KR100888320B1 - Iternal valve leak diagnosis apparatus and method the same - Google Patents

Abstract

An internal valve leak diagnosis apparatus and a method thereof are provided to determine leakage inside a valve regardless of various noises generated in surroundings by setting up an ultrasonic acoustic emission sensor in the valve. An acoustic emission sensor(110) is installed at a valve. An acoustic signal gathering unit(100) measures an acoustic signal level which is the duration or the mean value of an event by amplifying the signal detected with the acoustic emission sensor and comparing the specific signal with a set value. An interpreting display part(200) converts the acoustic signal, selected among original signals inputted from the acoustic signal gathering unit with frequency-discriminating, into a digital signal and indicates the digital signal.

Description

{Internal valve leak diagnosis apparatus and method the same}

The present invention relates to an apparatus and method for diagnosing a leak in a valve, and more particularly, to detect a leak state of a microfluid generated inside a valve using an ultrasonic acoustic emission sensor in a state in which ambient noise exists around the valve. The present invention relates to an apparatus and a method for diagnosing a leak in a valve.

In general, a large number of valves are used in industrial facilities such as power plants and chemical plants, and leakages from these valves can cause enormous losses in the operation of industrial facilities, including enormous energy loss, loss of cooling of facilities, and release of toxic and radioactive materials. It will cause an accident. In order to diagnose the internal leakage of such a valve, conventionally, a flow level reduction, an inlet / outlet pressure difference using a pressure gauge, a temperature and humidity measurement method, and the like are used.

However, the above-described conventional method for diagnosing internal leakage of a valve is in need of improvement because the diagnosis procedure is complicated and it is difficult to trust in terms of measurement accuracy due to indirect measurement.

Actual industrial facilities consist of valves of various types and quantities, so the diagnosis must be quick and even a small amount of leakage can be diagnosed. Therefore, conventionally, in the measurement method, the acoustic emission nondestructive testing method using a high sensitivity acoustic sensor capable of rapid diagnosis, high reliability of the diagnosis result, and precise diagnosis is widely used.

However, in the valves used in many industrial sites such as power plants and chemical plants, most of the valves are to diagnose a state where a small amount of fluid leakage occurs before a large amount of fluid leakage occurs in terms of safety accidents and energy saving. It is required.

However, in the acoustic emission nondestructive testing method using the acoustic sensor, there is always a background noise (BGN) of constant or fluctuating magnitude generated from the surrounding equipment, so that it is compared with the level of the detected acoustic signal and mixed detection There is a need for a signal separation procedure for inherent valve leakage. That is, when the differential pressure between the valve inlet and outlet is large and the amount of leakage is large, the acoustic signal at the time of leakage is much larger than the ambient noise, so it is easy to identify them.However, when the differential pressure is small and the amount of leakage is small, the ambient noise is mixed by the acoustic signal It is difficult to accurately identify the unique sound signal according to the occurrence. Therefore, there is a demand for a method for accurately separating and detecting a leakage sound inside a valve regardless of mixing ambient noise in a state where a small amount of fluid leakage occurs.

The purpose of the present invention is to simulate a small amount of fluid leakage occurring inside an actual valve and to simulate ambient noise similar to a site such as a power plant, so that even if a small amount of fluid leaks, a unique sound is identified regardless of ambient noise. The present invention provides a method and apparatus for diagnosing internal leakage of a valve, which is a kind of acoustic emission nondestructive testing method using a sensor.

Leak diagnosis apparatus inside the valve according to the present invention and the acoustic sensor is installed in the valve; An acoustic signal collector which amplifies the signal detected by the acoustic sensor and measures an acoustic signal level which is a duration or an average value of an event which is a specific occurrence signal; And an analysis display unit for receiving the original signals classified by the sound signal collecting unit and performing the A / D conversion of the selected sound signals, analyzing and displaying them, and outputting the necessary alarms. .

The sound signal collecting unit compares a preamplifier for passing only a set frequency signal among the sound signals input from the sound sensor, and compares a signal input from the preamplifier with a threshold value of a set signal size to the sound signal level, duration, and arrival of the event. It is desirable to include a local process that finds time.

In this case, the acoustic sensor is preferably composed of a first acoustic sensor and a second acoustic sensor attached to the valve and spaced apart from the valve in the valve.

