- Email: [email protected]
Initiation condition and mode of back discharge
Journal of Electrostatics, 4( 1977/1978) 35--52 © Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands
35
INITIATION CONDITION AND MODE OF BACK DISCHARGE
SENICHI MASUDA and AKIRA MIZUNO
Department of Electrical Engineering, University of Tokyo, 7--3--1, Hongo, Bunkyo-Ku, Tokyo (Japan) (Received November 8, 1976; in revised form January 21, 1977)
Summary Modes of back discharge occurring in the electrostatic precipitator were studied using, instead of a dust layer, the model samples of glass and mica plates with a pinhole, and tissue papers, It was confirmed that back discharge started to occur when the apparent field strength in the sample layers exceeded its breakdown field strength. Back discharge became a streamer corona under atmospheric conditions. It could be classified into a space streamer mode, a surface streamer mode and a mixed streamer mode, depending upon the field distribution around the breakdown point in the sample layers. The first and the third modes occurred when the field strength in the air gap, Ea, exceeded about 5 kV/cm, and positive ions were generated in the whole gas space. The second mode appeared when E a was lower than about 5 kV/cm, and ion generation was limited to the near surface region. Among the factors affecting the back discharge, dust resistivity was the most important. For low dust resistivity, space streamers tended to develop from the breakdown points when the applied voltage was raised. For high dust resistivity, on the other hand, the number of breakdown points increased, and surface discharge was pronounced. A remarkable difference in modes was observed when using positive corona. Neither space streamer nor surface discharge occurred and the flashover voltage was higher than that with negative corona.
1. I n t r o d u c t i o n Back discharge is one o f t h e m o s t difficult p r o b l e m s in electrostatic prec i p i t a t o r s impairing their p e r f o r m a n c e in m a n y industrial plants [1]. This is an a b n o r m a l kind o f discharge w h i c h is triggered b y b r e a k d o w n in a high resistivity d u s t layer d e p o s i t e d o n the collecting e l e c t r o d e and w h i c h lowers the flashover voltage, r e d u c i n g particle charge and causing a severe d r o p in c o l l e c t i o n efficiency. The n a t u r e o f b a c k discharge d e p e n d s o n m a n y f a c t o r s such as the electrical p r o p e r t i e s o f the d u s t layer and the c h e m i c a l p r o p e r t i e s o f the particles themselves, and its f o r m is v e r y c o m p l i c a t e d . T h e r e f o r e , m o r e intensive and basic investigations are required t o solve the b a c k discharge p r o b l e m , a n d also t o assess p r e c i p i t a t o r p e r f o r m a n c e , w h e n b a c k discharge occurs. Back discharge o c c u r r i n g w h e n using negative c o r o n a can be classified into t w o m a j o r discharge m o d e s . One is t h e s t r e a m e r m o d e , o c c u r r i n g w i t h high
36 gas density; the other, the glow corona mode, occurring with low gas density. Normally, streamers, formed at the breakdown point on the layer, proceed into the gas space towards the discharge electrode or to the accumulated charges on the dust surface, or in both directions, depending upon the field distribution around the starting point. It is appropriate to classify the streamer mode into three sub-modes: space streamer mode, surface streamer mode and mixed streamer mode. The third one appears in most of the practical cases. In this paper, an experimental study of the back discharge of the streamer mode carried out under atmospheric pressure and room temperature is reported. Studies of the back discharge of a glow corona mode will be reported separately [ 2]. 2. Initiation condition and initial mode of back discharge At first, the initiation condition and the initial mode of back discharge was studied using a needle-to-plane electrode system, located inside a thermostat where humidity could also be controlled. It was tested whether the initiation of back discharge was governed by the breakdown field strength of a layer, Eds , measured separately using parallel plane electrodes. In order to change the value of Eds of a sample, two glass plates, each having a pinhole, were used on top of one another as the layer sample, as indicated in Fig. 1. By altering the position of one plate and thus changing the distance between the holes, the value of -Eds could be changed. The resistivity of the glass plate, Pd, was 6 X 10 ~ ohm cm and the diameter of the pinhole was 0.5 mm. The thickness of one plate was 2.0 mm. An image intensifier tube (EMI, type 9912) was used at its maximum gain (about 106) to observe a o
D,C.H,V, NEEDLE ELECTRODE
T E
TEST SAMPLE (DOUBLE PLATES) "---PLANE ELECTRODE
~R=]OOmm ~ ~ A S U R I N G
ELECTRODE
Fig.1 Electrode system for studying back discharge.
37
20 --
AIR L O A D - - /
/
+
/i/J
10 7
~ 5
GAP = 60 (ram)
/ 2 ~
B3
B
/'
Bl~- ~K~7 A3
~ 3 ,////'AV AI •' p - -/
,/)~
1: Eds = 15.7 (kV/cm)
/~
2: Eds = 25.1 (kV/cm) 3: Eds : 39.0 (kV/cm)
1
,
m~
,
i
i0
. . . .
,
. . . .
t
15 20 V(KV)
~
I
30
Fig.2. Voltage-current curves under back discharge condition for different values of breakdown field strength. (A pair of glass plates, each having a single pinhole; sample resistivity P d = 6 × 10" ohm cm.)
back discharge glow at its initiation. Current pulse was observed at the same time by a cathode ray oscilloscope with a band width of 10 MHz. The breakdown field strength of the layer in corona field at the initiation of back discharge, Eds, was estimated as follows [3]. Voltage--currentdensity (V--J) curves with the layer for various values of Eds were measured, where J represents the average current density at the measuring electrode. They are shown by the solid curves labelled 1, 2 and 3 in Fig.2 corresponding to an electrode gap of 60 mm. In the same figure, the air load V--J curve (without glass plates) is denoted by a d o t t e d line, the plane electrode being raised to the position of the surface when the glass plates were present, i.e. a gap of 56 mm. The voltage across the glass plates, A V, is given by A V = V- V', where V and V' are the electrode voltages corresponding to the same value of current density, J, with and without the glass plates, respectively. With the increase in applied voltage from zero, a Trichel pulse current appeared at the measuring electrode when corona started at the needle electrode. With the voltage further increased, the repetition frequency of the Trichel pulse current increased and a d.c. current c o m p o n e n t appeared as shown in Fig. 3(a). This d.c. current c o m p o n e n t also increased when the voltage was increased. When the points Ai (i = 1--3) (Fig. 2) were reached, a feeble b u t continuous glow appeared at the pinhole (Fig. 3(b)(2)), leading to a slight nonlinear increase in current. Breakdown of the layer at the pinhole occurred at this point. The current waveform at this point is shown in Fig.3(b)(1). We t o o k this point as the initiation point of back discharge. From the value of A V at this point, (4 V)0, and the layer thickness, t, we obtain E'ds = (A V)o /t. The values of E'ds thus obtained at the points A1, A2 and A3 in Fig. 2 '
38
x 10-8 A/cm 2
(a) Trichel pulse
( 12 kV, 2.0 x 10-8 A/cm2 )
2 x 10-8
(1) Current (b) 0nset-glow
( 16 kV, 4.0 x 10-8 A/cm2 )
(2) Photograph obtained with intensifier
x 10-8 A/cm2
(2) photograph obtained
(1) Current (c) Onset-streamer
( 18 kV, I . I
x 10-7 A/cm 2 )
with intensifier
39
1.5 x 10-7 A/cm2
(I) Current (d) Streamer
(2) Photograph obtained
( 21 kV, 4.0 x 10-7 A/cm2 )
with i n t e n s i f i e r
x 10-8 A/cm2
(e) Repetitive breakdown (Mica plate)
( I0 kV, 4.0 x 10-8 A/cm2 )
Fig.3. Current waveforms and modes of back discharge.
