JPS58136917A - Monitoring method of combustion in boiler furnace - Google Patents

Abstract

PURPOSE:To contrive to save labor for monitoring operation by a method wherein the burning state in a combustion furnace is detected by an image-sensor a plurality of times and the frequency of occurrence of detection signals having a predetermined brightness within a predetermined period of time is counted in order to monitor the state in the furnace based upon the counted value. CONSTITUTION:An industrial television camera 4, for example, is mounted on the wall within the furnace 2 in order to monitor the combustion state in the furnace 2 of a boiler. Video signals 5 from the industrial television camera 4 are obtained through an A/D converter 6 by an electronic computer 8 in order to obtain the sum of the parts which are equal to one another between every picture by means of an integral function 11. Next, the mean value of the video signals is calculated at a mean value calculating function 12 so as to obtain a mean video signal, which is divided into partial pictures again in order to compare each partial picture with the set values 15 corresponding to the pre-set gradation. When the gradation of partial pictures exceeds the set value 15 in the comparison, the positional distribution of subject partial pictures is calculated in order to output the treatment result thereto to an output unit 10.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a device for monitoring the combustion state inside a boiler furnace, and particularly to a method for monitoring the combustion state inside an IC boiler furnace, which is applicable to boilers for thermal power generation.

Conventionally, the combustion state inside a boiler furnace was monitored using in-furnace monitoring IT.
Methods such as combustion monitoring using V, inspection from @1 window, and management of heat transfer tube temperature are mainly used.

As for automatic control devices, automatic burner ignition and fire extinguishing judgment using flame V-metameters have been rarely used to protect furnaces.

Moreover, the combustion conditions that change over time during commercial operation cannot be quantitatively evaluated, and operators have no choice but to rely on their experience and intuition, making it impossible to adequately deal with the combustion conditions.

Conventionally, as a method for monitoring abnormal combustion in the furnace, the output signal of a TV camera that monitors the condition 1M inside the furnace is adjusted to correspond to the fluctuations in the video signal caused by flame fluctuations in a normal combustion condition, using the following time constant. A reference value for a normal combustion state is determined from the average value of the integrated values, and this reference value is compared with the integrated value, and if the difference exceeds a certain value, a warning is issued. However, the problem with this method is that even if abnormal combustion occurs, no alarm will be issued as long as the integral value is within the standard value.

In addition, as a similar method, extract the video signal corresponding to a predetermined part in F, and then integrate it with a predetermined time constant in response to the fluctuation of the video signal in the normal combustion state, and combine the integral value with the reference value (integral value There is a method in which a predetermined value is subtracted from the ratio value) and an alarm is issued if this difference is greater than a certain value. Since the setting of p is determined by the amount of change in fluctuation, malfunctions may occur.

A similar method is to integrate the Eirin signal by the fluctuation time constant, derive a reference value from this value, and compare it with the integrated value, but this also can cause false alarms if the furnace output changes. The problem with this is that it may sometimes emit alarms or fail to detect abnormalities.

This is a common problem with the conventional methods described above.

The ninth step in integrating the video signal is that abnormalities can only be detected by the increase/decrease in the integral value and its rate of change, and the next step is to know what kind of abnormal state will occur. In addition, there is an essential problem that the increase and decrease are canceled out during the integration process, and in the next case, even if an abnormality occurs, no alarm is issued.

Furthermore, it was not possible to obtain information under normal combustion conditions.

SUMMARY OF THE INVENTION An object of the present invention is to provide a method for automatically evaluating and extracting a flame image pattern to determine the combustion state inside a boiler furnace.

The key point of the present invention is that when the image detected by an ITV camera or an infrared camera has fluctuations, the heat absorption in the furnace is related to the average brightness (gradation) distribution.
The aim is to monitor the combustion state in the furnace using the A degree.

Specifically, the brightness of a video signal detected by an ITV, etc. is divided into multiple brightness ranges, and the frequency of which of the divided brightness ranges the brightness of the detected video signal corresponds to is counted and calculated. The combustion condition in the furnace is monitored by the brightness after the caulking exceeds the set frequency.

The following is a 1k15 visit to the implementation of this invention! I will clarify. In FIG. 1, the purpose of the boiler is to combust fuel supplied from a burner 1 with a fire F2, and use the generated heat to convert water vapor in a heat exchanger tube 3 into saturated steam (or water into steam).

Therefore, in order to monitor the combustion state in the furnace 2,
For example, a method for attaching an ITV camera (or infrared camera, etc.) to the wall inside the 4t-furnace 3 will be described. The present embodiment uses a video signal tube from, for example, an ITV camera installed in this manner. Video signal 5'i from ITV Kamen
: jIiL9 is input to the electronic computer 8 via the A/D converter 6, and the processing result 9 is output to the output device 10.