Preferably, the preamplifier includes a high pass filter through which only a frequency signal of 20 kHz or more is passed, and the local process includes a power supply unit supplying 24 V of power to the preamplifier.

In addition, the analysis display unit transmits the multiple channels at the same time to control only the selected channel among the multi-channel by the master controller, an A / D converter for A / D conversion of the acoustic signal selectively received by the multi-transmitter and And a master controller for collecting and analyzing data from the A / D converter and the sound signal collecting unit, an analog recorder for recording the level of the sound signal output from the sound signal collecting unit, and a result analyzed by the master controller. It is preferable to include an alarm generator for generating an alarm when is exceeded a preset condition.

According to the present invention, there is provided a method of diagnosing a leak in a valve, the method comprising: setting a threshold value of an acoustic signal and an initial condition of a measurement time; Measuring a sound level and a duration of the sound signal; Comparing the measurement time with the duration; And performing correlation analysis of the acoustic signal.

At this time, it is preferable to generate an alarm when the result of performing the correlation analysis of the acoustic signal is within the range of the sound arrival time difference L / C.

According to the present invention, regardless of various noises generated from the surrounding equipment, there is an effect of early diagnosis of leakage of the inside of the valve to prevent safety accidents.

Therefore, it is possible to reduce the enormous energy that can be caused by the leakage failure in the installation, there is an advantage that the stable operation of the industrial equipment is possible.

Figure 1 is a block diagram of the simulation apparatus for the ambient noise excitation and valve internal leakage for measuring the acoustic signal applied to the present invention.

Since the acoustic signal caused by the leaked sound depends on the inherent frequency characteristic of the acoustic sensor, in this simulation, the acoustic sensor (S1) (S2) with a resonance frequency of 70 Hz is used, and the signal analysis ranges up to 100 Hz. Analyze with

The valve 10 uses a 2-inch gate valve and the internal fluid supplies nitrogen gas by means of a nitrogen gas flow control valve and a flange.

Nitrogen gas has almost the same sound leakage characteristics as steam and is easier to handle than water.

The leaky state of the valve was made by attaching an aluminum tape of about 0.05 mm in thickness and 4.3 mm in width to the valve seat surface so that a small amount of fluid leakage occurred in a gap between the valves and the slits.

The leakage of the fluid is changed to the valve differential pressure, and the pressure at the valve outlet is made to atmospheric pressure. In the experiment, a differential pressure is generated in the valve to generate a leak, and signal measurement is performed by the first acoustic sensor S1 and the second acoustic sensor S2 attached to the valve periphery and the pipe. The first acoustic sensor S1 is installed at a position 0.5 m away from the center line of the valve body 11, and the second acoustic sensor S2 is installed at a position 1 m away from the first acoustic sensor S1, thereby providing a constant interval. Keep it.

In order to maintain a proper level (amplitude) of the ambient noise, the pre-amplifier gives white noise to the preamplifier and simulates them as the ambient noise level. In this case, the white noise is performed by giving an independent noise signal to each channel.

2 is a block diagram of a sound signal duration measurement and correlation analysis device for directly measuring a sound signal applied to the present invention and performing correlation measurement between signals.

The embodiment of the present invention is composed of 16 channels, and two acoustic sensors are attached to one valve, so that eight valves can be measured at the same time.

Embodiments of the present invention can be broadly divided into an AE data acquisition part 100 and an analysis & display part 200.

The acoustic signal collecting unit 100 includes an acoustic sensor 110, a preamplifier 120, a local process 130, and the analysis display unit 200 includes a multi-plexer 210 and a high-speed A / D. The converter 220, the master controller 230, the analog recorder 240 and the alarm generator 250 is configured.

The acoustic signal collecting unit 100 amplifies the signal detected by the 16-channel acoustic sensor, measures the duration of the event (specific occurrence signal), the average value (sound signal level), and the like, and analyzes these results. 200) to output the result.

The acoustic sensor 110 converts the acoustic signal into an electrical signal. Titanium zirconate soft magnetic (PZT) series materials are used for the acoustic transducer, and the maximum operating temperature is about 175 캜.

The preamplifier 120 amplifies the output of the acoustic sensor 110, converts it to a low impedance (impedance) to enable long-range propagation, and sends it to the local process 130.

