1). The negative ions exiting the glow corona discharge together with
the charged particles drift towards the collecting plate.
If the glowcorona discharge current and layer resistivity are high enough, a positive back
corona discharge is formed in the layer. From the dust layer positive ions exit the backcorona
discharge and drift towards the emitter electrode discharging previously charged negative dust
particles. Besides positive ions may attach to neutral dust particles, causing their drift
backwards to the emitting electrodes.
3
ICESP X – June 2006
BackCorona Model
l
Collecting plate
Particle layer
J
U d
U
l
Emitter
electrode
U
a
Fig.1 Typical electrostatic precipitator configuration
2.1 Backcorona inception level
The backcorona discharge is formed by a series of microdischarges in the air spaces
between the dust partcles deposited on the collecting plates. It forms when an excessive
electric field is induced in the dust layer by the current flowing through its surface. A micro
discharge is associated to the collapse of the void spaces between the deposited particles
(Fig.2).
Fig.2 Schematic representation of the equipotential lines within the collected dust layer,
before (a) and after (b) the discharge in void spaces
The electric field threshold E
t
for backcorona inception depends on different parameters,
including particle geometrical shape and compactness inside the dust layer. Its value is very
difficult to be measured, as particle layer condition and resistivity inside an electrostatic
precipitator are strongly variable.
It is important to point out that the electric field for backcorona inception is not the dielectric
breakdown field of the layer, whose value is around 12 10
6
V/m [1][6], near the dielectric
breakdown of air (about 2.6 10
6
V/m).
In order to quantify the threshold value, the average electric field in the layer E
l
can be
calculated, as the product of the layer resistivity ? ?and current density J on the collecting
plates:
J E
l
ρ · (Eq.1)
4
ICESP X – June 2006
BackCorona Model
In the literature there are no published values of the average electric field threshold Et;
however, a layer threshold resistivity of about 2 10
10
Ohm.cm is commonly indicated. By
assuming an average current density on the collecting plates around 2.5  5.0 mA/m2, it is
possible to define an electric field threshold of the order of 0.5  1 10
5
V/m. This value is
confirmed by model theoretical approximations [12] and by the results described hereafter.
2.2 Backcorona current density
The estimation of the backcorona current density J
b
is based on the electric field in the
particle layer E
l
. A typical quadratic relationship has been assumed, corrected for the particle
layer thickness l:
( )
2 4 . 0
t l b b
E E l k J − · (Eq.2)
The value of the constant k
b
(3 10
6
) as well as the value of the power of the particle layer
thickness (0.4), have been optimized on the basis of SRI resistivity probe measurements and
on the laboratory and industrial measurements described in the next chapter.
This current calculation has been kept as simple as possible; it describes the steady state
regime, and does not take into account the dynamic behavior of backcorona [5][10].
2.3 Voltage current characteristics
Modeling voltage current characteristics in an electrostatic precipitator with backcorona
needs a representation of the particle layer. Particle layer is characterized by his electrical
resistivity. From an electrical point of view how the backcorona discharge interact with
particle layer resistivity? In our model, backcorona current adds the current produced by the
negative glowcorona discharge and do participate to the voltage drop in the particle layer.
J ’
V
a
V
d

J
CE
J +
V
c
+
J ’
U
a
U
d d

J
J
b
l
+
U
Fig.3 Electrical circuit model describing voltage current characteristics with backcorona
The increased voltage drop creates a higher electric field that increases backcorona current
density and once more modifies the voltage drop. We have a positive feedback effect that is a
typical phenomenon observed in the presence of backcorona.
It is possible to identify the voltagecurrent electrostatic precipitator working point including
the positive feedback using the equations presented in the previous chapter, where the
5
ICESP X – June 2006
BackCorona Model
function J(U
d
) is the voltagecurrent characteristics without the particle layer and backcorona
(see Fig.3 and Eq.3).
¹
¹
¹
'
¹
+ ·
·
+ ·
l E U U
J E
J U J J
l d a
l
b d
'
) ( '
ρ (Eq.3)
The computed influence of particle layer resistivity on the voltagecurrent characteristics are
shown in Fig. 4, for a defined particle layer thickness l of 0.01 cm and using a function J(U
d
)
taken by the Laboratory ESP described in [5]. The resistivity value does have a great
influence on voltagecurrent characteristics.
The characteristics for low dust resistivity are virtually identical with that obtained with clean
plate. As the resistivity value increase the current value increase too. The current rise can be
so high that it limits the achievable applied voltage and for high resistivity values voltage
current characteristics go from inception to spark with a variation of the applied voltage of a
few kilovolts.