were compared with those of Pd X J0, where Pd is the layer resistivity and J0 the current density at the corresponding points. A good agreement was noted between E'ds and Pd X Jo as indicated in Table 1. The values of Eds measured using parallel plane electrodes are also given. The values of E'ds agreed well with these values. This suggests that the breakdown of the layer occurred at a layer field strength nearly equal to Eds measured by a parallel plane electrode system. The continuous glow mode of back discharge at its initial stage should be considered as a kind of glow discharge. Hence, this should be refered to as the "onset-glow-mode". It should be distinguished from the more intense "steady-glow-mode" appearing under the conditions to be reported separately [2]. With further increase in voltage, very small surface streamers randomly appeared at the limited region around the upper edge of the pinhole (Fig.3 (c)(2)) at points Bi (i = 1--3) in Fig.2, corresponding to J = 0.5--1.0X 10 -7 A/cm 2. (The expression of current density in A/cm 2 lost its sense here, since
40 TABLE 1 Comparison of Eds , E'ds and Pd X Jo Curve
Initiation point
(A V)o(kV) 5.6
J0 (A/cm~) E'ds = (4 V)o/t (kV/cm)
Pd × Jo
Eds (kV/cm)
2 . 5 × 1 0 -8
15.0
15.7
(kV/cm)
1
A1
2
A2
9.2
4.1 X 1 0 -8
23.0
24.6
25.1
3
As
14.0
6 . 9 × 1 0 -8
35.0
41.4
39.0
14.0
most of the current passed through the pinhole hereafter.) Figure 3(c)(1) shows a current pulse of this streamer, which should be refered to as an "onset-streamer". The Trichel pulse current was still observed to exist. When the voltage was raised slightly above points Bi, space streamers and large surface streamers occurred from the pinhole at the points Ci (i = 1--3) in Fig. 2 (see Fig.3(d)(2)), accompanied by large current pulses (Fig.3(d)(1)). A more intense rise in current occurred b e y o n d these streamer starting voltages. It should therefore be noted that the criterion for the occurrence of the layer breakdown should be clearly distinguished from that for the occurrence of streamers which are the essential cause for rapid current increase. In the field measurements of the V--J curves, only the initiation points of streamers could be detected because of the much higher signal-tonoise ratio expected. For a glass plate with a pinhole as described above, the onset-glow appeared at the layer breakdown voltage. However, when a sufficiently high resistivity layer, such as a mica plate with a pinhole (Pd ~ 101s ohm cm), was used, a random breakdown occurred at first. With a slight increase in voltage, it was followed by a repetitive breakdown, as indicated in Fig.3(e). This was then followed by a stable onset-glow. Hence, the layer breakdown voltage, Vb, was different from the starting voltage of the onset-glow, Vo, in this case. The streamer discharge in the gas space is followed by a flashover occurring at a voltage much lower than that without back discharge. Thus, there are four major critical voltages for back discharge under atmospheric conditions; layer breakdown voltage Vb, onset-glow starting voltage Vo, streamer start~ ing voltage Vst and, finally, flashover voltage V8. The random breakdown, onset-glow, and onset-streamer, constitute an initial stage of back discharge where the current rise still remains comparatively low. This stage should be referred to as the "onset-stage". 3. Back discharge in streamer mode
With the increase in voltage beyond the point Ci in Fig.2, streamers are emitted either into the gas space towards the discharge electrode or along the surface of the layer, or in both directions. Hence, back discharge in this
41
mode should, be referred to as the "streamer m o d e " , or, more specifically, the space streamer mode, surface streamer mode and mixed streamer mode as a combination of the former two. When the voltage was raised further, the space streamer proceeded towards the discharge electrode and it bridged across the electrode gap until it finally turned into flashover. It could be expected that the most essential factors deciding the respective mode of streamers would be the strengths of the tangential and the vertical field around the breakdown point of the layer, as well as the corona current. Thus, these effects were studied separately. The detailed mechanism of propagation for these streamers will be discussed in another paper [4].
3.1 Effects of vertical field and current Along with the study of the effect of the vertical field, Ea, that of the corona current density, J, was also investigated. These two factors, E a and J, are closely coupled to each other in an actual precipitator, while their magnitudes at back discharge initiation differ greatly from one case to another, depending upon the dust layer resistivity as described in Section 3.3. To investigate the effects of these two factors separately, a grid electrode was inserted between the needle-and-plane electrodes as shown in Fig.4. A transient fluctuation in the grid electrode potential was eliminated by using a condenser of 0.5 pF capacity connected in parallel to its H.V. source. By the change in needle electrode voltage Va and grid electrode potential Vg, the vertical field strength E a and current density J could be varied independently. The value of E a w a s calculated from the ratio Vg / (grid-to-plane spacing). In these experiments, two glass plates each having a pinhole (0.5 mm in diameter) were used as before. The resistivity was 6 × 10 ~ ohm cm and the breakdown field strength was 20.7 kV/cm. Figure 5 shows curves of the current density J plotted against the voltage of the needle electrode Va with the grid potential Vg kept constant. The mode and current waveform of back discharge Ya
~NEEDLEELECTRODE / GRIDELECTRODE --/. . . . . .
_oVg
I Ea 1 PN I HOLE-~ t/-GLASSPLATES I F-R = I°°
mm~
PNE /
ELECTROD LE
T
I
Fig.4. Electrode system for studying the e f f e c t s o f vertical field and c u r r e n t .
42
/•,-Vg : 16 kV
10-6
/!
~,
~ 10-7 10 -8
, 10
,
i
15 Va
I
(kV) 20
30
I
I
40
Fig.5. Voltages-current-density curves for different values of grid potential Yg (of. Fig.4). under atmospheric conditions were observed with the aid of an image intensifier tube and a cathode ray oscilloscope. From these observations, and the curves in Fig. 5, the mode diagram of back discharge was depicted on an Ea-J domain as shown in Fig.6(a). No back discharge occurred in region I because of the low current density. When the current exceeded a certain value at which the layer breakdown condition described before was fulfilled, back discharge in the onset-glow mode occurred in Region II (see Fig.3(b)(2)). The further increase in current resulted in the onset-streamers occurring around the edge of the pinhole in Region III (see Fig.3(c)(2)). It should be noted that the critical current densities for the transitions between Regions I, II and III were nearly constant respectively independent of Ea, as is shown by the flat curves A and B. The two Regions II and III should be referred to as the "onset-stage" region. With further increase in current beyond the other critical curves C and D, back discharge in the streamer modes (surface and mixed streamer modes) t o o k place in Regions IV and V. Region IV, for a lower value of Ea, is the surface streamer region where the surface streamer mode was predominant and space streamers were few (see Fig.6(b)). In this region, current density J saturated at Curve E because of space-charge limitation (see Fig.5), and no flashover could be obtained between the grid and the plane electrodes. Whereas, in Region V, when Ea exceeded 5 kV/cm, both the surface and space streamers occurred to form the mixed streamer mode (see Fig.7(b)). Again the critical current density for the transition from Region III to Regions IV and V was nearly constant, except for a corner area G. When J in Region V exceeded curve F, the streamer turned into a flashover The critical value of the field strength between Regions IV and V (Curve H)
43
160Ia5-
FLASHOVERs~/I" /J~l
,, NOFLASHOVER\ X%'
-~/
V
I
REGION
I
l" x STREAMER~. REGION ~
!
10- ~'"~
C B
) I
III ONSET-STREAMERREGION I x--X--x~x~X~x..... ~j II ONSET-GLOWREGION"" I I NOBACKDS I CHARGE
10-8
I
I
I
2
I
4 Eo
I
I
I
6
I I
8
I
I
I
i0
( kV/cm )
Fig.6. E f f e c t o f tangential field and current density on mode of back discharge. (a) Mode
diagram in field-current domain. (b) Photograph of back discharge in surface streamer Region IV, ( E a = 4.0 k V / c m , J = 5.0 × 10 -7 A/cm2).
was about 5 kV/cm under atmospheric conditions, which had been taken as a criterion for the occurrence of streamer under these conditions. It should, however, be noted t h a t the initiation and growth of space and surface streamers are also governed by current density J.
