JPH05187604A - Boiler tube leakage inspection method and device thereof - Google Patents

Abstract

PURPOSE:To make it possible to detect a tube leakage with high accuracy by frequency-analyzing data of furnace wall vibration based on the sound of steam and hot water ejected from the tube. CONSTITUTION:A CPU 23 measures each steady data transmitted from each vibration sensor 6 in terms of the whole 24 channels when a boiler is operating in a normal fashion and stores the vibration data as 3M word digital data in a memory circuit 21. Then, a boiler tube leakage is measured and vibration data thus similarly obtained is stored in the memory circuit 21. After the measurement is over, the measurement data is frequency-analyzed by a frequency analysis circuit one after another. The data is compared with the previously measured steady data of the channels and the changed portion data is stored in a disk file device 24. Then, fuzzy processing is carried out based on the changed portion data. This construction makes it possible to eliminate the need for sensibility correction per sensor.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a tube leak inspection method and inspection device for a boiler used in a thermal power plant or the like.

[0002]

2. Description of the Related Art Conventionally, a plurality of boiler tubes are used in a boiler furnace building of a thermal power plant. However, if a leak occurs in the tubes and the discovery is delayed, the damage to the equipment may be expanded. is there.

Therefore, conventionally, as shown in FIG. 10, an acoustic emission sensor 42 such as an AE sensor or a microphone is directly attached to a tube 41 in a boiler 40 or via a waveguide rod so that a leak sound generated when a leak occurs is generated. It was detected by the sensor 42, and the detected signal was sent as an analog signal to the signal processing device by a cable, and the occurrence of leak was detected by the processing device.

[0004]

By the way, in the above-mentioned conventional apparatus, since the sensor 42 is directly attached to the tube through which the high temperature steam passes, it is necessary to use an expensive sensor specially designed for heat resistance, and the sensor 42 is attached. There was a problem that it was inefficient because it was necessary to stop the operation of the boiler during construction.

Further, since the detection signal is processed in the state of an analog signal, it is easily affected by external noise and the accuracy of measurement cannot be improved, and a large number of sensors 42 are used.
However, there is a problem in that a large amount of cables are required and the cost of the entire inspection device increases.

According to the present invention, a plurality of vibration sensors are installed on the outer wall of the furnace ridge, and when a leak occurs in the tube in the furnace, the vibration sensor detects vibration of the furnace wall due to jet noise of steam, hot water, etc. from the tube. Then, it aims at providing the method and apparatus which detect a tube leak by carrying out frequency analysis of the vibration data detected by this sensor.

[0007]

In order to solve the above problems, the present invention firstly converts vibration data from a plurality of vibration sensors installed on the outer wall of a boiler furnace building into digital data, Digital data is sent to the control device via an optical fiber cable, and the digital data of each of the vibration sensors input into the control device is frequency-analyzed and steady data for all vibration sensors during normal boiler operation and any Measure the measurement data at the time of measurement, measure the change between the high frequency specific frequency and the low frequency specific frequency with these steady data and measurement data, and measure the change in the boiler furnace building based on these change data. A method for inspecting a tube leak of a boiler, characterized in that a leak of a disposed tube is determined, and secondly, after determining a leak, The method for determining a leak of a boiler according to the above-mentioned 1), characterized in that the leak is judged again when the break lasts for a predetermined time. Thirdly, a criterion for judging leak, noise and load fluctuation is set. A membership function for determining the grade of the variation data for each of these judgment criteria is set in advance, the grade of the variation data is determined based on these membership functions, and the boiler is based on the determined grade value. A method for inspecting a tube leak of a boiler according to the first or second aspect, characterized in that a leak of the tube is determined.
Further, after determining that a leak has occurred, the location of the leak is detected based on the variation data of each vibration sensor. Boiler tube leak inspection method, No. 5
In addition, a plurality of vibration sensors are installed on the outer wall of the boiler furnace building where a plurality of tubes are arranged inside, and the vibration data from these sensors is converted into digital data.
A D converter is installed in the furnace building, and a control device is installed at a predetermined distance from the furnace building to connect with the AD converter via an optical fiber cable, and the optical fiber cable is installed in the control device. A tube of a boiler, comprising: a frequency analysis circuit for frequency-analyzing digital data sent via the tube; and a determination means for determining the leakage of the tube based on the data analyzed by the frequency analysis circuit. Leak inspection device, 6th
In addition, the AD converter is connected to a plurality of sensors and sends out a signal of a vibration sensor designated based on a command from a control device. The control circuit causes the measurement data of each of the plurality of vibration sensors to be measured. A memory circuit for storing
In the fifth aspect, a frequency analysis circuit for performing frequency analysis for each data stored in the memory circuit and a storage device for storing analysis data of all vibration sensors analyzed by the circuit are provided. It is composed of the tube leak inspection device of the described boiler.