0
0.1
0.2
0.3
0.4
0.5
0.6
20 30 40 50 60 70 80
Applied voltage (kV)
C
u
r
r
e
n
t
(
u
A
)
1.00E+09
1.00E+10
6.00E+10
1.00E+11
6.00E+11
Fig. 4 Voltagecurrent characteristics as functions of layer resistivity
Backcorona current varies also as function of particle layer thickness. Fig.5 shows voltage
current characteristics with different layer thickness using the same configuration of Fig.4 and
6x10
10
Ohm.cm layer resistivity. As particle layer thickness increase the backcorona current
is higher. The increase of air spaces within the collected dust layer subject to backcorona
explains this phenomenon.
Current increase as function of particle layer thickness has been validated on experimental
data presented in [5], and shown in Fig.5.
Using this model inside a Electrostatic Precipitator modeling software ORCHIDEE 2 [11] we
were able to successfully represent voltagecurrent characteristics variation as function of dust
layer thickness and properties in an industrial environment. The presented model agrees with
experimental backcorona available data and well represents qualitative backcorona effects.
Layer
resistivity
(Ohm cm)
6
ICESP X – June 2006
BackCorona Model
0
100
200
300
400
500
C
u
r
r
e
n
t
(
m
A
)
Applied voltage (kV)
0
100
200
300
400
500
C
u
r
r
e
n
t
(
m
A
)
Layer 0 cm
Layer 0.02 cm
Layer 0.06 cm
Layer 0.1 cm
0 10 20 30 40 50 60 70 80
Applied Voltage (kV)
Computed
Experimental
0
100
200
300
400
500
C
u
r
r
e
n
t
(
m
A
)
Applied voltage (kV)
0
100
200
300
400
500
C
u
r
r
e
n
t
(
m
A
)
Layer 0 cm
Layer 0.02 cm
Layer 0.06 cm
Layer 0.1 cm
0 10 20 30 40 50 60 70 80
Applied Voltage (kV)
Computed
0
100
200
300
400
500
C
u
r
r
e
n
t
(
m
A
)
Applied voltage (kV)
0
100
200
300
400
500
C
u
r
r
e
n
t
(
m
A
)
Layer 0 cm
Layer 0.02 cm
Layer 0.06 cm
Layer 0.1 cm
0 10 20 30 40 50 60 70 80
Applied Voltage (kV)
Applied voltage (kV)
0
100
200
300
400
500
C
u
r
r
e
n
t
(
m
A
)
Layer 0 cm
Layer 0.02 cm
Layer 0.06 cm
Layer 0.1 cm
0 10 20 30 40 50 60 70 80
Applied Voltage (kV)
Computed
Experimental
Fig. 5 Comparison between computed and experimental VoltageCurrent characteristics, as
functions of particle layer thickness
2.4 Graphical representation of the voltage current working point
It is possible to have a graphical representation and solution of the voltage current working
point. Rewriting model equations (Eq.3) as shown in (Eq.4), it is possible to show graphically
how the applied voltage (U
a
) cannot be applied if there is a strong backcorona and how two
equilibrium voltages can be associated to given current density.
) (
1
2
1 3
l a
l
b
U U J J where
U
l
J
J J J
− ·
¹
¹
¹
'
¹
·
+ ·
ρ
(Eq.4)
Fig.6 shows the relationship between voltage drop in the particle layer (U
l
, xaxe ) and the
current density (J, yaxe) of each equation. Voltage drop and current density are computed for
a typical configuration, using a standard quadratic J(U
d
) function and for a given fixed applied
voltage U
a
(35 kV).
Curve 1 point up the glowcorona current density variation J
1
. For a given applied voltage and
in the absence of backcorona the glowcorona current density lowers as the voltage drop
increase.
7
ICESP X – June 2006
BackCorona Model
Curve 2 shows the voltagecurrent density characteristics given by the particle layer
resistivity. The linear curve slope is function of the particle layer resistivity: the higher the
slope, the lower the resistivity.
The crossing point of curve 1 and 2 is the electrostatic precipitator working point and fix the
voltage drop in the layer. We can imagine graphically how a resistivity variation (i.e. a
different slope in curve 2) influences the electrostatic precipitator electrical conditions.