3. 2 Effect of tangential field In the present case where the surface resistivity of the layer is extremely high, the surface charge would be firmly bound to its original position. In this case, the tangential field around the breakdown point will become a function of the surface-charge density on the layer, o0, at the instant of breakdown at which the potential of the breakdown point becomes almost zero. The value of a0 in turn is given by eEds where e is the dielectric constant of the layer. If a0 is sufficiently high, the breakdown of the point will directly trigger the surface streamer. In the opposite case, onset-glow appears prior to the occurrence of surface streamer, as long as the vertical field strength in the gas space is not sufficiently high for the space streamer to be triggered. Such a high vertical field strength does n o t normally exist at the initiation of back discharge, unless the layer resistivity is in the range of 5 X 10 '0 - - 1 0 " ohm cm as discussed later. Thus, the effect of o0 on back discharge in streamer mode was studied. Two glass plates were used, as before, so t h a t Eds and, hence, a0 could be changed. Photographs of the back discharge for two values of breakdown strength are shown in Fig.7. When
44
e
(a)
Eds
LOW
(b)
Eds
HIGH
Eds = 13.8 (kV/cm)
Eds = 33.8 (kV/cm)
V
: 30
(kV)
V
: 40
(kV)
I
= 29
(~A)
I
= 23
(.A)
Fig.7. Effect of tangential field on back discharge in the m i x e d streamer mode. (A pair of glass plates, each having a single pinhole; Electrode gap = 50 ram).
Eds was 13.8 k V / c m , a space s t r e a m e r was d o m i n a n t p r o c e e d i n g to the discharge e l e c t r o d e (Fig.7(a)). When Eds was 33.8 k V / c m , the m i x e d s t r e a m e r m o d e appeared, where a r e m a r k a b l e surface s t r e a m e r in the vicinity o f the pinhole c o u l d be observed (Fig.7(b)). This was because the tangential field strength b e c a m e larger as o0 increased. The surface discharge b e c a m e especially d o m i n a n t w h e n the Value o f a0 e x c e e d e d a b o u t 5 X 10 -9 C / c m 2. 3. 3 Effect o f dust resistivity A tissue p a p e r was used as a sample in this e x p e r i m e n t . This was because its a p p a r e n t resistivity Pd c o u l d easily be changed f r o m 108 t o 1014 o h m c m b y adjusting t h e a m b i e n t h u m i d i t y . Thus, the e f f e c t o f Pd on t h e b a c k discharge m o d e at r o o m t e m p e r a t u r e was studied. V o l t a g e - - c u r r e n t - d e n s i t y curves f o r d i f f e r e n t values o f Pd, ranging f r o m 6 X 109 to 2 X 10 is o h m cm, are s h o w n in Fig.8 where the e l e c t r o d e gap was k e p t at 60 mm. P h o t o g r a p h s o f the back discharge for t h r e e d i f f e r e n t values o f Pd are s h o w n in Fig.9, where the values o f J were o f the same order. When the resistivity was 6 X 109 o h m cm {Curve 1 in Fig.8), n o b a c k discharge o c c u r r e d until flashover t o o k place at V = 65 kV and J = 7.6 p A / c m 2 .
45
FLASHOVER
10-1
/~. /
/
/,.~"
~
vs6s =
/71/
o.1
/ / c// /i/ //,~
gl.
7
. A
.... >
2: Pd = 0.9 x lO:~oh . . . . )
/Zd
0,01
T,,
3: Pd- 1.6 x lOlZ(oh....) ,
A.
.
,
10 V
(kv)
J =7.6 (pA/cm2)
,
,
I
15
20
,
I
30
,
I
40
(KV)
Fig.8. Effect of dust resistivity Pd on voltage-current-density curves under back discharge condition when negative corona is used. (Tissue paper, 1.0 m m in thickness.)
For the case of a needle-to-plane electrode system and the experimental conditions investigated, the initiation condition of back discharge, Eds = Pd X J0, was n o t satisfied prior t o the occurrence of flashover when Pd was lower than a b o u t 5 X 10 '° ohm cm. In o t h e r words, the initiation voltage o f back discharge was higher than the flashover voltage o f the gap because of t o o low a value of Pd. When the value of Pd slightly exceeded this critical value, space streamers occurred as soon as the layer broke down, owing to the large voltage drop across the gas space. For instance, when the resistivity was 0.9 X 1 0 " o h m cm (Curve 2), the streamer starting voltage Vst was a b o u t 27 kV. The n u m b e r of b r e a kdow n points was less, and streamers proceeded into space towards the discharge electrode, as shown in Fig.9(a). The occurrence of space streamers lowered the flashover voltage 118 t o a great extent. It was observed that, when Pd was between a b o u t 5 X 10 l° and 0.9 X 10 H ohm cm, excessive sparking t ended to occur. In this range o f Pd, Vst would be lowered with the increase in Pd, so t h a t it finally becomes lower than Vs as in the case o f Curve 2 in Fig. 8. A slight increase in voltage b e y o n d Vst would cause flashover because Vst remained still close to Vs. For Pd higher than 10 '2 ohm cm (Curves 3 and 4), the back discharge streamers started to occur at a m uch lower voltage and c ur r e nt density. There was a larger interval between Vst and ITs so th at the excessive sparking disappeared. There were m ore breakdown points with a general glow surrounding each point. In this case, a surface glow d o m i n a t e d and space streamers were few. This t e n d e n c y became pron o u nced with the increase in Pd (Fig.9(c)).
46
(a) Pd = 0.9 x IOII (ohm-cm) J = 3.2 ~ / c m 2)
(b) Pd = 1.6 x 1012 (ohm-cm) J = 5.5
OxA/cm2)
g
(c) Pd = 2.0 x 1013 (ohm-era)
J = 2.2 Fig.9. Effect of dust resistivity on back discharge mode. (Tissue paper).
(,wA/cm2)
47
The different discharge modes were caused by the difference in the ratio of the voltage drop across the dust layer to that across the gas space. If the resistivity was high, the voltage drop across the dust layer was high even at a low current density on the initial stage of back discharge, whereas the vol, tage drop across the gas space was low. As a result, the development of a space streamer was suppressed, and a surface discharge occurred. In this case, many weak points broke down and the current increased readily w i t h o u t excessive sparking. However, when voltage was raised, the space streamer occurred also in this case, taking the form of a general glow bridging across the gap. A more severe increase in current at this later stage. It can be seen that flashover occurred almost at the same voltage in spite of a large difference in Pd, once back discharge occurred. This agrees well with the results of Penney and Graig [ 5], i.e. the flashover voltage of back discharge was n o t affected by the value of resistivity. This flashover voltage was almost half the value of that under a non-back-discharge condition.
3. 4 Charging efficiency in different regions For negative corona, back discharge is a source of positive ions which produce a bipolar ion atmosphere in the gas space. The effect on particle charging, however, is different, depending on the mode of back discharge. In the surface streamer mode, the ion source is considered to be surface-like, but
/
D
i oVa i
Omm - - 3 0 mm~
I I
F
D:
DISCHARGEELECTRODE (SAW TOOTH ELECTRODE)
G:
GRID ELECTRODE
P:
PLANE ELECTRODE
M:
MEASURINGELECTRODE
T:
MICA PLATE WITH PINHOLES
B:
STEEL BALL WITH 3.0 mm IN DIAMETER
F:
FARADAYCAGE
Fig.lO. Electrode system for measuring particle charging.