[0008]

According to the first method, the steady data in which the boiler is normally operating is compared with the measurement data at the time of measurement, and the high frequency specific frequency and the low frequency specific frequency of both data are compared. Since the amount of change is measured and the determination is made based on the amount of change data, it is possible to absorb variations in the sensitivity of each sensor, and there is no need to perform sensitivity correction for each sensor.

According to the second method, it is possible to judge that "leak has occurred" only when the leak has continued for a predetermined time, and it is possible to make a more reliable judgment.

According to the third method, the grade of the change is determined by the membership function for each criterion of the leak, noise and load fluctuation, and the leak is determined by the so-called fuzzy inference based on the determined grade value. Therefore, it is not necessary to set a detailed judgment standard for each sensor or each boiler, and it is easy to transfer to another boiler or another device.

According to the fourth method, it is possible to detect at which position in the furnace building the leak is occurring based on the variation data of each vibration sensor.

According to the apparatus described in the fifth aspect, each vibration sensor installed in the boiler furnace building detects the vibration of the furnace wall based on the sound generated from the tube in the furnace building, and the detected vibration data is AD. It is converted into digital data by the converter, sent out to the control device through the optical fiber cable, frequency-analyzed by the frequency analysis circuit in the control device, and based on the analysis result, the determination means causes the current tube to leak. Is determined. Therefore, it is not necessary to install the sensor directly on the tube, and since the vibration data is once converted into digital data, it is not affected by external noise.

According to the sixth device, since a plurality of vibration sensors can be collectively connected to the AD converter, the number of connection cables between the vibration sensor and the control device can be reduced and the cost can be reduced. Further, each vibration data sent by the optical fiber cable is temporarily stored in the memory circuit, frequency analysis is performed by the frequency analysis circuit for each stored data, and all the frequency-analyzed data is stored in the storage device. It is possible to accurately determine whether a leak is currently occurring in the tube based on all the stored analysis data.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT An embodiment of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 shows the overall construction of a thermal power plant in which the tube leak detection device of the present invention is installed. 1 is a boiler furnace building, 2 is a boiler in the furnace building 1, 3
Is a boiler smoke path, 4 is a boiler tube through which high temperature steam generated by boiler heat passes, and a plurality of vibration sensors are provided at predetermined positions (position A in FIG. 1) on the smoke path side of the outer wall of the furnace ridge 1 corresponding to the boiler tube 4. Is provided. Reference numeral 5 denotes a control room installed at a position about 300 m away from the furnace building 1, which is equipped with a control device internally connected to the vibration sensor by an optical fiber cable. You can monitor for leaks.

FIG. 2 is a perspective view of the boiler furnace building 1 showing the installation positions of the vibration sensors. The outer walls of the three floors 3a to 3c on the smoke path 3 side of the furnace building 1 have eight floors each. A total of 24 vibration sensors 6, ... Are installed. The vibration sensor 6, ... uses a piezoelectric accelerometer with a built-in preamplifier. As shown in FIG. 3, the M5 tap 7 is driven from the outside of the furnace wall until its tip contacts the inner wall of the furnace, and the tap 7 An adapter 9 is screwed onto the rear end of the adapter through an SM nut 8 and fitted and fixed to the end of the adapter 9.