Next, the processing functions in the electronic computer 8 will be described using FIGS. 2 to 4! I will clarify. Moreover, the same effects can be obtained even if each processing function is configured by hardware. A summary of the processing functions within the computer 8 is shown in FIG. 2. In FIG. It will be done. The captured video signal 7 is used to calculate the sum of the same portions for each screen using an integration function 11, and average values of the video signals are calculated using average value calculators 11 and 12 to obtain an average video signal.

An example of another method for obtaining an average video signal is shown in FIG. In FIG. 3, the A/f) converter 6 converts the signal into a digital signal, and then the video signal 7t-1 video signal splitter t! 1
6, the screen is arbitrarily divided into partial screens as shown in FIG. Next, any part @@t1. fc Each of the video signals applied to each partial screen is processed by a histogram calculation function 17 that calculates a histogram of a fixed number of screens. The basic pattern of a histogram is shown in Figure 6 (
匈, (b), (C).

Each (d) is divided into four species, and those r! This corresponds to the partial human image 11C shown in FIG. Each part obtained in this way
The histogram for the floor of the 1ii surface is compared with the preset value 19 for the @ degree using the comparison - wear 18, and the average gradation of the portion exceeding the setting [19 is determined for that portion Ij! By setting the gradation to 1iiiI, an average video signal is calculated.

Nine, as another method, set value 19 in the above method.
It is also an effective method to divide the average gradation of the part that is displayed in 'iMi' into the part where 9 exceeds the set value of 19 and the gradation before and after it, and to convert it into binarization and convert it into 300 million. Also,
f & stationary 19? The most frequent floor of the exceeded part g4t-
A method to obtain an average video signal.

Alternatively, the floor IA is classified as follows. (9iIl
For example, in Figure 6 (when dividing into 4 classes by M floor - like Hiroshima) no = ``t l'1tXi is A class ml+
1 to m, tt is B 〃m 1 to m, floor is C class ms to M'Il'lsD When you enter.

If 9 falls in class A, set 1 gradation m1°.

(me 'hmto 'hmt) If you enter By Shirasu, 1st floor tl14 ”me t-setting.

(m++1sm*oSmm) If you enter class C, you will be on the 1st floor i! Set II m no.

(fnl+1s, mso<m5) If you enter D class, the 1st floor GM46T- setting.

(ms+ISlSm4os) If yes, setting the floor as described above and using it as an average video signal is also an effective method for determining the average video signal.

The nine-average video polarization obtained using the various methods described above is
The average video signal division function 13 divides it into parts and planes again.
Each divided partial screen is set in advance.
Compare with the setting value 15 for . At this time, the position distribution of the partial screen where the first floor - exceeds the set value 15 is determined. Next, this processing result 9t is outputted to the output device 110.

In the above method, when calculating the average video gI, there is a condition that the load remains constant. This is because when the load is rising or falling, or when moving from one load to another, the fuel supply to the burner changes, and as a result, the size, shape, etc. of the flame changes. This is because the gradation of the entire m-plane changes accordingly. During this rounding, the set value 15 must be changed every time the load changes.

Therefore, a method for compensating for the above drawbacks will be explained.

#Other embodiments of the invention! Shown in Figure 4. s4 diagram.

Basically, the size of the flame is the same as the s3 diagram.

By calculating the set value 15 for the new gradation due to the change in shape using the average video signal, the flame can be captured well even during load fluctuations.

In FIG. jI4, for example, using the average video signal obtained by the average value calculation function 12, the setting value calculation function 20 for one gradation calculates the setting value for the gradation using equation (1).

A: Set value for gradation B 2 Number of legs N Number of divisions of knee peak can be monitored.

In addition, by calculating the difference between the average video signal and the video signal that is not included in the averaging process, the flame fluctuation f&FjLt-
Information on burner air volume adjustment, etc. can be grasped:
You can also get it.

Furthermore, by memorizing the flame shape under normal conditions,
Abnormal boiler combustion can be detected automatically.

Also, in Figure 1, the signal inside the furnace is output as a video signal using an ITV camera, and the signal is red (R).

Edge (G), truth (B) K separated, brightness f of separated signal
The inside of the furnace may be monitored based on (gradation) frequency. For example, as shown in FIG. 8(a), the signal 5f:R.

The signals are separated by the G and B separation means 50, and then passed through the corresponding signal conversion circuits 52, 54, and 56, and the changeover switch 58 selects one of the signals and outputs it to the A/Di converter. You can use any method you like to enter it. If the data 52 to 56 are A/D converted and stored sequentially at this stage, the A/D converter 6 is not necessary. In this way, when 52 to 56 include the storage device 1ilt, it is convenient because after monitoring with 1, for example, the B signal, it is possible to analyze again with another signal (if you ask, the G signal). In general, when the combustion condition in the furnace is sudden, the color component of the flame changes depending on the load condition, etc., so depending on the furnace condition, the changeover switch 58 is selected by the 1 selection circuit 60.
good.