48 in the space or mixed streamer modes, ion generation in space may occur. This was confirmed by measuring particle charge using the electrode system as shown in Fig.10 [6]. This system enabled the change in back discharge mode by changing the field strength between the grid and plane electrodes, Ea (see Fig.6). A steel ball, 3.0 mm in diameter, was dropped between the plane and grid electrodes and its saturation charge was measured by a Faraday cage. The distance between the plane and grid electrodes was 50 mm, grid to discharge electrodes 30 mm, and a mica plate having many pinholes was used as a layer. Figure 11 is an example of the results obtained, showing the saturation charge of a steel ball as a function of its position d from the plane electrode. The values indicated in the bracket represent its ratio to the theoretical saturation charge due to monopolar ions, calculated from Pauthenier's equation [7]. In the surface streamer region (Curve 1), the value of charge was about 90% less than the theoretical value but the sign of particle charge remained the same as that of the discharge electrode. The value of charge decreased as the particle crossed nearer the plane electrode. This result indicates that the back discharge of this mode can be considered as a surface-like ion source so that the density of positive ions decreased into the space. In the mixed streamer region, where space streamer was pronounced {Curve 2), particle charge scattered largely around its average value, which was a fairly high positive value and almost the same regardless of position. This (2) MIXED STREAMERREGION Ea : 6.0 (kV/cm) J = 2.0 (,~A/cm2)
+I0
+8 +6
~+4
I+2
(3) TRANSITION REGION
Ea = 4.6 (kV/cm) J 1.7 (~A/cm ~)
'o x
0
o
12
Q -4
_.~/5.4 %)
~-- . . . .
T
( I ) SURFACE STREAMERREGION] 9.q ~) Ea - 3.6 (kV/cm) J = 1.6 (#A/cm 2)
i
I
I
I
i
i0
20
30
40
50
d (mm)
Fig.11. Saturation charge versus position d for different back discharge modes.
49
result might indicate that positive ions were generated abundantly inside the whole space and played a d o m i n a n t role in particle charging. The effect of the streamer tip colliding with a particle might also be a factor. Curve 3 represents the transition region between the foregoing two. In this case, particle charges were also positive but as low as in the case of Curve 1. 3. 5 Back discharge with positive corona It was observed t h a t the mode of back discharge with positive corona at the needle was completely different, as shown in Fig.12. Tissue paper was used and the electrode gap was 60 ram. Voltage--current-density curves are shown in Fig.13 for various values of Pd. In this case, breakdown points were distributed uniformly on the surface, no space or surface streamers could be observed and the discharge mode was only glow-mode independent on resistivity. The abnormal increase in current was small and the flashover voltage, when back discharge occurred, was approximately 1.5 times higher than that for back discharge with negative corona at the needle. The relationship between the flashover voltage of back discharge, Vs, and gap tlistance d is shown in Fig.14 for the positive and negative coronas. Vs of the positive corona was higher than that of the negative corona for a gap distance range of 1.0--10.0 cm. The flashover voltage of the positive corona under back discharge condition was also higher than that without back discharge when the layer was removed (air load). The mechanism for this behaviour is considered to be due to a stable nature of negative glow corona at the breakdown point, and to
r i g . 1 z ~ a c k discharge under positive corona point. (Tissue paper, V ffi + 4 0 k V , J = 2.8 X 10 -4 A / c m 2, P d = 1013 o h m c m . )
50
i~[
°~
GAP: 60(~)
4/,/~
l: A,~ LoAo
/'7
IOII (oh.... ) 3: /:~1= 1012 (oh.... )
2: ,°d=
I"
F
j/
3 2
~i~
/0/~ i
/,4/1
....
v
o, ol ¼
~
7
i0 v
15
20
30
q0 50
(+kV)
Fig.13. Voltage--current-density curves under back discharge condition w h e n positive corona is used. (Tissue paper, 1.0 m m in thickness)
60 50 40 v
3O
L/
d :=" 20
0
/
I
i
I
i
I
I
I
50 ELECTRODE GAP d (mm)
i
I
I
100
Fig.14 Flashover voltage versus gap distance under back discharge condition with positive and negative coronas. (Tissue paper, Pd = 1 . 2 X 1 0 ~1 ohm cm, 1.0 mm in thickness.) the positive c o r o n a at t h e needle tip b e i n g c o n v e r t e d t o H e r m s t e i n ' s g l o w c o r o n a [8]. The l a t t e r m a y result f r o m a c o p i o u s s u p p l y o f negative ions f e d to t h e needle e l e c t r o d e , f r o m which e l e c t r o n s w o u l d be shed to f o r m a c o n t i n u o u s a n d stable positive glow discharge at t h e n e e d l e tip.
51 4. Conclusions From the experimental studies described above, using the model samples of tissue papers, glass and mica plate, the following results concerning the effects of apparent resistivity and breakdown field strength on back discharge were obtained. (1) With the negative corona at the needle, the layer breakdown started to occur when Eds = Pd X J is fulfilled. It was followed by the onset-glow mode occurring with a slight increase in voltage. A rapid increase in current, however, occurred only when the streamers started to occur at a critical voltage Vst. Thus, there are four major critical voltages for back discharge under atmospheric conditions; layer breakdown voltage Vb, onset-glow starting voltage Vo, streamer starting voltage Vst and flashover voltage Vs. For the case of the electrode system investigated, the initiation condition of back discharge may n o t be fulfilled prior to the occurrence of flashover when Pd does not exceed a b o u t 5 X 10 '° ohm cm. When Pd is in the range of 5 X 10 ~° - - 1 0 " ohm cm, Vst <- Vs, so that excessive sparking occurs. When Pd > 10': ohm cm, Vst becomes sufficiently lower than Vs so that excessive sparking disappears, b u t an abnormal increase in current occurs. (2) There are three sub-modes in the streamer mode, depending u p o n the field distribution around the breakdown point in the sample layers: space streamer mode, surface streamer mode and mixed streamer mode. This, in turn, is a function of Pd, Eds, Ea and J. In the space streamer mode, positive ion generation in space occurs and particles aquire a fairly high positive charge. Whereas in the surface streamer mode, positive ion generation is limited to the surface region and the sign of particle charge is the same that of the needle electrode. In most of the actual cases, however, the mixed streamer m o d e appears. (3) With positive corona at the needle, the back discharge mode is completely different. The flashover voltage is higher than that under the back discharge condition with the negative corona.
Acknowledgements This research was sponsored by the Ministry of Education, Japan, as its Special Research Project (I) (Project No.011914). The authors are gratefully indebted for its support. Thanks are also due to Mr Masao Kuroda for his help in some of the experiments. Notation
Ea
vertical field strength in the gap
Eds breakdown field strength of the layer measured by parallel plane electrodes
E'ds breakdown field strength of the layer in corona field
52
J J0
current density at the measuring electrode current density at the initiation of back discharge Va discharge electrode voltage Vb layer breakdown voltage Vg grid electrode potential Vo onset-glow starting voltage Vs flashover voltage Vst streamer starting voltage e dielectric constant of the layer o0 surface charge density at the instant of breakdown Pd apparent dust resistivity References i 2 3 4 5 6 7 8
S.Masuda, Recent progress in electrostatic precipitation, Static Electrification 1975, Institute of Physics Conference Series, No. 27 (1975) 154. S. Masuda and A. Mizuno, Flashover measurements of back discharge, J. Electrostatics, to be published. S. Masuda, Reverse ionization phenomena in electrostatic precipitator, J. Inst. Electr. Eng. Jpn., 35(102) (1960) 1482. S. Masuda and A. Mizuno, Light measurement of back discharge, J. Electrostatics, 2 (1977) 375. G.W. Penny and S.E. Craig, Sparkover as influenced by surface conditions in d.c. corona, Am. Inst. Electr. Eng., Trans. 79 (1960) 112--118. M. Kuroda, Graduation Thesis, Univ. of Tokyo, 1975. M. Pauthenier, Moreau-Hanot, Rev. Gen. Electr. XIV (18) (1932) 583. L.B. Loeb, Electrical Coronas, Univ. of Calif. Press, 1965, p. 95.