0
0.05
0.1
0.15
0.2
0.25
0.3
0 0.5 1 1.5 2 2.5 3 3.5 4
Ul (kV)
J
(
m
A
/
m
2
)
_
1: glowcorona
2: particle layer resistivity
3: glowcorona + backcrona
Fig.6 Graphical representation of the voltage current working point
If we add the current density produced by the backcorona to the curve 1 it becomes curve 3
(Eq.4). For a low voltage drop in the layer the backcorona is not incepted and current the
current density is null, as the voltage creates a higher electric field in the layer the back
corona current density increase as a quadratic function of U
l
.
The crossing point of curve 2 and 3 shows the electrostatic precipitator working point. If we
extrapolate the curves we can see that there is actually two crossing point of curve 2 and 3
corresponding to two possible working points in the electrostatic precipitator.
If the backcorona current is high it is actually possible that no crossing point are possible
between curve 2 and 3. In this situation the theoretical given applied voltage cannot be
applied. This corresponds to the experimental data showing a voltage limitation given by a
strong backcorona.
3. BACKCORONA EFFECT ON COLLECTION EFFICIENCY
3.1 Efficiency loss by positive particle charging
The backcorona current adds to the glow corona current, because the positive and negative
ions move in opposite directions. On the contrary, the positive and negative charging effects
on the particles subtract each other, leading to an average lower particle charge and lower
particle collection efficiency.
The efficiency loss given by positive ions and by effects in particles need to be defined.
Performance degradation due to back corona has the following characteristics [7]:
8
ICESP X – June 2006
BackCorona Model
(a) Zero degradation at currents below that required for back corona formation.
(b) Increasing degradation with increasing current above that required for back corona
formation.
(c) The degradation is limited at currents above a saturation level.
The equation we use to link the efficiency without ? ?and with the positive ions effects ? ' is:
η η
,
_
¸
¸
+
−
− + · ′
b
b
d d
J J
J J
k k ) 1 ( (Eq.5)
where J are J
b
are the current density given by glowcorona and backcorona discharge and k
d
is a coefficient optimized on experimental data[8][9].
The equation effects are twofold. It increases the efficiency degradation induced by the
injected positive ions and defines an efficiency loss lower limit given by the constant k
d
. The
efficiency loss lower limit is typical of industrial electrostatic precipitator where the power
supply control and the great numbers of collecting electrodes do not allow the backcorona
discharge to attend the equilibrium conditions where the efficiency goes to zero. In this case
we do use a k
d
coefficient equals to 0.65.
In laboratory conditions and with a fixed applied voltage the saturation level goes near zero
and k
d
in stable conditions is null.
3.2 Efficiency computation
In order to compute the efficiency we used an iterative computation described and
implemented in ORCHIDEE [1214], comprehensive electrostatic precipitator modeling
software. The iterations related to backcorona simulation are the following (see also Fig. 7):
1. Simulation of the electrofilter collection with a null particle layer thickness. The
particle collection strongly depends on the effective collection voltage U
d
that in this
case is equal to the applied voltage U
a
.
2. Computation of particle layer dynamics using the computed collection efficiency as
well as the estimated influence of rapping and reentrainment [15]. A useful result is
the particle layer thickness.
3. Based on particle layer thickness estimation of the voltage drop U
l
at the layer.
4. Computation of the glowcorona current and, if there are the inception conditions,
backcorona current.
5. From this point a new computation cycle starts. Electrofilter collection is computed
using the effective collection voltage (U
d
= U
a
U
l
). The collection efficiency includes
the loss due to the backcorona if this discharge is active on the layer.
6. etc.
9
ICESP X – June 2006
BackCorona Model
ESP
Simulation
Layer dynamics
Layer thickness
ESP
efficiency
Resistivity
Layer volt.drop
Back  corona
current
Glow  corona
current
Fig. 7 Nonlinear coupling between the different physical processes in ORCHIDEE 2
The needed iteration cycle depends on dust layer influence on efficiency. In most
configurations little cycles allow a good estimation of collection efficiency.
We applied the efficiency computation to an industrial configuration (Le Havre) where the
collecting distance between plates allow a comparison with migration velocity found in
literature [8][9]. Fig. 8 shows a comparison of efficiency as function of particle layer
resistivity. It is a qualitative comparison as the electrostatic precipitator details used in
literature were not available. It can be seen that starting at a resistivity of about 10
10
Ohm cm
the efficiency start lowering and saturate at 10
12
Ohm cm.
Even if further quantitative data are needed to fully validate the model, we have for the first
time an electrostatic precipitator model that can estimate the efficiency of an industrial
electrostatic precipitator as function of coal resistivity.