FIG. 4 is a block diagram showing the overall construction of the detection apparatus of the present invention. In the figure, 10a to 10c are AD conversion terminals installed on the wall surface of the furnace building 1, and the eight vibration sensors 6, ... Of each floor are 10 m to the AD conversion terminals 10a to 10c, respectively.
It is connected by a low noise cable L of 20 m to 20 m. These AD conversion terminals 10a and 10b are for converting the detection signals (analog signals) from the respective sensors 6, ... To digital signals, and are specifically configured as shown in FIG. Since the AD conversion terminals 10a to 10c have the same configuration, only the configuration of the terminal 10a is shown in FIG.

In the figure, 11 is an input buffer amplifier for amplifying a detection signal input from each of the eight sensors 6, ...
The input selection circuit receives the signal from the CPU and selects the input based on the command from the CPU. The 13 removes unnecessary low frequency when the boiler is operating, and high frequency (48 KHz or more) unnecessary for sampling. A filter circuit for removing the signal, 14 is a 16-bit AD converter with a built-in digital filter for converting the signal from the filter 13 into a digital signal, and 15 is a pulse-shaped detection signal converted into a digital signal by the AD converter 14. The electro-optical converter 16 modulates the optical signal and converts it into an optical signal and sends it to the control device 18 through the optical fiber cable F, and 16 is a C in the control device 18.
A photoelectric converter for converting a channel designation signal sent from the PU 23 via the optical fiber cable F into an electric signal, and 17 is an interface for sending the command signal to the selection circuit 12.

Reference numeral 18 (see FIG. 4) is a control device in the control room 5 which is connected to the AD conversion terminals 10a to 10c by an optical fiber cable F. In the control device 18, 19 and 20 are connected to the three optical fiber cables F from the AD conversion terminals 10a to 10c, and an electro-optical converter and a photoelectric converter for performing necessary photoelectric / electro-optical conversion with each terminal. Reference numerals 21 and 21 are memory circuits having a memory of 3M words, and the detection data for each vibration sensor 6 sent from the AD conversion terminals 10a to 10c based on the instruction of the CPU 23 is 100 at maximum.
It has a function of storing at a sampling frequency of KHz for 10 seconds.

Reference numeral 22 is a frequency analysis circuit for frequency-analyzing the detection data stored in the memory circuit 21 by FFT calculation based on an instruction from the CPU 23, and is constituted by a so-called digital signal processor (DSP).

Reference numeral 23 denotes a CPU (judgment means) which stores a procedure relating to a series of operations of the present invention, which gives a command to each circuit according to the stored operation procedure and is analyzed by the frequency analysis circuit 22. Fuzzy inference is performed by the preset membership function based on the data
The leak of the boiler tube 4 is detected.

Reference numeral 24 is a disk file device for storing the analysis data of each channel frequency-analyzed by the frequency analysis circuit 22, 25 is a CRT display device for displaying the monitoring state of the boiler tube 4, and the judgment result, and 26 is each. The printer device outputs frequency analysis data for each channel.

Next, the operation of the present invention will be described with reference to the flowcharts of FIGS. First, the CPU 23 sends a channel designation signal in order to measure steady data of each of the vibration sensors 6 ... When the boiler is operating normally. The signal is sent to the AD conversion terminal 10a via the electro-optical converter 20 and the optical fiber cable F (P1).

The specified signal is converted into an electric signal by the photoelectric converter 16 of the AD conversion terminal 10a and demodulated, and then sent to the input selection circuit 12 through the interface 17. On the basis of the designated signal, the circuit 12 designates a signal of a designated specific channel (see FIG. 6b) among the signals from the vibration sensor 6, ... It is sent to the circuit 13. The signal is a detection of vibration when the boiler is operating normally.

After the unnecessary frequency is cut by the circuit 13 and the signal is converted into digital data by the AD converter 14,
The signal is pulse-modulated by the electro-optical converter 15 and transmitted as an optical signal to the optical fiber cable F. This optical signal is the control device 1
After being received at 8 and converted into an electric signal by the photoelectric converter 19 and demodulated, the memory circuit 21 based on a command from the CPU 23.
It is stored in (P2).