35
INITIATION CONDITION AND MODE OF BACK DISCHARGE
SENICHI MASUDA and AKIRA MIZUNO
Department of Electrical Engineering, University of Tokyo, 7--3--1, Hongo, Bunkyo-Ku, Tokyo (Japan) (Received November 8, 1976; in revised form January 21, 1977)
Summary Modes of back discharge occurring in the electrostatic precipitator were studied using, instead of a dust layer, the model samples of glass and mica plates with a pinhole, and tissue papers, It was confirmed that back discharge started to occur when the apparent field strength in the sample layers exceeded its breakdown field strength. Back discharge became a streamer corona under atmospheric conditions. It could be classified into a space streamer mode, a surface streamer mode and a mixed streamer mode, depending upon the field distribution around the breakdown point in the sample layers. The first and the third modes occurred when the field strength in the air gap, Ea, exceeded about 5 kV/cm, and positive ions were generated in the whole gas space. The second mode appeared when E a was lower than about 5 kV/cm, and ion generation was limited to the near surface region. Among the factors affecting the back discharge, dust resistivity was the most important. For low dust resistivity, space streamers tended to develop from the breakdown points when the applied voltage was raised. For high dust resistivity, on the other hand, the number of breakdown points increased, and surface discharge was pronounced. A remarkable difference in modes was observed when using positive corona. Neither space streamer nor surface discharge occurred and the flashover voltage was higher than that with negative corona.
1. I n t r o d u c t i o n Back discharge is one o f t h e m o s t difficult p r o b l e m s in electrostatic prec i p i t a t o r s impairing their p e r f o r m a n c e in m a n y industrial plants [1]. This is an a b n o r m a l kind o f discharge w h i c h is triggered b y b r e a k d o w n in a high resistivity d u s t layer d e p o s i t e d o n the collecting e l e c t r o d e and w h i c h lowers the flashover voltage, r e d u c i n g particle charge and causing a severe d r o p in c o l l e c t i o n efficiency. The n a t u r e o f b a c k discharge d e p e n d s o n m a n y f a c t o r s such as the electrical p r o p e r t i e s o f the d u s t layer and the c h e m i c a l p r o p e r t i e s o f the particles themselves, and its f o r m is v e r y c o m p l i c a t e d . T h e r e f o r e , m o r e intensive and basic investigations are required t o solve the b a c k discharge p r o b l e m , a n d also t o assess p r e c i p i t a t o r p e r f o r m a n c e , w h e n b a c k discharge occurs. Back discharge o c c u r r i n g w h e n using negative c o r o n a can be classified into t w o m a j o r discharge m o d e s . One is t h e s t r e a m e r m o d e , o c c u r r i n g w i t h high
36 gas density; the other, the glow corona mode, occurring with low gas density. Normally, streamers, formed at the breakdown point on the layer, proceed into the gas space towards the discharge electrode or to the accumulated charges on the dust surface, or in both directions, depending upon the field distribution around the starting point. It is appropriate to classify the streamer mode into three sub-modes: space streamer mode, surface streamer mode and mixed streamer mode. The third one appears in most of the practical cases. In this paper, an experimental study of the back discharge of the streamer mode carried out under atmospheric pressure and room temperature is reported. Studies of the back discharge of a glow corona mode will be reported separately [ 2]. 2. Initiation condition and initial mode of back discharge At first, the initiation condition and the initial mode of back discharge was studied using a needle-to-plane electrode system, located inside a thermostat where humidity could also be controlled. It was tested whether the initiation of back discharge was governed by the breakdown field strength of a layer, Eds , measured separately using parallel plane electrodes. In order to change the value of Eds of a sample, two glass plates, each having a pinhole, were used on top of one another as the layer sample, as indicated in Fig. 1. By altering the position of one plate and thus changing the distance between the holes, the value of -Eds could be changed. The resistivity of the glass plate, Pd, was 6 X 10 ~ ohm cm and the diameter of the pinhole was 0.5 mm. The thickness of one plate was 2.0 mm. An image intensifier tube (EMI, type 9912) was used at its maximum gain (about 106) to observe a o
D,C.H,V, NEEDLE ELECTRODE
T E
TEST SAMPLE (DOUBLE PLATES) "---PLANE ELECTRODE
~R=]OOmm ~ ~ A S U R I N G
ELECTRODE
Fig.1 Electrode system for studying back discharge.
37
20 --
AIR L O A D - - /
/
+
/i/J
10 7
~ 5
GAP = 60 (ram)
/ 2 ~
B3
B
/'
Bl~- ~K~7 A3
~ 3 ,////'AV AI •' p - -/
,/)~
1: Eds = 15.7 (kV/cm)
/~
2: Eds = 25.1 (kV/cm) 3: Eds : 39.0 (kV/cm)
1
,
m~
,
i
i0
. . . .
,
. . . .
t
15 20 V(KV)
~
I
30
Fig.2. Voltage-current curves under back discharge condition for different values of breakdown field strength. (A pair of glass plates, each having a single pinhole; sample resistivity P d = 6 × 10" ohm cm.)
back discharge glow at its initiation. Current pulse was observed at the same time by a cathode ray oscilloscope with a band width of 10 MHz. The breakdown field strength of the layer in corona field at the initiation of back discharge, Eds, was estimated as follows [3]. Voltage--currentdensity (V--J) curves with the layer for various values of Eds were measured, where J represents the average current density at the measuring electrode. They are shown by the solid curves labelled 1, 2 and 3 in Fig.2 corresponding to an electrode gap of 60 mm. In the same figure, the air load V--J curve (without glass plates) is denoted by a d o t t e d line, the plane electrode being raised to the position of the surface when the glass plates were present, i.e. a gap of 56 mm. The voltage across the glass plates, A V, is given by A V = V- V', where V and V' are the electrode voltages corresponding to the same value of current density, J, with and without the glass plates, respectively. With the increase in applied voltage from zero, a Trichel pulse current appeared at the measuring electrode when corona started at the needle electrode. With the voltage further increased, the repetition frequency of the Trichel pulse current increased and a d.c. current c o m p o n e n t appeared as shown in Fig. 3(a). This d.c. current c o m p o n e n t also increased when the voltage was increased. When the points Ai (i = 1--3) (Fig. 2) were reached, a feeble b u t continuous glow appeared at the pinhole (Fig. 3(b)(2)), leading to a slight nonlinear increase in current. Breakdown of the layer at the pinhole occurred at this point. The current waveform at this point is shown in Fig.3(b)(1). We t o o k this point as the initiation point of back discharge. From the value of A V at this point, (4 V)0, and the layer thickness, t, we obtain E'ds = (A V)o /t. The values of E'ds thus obtained at the points A1, A2 and A3 in Fig. 2 '
38
x 10-8 A/cm 2
(a) Trichel pulse
( 12 kV, 2.0 x 10-8 A/cm2 )
2 x 10-8
(1) Current (b) 0nset-glow
( 16 kV, 4.0 x 10-8 A/cm2 )
(2) Photograph obtained with intensifier
x 10-8 A/cm2
(2) photograph obtained
(1) Current (c) Onset-streamer
( 18 kV, I . I
x 10-7 A/cm 2 )
with intensifier
39
1.5 x 10-7 A/cm2
(I) Current (d) Streamer
(2) Photograph obtained
( 21 kV, 4.0 x 10-7 A/cm2 )
with i n t e n s i f i e r
x 10-8 A/cm2
(e) Repetitive breakdown (Mica plate)
( I0 kV, 4.0 x 10-8 A/cm2 )
Fig.3. Current waveforms and modes of back discharge.