75%
80%
85%
90%
95%
100%
9 10 11 12 13
Log10 Resistivity (Ohm cm)
E
f
f
i
c
i
e
n
c
y
Parker
Oglesby A
Le Havre
Fig. 8 Efficiency as function of particle layer resistivity
10
ICESP X – June 2006
BackCorona Model
4. CONCLUSIONS
This paper presents for the first time a model reproducing the backcorona effects on voltage
current characteristics and on electrostatic precipitator efficiency.
The strict connection between the discharge at the emitter electrode and backcorona is
quantitatively estimated. A simple electric scheme explains the positive feedback given by the
particle layer resistivity on backcorona current. Efficiency loss estimation makes use of the
backcorona electrical properties calculations and is coupled with the overall collection
simulation.
The model simulation is implemented in a userfriendly electrostatic precipitator model
(ORCHIDEE) already in use by coal power plant personnel [12]. Operating personnel can
contribute to dust emission control on their unit
Practical application of this model including : estimating the impact of different type of coal
on efficiency and making the right choice in coal blending, distinguishing voltagecurrent
variation associated with faults or to particle resistivity variation, etc.
More studies can complete and refine this model as well as giving new uses. Future back
corona model applications include:
• Better powersupply controls making use of detailed modeling of glow and back
corona discharge and allowing higher collection efficiency;
• Simulation in environment where backcorona is a known problem: cement plant
electrostatic precipitators and Surface coating equipment
• Model extension to describe the more general backdischarge [10] of which back
corona is the low current part
4 REFERENCES
[1] H.J. WHITE, Industrial electrostatic precipitation, Addison Wesley, 1963
[2] S. MASUDA, A. MIZUNO: “Initiation condition and mode of backdischarge”, Journal
of Electrostatics 4, 1977
[3] S. MASUDA, A. MIZUNO: “Flashover measurements of backdischarge”, Journal of
Electrostatics 4, 1978
[4] R.J. TRUCE, “Back Corona and its Effects on the Optimisation of electrostatic
Precipitator Energisation control”, Seventh International Clean Air Conference, August
2428, 1981
[5] G. BACCHIEGA, I. GALLIMBERTI, V. ARRONDEL, PH. RAIZER, J. LECOINTRE, M.
HAMLIL, “Static and dynamic backcorona characteristics”, IX ICESP, Kruger – South
Africa, 2004.
[6] A. JAWOREK, T. CZECH, E. RAJCH, M. LACKOWSKI, “Laboratory studies of back
discharge in fly ash”, Journal of Electrostatic, 64 (2006),p 306317
[7] C. CHESMOND, R.J. TRUCE, “Using Electrostatic Precipitators to Collect Highly
Resistive Dust”, Institute of Engineers Australia, Queensland Technical Papers. Vol. 26,
No.9, June 1985, pp 1620.
[8] S. OGLESBY, G. NICHOLS: “Electrostatic precipitation”, Dekker, New York, 1978
[9] K. R. PARKER, “Applied Electrostatic Precipitation”, Chapman & Hall, London  UK.
1997
[10] P.A. LAWLESS, E. SPARKS, “A mathematical model for calculating effects of back
corona in wireduct electrostatic precipitators”, Journal of Applied Physics  January
1980  Volume 51, Issue 1, pp. 242256
11
ICESP X – June 2006
BackCorona Model
[11] S. CRISTINA, M. FELIZIANI, “Method for computing backcorona in an electrostatic
precipitator”, III ICESP, Abano – Italy, 1987
[12] V. ARRONDEL, N. CARAMAN, M. HAMLIL, G. BACCHIEGA, I. GALLIMBERTI,
“ORCHIDEE: A friendly tool to assist the operator and the engineering in the control of
dust emission”, X ICESP, Cairns – Australia, 2006.
[13] V. ARRONDEL, G. BACCHIEGA, I. GALLIMBERTI, "ESP modelling: from University
to Industrial Application", VIII ICESP, Birmingham – USA, 2001
[14] V. ARRONDEL, J. SALVI, I. GALLIMBERTI, G. BACCHIEGA, “ORCHIDEE:
Efficiency Optimisation of Coal Ash Collection in Electrostatic Precipitators”, IX ICESP,
Kruger – South Africa, 2004.