In addition, Fig. 8(b) shows a good signal collected by ITV and filtered to 4.
It is also possible to use a method of detecting only a specific color signal that is transmitted through 2. By switching the filter, 1, and 42t according to the color, G, and B, it is possible to detect a signal corresponding to a specific color.

In addition, smooth and defined wavelengths (for example B) can also be used.

When performing image sensing for wavelengths (G, 几), it is necessary to set the setting value xsri for each wavelength.

When estimating and diagnosing the flame shape in the furnace, it is necessary to store the reference flame shape signal ti as the caulking setting value 15.

According to the present invention, it is possible to monitor the stable combustion state of the furnace even when there is little flame fluctuation.

More specifically, the following effects can be obtained.

fl) Flame fluctuation r't-Easy to know,
It can be used as burner air amount joint information.

(2) By extracting the parts with high brightness (gradation), it is possible to understand the flame as a pattern and analyze its shape.

(32) Since it relies on the intuition and experience of the operator, the burden on the operator can be reduced.

(4) By memorizing the normal flame shape, abnormal combustion can be automatically detected.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing an embodiment of the present invention, Fig. 2 is a diagram showing an example of the processing function of the embodiment of the invention, and Fig. 3 is another example of the processing function of the embodiment of the invention. FIG. 4 is a diagram showing another embodiment of the present invention. FIG. 5 is a diagram showing a video signal when the UII plane is arbitrarily divided to obtain an average video signal, and FIG. 6 is a diagram showing a basic pattern of a histogram for gradations. FIG. 7 is a diagram showing an example of the position corresponding to each hisdog 2m in FIG. 6, and FIG. 8 is a diagram showing another embodiment. l... Burner, 2... Furnace, 3... Heat exchanger tube, 4...
...ITV camera, 5...video signal, 6...A/I
) converter, 7...Eirin signal converted to digital signal, 8...electronic computer, 9...processing result, lO...
Output device, 11... Integration function, 12... Average value calculation function, 13... Average video signal division function, 14... Comparison function, 15... Setting value for gradation, 16... Video signal division function, 17... Histogram calculation function, 1
8... Comparison function, 19... Setting value for frequency, 2
0... Calculation function of setting value for gradation. ``Shishiko' thing 6 Figure (α) (su) γ everyone Souji Zoro figure (me) Floor Wani

Claims (1)

[Claims] 1. In a method for monitoring the combustion status of a combustion furnace, the combustion status in the furnace is detected multiple times using an image sensor, and the predetermined brightness of the cough detection signal is detected within a predetermined period of time. A combustion status monitoring method that involves counting the frequency of occurrences and monitoring the inside of the furnace based on the needle value. 2. The combustion status monitoring method according to the above-mentioned patent claim O@s, paragraph 1, characterized in that the detection signal is divided into a plurality of luminances, and the detection frequency is counted for each of the divided luminances. 3. In the above-mentioned OmWA @ Section 1, the combustion status 5 is to detect the combustion status of nine predetermined areas west of the cough furnace using the image sensor.
1i viewing method. 4. The combustion situation described in the third aspect of the patent claim, characterized in that a plurality of s-minute regions are set in the furnace, and the bright straight IK is counted for each set partial region. Monitoring method. 56 In the claims 1 to 31A, a smoothing reference count IIAlt is set, and the combustion in the smelting furnace is determined depending on whether the frequency of coughing exceeds the frequency setting value of the reference needle a. A combustion status monitoring method characterized by diagnosing the status. 6. In the claim 4, the number of smoothing stitches Sat is set for each of a plurality of partial areas, and it is determined whether the number of brightness stitches (mm) for each of the plurality of areas exceeds the cough reference counting frequency. Combustion status monitoring method that uses the q# value to diagnose 19 furnaces in combination. 7. Combustion status monitoring according to claim 6, characterized in that the flame shape in the furnace is estimated and diagnosed based on a signal with a brightness exceeding a reference number of needles Mft defined in multiple cough regions. Method. 8. In claim 117, the estimated next flame shape is diagnosed by comparing the flame shape stored in the smooth memory and the flame shape added by cough #1. -Characteristic combustion status monitoring method. 9. The range of the q# pestle test IE! The combustion status monitoring method as described in item 1, wherein the t% sign is to change the base * number of stitches @'gL set value according to the operating status of the cough oven. 10. The combustion status monitoring method as set forth in claim 1, in which the luminance frequency WLt-counting of the signal of a predetermined wavelength among the signals from the image sensor is calculated as the tq# value. 11. In the above-mentioned claim 8j110. A combustion status monitoring method in which the brightness frequency is calculated as t** for each signal of a smoothly determined wavelength of nine Australian numbers.