were compared with those of Pd X J0, where Pd is the layer resistivity and J0 the current density at the corresponding points. A good agreement was noted between E'ds and Pd X Jo as indicated in Table 1. The values of Eds measured using parallel plane electrodes are also given. The values of E'ds agreed well with these values. This suggests that the breakdown of the layer occurred at a layer field strength nearly equal to Eds measured by a parallel plane electrode system. The continuous glow mode of back discharge at its initial stage should be considered as a kind of glow discharge. Hence, this should be refered to as the "onset-glow-mode". It should be distinguished from the more intense "steady-glow-mode" appearing under the conditions to be reported separately [2]. With further increase in voltage, very small surface streamers randomly appeared at the limited region around the upper edge of the pinhole (Fig.3 (c)(2)) at points Bi (i = 1--3) in Fig.2, corresponding to J = 0.5--1.0X 10 -7 A/cm 2. (The expression of current density in A/cm 2 lost its sense here, since
40 TABLE 1 Comparison of Eds , E'ds and Pd X Jo Curve
Initiation point
(A V)o(kV) 5.6
J0 (A/cm~) E'ds = (4 V)o/t (kV/cm)
Pd × Jo
Eds (kV/cm)
2 . 5 × 1 0 -8
15.0
15.7
(kV/cm)
1
A1
2
A2
9.2
4.1 X 1 0 -8
23.0
24.6
25.1
3
As
14.0
6 . 9 × 1 0 -8
35.0
41.4
39.0
14.0
most of the current passed through the pinhole hereafter.) Figure 3(c)(1) shows a current pulse of this streamer, which should be refered to as an "onset-streamer". The Trichel pulse current was still observed to exist. When the voltage was raised slightly above points Bi, space streamers and large surface streamers occurred from the pinhole at the points Ci (i = 1--3) in Fig. 2 (see Fig.3(d)(2)), accompanied by large current pulses (Fig.3(d)(1)). A more intense rise in current occurred b e y o n d these streamer starting voltages. It should therefore be noted that the criterion for the occurrence of the layer breakdown should be clearly distinguished from that for the occurrence of streamers which are the essential cause for rapid current increase. In the field measurements of the V--J curves, only the initiation points of streamers could be detected because of the much higher signal-tonoise ratio expected. For a glass plate with a pinhole as described above, the onset-glow appeared at the layer breakdown voltage. However, when a sufficiently high resistivity layer, such as a mica plate with a pinhole (Pd ~ 101s ohm cm), was used, a random breakdown occurred at first. With a slight increase in voltage, it was followed by a repetitive breakdown, as indicated in Fig.3(e). This was then followed by a stable onset-glow. Hence, the layer breakdown voltage, Vb, was different from the starting voltage of the onset-glow, Vo, in this case. The streamer discharge in the gas space is followed by a flashover occurring at a voltage much lower than that without back discharge. Thus, there are four major critical voltages for back discharge under atmospheric conditions; layer breakdown voltage Vb, onset-glow starting voltage Vo, streamer start~ ing voltage Vst and, finally, flashover voltage V8. The random breakdown, onset-glow, and onset-streamer, constitute an initial stage of back discharge where the current rise still remains comparatively low. This stage should be referred to as the "onset-stage". 3. Back discharge in streamer mode
With the increase in voltage beyond the point Ci in Fig.2, streamers are emitted either into the gas space towards the discharge electrode or along the surface of the layer, or in both directions. Hence, back discharge in this
41
mode should, be referred to as the "streamer m o d e " , or, more specifically, the space streamer mode, surface streamer mode and mixed streamer mode as a combination of the former two. When the voltage was raised further, the space streamer proceeded towards the discharge electrode and it bridged across the electrode gap until it finally turned into flashover. It could be expected that the most essential factors deciding the respective mode of streamers would be the strengths of the tangential and the vertical field around the breakdown point of the layer, as well as the corona current. Thus, these effects were studied separately. The detailed mechanism of propagation for these streamers will be discussed in another paper [4].
3.1 Effects of vertical field and current Along with the study of the effect of the vertical field, Ea, that of the corona current density, J, was also investigated. These two factors, E a and J, are closely coupled to each other in an actual precipitator, while their magnitudes at back discharge initiation differ greatly from one case to another, depending upon the dust layer resistivity as described in Section 3.3. To investigate the effects of these two factors separately, a grid electrode was inserted between the needle-and-plane electrodes as shown in Fig.4. A transient fluctuation in the grid electrode potential was eliminated by using a condenser of 0.5 pF capacity connected in parallel to its H.V. source. By the change in needle electrode voltage Va and grid electrode potential Vg, the vertical field strength E a and current density J could be varied independently. The value of E a w a s calculated from the ratio Vg / (grid-to-plane spacing). In these experiments, two glass plates each having a pinhole (0.5 mm in diameter) were used as before. The resistivity was 6 × 10 ~ ohm cm and the breakdown field strength was 20.7 kV/cm. Figure 5 shows curves of the current density J plotted against the voltage of the needle electrode Va with the grid potential Vg kept constant. The mode and current waveform of back discharge Ya
~NEEDLEELECTRODE / GRIDELECTRODE --/. . . . . .
_oVg
I Ea 1 PN I HOLE-~ t/-GLASSPLATES I F-R = I°°
mm~
PNE /
ELECTROD LE
T
I
Fig.4. Electrode system for studying the e f f e c t s o f vertical field and c u r r e n t .
42
/•,-Vg : 16 kV
10-6
/!
~,
~ 10-7 10 -8
, 10
,
i
15 Va
I
(kV) 20
30
I
I
40
Fig.5. Voltages-current-density curves for different values of grid potential Yg (of. Fig.4). under atmospheric conditions were observed with the aid of an image intensifier tube and a cathode ray oscilloscope. From these observations, and the curves in Fig. 5, the mode diagram of back discharge was depicted on an Ea-J domain as shown in Fig.6(a). No back discharge occurred in region I because of the low current density. When the current exceeded a certain value at which the layer breakdown condition described before was fulfilled, back discharge in the onset-glow mode occurred in Region II (see Fig.3(b)(2)). The further increase in current resulted in the onset-streamers occurring around the edge of the pinhole in Region III (see Fig.3(c)(2)). It should be noted that the critical current densities for the transitions between Regions I, II and III were nearly constant respectively independent of Ea, as is shown by the flat curves A and B. The two Regions II and III should be referred to as the "onset-stage" region. With further increase in current beyond the other critical curves C and D, back discharge in the streamer modes (surface and mixed streamer modes) t o o k place in Regions IV and V. Region IV, for a lower value of Ea, is the surface streamer region where the surface streamer mode was predominant and space streamers were few (see Fig.6(b)). In this region, current density J saturated at Curve E because of space-charge limitation (see Fig.5), and no flashover could be obtained between the grid and the plane electrodes. Whereas, in Region V, when Ea exceeded 5 kV/cm, both the surface and space streamers occurred to form the mixed streamer mode (see Fig.7(b)). Again the critical current density for the transition from Region III to Regions IV and V was nearly constant, except for a corner area G. When J in Region V exceeded curve F, the streamer turned into a flashover The critical value of the field strength between Regions IV and V (Curve H)
43
160Ia5-
FLASHOVERs~/I" /J~l
,, NOFLASHOVER\ X%'
-~/
V
I
REGION
I
l" x STREAMER~. REGION ~
!
10- ~'"~
C B
) I
III ONSET-STREAMERREGION I x--X--x~x~X~x..... ~j II ONSET-GLOWREGION"" I I NOBACKDS I CHARGE
10-8
I
I
I
2
I
4 Eo
I
I
I
6
I I
8
I
I
I
i0
( kV/cm )
Fig.6. E f f e c t o f tangential field and current density on mode of back discharge. (a) Mode
diagram in field-current domain. (b) Photograph of back discharge in surface streamer Region IV, ( E a = 4.0 k V / c m , J = 5.0 × 10 -7 A/cm2).
was about 5 kV/cm under atmospheric conditions, which had been taken as a criterion for the occurrence of streamer under these conditions. It should, however, be noted t h a t the initiation and growth of space and surface streamers are also governed by current density J.