[15] V. ARRONDEL, N. CARAMAN, M. HAMLIL, G. BACCHIEGA, I. GALLIMBERTI,
“Development of an industrial model of rapping effect on the collecting efficiency”, X
ICESP, Cairns – Australia, 2006.
and layer resistivity and thickness. If the glowcorona discharge current and layer resistivity are high enough. they were based on a finite difference [10] or finite element [11] simulation of the glowcorona discharge. causing their drift backwards to the emitting electrodes. The negative ions exiting the glow corona discharge together with the charged particles drift towards the collecting plate. BACKCORONA DISCHARGE MODEL In a typical electrostatic precipitator configuration a negative applied voltage Ua between an emitter electrode and a collecting plate produce a negative glowcorona discharge at the emitter electrode (Fig. a positive backcorona discharge is formed in the layer. Besides positive ions may attach to neutral dust particles. it has been possible to reproduce the change in the voltagecurrent characteristics induced by backcorona. Its presence inside of an electrostatic precipitator is usually associated to a threshold in the resistivity of the particle layer deposited on the plates. Previous mathematical models of backcorona inside the precipitator duct computed the voltagecurrent characteristics without taking into account the particle layer resistivity and voltage drop. its characteristics and effects are often described only on qualitative basis. particle layer thickness and electrical conditions. The model results successfully reproduce the efficiency losses described in literature [1][8]. It is experimentally evidenced by a strong bend in the voltagecurrent characteristics. and it results one order of magnitude lower than the dielectric particle layer breakdown threshold [1][6]. including the backdischarge ion production associated to the glow corona current density. From the dust layer positive ions exit the backcorona discharge and drift towards the emitter electrode discharging previously charged negative dust particles. 2. Finally. it has been possible to estimate the efficiency loss created by backcorona. Using this value in an equivalent electrical circuit of the dust layer. The value of the backcorona discharge current is defined on the basis of the overvoltage with respect to the inception electric field. 2 ICESP X – June 2006 BackCorona Model . which has been the subject of a limited number of investigations [19].1. The model highlights the coupling between glowcorona and backcorona. using the modified voltagecurrent characteristics into an iterative method that takes into account the reciprocal influence between efficiency. This paper describes a simple backcorona model describing the nonlinear effects of layer resistivity and thickness.1). Back corona inception threshold is linked to the average electric field in the particle layer. The influence of the positively charged ions produced by backcorona on precipitation efficiency has been estimated through the reduction of the particle charging rate. INTRODUCTION The back corona discharge is a complex nonlinear phenomenon.
In order to quantify the threshold value. It forms when an excessive electric field is induced in the dust layer by the current flowing through its surface. the average electric field in the layer El can be calculated.1 Backcorona inception level The backcorona discharge is formed by a series of microdischarges in the air spaces between the dust partcles deposited on the collecting plates.2 Schematic representation of the equipotential lines within the collected dust layer. before (a) and after (b) the discharge in void spaces The electric field threshold Et for backcorona inception depends on different parameters. It is important to point out that the electric field for backcorona inception is not the dielectric breakdown field of the layer.1) 3 ICESP X – June 2006 BackCorona Model . Its value is very difficult to be measured. near the dielectric breakdown of air (about 2.6 106 V/m).1 Typical electrostatic precipitator configuration 2. Fig. A microdischarge is associated to the collapse of the void spaces between the deposited particles (Fig.2). as particle layer condition and resistivity inside an electrostatic precipitator are strongly variable.Ua Ul Particle layer Emitter electrode Collecting plate Ud J l Fig. whose value is around 12 106 V/m [1][6]. as the product of the layer resistivity ? ?nd current density J on the collecting a plates: E l = ρJ (Eq. including particle geometrical shape and compactness inside the dust layer.