An interval of about 300 m is suitable between the preamplifier 120 and the local process 130. In addition, in order to remove noise in the low frequency region, a high pass filter is built in to receive only a frequency signal higher than about 20 kHz, which is the main noise level of a plant such as a power plant, and the power of the preamplifier 120 is a local process. To be supplied by 130.

The local process 130 obtains the duration of an event by comparing the threshold value of a predetermined signal size by algebraically or linearly amplifying the signal from the preamplifier 120 by a frequency discrimination by a filter. Processing such as calculation is performed. The result of the average value, peak value, and duration obtained as described above is output to the display apparatus or the output apparatus. A power supply unit for supplying power of 24V to the preamplifier 120 is provided.

The analysis display unit 200 receives the original signal (acquisition data) classified by the frequency of the filter of the local process 130, selects the acoustic signal of two channels among them, and performs A / D conversion (analog to digital). Simultaneously with the conversion, data are received from the local process 130 and the A / D converter 220, and the data are interpreted and displayed, and a necessary alarm is output.

The multi-plexer 210 can transmit signals simultaneously to multiple channels of up to 16 channels, and two channels selected from the 16 channels of input are selected by the control of the master controller 230.

The high speed A / D converter 220 converts a signal selected and received by the multi transmitter 210 to high speed A / D and transmits the signal to the master controller 230.

The master controller 230 is a central processing unit that controls the operation of the entire apparatus. The master controller 230 collects and analyzes data from the local process 130 and the A / D converter 220, and the result is set in advance. In case of exceeding the alarm generator 250 to emit an alarm.

The analog recorder 240 records the level (including BGN) of the sound signal output from the local process 130.

3A to 3F are flowcharts and graphs for explaining a method of diagnosing a valve leakage by measuring a sound signal duration and a correlation analysis applied to the present invention.

Among the acoustic sensors 110, the first acoustic sensor S1 is located outside the valve body 11 closest to the valve disc, and the second acoustic sensor S2 is 0.5 to a distance L from the first acoustic sensor S1. Attach so that it is about 1.0 m. Acoustic signals obtained from both sensors S1 and S2 are amplified by a pre-amplifier 120 and a LOG (logarithm) amplifier. Measure the duration of the event above the threshold level. These measurement results are output to the leak determination unit (control unit) at every preset measurement time, and the presence or absence of abnormality is determined.

In the method of diagnosing valve leakage by acoustic signal correlation analysis, first, an initial condition necessary for measurement such as a threshold level and a measurement time is set.

The sound level and duration of the measured sound signal event are then measured.

The measurement of the duration of the event is to calculate the sum of the times when the signal level exceeds the threshold between the measurement times To, as shown in Figs. 3C and 3D. If there is no abnormality in the signal, it becomes To> D1 + D2 + .. + Dn as in the signal waveform of FIG. 3C. However, if there is an abnormality, the signal level exceeds all thresholds between To times as shown in FIG. 3D. To = D

By using this method, it is possible to measure valve leakage regardless of ambient noise. Therefore, in comparison with the measurement time and the duration, in the case of To > D1 + D2 + ... Dn, it is determined that there is no abnormality, and the steady state measurement is repeated. However, when To = D, it is impossible to know what abnormality is. Therefore, the acoustic signal correlation analysis of the first acoustic sensor S1 and the second acoustic sensor S2 is performed.

After measuring the signal arrival time difference between the two signals, the waveforms of FIGS. 3E and 3F are observed, and the sound velocity of the sound signal received from the sensor and the interval L between the two sensors S1 and S2. ) L / C, which is a ratio between C, that is, the time difference between the signal arrival times at which the generated acoustic signals reach both sensors, and the presence or absence of an event from the L / C values obtained within a predetermined time period are investigated. do.

That is, if an event occurs in the signal arrival time difference L / C within the set time, it is determined that there is an acoustic source due to leakage from the valve abnormally, and if an event does not appear in the signal arrival time difference L / C within the set time. Normally, it is determined that there is no acoustic source due to leakage from the valve. At this time, even if it is normal, the process returns to the measurement command group in the normal state and repeats the judgment in the above order.

The present invention uses the three measurement elements of the correlation analysis of the acoustic level, the event duration, and the signal delivery time difference, and performs the double check of the duration and correlation analysis of the event to perform leakage determination and leakage by the acoustic signal level. You can finally check the degree.

4A and 4B are explanatory views of a sound signal duration measuring method applied to the present invention.