Thereafter, the CPU 23 again sends the next channel designation signal and the same operation is performed, and the vibration data for all 24 channels is stored in the memory circuit 21 as digital data of 3M words (P3).

Next, the stored data is sequentially sent to the frequency analysis circuit 22 according to a command from the CPU 23, and the circuit 2
2 for frequency analysis (P4). That is, the steady data in FIG. 6c is measured. Thereafter, the frequency-analyzed steady data is stored in the disk file device 24 (P5). Then, the measurement is performed for all the vibration sensors 6, ... (P6) by the above procedure, and the steady data obtained from the all vibration sensors 6 ,.

Next, the leak of the boiler tube is measured (see FIG. 7). First, the CPU 23 sends a channel designation signal for designating a specific channel to the AD conversion terminals 10a to 10c, as in the case of the steady data measurement (P7). The command signal is sent to the AD conversion terminals 10a to 10c, the measurement signal of the specific vibration sensor 6 is converted into a digital signal, and then the optical fiber cable F
Is sent to the control unit 5 via the above-mentioned predetermined modulation processing and is similarly stored in the memory circuit 21 (P8), and such data acquisition is performed for all 24 channels (P8).
9). Thereafter, the measured data is sequentially frequency-analyzed by the frequency analysis circuit 22 (P10) and compared with the previously measured steady data of the same channel (see FIG. 6c).

That is, the changes x ₁ and x ₂ (ratio = measurement data / steady data) between the steady data and the measurement data at the low frequency level (10 KHz) and the high frequency level (30 KHz) are calculated (P11), respectively. The change amount data is stored in the disk file device 24 (P12). The above-mentioned change is obtained by subtraction because each measured value is expressed in [dB]. Then, the change amount is calculated for all the vibration sensors (P13), and the change amount data for all the sensors is stored in the disk file device 24.
Remember.

The measurement data is taken in every 5 minutes, and it is possible to store the change data for the past several days in the disk file device 24.

Next, fuzzy processing is performed based on the variation data stored in the file device. CPU23
Has a fuzzy inference circuit shown in FIG. That is,
The leak determination circuit 27 for the leak determines the grade of the variation data by the membership function M1 of the high frequency level difference, the membership function M2 of the low frequency level difference, and the membership function M3 of the duration, and the noise related noise The determination circuit 28 determines the grade of the variation data by the membership function M4 of the high frequency level difference and the membership function M5 of the low frequency level difference, and the load variation determination circuit 29 concerning the load variation determines the member of the high frequency level difference. The grade of the variation data is determined by the ship function M6 and the membership function M7 of the low level difference. Then, the MINI circuits 30, 31, 32, and M based on the grades determined from these three viewpoints regarding the high-frequency level difference, low-frequency level difference, and leak duration of the variation data.
The AX circuit 33 is configured to obtain an inference result.

The membership functions M1 to M7 are
The inventors of the present invention have prepared it based on the judgment criteria shown in Table 1 from the results of actual measurement of leak, load fluctuation, and noise at a thermal power plant.

[0032]

[Table 1]

That is, regarding the leak, rule a in Table 1 is used.
As shown in, the change in the low frequency level difference (around 10 KHz) is slightly large, and the high frequency level difference (around 30 KHz).
If the change of is large and the duration is long, it can be inferred that it is a "leak". The type of leak can be inferred as shown in rules b to f depending on the degree of each membership.

Regarding load fluctuation, as indicated by rule g in Table 1, if the change in the low frequency is slightly small and the change in the high frequency is equal or small, it is inferred to be "load change". be able to. Furthermore, the types of load fluctuation can be inferred as shown in rules h to j depending on the degree of each membership.

Regarding noise, it can be inferred that "there is a suspicion of artificial noise" when there is a large change in the low frequency and no change in the high frequency as indicated by rule k in Table 1.

