JPS61187907A - Process for controlling flock formation in water purification plant - Google Patents

Abstract

PURPOSE:To form easily settled flock in a flock forming basin of a water purification plant by measuring the area and circumferential length of the flock by the picture processing, and reducing the amt. of flocculant to be injected when the measured values exceed regulated values. CONSTITUTION:Crude water is introduced into a quick mixing basin 20 where a flocculant injected from a flocculant injector 22 is stirred with a stirrer 21. Stirring paddles 31 are installed to the flock forming basin 30, thus, fine flocks are flocculated to each other. The picture image of the flock 34 is converted to electric signals by an image pick up device 100, and the area S and the circumferential length L of the flock 34 are operated by a picture processing device 110. Then, the amt. of flocculant to be injected is controlled basing on the signal of the operator 210. The flock 34 is poured into a settling basin 40, where it is settled.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a method for controlling the formation of 70 flocs in a t-th water purification plant, in which the formation of flocs in a floc formation pond of a water purification plant is performed well.

[Background of the invention]

Since the particle size of suspended matter in raw water is small at a water treatment plant, flocculant F is injected to make it aggregate, form flocs, and cause it to settle. Next, floc formation pond (mixing pond)
It is necessary to monitor the formation of flocs in K.

Conventionally, flocs have been visually monitored by water treatment plant maintenance managers only several times a day. However, since it depends on visual inspection, even if the size of the flocs can be qualitatively determined, the density of the flocs cannot be measured.

It is impossible to know whether the floc is likely to settle in the Tonoike Pond or not.

On the other hand, industrial TV cameras these days? (ITV) has also been used to monitor floc groups in floc formation ponds. As described in Japanese Patent Application Laid-Open No. 54-143296, it has also been proposed to use a photoelectric conversion device to extract electrical signals according to the shape of the flocs and monitor them through image processing.

However, if the particle size of the floc is large, it does not mean that agglomeration is being performed well, and it is not possible to form the floc optimally only by recognizing the shape t-of the floc, such as the particle size of the floc.

That is, even if the flocs are large, if the degree of aggregation is low, the sedimentation speed in the sedimentation tank will be slow, and furthermore, micro flocs will be left behind during the sedimentation process and will flow into the filter cell. If too much flocculant is injected into this, the amount will increase next time. Conversely, even if the flocs have a high degree of aggregation, it becomes difficult for the flocs to settle in the settling tank when the flocs are small. This often occurs when the flocculant is insufficient or the stirring is weak.

As described above, there is a drawback in that it is not possible to optimally form flocula just by knowing the size of the flocs.

[Purpose of the invention]

An object of the present invention is to provide a method for controlling floc formation in a water purification plant that can form flocs with good sedimentation properties in a floc formation pond (mixing pond) of a water purification plant.

[Summary of the Invention] The feature of the present invention is that the area and perimeter of flocs in a floc formation pond are measured by image processing, and the floc shape index (
= Perimeter ÷ Area) and perform trap control to reduce the number of flocculant injectors whose floc area exceeds a predetermined value and the shape index exceeds a predetermined value.

[Embodiments of the invention]

FIG. 1 shows an embodiment of the present invention.

In FIG. 1, the taken water is led to the rapid mixing pond 20 via the landing water divided by 10. A stirrer 21 is provided in the rapid mixing pond 20. The nine flocculants injected from the flocculant injector 22 are stirred by the agitator 21.

Companion paddles 31A and 31B are placed in the floc formation pond 3°.
, 31C are provided. The stirring paddles are rotationally driven by stirring motors 32A, 32B, and 32C, respectively. A rectifying wall 33A in which a large number of holes are bored between stirring paddles 31A, 31B, and 31C.

It is separated by 33B. The fine flocs flowing from the rapid mixing pond 20 are sequentially stirred in the floc forming pond 30 by the stirring paddles 31A, 31B, and 31C. As a result, the minute flocs collide and coalesce to become flocs 34. While the flocs 34 pass through the floc formation pond 30, the particle size gradually increases.

The Qiao Wei Soo 100 converts the image of the flock 34 into an electrical signal using an industrial television camera (not shown).

Image processing device 110H performs image processing on the floc 34 based on the electric signal obtained from the image pickup device 100.
340 Area S and perimeter length are calculated.

Arithmetic device (control computer) 210 and image processing device 110
The shape index K of 70 is calculated from the specific area S and the perimeter length.
is calculated using the following formula.

K=L-;S (1) Coagulant injection t is made to the coagulant injection machine 22 based on a command signal from the control computer 210.

The flocs 34 formed in the floc formation basin 30 flow into the settling basin 40 where they are settled and removed. The supernatant water from which the flocs 34 have been removed is injected into the filtration cell 50. The remaining minute flocs that were not removed in the sedimentation tank 40 are filtered out by the filtration cell 50. The supernatant water flowing out of the filtration cell 50 is supplied to consumers through a drainage pond (not shown), a reservoir (not shown), and the like.

The detailed configuration of the imaging device 100 and the image processing device 110 is shown in the tsz diagram.