Particle layer is characterized by his electrical resistivity. 2.In the literature there are no published values of the average electric field threshold Et. 4 (E l − Et )2 (Eq. corrected for the particle layer thickness l: J b = k b l 0.3 Voltage current characteristics Modeling voltage current characteristics in an electrostatic precipitator with backcorona needs a representation of the particle layer.2 Backcorona current density The estimation of the backcorona current density Jb is based on the electric field in the particle layer El. backcorona current adds the current produced by the negative glowcorona discharge and do participate to the voltage drop in the particle layer. We have a positive feedback effect that is a typical phenomenon observed in the presence of backcorona. a layer threshold resistivity of about 2 1010 Ohm. 2. From an electrical point of view how the backcorona discharge interact with particle layer resistivity? In our model.0 mA/m2. it is possible to define an electric field threshold of the order of 0. and does not take into account the dynamic behavior of backcorona [5][10].cm is commonly indicated. have been optimized on the basis of SRI resistivity probe measurements and on the laboratory and industrial measurements described in the next chapter. This current calculation has been kept as simple as possible. it describes the steady state regime. A typical quadratic relationship has been assumed. where the 4 ICESP X – June 2006 BackCorona Model . It is possible to identify the voltagecurrent electrostatic precipitator working point including the positive feedback using the equations presented in the previous chapter. By assuming an average current density on the collecting plates around 2. however.3 Electrical circuit model describing voltage current characteristics with backcorona The increased voltage drop creates a higher electric field that increases backcorona current density and once more modifies the voltage drop. This value is confirmed by model theoretical approximations [12] and by the results described hereafter.5 .5.5 .4).1 105 V/m.2) The value of the constant kb (3 106 ) as well as the value of the power of the particle layer thickness (0. J’  J U Va ++ U Vd Jb+ J CE U c Vl Fig.
4 0.3 0.2 0. 5 ICESP X – June 2006 BackCorona Model . The resistivity value does have a great influence on voltagecurrent characteristics.3). As particle layer thickness increase the backcorona current is higher.5 shows voltagecurrent characteristics with different layer thickness using the same configuration of Fig. Fig. J ' = J (U d ) + J b El = ρ J ' U = U + E l a d l (Eq.5.00E+11 Current (uA) Layer resistivity (Ohm cm) Fig.3) The computed influence of particle layer resistivity on the voltagecurrent characteristics are shown in Fig.4 and 6x1010 Ohm. The presented model agrees with experimental backcorona available data and well represents qualitative backcorona effects.1 0 20 30 40 50 Applied voltage (kV) 60 70 80 1. 4. The characteristics for low dust resistivity are virtually identical with that obtained with clean plate.00E+10 6. and shown in Fig.00E+09 1.3 and Eq.00E+10 1.00E+11 6. The current rise can be so high that it limits the achievable applied voltage and for high resistivity values voltagecurrent characteristics go from inception to spark with a variation of the applied voltage of a few kilovolts. 0.6 0. The increase of air spaces within the collected dust layer subject to backcorona explains this phenomenon.01 cm and using a function J(Ud) taken by the Laboratory ESP described in [5].function J(Ud) is the voltagecurrent characteristics without the particle layer and backcorona (see Fig.cm layer resistivity. for a defined particle layer thickness l of 0. Using this model inside a Electrostatic Precipitator modeling software ORCHIDEE 2 [11] we were able to successfully represent voltagecurrent characteristics variation as function of dust layer thickness and properties in an industrial environment. As the resistivity value increase the current value increase too.5 0. Current increase as function of particle layer thickness has been validated on experimental data presented in [5]. 4 Voltagecurrent characteristics as functions of layer resistivity Backcorona current varies also as function of particle layer thickness.
Curve 1 point up the glowcorona current density variation J1 . xaxe ) and the current density ( yaxe) of each equation.3) as shown in (Eq. as functions of particle layer thickness 2.4 Graphical representation of the voltage current working point It is possible to have a graphical representation and solution of the voltage current working point. 6 ICESP X – June 2006 BackCorona Model .500 400 Current (mA) 300 200 Experimental 100 0 500 Layer 0 cm 400 Applied voltage (kV) Layer 0.1 cm 200 Computed 100 0 0 10 20 30 40 50 60 70 80 Applied Voltage (kV) Fig. a typical configuration. For a given applied voltage and in the absence of backcorona the glowcorona current density lowers as the voltage drop increase.02 cm Layer 0. Rewriting model equations (Eq.6 shows the relationship between voltage drop in the particle layer (Ul. Voltage drop and current density are computed for J.4).06 cm Current (mA) 300 Layer 0. using a standard quadratic J(Ud) function and for a given fixed applied voltage Ua (35 kV). 5 Comparison between computed and experimental VoltageCurrent characteristics.4) Fig. it is possible to show graphically how the applied voltage (Ua) cannot be applied if there is a strong backcorona and how two equilibrium voltages can be associated to given current density. J 3 = J1 + J b where J 1 = J (U a − U l ) J =ρ l 2 Ul (Eq.