Referring to FIG. 2, the duration D and the sound signal level of the sound signal event measured in the local process 130 after the measurement is started are acquired by the master controller 230 for the total To of the measurement time. At this time, the duration sent to the master controller 230 is the sum of the time D1, D2 ....... Dn when the original signal of the acoustic signal is received during the To time and the acoustic signal exceeds the threshold. (③ + ④ + ⑤ or ⑩). If there is no abnormality in the signal, as in the signal waveform of FIG. 4A, To> D1 + D2 + ..... + Dn. However, if there is an abnormality, as shown in FIG. To = D

The duration and the value of the sound signal level are displayed on the diagnostic screen of the master controller 230. If the duration is equal to the measurement time To, it can be regarded that there is any abnormality in the corresponding channel. If there is no abnormality in the duration, check whether there is a change in the operating conditions of industrial plants such as power plants, and if there is a change, reset each threshold value and repeat the measurement of the duration and sound signal level. .

5 is a principle diagram of acoustic signal correlation analysis according to the present invention.

The measuring principle of the correlation analysis used in the present invention can be simply expressed as follows. When the acoustic sensors S1 and S2, which are two acoustic sensors, are installed in the valve at an interval L, the output of the acoustic signals received from the sound source can be represented by the following equations (1) and (2), respectively.

V1 (t) = x (t) + n1 (t) (1)

V2 (t) = αx (t-τ1) + n2 (t) (2)

Where Vi (t) is the signal received from the acoustic sensor Si

x (t): signal from sound source

ni (t): disturbance other than x (t) received by acoustic sensor Si

α: Attenuation coefficient of signal x (t) between acoustic sensors S1, S2

t: time

τ1: Signal arrival time difference between acoustic sensors S1 and S2 (= L / C)

C: sound speed

Now, x (t), n1 (t), and n2 (t) are mutually uncorrelated and can be expressed by the following expression (3) if the two acoustic sensor outputs are correlated.

Here, C ₁₂ (τ): cross correlation function of V1 (t), V2 (t)

τ: random time delay

Here, if the sound source is in the valve, x (t) and αx (t-τ1) are correlated and the correlation coefficient C ₁₂ (τ) / [C ₁₂ ] max of the outputs of the two acoustic sensors is τ1 = L / It is the maximum in C. That is, in the case where the distance L of the acoustic sensor is determined and the valve is a sound source, the correlation coefficient becomes maximum at a time delayed by the acoustic arrival time tau 1. The time τ1 at which the correlation coefficient is maximized depends on the direction in which sound propagates through the pipe connected to the valve, and the maximum value at the position as shown in the drawing only when the propagation direction of the sound is S1-> To appear. As a result, the sound from the reverse direction can be removed by the ambient noise.

6 is a diagram illustrating an example of a diagnosis screen at the time of measuring a duration and a leakage sound level according to the present invention.

In the drawing, part 1 indicates an area representing the diagnostic channel CH, sound signal level BGN, threshold REF, and duration D, and part 2 indicates supply voltage VOLT and sound signal measurement time. It is an area indicating (To) and the current time (REAL TIME). The part of ③ indicates unused diagnostic channel. In the state where there is no leakage sound, the acoustic signal level BGN is 49 dB, which is 6 dB lower than the threshold 55 dB, and the duration is also 0.

When the gas is leaked from the inside of the valve in such a state, the leakage sound level exceeds the threshold value, the duration D of the event is equal to the measurement time To = 60 sec, and the state is abnormal. Therefore, the channel with the abnormality turns to the blinking display, and the processing moves to the cross correlation analysis.

7 is an exemplary view showing a diagnosis screen when there is a valve leakage by correlation analysis according to the present invention.

In the figure, the left part is an acoustic signal waveform indicating the result of the correlation analysis, and the right part is an area indicating the diagnostic channel CH, the acoustic signal level BGN, the threshold value REF, and the duration D. The left part shows the acoustic signal arrival time (L / C), measurement time (To) and supply voltage (A / D) of the measurement site.

The cross-correlation coefficient between channel 1 (CH1) and channel 2 (CH2) is calculated, and whether the acoustic arrival time difference (L / C) is within the range of that value compared with a preset value (0.26 to 0.36 msec). Check it. The diagnosis screen shows the case of valve leakage because the sound arrival time difference is 0.33 msec and falls within the preset sound arrival time difference range. Each acoustic sensor determines that a valve leakage sound has been detected and sounds an alarm buzzer so that the necessary information of the abnormal channel is stored in the memory of the master controller.