3. 2 Effect of tangential field In the present case where the surface resistivity of the layer is extremely high, the surface charge would be firmly bound to its original position. In this case, the tangential field around the breakdown point will become a function of the surface-charge density on the layer, o0, at the instant of breakdown at which the potential of the breakdown point becomes almost zero. The value of a0 in turn is given by eEds where e is the dielectric constant of the layer. If a0 is sufficiently high, the breakdown of the point will directly trigger the surface streamer. In the opposite case, onset-glow appears prior to the occurrence of surface streamer, as long as the vertical field strength in the gas space is not sufficiently high for the space streamer to be triggered. Such a high vertical field strength does n o t normally exist at the initiation of back discharge, unless the layer resistivity is in the range of 5 X 10 '0 - - 1 0 " ohm cm as discussed later. Thus, the effect of o0 on back discharge in streamer mode was studied. Two glass plates were used, as before, so t h a t Eds and, hence, a0 could be changed. Photographs of the back discharge for two values of breakdown strength are shown in Fig.7. When
44
e
(a)
Eds
LOW
(b)
Eds
HIGH
Eds = 13.8 (kV/cm)
Eds = 33.8 (kV/cm)
V
: 30
(kV)
V
: 40
(kV)
I
= 29
(~A)
I
= 23
(.A)
Fig.7. Effect of tangential field on back discharge in the m i x e d streamer mode. (A pair of glass plates, each having a single pinhole; Electrode gap = 50 ram).
Eds was 13.8 k V / c m , a space s t r e a m e r was d o m i n a n t p r o c e e d i n g to the discharge e l e c t r o d e (Fig.7(a)). When Eds was 33.8 k V / c m , the m i x e d s t r e a m e r m o d e appeared, where a r e m a r k a b l e surface s t r e a m e r in the vicinity o f the pinhole c o u l d be observed (Fig.7(b)). This was because the tangential field strength b e c a m e larger as o0 increased. The surface discharge b e c a m e especially d o m i n a n t w h e n the Value o f a0 e x c e e d e d a b o u t 5 X 10 -9 C / c m 2. 3. 3 Effect o f dust resistivity A tissue p a p e r was used as a sample in this e x p e r i m e n t . This was because its a p p a r e n t resistivity Pd c o u l d easily be changed f r o m 108 t o 1014 o h m c m b y adjusting t h e a m b i e n t h u m i d i t y . Thus, the e f f e c t o f Pd on t h e b a c k discharge m o d e at r o o m t e m p e r a t u r e was studied. V o l t a g e - - c u r r e n t - d e n s i t y curves f o r d i f f e r e n t values o f Pd, ranging f r o m 6 X 109 to 2 X 10 is o h m cm, are s h o w n in Fig.8 where the e l e c t r o d e gap was k e p t at 60 mm. P h o t o g r a p h s o f the back discharge for t h r e e d i f f e r e n t values o f Pd are s h o w n in Fig.9, where the values o f J were o f the same order. When the resistivity was 6 X 109 o h m cm {Curve 1 in Fig.8), n o b a c k discharge o c c u r r e d until flashover t o o k place at V = 65 kV and J = 7.6 p A / c m 2 .
45
FLASHOVER
10-1
/~. /
/
/,.~"
~
vs6s =
/71/
o.1
/ / c// /i/ //,~
gl.
7
. A
.... >
2: Pd = 0.9 x lO:~oh . . . . )
/Zd
0,01
T,,
3: Pd- 1.6 x lOlZ(oh....) ,
A.
.
,
10 V
(kv)
J =7.6 (pA/cm2)
,
,
I
15
20
,
I
30
,
I
40
(KV)
Fig.8. Effect of dust resistivity Pd on voltage-current-density curves under back discharge condition when negative corona is used. (Tissue paper, 1.0 m m in thickness.)
For the case of a needle-to-plane electrode system and the experimental conditions investigated, the initiation condition of back discharge, Eds = Pd X J0, was n o t satisfied prior t o the occurrence of flashover when Pd was lower than a b o u t 5 X 10 '° ohm cm. In o t h e r words, the initiation voltage o f back discharge was higher than the flashover voltage o f the gap because of t o o low a value of Pd. When the value of Pd slightly exceeded this critical value, space streamers occurred as soon as the layer broke down, owing to the large voltage drop across the gas space. For instance, when the resistivity was 0.9 X 1 0 " o h m cm (Curve 2), the streamer starting voltage Vst was a b o u t 27 kV. The n u m b e r of b r e a kdow n points was less, and streamers proceeded into space towards the discharge electrode, as shown in Fig.9(a). The occurrence of space streamers lowered the flashover voltage 118 t o a great extent. It was observed that, when Pd was between a b o u t 5 X 10 l° and 0.9 X 10 H ohm cm, excessive sparking t ended to occur. In this range o f Pd, Vst would be lowered with the increase in Pd, so t h a t it finally becomes lower than Vs as in the case o f Curve 2 in Fig. 8. A slight increase in voltage b e y o n d Vst would cause flashover because Vst remained still close to Vs. For Pd higher than 10 '2 ohm cm (Curves 3 and 4), the back discharge streamers started to occur at a m uch lower voltage and c ur r e nt density. There was a larger interval between Vst and ITs so th at the excessive sparking disappeared. There were m ore breakdown points with a general glow surrounding each point. In this case, a surface glow d o m i n a t e d and space streamers were few. This t e n d e n c y became pron o u nced with the increase in Pd (Fig.9(c)).
46
(a) Pd = 0.9 x IOII (ohm-cm) J = 3.2 ~ / c m 2)
(b) Pd = 1.6 x 1012 (ohm-cm) J = 5.5
OxA/cm2)
g
(c) Pd = 2.0 x 1013 (ohm-era)
J = 2.2 Fig.9. Effect of dust resistivity on back discharge mode. (Tissue paper).
(,wA/cm2)
47
The different discharge modes were caused by the difference in the ratio of the voltage drop across the dust layer to that across the gas space. If the resistivity was high, the voltage drop across the dust layer was high even at a low current density on the initial stage of back discharge, whereas the vol, tage drop across the gas space was low. As a result, the development of a space streamer was suppressed, and a surface discharge occurred. In this case, many weak points broke down and the current increased readily w i t h o u t excessive sparking. However, when voltage was raised, the space streamer occurred also in this case, taking the form of a general glow bridging across the gap. A more severe increase in current at this later stage. It can be seen that flashover occurred almost at the same voltage in spite of a large difference in Pd, once back discharge occurred. This agrees well with the results of Penney and Graig [ 5], i.e. the flashover voltage of back discharge was n o t affected by the value of resistivity. This flashover voltage was almost half the value of that under a non-back-discharge condition.
3. 4 Charging efficiency in different regions For negative corona, back discharge is a source of positive ions which produce a bipolar ion atmosphere in the gas space. The effect on particle charging, however, is different, depending on the mode of back discharge. In the surface streamer mode, the ion source is considered to be surface-like, but
/
D
i oVa i
Omm - - 3 0 mm~
I I
F
D:
DISCHARGEELECTRODE (SAW TOOTH ELECTRODE)
G:
GRID ELECTRODE
P:
PLANE ELECTRODE
M:
MEASURINGELECTRODE
T:
MICA PLATE WITH PINHOLES
B:
STEEL BALL WITH 3.0 mm IN DIAMETER
F:
FARADAYCAGE
Fig.lO. Electrode system for measuring particle charging.