An industrial television (ITV) fixed in an airtight container 120
A 130d close-up lens 131 is used to view the floc formation pond 30 through a clear observation window 121 made of a transparent material such as glass.
The image of the flock 34 inside is enlarged and recognized. The wiper 122 driven by the wiper drive device 123 operates periodically to remove dirt from the surface of the observation window 1210. A back screen 124 is provided next to recognize the flock group with high contrast and accuracy. Screen 12
4 is fixed to the airtight container 120 and installed in front of the observation window 121 via nine-pack screen fixtures 124A and 124B. The back screen 124 is preferably of a dark color so that white flocks can be recognized with high contrast and accuracy. A light-shielding cover 142 is provided next to darken the surroundings and provide illuminance under a certain condition only by the illumination device 140. By providing the light-shielding cover 142, the images of the flock group will not be affected by changes in illuminance during the same song. - It is practical to install a plurality of illumination devices 140 to irradiate the flock from multiple directions. Note that the lighting device 1405C can also be controlled by the lighting controller 141 according to changes in ambient illuminance without providing the light-shielding cover 142. The baffle plates 125A, 125B, and 125C are provided at the ninth position in order to suppress the movement of water as much as possible within the floc formation pond 30 and not to block the illumination.

By suppressing the movement of water, images of flocks in a flowing state can be recognized with high accuracy.

Flock image information 1I captured by ITV130
Image recognition device t150 via TV controller 132
sent to. Image recognition control device 160 is image recognition device 1
50 controls the frequency of floc recognition. Generally, the floc formation situation does not change rapidly in a short period of time. Therefore, it is sufficient to carry out the floc monitoring operation once every 10 minutes to once every hour.

A process in which image information of a flock group is subjected to signal processing within the image recognition device 150 will be described below.

FIG. 3(a) shows an image plane of a flock group recognized by the ITV 130. Since the flock group is white, the brightness level is high, and since the background back screen 124 is black, the brightness level of the water is low.

Since the flock group is a grayscale image, it is actually flock 3.
The boundary between 4 and the water part is not clear. Figure f43 (a) shows only the outline of the friend flock group for simplicity.

@3 Figure (b) shows the distribution of brightness levels when scanning the AA' line on the screen of Figure (a). Brightness level is 256
The brightness level is displayed in stages, with the brightness level increasing toward the top of the vertical axis and decreasing toward the bottom. flock 3
4 is white, so the brightness level is higher. Figure 3 (b
), the downward trough represents 70 points. The brightness distribution shown in FIG. 3(b) is binarized based on the brightness level threshold of the BB' line. A pixel at a brightness level higher than the threshold is set to level 112, and a pixel at a brightness level below the threshold is set to level 10. FIG. 3(C) shows a signal waveform obtained by binarizing the screen shown in FIG. 3(a)K along line AA'.

By subjecting the entire screen of FIG. 3(a) to the aforementioned binarization process, a z-valued screen as shown in FIG. 3(d) K is obtained. 300 in FIG. 3(d) indicates each pixel. The sum of 300 pixels that each flock has (N
s+, i=1...n) to the area of each floc (8
t, i=1.2. ... n) is calculated using the following formula.

81 =CHXNs+ (2
) Actual area corresponding to C1=1 pixel A signal of the calculated area is output from the image recognition device 150.

Specific pressure, peripheral length of each floc (Ll, 3=1...n
)H First, among the pixels corresponding to each flock, in Figure 3 (
As shown in e), the sum of pixels located around the 7Cf block is added. Then, using the added value NLI, the perimeter L1 is calculated using the following equation.

L1=Cx

Ks = L ÷ St ・・・・・・(4
) The smaller 1 is in the shape index obtained by formula (4), the more circular the floc shape becomes, and the more round the floc shape becomes. This will be explained with reference to the figures.

FIG. 4(a) i shows a case where the amount of coagulant injected is insufficient.

Due to the lack of flocculant 60, the flocculation effect is weak, and only a small amount of suspended particles 61 are bonded together, but cannot grow into large flocs. At this point, the sedimentation speed of the flocs becomes slow, and the flocs cannot settle completely in the sedimentation tank, causing clogging of the filtration cell. At this time,
Although the formed flocs are small, they are densely bonded to each other and the shape index Ks (=Ls/S+) becomes small.

FIG. 4(b) shows the case where the amount of coagulant injected is appropriate.

Sufficient flocculant 60 to tightly bond a large amount of suspended particles 61
The flocs 34 formed by injecting the flocs grow large and have a high apparent density, resulting in flocs with good sedimentation properties. At this time, since the flocs formed are large and the degree of bonding of the suspended particles 61 is also dense, the shape index of 1 becomes small.

FIG. 4(C) shows a case where the amount of coagulant injected is excessive.