4). This corresponds to the experimental data showing a voltage limitation given by a strong backcorona.e.6 Graphical representation of the voltage current working point If we add the current density produced by the backcorona to the curve 1 it becomes curve 3 (Eq. the positive and negative charging effects on the particles subtract each other. because the positive and negative ions move in opposite directions. 3. In this situation the theoretical given applied voltage cannot be applied. 0.5 3 3. 3. Performance degradation due to back corona has the following characteristics [7]: 7 ICESP X – June 2006 BackCorona Model . We can imagine graphically how a resistivity variation (i.2 0.05 0 0 0. The crossing point of curve 1 and 2 is the electrostatic precipitator working point and fix the voltage drop in the layer.3 1: glowcorona 0. as the voltage creates a higher electric field in the layer the backcorona current density increase as a quadratic function of Ul.Curve 2 shows the voltagecurrent density characteristics given by the particle layer resistivity.1 BACKCORONA EFFECT ON COLLECTION EFFICIENCY Efficiency loss by positive particle charging The backcorona current adds to the glow corona current. If we extrapolate the curves we can see that there is actually two crossing point of curve 2 and 3 corresponding to two possible working points in the electrostatic precipitator.1 0.5 2 2. a different slope in curve 2) influences the electrostatic precipitator electrical conditions. leading to an average lower particle charge and lower particle collection efficiency.15 0. The crossing point of curve 2 and 3 shows the electrostatic precipitator working point. On the contrary.5 1 1. The linear curve slope is function of the particle layer resistivity: the higher the slope. For a low voltage drop in the layer the backcorona is not incepted and current the current density is null.25 2: particle layer resistivity 3: glowcorona + backcrona J (mA/m2)_ 0. The efficiency loss given by positive ions and by effects in particles need to be defined. the lower the resistivity.5 4 U l (kV) Fig. If the backcorona current is high it is actually possible that no crossing point are possible between curve 2 and 3.
The equation effects are twofold. comprehensive electrostatic precipitator modeling software. 4. In this case we do use a kd coefficient equals to 0. (b) Increasing degradation with increasing current above that required for back corona formation.65. The efficiency loss lower limit is typical of industrial electrostatic precipitator where the power supply control and the great numbers of collecting electrodes do not allow the backcorona discharge to attend the equilibrium conditions where the efficiency goes to zero. 6. 5. The iterations related to backcorona simulation are the following (see also Fig.5) where J are Jb are the current density given by glowcorona and backcorona discharge and kd is a coefficient optimized on experimental data[8][9]. 3.(a) Zero degradation at currents below that required for back corona formation. Electrofilter collection is computed using the effective collection voltage (Ud= UaUl). A useful result is the particle layer thickness. Computation of particle layer dynamics using the computed collection efficiency as well as the estimated influence of rapping and reentrainment [15]. Based on particle layer thickness estimation of the voltage drop Ul at the layer. 7): 1. 8 ICESP X – June 2006 BackCorona Model .2 Efficiency computation In order to compute the efficiency we used an iterative computation described and implemented in ORCHIDEE [1214]. In laboratory conditions and with a fixed applied voltage the saturation level goes near zero and kd in stable conditions is null. etc. The collection efficiency includes the loss due to the backcorona if this discharge is active on the layer. Computation of the glowcorona current and. 3. It increases the efficiency degradation induced by the injected positive ions and defines an efficiency loss lower limit given by the constant kd. (c) The degradation is limited at currents above a saturation level. if there are the inception conditions. The particle collection strongly depends on the effective collection voltage Ud that in this case is equal to the applied voltage Ua. backcorona current. 2. From this point a new computation cycle starts. Simulation of the electrofilter collection with a null particle layer thickness. The equation we use to link the efficiency without ? ?nd with the positive ions effects ? ' is: a J − Jb η η ′ = k d + (1 − kd ) J + Jb (Eq.
We applied the efficiency computation to an industrial configuration (Le Havre) where the collecting distance between plates allow a comparison with migration velocity found in literature [8][9]. 7 Nonlinear coupling between the different physical processes in ORCHIDEE 2 The needed iteration cycle depends on dust layer influence on efficiency. 100% 95% Efficiency 90% 85% Parker Oglesby A Le Havre 80% 75% 9 10 11 Log10 Resistivity (Ohm cm) 12 13 Fig. 8 Efficiency as function of particle layer resistivity 9 ICESP X – June 2006 BackCorona Model . Fig. Even if further quantitative data are needed to fully validate the model. It can be seen that starting at a resistivity of about 1010 Ohm cm the efficiency start lowering and saturate at 1012 Ohm cm. we have for the first time an electrostatic precipitator model that can estimate the efficiency of an industrial electrostatic precipitator as function of coal resistivity. In most configurations little cycles allow a good estimation of collection efficiency.drop Layer dynamics Layer thickness Fig.Glowcorona current Backcorona current ESP Simulation ESP efficiency Resistivity Layer volt. 8 shows a comparison of efficiency as function of particle layer resistivity. It is a qualitative comparison as the electrostatic precipitator details used in literature were not available.