48 in the space or mixed streamer modes, ion generation in space may occur. This was confirmed by measuring particle charge using the electrode system as shown in Fig.10 [6]. This system enabled the change in back discharge mode by changing the field strength between the grid and plane electrodes, Ea (see Fig.6). A steel ball, 3.0 mm in diameter, was dropped between the plane and grid electrodes and its saturation charge was measured by a Faraday cage. The distance between the plane and grid electrodes was 50 mm, grid to discharge electrodes 30 mm, and a mica plate having many pinholes was used as a layer. Figure 11 is an example of the results obtained, showing the saturation charge of a steel ball as a function of its position d from the plane electrode. The values indicated in the bracket represent its ratio to the theoretical saturation charge due to monopolar ions, calculated from Pauthenier's equation [7]. In the surface streamer region (Curve 1), the value of charge was about 90% less than the theoretical value but the sign of particle charge remained the same as that of the discharge electrode. The value of charge decreased as the particle crossed nearer the plane electrode. This result indicates that the back discharge of this mode can be considered as a surface-like ion source so that the density of positive ions decreased into the space. In the mixed streamer region, where space streamer was pronounced {Curve 2), particle charge scattered largely around its average value, which was a fairly high positive value and almost the same regardless of position. This (2) MIXED STREAMERREGION Ea : 6.0 (kV/cm) J = 2.0 (,~A/cm2)
+I0
+8 +6
~+4
I+2
(3) TRANSITION REGION
Ea = 4.6 (kV/cm) J 1.7 (~A/cm ~)
'o x
0
o
12
Q -4
_.~/5.4 %)
~-- . . . .
T
( I ) SURFACE STREAMERREGION] 9.q ~) Ea - 3.6 (kV/cm) J = 1.6 (#A/cm 2)
i
I
I
I
i
i0
20
30
40
50
d (mm)
Fig.11. Saturation charge versus position d for different back discharge modes.
49
result might indicate that positive ions were generated abundantly inside the whole space and played a d o m i n a n t role in particle charging. The effect of the streamer tip colliding with a particle might also be a factor. Curve 3 represents the transition region between the foregoing two. In this case, particle charges were also positive but as low as in the case of Curve 1. 3. 5 Back discharge with positive corona It was observed t h a t the mode of back discharge with positive corona at the needle was completely different, as shown in Fig.12. Tissue paper was used and the electrode gap was 60 ram. Voltage--current-density curves are shown in Fig.13 for various values of Pd. In this case, breakdown points were distributed uniformly on the surface, no space or surface streamers could be observed and the discharge mode was only glow-mode independent on resistivity. The abnormal increase in current was small and the flashover voltage, when back discharge occurred, was approximately 1.5 times higher than that for back discharge with negative corona at the needle. The relationship between the flashover voltage of back discharge, Vs, and gap tlistance d is shown in Fig.14 for the positive and negative coronas. Vs of the positive corona was higher than that of the negative corona for a gap distance range of 1.0--10.0 cm. The flashover voltage of the positive corona under back discharge condition was also higher than that without back discharge when the layer was removed (air load). The mechanism for this behaviour is considered to be due to a stable nature of negative glow corona at the breakdown point, and to
r i g . 1 z ~ a c k discharge under positive corona point. (Tissue paper, V ffi + 4 0 k V , J = 2.8 X 10 -4 A / c m 2, P d = 1013 o h m c m . )
50
i~[
°~
GAP: 60(~)
4/,/~
l: A,~ LoAo
/'7
IOII (oh.... ) 3: /:~1= 1012 (oh.... )
2: ,°d=
I"
F
j/
3 2
~i~
/0/~ i
/,4/1
....
v
o, ol ¼
~
7
i0 v
15
20
30
q0 50
(+kV)
Fig.13. Voltage--current-density curves under back discharge condition w h e n positive corona is used. (Tissue paper, 1.0 m m in thickness)
60 50 40 v
3O
L/
d :=" 20
0
/
I
i
I
i
I
I
I
50 ELECTRODE GAP d (mm)
i
I
I
100
Fig.14 Flashover voltage versus gap distance under back discharge condition with positive and negative coronas. (Tissue paper, Pd = 1 . 2 X 1 0 ~1 ohm cm, 1.0 mm in thickness.) the positive c o r o n a at t h e needle tip b e i n g c o n v e r t e d t o H e r m s t e i n ' s g l o w c o r o n a [8]. The l a t t e r m a y result f r o m a c o p i o u s s u p p l y o f negative ions f e d to t h e needle e l e c t r o d e , f r o m which e l e c t r o n s w o u l d be shed to f o r m a c o n t i n u o u s a n d stable positive glow discharge at t h e n e e d l e tip.
51 4. Conclusions From the experimental studies described above, using the model samples of tissue papers, glass and mica plate, the following results concerning the effects of apparent resistivity and breakdown field strength on back discharge were obtained. (1) With the negative corona at the needle, the layer breakdown started to occur when Eds = Pd X J is fulfilled. It was followed by the onset-glow mode occurring with a slight increase in voltage. A rapid increase in current, however, occurred only when the streamers started to occur at a critical voltage Vst. Thus, there are four major critical voltages for back discharge under atmospheric conditions; layer breakdown voltage Vb, onset-glow starting voltage Vo, streamer starting voltage Vst and flashover voltage Vs. For the case of the electrode system investigated, the initiation condition of back discharge may n o t be fulfilled prior to the occurrence of flashover when Pd does not exceed a b o u t 5 X 10 '° ohm cm. When Pd is in the range of 5 X 10 ~° - - 1 0 " ohm cm, Vst <- Vs, so that excessive sparking occurs. When Pd > 10': ohm cm, Vst becomes sufficiently lower than Vs so that excessive sparking disappears, b u t an abnormal increase in current occurs. (2) There are three sub-modes in the streamer mode, depending u p o n the field distribution around the breakdown point in the sample layers: space streamer mode, surface streamer mode and mixed streamer mode. This, in turn, is a function of Pd, Eds, Ea and J. In the space streamer mode, positive ion generation in space occurs and particles aquire a fairly high positive charge. Whereas in the surface streamer mode, positive ion generation is limited to the surface region and the sign of particle charge is the same that of the needle electrode. In most of the actual cases, however, the mixed streamer m o d e appears. (3) With positive corona at the needle, the back discharge mode is completely different. The flashover voltage is higher than that under the back discharge condition with the negative corona.
Acknowledgements This research was sponsored by the Ministry of Education, Japan, as its Special Research Project (I) (Project No.011914). The authors are gratefully indebted for its support. Thanks are also due to Mr Masao Kuroda for his help in some of the experiments. Notation
Ea
vertical field strength in the gap
Eds breakdown field strength of the layer measured by parallel plane electrodes
E'ds breakdown field strength of the layer in corona field
52
J J0
current density at the measuring electrode current density at the initiation of back discharge Va discharge electrode voltage Vb layer breakdown voltage Vg grid electrode potential Vo onset-glow starting voltage Vs flashover voltage Vst streamer starting voltage e dielectric constant of the layer o0 surface charge density at the instant of breakdown Pd apparent dust resistivity References i 2 3 4 5 6 7 8
S.Masuda, Recent progress in electrostatic precipitation, Static Electrification 1975, Institute of Physics Conference Series, No. 27 (1975) 154. S. Masuda and A. Mizuno, Flashover measurements of back discharge, J. Electrostatics, to be published. S. Masuda, Reverse ionization phenomena in electrostatic precipitator, J. Inst. Electr. Eng. Jpn., 35(102) (1960) 1482. S. Masuda and A. Mizuno, Light measurement of back discharge, J. Electrostatics, 2 (1977) 375. G.W. Penny and S.E. Craig, Sparkover as influenced by surface conditions in d.c. corona, Am. Inst. Electr. Eng., Trans. 79 (1960) 112--118. M. Kuroda, Graduation Thesis, Univ. of Tokyo, 1975. M. Pauthenier, Moreau-Hanot, Rev. Gen. Electr. XIV (18) (1932) 583. L.B. Loeb, Electrical Coronas, Univ. of Calif. Press, 1965, p. 95.