8 is an exemplary view showing a diagnosis screen when there is no valve leakage by correlation analysis according to the present invention.

The display form of the figure is the same as that of FIG. However, FIG. 8 is an example in which each acoustic sensor detects mutually unrelated sounds other than the valve internal leakage sound, and shows a result of generating white noise unrelated to each channel. In this case, the correlation coefficient does not show a clear peak (signal value of maximum magnitude), and the sound arrival time difference L / C does not fall within a preset range (0.26 to 0.36 msec). Therefore, in the case of FIG. 8, the sound signal level BGN exceeds the threshold but the sound source exists outside the valve.

As in the above-described embodiment, the present invention uses a ultrasonic acoustic emission sensor in the presence of ambient noise around the valve, the leakage state of the minute fluid generated inside the valve Apparatus and method for determining the circumference of the valve through the measurement and interpretation of three measurement elements: the acoustic signal level, the duration and the time of arrival of the signal. Provided are an apparatus and a method for accurately determining whether a leak is present in a valve regardless of ambient noise present.

1 is a block diagram of a valve internal leakage simulation apparatus applied to the present invention.

Figure 2 is a block diagram of the acoustic signal duration measurement and correlation analysis device applied to the present invention.

3a to 3f is a flow chart and graph of the valve leakage diagnosis method by measuring the acoustic signal duration and correlation analysis applied to the present invention.

4A and 4B are explanatory views of a sound signal duration measuring method applied to the present invention.

5 is a sound signal correlation analysis principle applied to the present invention.

Figure 6 is an illustration of a diagnostic screen at the time of measuring the duration and leakage sound level applied to the present invention.

7 is a diagram illustrating a diagnosis screen when there is a valve leakage by correlation analysis according to the present invention.

Figure 8 is an illustration of a diagnostic screen when there is no valve leakage by the correlation analysis according to the present invention.

<Description of Symbols for Main Parts of Drawings>

10; Valve 11; Valve body

100; Sound signal collectors 110, S1, and S2; Acoustic sensor

120; Preamplifier 130; Local Process

200; Analysis display section 210; Multi transmitter

220; A / D converter 230; Master controller

240; Analog recorder 250; Alarm generator

Claims (8)

An acoustic sensor installed in the valve; An acoustic signal collector which amplifies the signal detected by the acoustic sensor and measures an acoustic signal level which is a duration or an average value of an event which is a specific occurrence signal; And an analysis display unit which receives the original signals classified by the sound signal collection unit, performs an A / D conversion on the selected ones, analyzes them, displays them, and outputs the necessary alarms. Valve internal leak diagnosis device. The method of claim 1, The sound signal collecting unit A pre-amplifier for passing only a set frequency signal among the sound signals input from the sound sensor, and a local process for comparing the signal input from the pre-amplifier with a threshold value of a set signal size to obtain an acoustic signal level, duration and arrival time of the event Valve internal leakage diagnostic apparatus comprising a. The method according to claim 1 or 2, The acoustic sensor is a valve leakage diagnosis device, characterized in that consisting of the first acoustic sensor and the second acoustic sensor attached to the valve and spaced around the valve in the valve. The method of claim 2, The preamplifier has a built-in high-pass filter for passing only a frequency signal of 20 kHz or more leakage diagnostic device. The method of claim 2, And the local process is provided with a power supply for supplying 24V power to the preamplifier. The method of claim 1, The analysis display unit A multi-transmitter which transmits multiple channels simultaneously and controls only a selected channel among the multi-channels by a master controller, an A / D converter for A / D converting an acoustic signal selectively received by the multi-transmitter, and the A / D converter And a master controller for collecting and analyzing data from the sound receiving collection unit, an analog recorder for recording the level of the sound signal output from the sound receiving collecting unit, and a condition for which a result analyzed by the master controller is preset. And an alarm generator for generating an alarm if exceeded. Setting a threshold and a measurement time initial condition of the acoustic signal; Measuring a sound level and a duration of the sound signal; Steps to compare the measurement time and duration: And performing correlation analysis of the acoustic signal. The method of claim 7, wherein And an alarm is generated when the result of performing the correlation analysis of the acoustic signal is within the range of the acoustic arrival time difference (L / C).

Images