Journal of Electrostatic. New York. August 2428. MASUDA. M. R. 242256 10 ICESP X – June 2006 BackCorona Model . RAJCH. J.A. “Applied Electrostatic Precipitation”. 1978 [4] R. 1977 [3] S.J. A simple electric scheme explains the positive feedback given by the particle layer resistivity on backcorona current. Journal of Electrostatics 4. The strict connection between the discharge at the emitter electrode and backcorona is quantitatively estimated. HAMLIL. PH. M.p 306317 [7] C. G. “Static and dynamic backcorona characteristics”. etc.J. R. 1978 [9] K. Operating personnel can contribute to dust emission control on their unit Practical application of this model including : estimating the impact of different type of coal on efficiency and making the right choice in coal blending. V. E. Institute of Engineers Australia. pp 1620. Queensland Technical Papers.J. “Using Electrostatic Precipitators to Collect Highly Resistive Dust”. CONCLUSIONS This paper presents for the first time a model reproducing the backcorona effects on voltagecurrent characteristics and on electrostatic precipitator efficiency. RAIZER. IX ICESP. CHESMOND.9. 64 (2006). 1997 [10] P. JAWOREK.UK. 1963 [2] S. Kruger – SouthAfrica. T. CZECH. The model simulation is implemented in a userfriendly electrostatic precipitator model (ORCHIDEE) already in use by coal power plant personnel [12]. NICHOLS: “Electrostatic precipitation”. [6] A. E. LACKOWSKI. June 1985. LECOINTRE. Journal of Applied Physics . TRUCE. MIZUNO: “Flashover measurements of backdischarge”. SPARKS. • Simulation in environment where backcorona is a known problem: cement plant electrostatic precipitators and Surface coating equipment • Model extension to describe the more general backdischarge [10] of which backcorona is the low current part 4 REFERENCES [1] H. [8] S.January 1980 . Addison Wesley. A. ARRONDEL. Vol. Issue 1. Future backcorona model applications include: • Better powersupply controls making use of detailed modeling of glow and back corona discharge and allowing higher collection efficiency. MASUDA. Industrial electrostatic precipitation. distinguishing voltagecurrent variation associated with faults or to particle resistivity variation. 2004. LAWLESS. MIZUNO: “Initiation condition and mode of backdischarge”.Volume 51. More studies can complete and refine this model as well as giving new uses. GALLIMBERTI. Chapman & Hall. pp. BACCHIEGA. A. Dekker. “A mathematical model for calculating effects of back corona in wireduct electrostatic precipitators”. “Back Corona and its Effects on the Optimisation of electrostatic Precipitator Energisation control”. 1981 [5] G. Seventh International Clean Air Conference. Efficiency loss estimation makes use of the backcorona electrical properties calculations and is coupled with the overall collection simulation. 26. I. TRUCE. OGLESBY.4. WHITE. PARKER. “Laboratory studies of backdischarge in fly ash”. London . No. Journal of Electrostatics 4.
GALLIMBERTI. 2004. "ESP modelling: from University to Industrial Application". M. FELIZIANI. ARRONDEL. J. VIII ICESP. BACCHIEGA. 11 ICESP X – June 2006 BackCorona Model . IX ICESP. “Method for computing backcorona in an electrostatic precipitator”. HAMLIL. N. GALLIMBERTI. I. I. BACCHIEGA. [13] V. I. I.[11] S. G. Kruger – South Africa. G. 2001 [14] V. ARRONDEL. B ACCHIEGA. X ICESP. 2006. N. X ICESP. “Development of an industrial model of rapping effect on the collecting efficiency”. 2006. “ORCHIDEE: Efficiency Optimisation of Coal Ash Collection in Electrostatic Precipitators”. 1987 [12] V. Cairns – Australia. CARAMAN. M. ARRONDEL. CRISTINA. M. III ICESP. BACCHIEGA. GALLIMBERTI. Abano – Italy. GALLIMBERTI. G. Cairns – Australia. Birmingham – USA. SALVI. ARRONDEL. G. HAMLIL. [15] V. “ORCHIDEE: A friendly tool to assist the operator and the engineering in the control of dust emission”. CARAMAN.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.