Next, the fuzzy processing by the CPU 23 will be described with reference to the flowchart of FIG. 8 and FIG.
The CPU 23 takes out the change data of the specific channel from the disk file device 24, and these data are, for example, +5 dB for the high frequency level difference and -5 dB for the low frequency level difference.
, The high level difference is M1, M4, M6
The membership function of M2, M for the low level difference
When the membership functions of M5 and M7 are applied, the grades of M1 and M2 are 0.8, 0.
6, regarding noise, the grades of M4 and M5 are 0.3 and 0, respectively, and regarding load fluctuations, the grades of M6 and M7 are 0.2 and 1.0, respectively (P14). The MINI circuits 30, 31, 32 are used for each criterion.
When the minimum grade value is taken, leak is 0.6, noise is 0, and load fluctuation is 0.2 (P15). Then, the maximum grade is taken by the MAX circuit 33, which is 0.6 (P16). ). Therefore, it is determined that the criterion (leakage) for the maximum value of this grade is currently occurring, that is, "leakage is occurring".

In this case, the CPU 23 increments the counter circuit 35 incorporated therein by 1 (P17, 18) and judges the membership function M3 of the duration (P20).
That is, the leak duration is measured by the count value of the counter circuit 35. When it is determined in P17 that the grade of noise or load fluctuation is maximum and no leak has occurred, the counter is cleared to 0 (P19), and the duration is determined.

When the grade of the membership function M3 of the duration is 1.0, the MAX circuit 33 re-evaluates the grade value including the grade value (P21, 22). As a result, as shown in FIG. 9, when the grade 0.6 of the leak judgment circuit is the maximum,
It is judged that "leak has occurred"("abnormal") (P
23), that is displayed on the CRT display device 25. If the maximum value of the grades is the grade of the noise judgment circuit 28 and the grade of the load fluctuation judgment circuit 29, it is judged as "normal". In this case, the changed data after a lapse of a predetermined time is fetched again from the disk file device 24 (P25), and the fuzzy processing from P14 is performed again.

Next, when "abnormal" is detected, the place where the leak is occurring is determined. After determining that there is an abnormality in P23, the variation data of 10 KHz at that time is examined over all the vibration sensors 6, ..., And the vibration sensor showing the largest value of the variation data (for example, FIG. 2).
Check the installation location of the vibration sensor 6a) (P2)
6).

After that, the variation data of the vibration sensor 6a is compared with the variation data of the vibration sensors 6 around the sensor 6a, and based on the comparison data, the position of the tube in the furnace building Then, the position, depth, etc. of the leak are detected (P27). This makes it possible to confirm at which position in the furnace building the leak has occurred. In this case, the reason why the leak position is detected based on the change amount data of 10 KHz is that a signal of a low frequency has a lower directivity and is less affected by the direction of the leak ejected from the tube.

The various results detected in P23, P24, and P27, that is, whether or not a leak has occurred, the position where the leak has occurred, can be displayed on the CRT display device 25, and the measured frequency data can be displayed. Printer device 2 etc.
It can be output at 6. In FIG. 1, 35a is a coal bunker, 35b is an air preheater, 35c is various dust collectors,
5d is a conveyor and 35e is coal.

[0042]

[Effect] As described above, according to the present invention, the steady data in which the boiler is normally operating is compared with the measurement data at the time of measurement, and the high frequency specific frequency and the low frequency specific frequency of these data are compared. Since the amount of change is measured and the determination is made based on the amount of change data, it is possible to absorb variations in the sensitivity of each sensor, and there is no need to perform sensitivity correction for each sensor.

Moreover, since the grade of the change amount is determined by the membership function for each of the judgment criteria of leak, noise and load fluctuation, and the leak is determined by so-called fuzzy inference based on the decided grade value, each sensor is determined. Alternatively, it is not necessary to set a detailed determination standard for each boiler, it can be easily diverted to another boiler or another device, and the leak occurrence position can be detected.