Since the flocculant 60 is in excess, the excess flocculant 60 (positive ions) makes the surface charge of the suspended particles 61 (negative ions) positive, which becomes a repulsive force and weakens the bonding force between the suspended particles 61. Therefore, during aggregation, the intervals between suspended particles become long and sparse, and the bonding of suspended particles becomes uneven. As shown in FIG. 4(C), the shape of the flocs formed is more uneven than when the flocculant is optimally injected.

By the way, it is known that if the shape index has the same area as +i, a perfect circle will be the smallest.

Therefore, in the case of excess flocculant, the shape index is +(-L t
/'S plate) becomes larger.

As described above, flocs that grow large and have a small shape index of 1 can be said to be flocs suitable for sedimentation and removal.

Next, f! Using FIG. 45, FIG. 6, and FIG. 7, differences in image information when image processing is performed due to differences in coagulant injection rate are shown.

Now, when the floc formation does not proceed due to the shape of the flocs and the amount of coagulant injected is insufficient, and the flocs are small, the image information shown in FIG. 5(a) is obtained. However, as shown in the binary image of FIG. 5(b) obtained by binarizing this information, the area of the flocs is small, so the flocculant injection rate is increased to promote floc formation. When shown in terms of area, Figure 8 (
Horizontal axis: area Sr, vertical axis: area α with shape index +).

Next, Fig. 6(a) shows image information of a well-known spot with good sedimentation characteristics and flocs showing an optimal shape. The calculated area St is large and the shape index 1 is small. t48
This is region β in the figure.

Next, FIG. 7(a) shows image information of flocs when the flocculant injection rate is excessive. Although the size of the formed flocs is large, the overall unevenness is severe, Part 2
The shape index KI is calculated from the value of FIG. 7(b) and becomes larger. This becomes area r in FIG.

FIG. 9 shows an example of a processing flow focusing on the above-mentioned nine points.

The image capturing device 100 in FIG. 1 captures an image of the flock 34 (step S1 in FIG. 9), and the image processing device 110 captures an image of the flock 34.
Perform value calculation (step S2 in FIG. 9), and the area is 5It.
- Calculate (step 83 in Figure 9). In step S4, the control computer 210 compares the area S1 with the set value S (shown in FIG. 8), and if St<8-, issues a command to increase the flocculant injection rate to the flocculant injection machine 22 ( Step S9 in FIG. 9).

Furthermore, if Sl>S, the process moves to step $5.

In step S5, the image processing device 110 calculates the flock circumference L1 in line Vh, and from the result, the −-shaped index KI
is calculated by the control computer 210 (step 86).
. In step S7, the control computer 210 sets the shape index to 1 and the set value K.

(shown in Fig. 8), and when KI)K, a command to reduce the coagulant injection rate is given to the coagulant injection machine 22 (step 510 in Fig. 9).

Further, when KI<K, it is determined in step S8 that the floc formation situation is normal, and the series of steps is terminated.

In particular, when these steps S1 to S5 are repeated, the floc formation status is determined and the injection rate of the flocculant is controlled in accordance with the floc formation status.

In this way, flocs are formed by controlling the injection of the flocculant, and since the injection amount is changed depending on the area and shape index of the flocs, good flocs can be formed.

Note that since there are a large number of flocs, it is preferable to use a value such as 80% of the number of flocs to determine the area and shape index.

〔Effect of the invention〕

As explained above, according to the present invention, since the flocculant injection amount is controlled by constantly measuring the floc shape index, flocs with excellent sedimentation properties can be formed with the minimum flocculant injection amount. I can do it. With this in mind, it is possible to maintain a low load on the sedimentation tank and filtration system, which in turn makes it possible to save energy and labor in water purification plant maintenance and management, and improve reliability.

Although the above explanation describes the case where the injection of the flocculant is controlled using the shape index, it goes without saying that the injection amount of the flocculant may be determined based on the raw water turbidity and corrected using the shape index. It is.

[Brief explanation of the drawing]

Fig. 1 is a configuration diagram showing one embodiment of the present invention, and Fig. 2 is a block diagram showing an embodiment of the present invention.
A detailed configuration diagram of the main parts of the figure, Figure 3 is an explanatory diagram of image signal processing,
FIG. 4 is a diagram of the formation state of flocks, FIGS. 5 to 7 are screen diagrams of the flocks shown in FIG. 4, FIG. 8 is a characteristic diagram of shape indexes, and FIG. 9 is an operation flow diagram. 10...Water landing φ, 20...Rapid mixing pond, 22...
Coagulant injection machine, 30...Floc formation pond, 100...
・Photographer outfit! , 110... image processing device, 210...
Control calculation part 3 Figure <a)
(b) Guy Bo (a) ζb no kudzu 7ko89 Figure

Claims (1)

[Claims] 1. In a water treatment plant where a coagulant is injected into raw water and then introduced into a flocculation pond to grow flocs, the area and perimeter of flocs in the flocculation pond are measured by image processing, and the measured flocs are A shape index of the floc is determined from the area and perimeter, and control is performed to reduce the injection amount of the flocculant when the floc area is equal to or greater than a predetermined value and the shape index is equal to or greater than the measured value. Characteristic method for controlling floc formation in water treatment plants.