Further, since the vibration sensor is mounted on the wall surface of the furnace ridge, it is not necessary to directly install the sensor on the tube, there is no need to use a special sensor designed for heat resistance, and it is also necessary to stop the boiler at the time of mounting. Not efficient. Further, since the vibration data is once converted into digital data and sent to the control device by the optical fiber cable, it is not affected by external noise.

Furthermore, since a plurality of vibration sensors can be collectively connected to the AD converter, the number of connection cables between the vibration sensor and the control device can be reduced and the cost can be reduced.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing a thermal power plant to which a tube leak inspection device for a boiler according to the present invention is applied.

FIG. 2 is a perspective view of a boiler furnace building in which a vibration sensor is installed.

FIG. 3 is a side sectional view of an essential part of a state where a vibration sensor is installed in a boiler furnace building.

FIG. 4 is a block diagram of a tube leak inspection device.

FIG. 5 is a block diagram of an AD conversion terminal.

FIG. 6a is a flowchart of steady data measurement.

FIG. 6b is a characteristic diagram showing vibration data measured by a vibration sensor.

FIG. 6c is a frequency characteristic diagram obtained by frequency-analyzing vibration data.

FIG. 7 is a flowchart of measurement data measurement.

FIG. 8 is a flowchart showing fuzzy processing.

FIG. 9 is a block diagram showing a fuzzy processing function in the CPU.

FIG. 10 is a perspective view showing a conventional tube leak inspection device.

[Explanation of symbols]

 1 Boiler furnace building 4 Tube 6 Vibration sensor 18 Control device 21 Memory circuit 22 Frequency analysis circuit 23 CPU 24 Disk file device F Optical fiber cable M1 to M7 Membership function

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kazutoshi Nomura 1-6-18 Sanno, Hakata-ku, Fukuoka City Kyushu Measuring Instruments Co., Ltd.

Claims (6)

[Claims] 1. A vibration data from a plurality of vibration sensors installed on an outer wall of a boiler furnace building is converted into digital data, the digital data is sent to a control device through an optical fiber cable, and the digital data is sent to the control device. Frequency analysis is performed on the digital data of each of the vibration sensors to measure steady data during normal boiler operation and measurement data at any measurement for all vibration sensors. Boiler tube leak inspection characterized by measuring the amount of change between a specific frequency in the low range and a specific frequency in the low range and judging the leakage of the tubes installed in the boiler furnace building based on the data of these changes. Method. 2. The method for inspecting a tube leak in a boiler according to claim 1, wherein after the leak is judged, the leak is judged again when the leak lasts for a predetermined time. 3. A membership function for setting judgment criteria for leak, noise, and load fluctuation, and for determining the grade of the variation data for each of these judgment criteria is set in advance, and based on these membership functions, a membership function is set. The tube leak inspection method for a boiler according to claim 1 or 2, wherein a grade of the variation data is determined, and a leak of the boiler tube is determined based on the determined grade value. 4. The method according to claim 1, wherein the location of the leak is detected based on the change data of each vibration sensor after determining that the leak has occurred. Boiler tube leak inspection method. 5. A plurality of vibration sensors are installed on the outer wall of a boiler furnace building in which a plurality of tubes are arranged inside, and an AD converter for converting vibration data from these sensors into digital data is provided in the furnace building. Installed, and further installed a control device at a place distant from the furnace building by a predetermined distance, connected to the AD converter via an optical fiber cable, and sent to the control device via the optical fiber cable as digital data. A tube leak inspection apparatus for a boiler, comprising: a frequency analysis circuit for performing a frequency analysis of the above; and a determination unit for determining a leak of the tube based on the data analyzed by the frequency analysis circuit. 6. The AD converter is connected to a plurality of sensors and sends out a signal of a vibration sensor designated based on a command from a control device. The control circuit comprises:
A memory circuit that stores the measurement data of each of the plurality of vibration sensors, a frequency analysis circuit that analyzes the frequency of each of the data stored in the memory circuit, and a memory that stores the analysis data of all the vibration sensors analyzed by the circuit. 6. The boiler tube leak inspection apparatus for a boiler according to claim 5, further comprising: