JP7792359B2 - Method and system for identifying leakage phenomena using three-dimensional image analysis - Google Patents

Description

The present invention relates to a method and system for identifying leakage phenomena using three-dimensional image analysis, which is capable of identifying fluid leakage phenomena in pipelines.

In pipeline systems that transmit and distribute water, leaks and water hammer phenomena affect the longevity of the facilities. However, because these phenomena are unavoidable when using water, establishing inspection techniques for pipeline systems is essential from a maintenance and management perspective.

Under these circumstances, the technologies disclosed in Non-Patent Documents 1 and 2 consider establishing a detailed detection method for leaks in irrigation pipelines by adding information about the leak to the water hammer pressure caused by valve operation and using numerical simulations.

Furthermore, the technology disclosed in Non-Patent Document 3 considers a direct measurement method in which a capsule-type detection device is inserted into the pipeline and the location of the leak in the pipeline is identified from the leak sound picked up by the detection device.

Yohei Asada and four others, "Elucidation of the damping mechanism of pressure waveforms in pipelines," [online], Japan Society of Civil Engineers, [searched October 20, 2022], Internet <https://www.jstage.jst.go.jp/article/jscejhe/74/5/74_I_769/_article/-char/ja/> Yohei Asada and four others, "Study on the applicability of a leak detection method using energy attenuation in pipes," [online], Japan Society of Civil Engineers, [searched October 20, 2022], Internet <https://www.jstage.jst.go.jp/article/jscejhe/75/2/75_I_799/_article/-char/ja/> Isamu Asano and five others, "Development of a Capsule-Type Leak Detection Device for Pipelines," [online], Japan Society of Irrigation, Drainage and Forestry Engineering, [Retrieved October 20, 2022], Internet <https://www.jstage.jst.go.jp/article/jjsidre/86/6/86_513/_article/-char/ja/>

However, the methods disclosed in Non-Patent Document 1 and Non-Patent Document 2 require the installation of a water pressure gauge in the pipeline to measure the water pressure, and it is anticipated that installing a water pressure gauge will be difficult, particularly in existing aging pipelines.

Furthermore, the method disclosed in Non-Patent Document 3 requires the injection of a φ55mm x 165mm capsule with a built-in underwater microphone into the pipeline, making it particularly difficult to apply to pipelines with small diameters. In addition, the capsule must be retrieved, which poses challenges such as the need to process the existing pipeline for injection and retrieval, as well as the risk of being unable to retrieve the capsule in the unlikely event that it becomes impossible to do so.

The present invention aims to provide a method and system for identifying leakage phenomena using three-dimensional image analysis, which can identify fluid leakage phenomena in existing pipelines by capturing the deformation behavior of the pipeline's pipe material using image analysis techniques, without requiring any processing of the pipeline.

The present invention is a method for identifying fluid leakage phenomena in a pipeline, characterized by comprising at least an analysis surface installation step of installing an image analysis surface on the surface of a pipe material of the pipeline; an analysis surface photographing step of photographing the image analysis surface with an imaging means; a displacement amount measurement step of measuring the circumferential displacement of the pipe material based on the image photographed by the imaging means; and a leakage phenomenon determination step of performing an analysis based on the measured circumferential displacement amount, thereby making it possible to determine at least the presence or absence of a fluid leakage phenomenon.

The configuration of the present invention makes it possible to capture the deformation behavior of the pipe material of a pipeline and identify fluid leakage phenomena in the pipeline using image analysis techniques, without having to modify the existing pipeline, even in the limited space of a manhole, for example.

FIG. 1 is a flow diagram of a process from experiment to analysis and non-contact detection using a model pipeline according to the present invention. (a) is a table showing the specifications of the pipe material of the model pipeline, and (b) is a table showing the detailed breakdown of the cases studied in the experiment. FIG. 2 is a schematic diagram illustrating the configuration of a model pipeline. FIG. 10 is a schematic diagram illustrating an imaging mode for a pipeline. FIG. 1 is a schematic diagram illustrating a cylindrical coordinate system model for a pipeline. 10A is a visible image of the image measurement analysis plane A, and FIG. 10B is a three-dimensional analysis image of the image measurement analysis plane A. FIG. 1 is a schematic diagram showing measurement conditions in a model pipeline. 10A is a graph showing the water pressure fluctuations in the pipe and the circumferential displacement of the pipe body over time for Case A, and FIG. 10B is a graph showing the water pressure fluctuations in the pipe and the circumferential displacement of the pipe body over time for Case B. (a) shows the water pressure fluctuations in the pipes of Case C, and (b) shows the water pressure fluctuations in the pipes of Case D. 10A is a graph showing the water pressure fluctuations in the pipe and the circumferential displacement of the pipe body over time for Case E, and FIG. 10B is a graph showing the water pressure fluctuations in the pipe and the circumferential displacement of the pipe body over time for Case F. 10A is a graph showing the water hammer pressure in Case C and Case D over time, and FIG. 10B is a graph showing the frequency distribution of the water hammer pressure. 10A is a graph showing the circumferential displacement amounts of Case C and Case D in time series, and FIG. 10B is a graph showing the frequency distribution of the circumferential displacement amounts. 1A is a graph showing the water hammer pressure in Case E and Case F over time, and FIG. 1B is a graph showing the frequency distribution of the water hammer pressure. 10A is a graph showing the circumferential displacement amounts of Case E and Case F in time series, and FIG. 10B is a graph showing the frequency distribution of the circumferential displacement amounts. 10A and 10B are scalograms of water pressure fluctuations in the pipe and the circumferential displacement of the pipe body, showing the time-frequency characteristics of water hammer pressure for Case C and Case D, respectively. 10A and 10B are scalograms of water pressure fluctuations in the pipe and the circumferential displacement of the pipe body, showing the time-frequency characteristics of the circumferential displacement for Case C and Case D, respectively. 10A and 10B are scalograms of water pressure fluctuations in the pipe and the circumferential displacement of the pipe body, showing the time-frequency characteristics of water hammer pressure for Case E and Case F, respectively. 10A and 10B are scalograms of water pressure fluctuations in the pipe and the circumferential displacement of the pipe body, showing the time-frequency characteristics of the circumferential displacement for Case E and Case F, respectively.

Below, with reference to the drawings, an embodiment of the leakage phenomenon identification method and identification system using 3D image analysis of the present invention will be explained, taking as an example a method for identifying leakage phenomena in a water transmission/distribution pipeline.

(Model pipeline experiment)
Prior to the present invention, a model pipeline was fabricated and experiments were conducted to detect the presence or absence of simulated water leaks in the model pipeline and the behavior of the pipe body that occurs with valve operation in a non-contact manner using digital image correlation (hereinafter referred to as the "DIC method"), which is one of the image analysis techniques.

In other words, as shown in the flow diagram in Figure 1, in this experiment, different water pressure fluctuation cases were set up depending on the presence or absence of a leak and the valve opening degree, and the water pressure and image data acquisition were synchronized to measure water pressure fluctuations and pipe deformation (the pipe deformation behavior was evaluated by comparing it with the water pressure fluctuations inside the pipe).

The specifications of the pipe material for the model pipeline are as shown in Figure 2(a). The appearance of the model pipeline is as shown in Figure 3. A water tank (not shown) is installed at the most upstream point, and the flow rate is adjusted by operating valve B at the most downstream point. The water tank is fixed to a frame 2.9 m above ground level, ensuring a pressure head of approximately 7.8 m at the most downstream point.

The simulated leak in the model pipeline was reproduced by creating two 5mm diameter leak holes H on the downstream side, as shown in Figure 3. In addition, to measure the water pressure inside the pipe and analyze the deformation behavior of the pipe body, a water pressure gauge P and an image analysis surface A (described below) were installed at the positions shown in the figure.

Next, Figure 2(b) shows the cases examined in this experiment, with six different cases being set up, taking into account the presence or absence of simulated water leaks and the valve opening degree. The measurement conditions are as follows:

In Case A, valve B was closed, and the pipe was filled with water in a still water state, and the water pressure and deformation behavior of the pipe were measured for 60 seconds.
In Case B, simulated water leakage was initiated from two leak holes H 10 seconds after the start of the still water state, and the water pressure and deformation behavior of the pipe body were measured for a total of 60 seconds, continuing for the next 50 seconds.
In Case C, valve B was opened (opening angle 16.9°) 10 seconds after the start of the still water state to allow water to pass through, and when the water level in the tank dropped by about 1 m, valve B was closed to generate water hammer pressure, and the water pressure and deformation behavior of the pipe were measured until the pressure had decayed.
In Case D, simulated water leakage was initiated from two leak holes H 10 seconds after the start of the still water state, and 10 seconds later, valve B was opened (opening angle 16.9°) to allow water to pass through. When the water level in the tank dropped by approximately 1 m, valve B was closed to generate water hammer pressure, and the water pressure and deformation behavior of the pipe were measured until the pressure had decayed.
In Case E, valve B was opened (opening angle 5.6°) 10 seconds after the start of the still water state to allow water to pass through, and when the water level in the tank dropped by about 1 m, valve B was closed to generate water hammer pressure, and the water pressure and deformation behavior of the pipe were measured until the pressure had decayed.
In Case D, simulated water leakage was initiated from two leak holes H 10 seconds after the start of the still water state, and 10 seconds later, valve B was opened (opening angle 5.6°) to allow water to pass through. When the water level in the tank dropped by approximately 1 m, valve B was closed to generate water hammer pressure, and the water pressure and deformation behavior of the pipe were measured until the pressure had decayed.

For the water pressure measurements described above, small pressure sensors (HTV-100KP, manufactured by Senses) were installed in two locations: near the upstream water tank and near valve B at the downstream end. Recordings were made using a voltage data logger (WCR-4V, manufactured by T&D) (sampling: 100 Hz) and via synchronous connection in conjunction with image measurement of the tube deformation behavior (sampling: 20 Hz). In other words, the water pressure data obtained via synchronous connection was used in the analysis described below to enable cross-comparison with the image analysis results of the tube deformation behavior.

In addition, to detect the deformation behavior of the above-mentioned tube body without contact, the deformation behavior of the tube body is image-measured in the manner shown in Figure 4. The DIC method described above is used to identify the deformation behavior of the tube body in three dimensions. This method applies a random pattern to image analysis surface A placed on the surface of the tube material to be measured, and then photographs the movement of dots in this random pattern with a CCD camera, calculating the amount of displacement from the dynamics of the pixel groups in the digital image.

As shown in Figure 4, image analysis surface A, which has a random pattern applied to it, is placed at a position corresponding to the top of the pipe, and two CCD cameras 11 and 12 are installed so as to look down directly onto image analysis surface A. In this example, the random pattern is created as a rectangle measuring 55 mm (width) x 80 mm (height). The measurement conditions are a shutter speed of 5 ms, an aperture of 4, and a frame rate of 20 Hz, so that water pressure data can also be captured via synchronous connection.

In addition, because the brightness of the image analysis plane A significantly affects the analysis accuracy in the DIC method, measurements are taken after sunset when there is less external disturbance, and brightness is adjusted by installing lighting 20 consisting of a stand lamp and clip light as shown in the figure.

The software used to perform the DIC method is Vic Snap (Correlated Solutions) for measurements and Vic 3D (Correlated Solutions) for analysis. Furthermore, the circumferential displacement (rad) in a cylindrical coordinate system (represented as dTheta in the analysis results described below) is used as an evaluation index for the deformation behavior of the tube using the DIC method.

Figure 5 shows a cylindrical coordinate system model, and this indicator represents the amount of displacement in the circumferential direction θ from the random pattern on the image analysis surface A in the initial state. Furthermore, with respect to the positive direction of the tube axis direction Z, counterclockwise is the positive direction and clockwise is the negative direction.

Figure 6 shows an example of the analysis results of the above experiment. Time-series data on the amount of circumferential displacement at the center coordinates of the random pattern on image analysis surface A, enclosed by a rectangle, was extracted and analyzed.

(Analysis method)
The analysis method will be explained below according to the flow chart of FIG.

The calculation of the natural frequency based on the specifications, physical properties and installation conditions of the pipe body will be explained.
The natural frequency is calculated assuming that the vibration of the pipe is that of a beam. In calculating the natural frequency, as shown in Figure 7, the length of the pipe is set to approximately 7.4 m between the downstream bend and the upstream bend, and the boundary condition is that both ends are simply supported. The natural angular frequency ω ^* is expressed by the following equation 1.

Here, l is the length (m), E is the elastic modulus (N/m ² ), I is the second moment of area (m ⁴ ), ρ is the density (kg/m ³ ), and A is the cross-sectional area (m ² ). The natural frequency f ^* is expressed by the following equation 2 from the above natural angular frequency.

In this analysis, the specifications of the model pipeline shown in the table in Figure 2(a) were taken into consideration, with l = 7.4 (m), E = 3.33 x ¹⁰⁹ (N/ ^m2 ), I = 3.22 x ^10-6 ( ^m4 ), ρ = 1.43 x ¹⁰³ (kg/ ^m3 ), and A = 2.23 x ^10-3 ( ^m2 ), and the calculation results showed that the natural frequencies of mode 1 were 1.67 Hz and mode 2 were 6.67 Hz. (In the following analysis, the analysis will be carried out using 1.7 Hz and 6.7 Hz, respectively.)

Discrete Fourier transform will now be explained.
To evaluate the frequency characteristics of the water pressure fluctuations in the pipe and the circumferential displacement of the pipe, a discrete Fourier transform is performed. The discrete Fourier transform of a sampled signal x[n] (n = 0, 1, ..., N-1) of length N is shown in Equation 3 below.

Here, k is the frequency number associated with discretization, and the kth frequency fk is expressed by the sampling frequency fs using the following equation 4:

In this analysis, the sampling rate for image measurement and its synchronized water pressure measurement was 20 Hz, and 512 data points of water pressure and circumferential displacement, including water hammer pressure due to the blockage of valve B, were analyzed.

The continuous wavelet transform will now be described.
A continuous wavelet transform is performed to evaluate the time-frequency characteristics of water pressure fluctuations and circumferential displacement. This wavelet transform can obtain frequency information while preserving time information. The continuous wavelet transform for the time signal f(t) is shown in Equation 5 below (however, it is treated as discrete in the numerical calculations).

Here, a is the extension parameter, b is the position parameter, ψ(t) is the mother wavelet, and the asterisk represents the complex conjugate. In this analysis, the complex Morlet wavelet is used as the mother wavelet. The mother wavelet is shown in Equation 6 below.

(Results of the study)
The time series of water pressure fluctuations inside the pipe and the circumferential displacement of the pipe body will be explained.
Figures 8 to 10 show the water pressure fluctuations inside the pipe and the circumferential displacement of the pipe body over time for each of the aforementioned cases. As shown, in the still water state of Figure 8(a), it can be seen that the water pressure and circumferential displacement fluctuate steadily. Also, in the leaking state of Figure 8(b), it can be seen that the circumferential displacement fluctuates to a negative value as the water pressure drops when the leak begins. This is because the release of water pressure due to the leak causes the resultant water pressure inside the pipe to be biased toward the opposite side of the leak hole H, and this biases in the negative circumferential direction.

When water is flowing as shown in Figure 9 (valve opening 16.9°), the time-series fluctuations in circumferential displacement increase. Comparing Figure 9(c) and Figure 9(d), it can be seen that the amplitude of the fluctuations in circumferential displacement is small when water is flowing during a leak.

When water is flowing as shown in Figure 10 (valve opening 5.6°), the amplitude of the fluctuation in circumferential displacement is smaller than when water is flowing as shown in Figure 9 (valve opening 16.9°). When there is no water leakage as shown in Figure 10(e), the circumferential displacement fluctuates in a negative value while water is flowing, while when there is water leakage as shown in Figure 10(f), the tendency for the circumferential displacement to fluctuate in a positive value can be seen.

From the above, it can be seen that the presence or absence of water leakage and differences in flow conditions within the pipe depending on the valve opening degree can be detected non-contact from the time series of circumferential displacement of the pipe body obtained through image analysis.

(Frequency characteristics of water hammer pressure and circumferential displacement depending on whether or not there is a water leak)
Next, we focus on the water hammer pressure, which is the sudden rise in water pressure when valve B is closed, and evaluate the amount of water leakage and circumferential displacement. Here, we focus on cases C to F, and evaluate 512 data points (sampling: 20 Hz) extracted from 25.6 seconds from 8 seconds before valve closure.

Figures 11(a) to 14(a) show the circumferential displacement of the pipe body and water hammer pressure over time. As shown, the maximum water pressure was 134 kPa in Case C, 108 kPa in Case D, 79 kPa in Case E, and 80 kPa in Case F. Comparing the time series of water pressure shown in Figure 11(a) and Figure 13(a) reveals that the rise in water pressure was greater when the valve was closed in Cases C and D, where the valve opening was 16.9° and the valve was closed to a greater extent.

Furthermore, when comparing cases with and without leakage at the same valve opening, comparing Cases C and D, and Cases E and F, it can be seen that water pressure fluctuations decay more quickly when there is leakage.

When comparing the circumferential displacement of the pipe body in Figure 12(a) and Figure 14(a), it can be seen that in both cases the value of the circumferential displacement increases in the positive direction as the water pressure increases due to water hammer pressure.

Figures 11(b) to 14(b) show the frequency distribution of the circumferential displacement of the pipe body and the water hammer pressure. The frequency distribution of the water hammer pressure in Figures 11(b) and 13(b) shows a tendency for the amplitude spectrum to concentrate around 1.7 Hz. It can be seen that in cases with water leakage (Cases D and F), the amplitude spectrum value concentrated around 1.7 Hz is smaller than in cases without water leakage (Cases C and E).

In the frequency distribution of circumferential displacement shown in Figures 12(b) and 14(b), it can be seen that the amplitude spectrum tends to concentrate at the natural frequencies of 1.7 Hz (mode 1) and 6.7 Hz (mode 2), which were calculated taking into account the installation conditions of the model pipeline. This shows that the natural vibration of the pipe can be detected using the parameters of the circumferential displacement of the pipe obtained by image analysis.

The above results suggest that the frequency band around 1.7 Hz in the water pressure frequency distribution may be a frequency band originating from water hammer pressure generated by the pipe material and flow conditions under the experimental conditions. Water hammer pressure is assumed to be compressible, and its propagation speed depends on the physical properties and pipe thickness of the pipe material that constrains the water. Therefore, its frequency characteristics are not uniform and are likely to vary depending on the pipe material and flow conditions. Considering water hammer as a phenomenon caused by the combination of water and pipe material, monitoring the behavior of the pipe material may enable estimation of the characteristics of the flow conditions within the pipe. In other words, by focusing on the natural vibrations, which are the vibration characteristics of the pipe (combination of water and pipe material), it is possible to detect the frequency characteristics of water pressure fluctuations within the pipe. Therefore, by detecting the natural vibrations from the frequency distribution of the circumferential displacement of the pipe obtained through image analysis, it is possible to non-contactly identify the frequency characteristics of water hammer pressure within the pipe.

(Water hammer pressure and circumferential displacement attenuation depending on the presence or absence of water leakage in the time-frequency domain)
Next, Figures 15 to 18 show scalograms (wavelet transform results) of water pressure fluctuations inside the pipe and the circumferential displacement of the pipe body. That is, Figures 15 and 17 show the time-frequency characteristics of water hammer pressure, and Figures 16 and 18 show the time-frequency characteristics of circumferential displacement. This represents the time series shown in Figures 11(a) to 14(a) in the time-frequency domain using wavelet transform. This makes it possible to obtain frequency information while preserving the time information.

As a result of the analysis shown in Figures 15 to 18, the scalograms of water pressure fluctuations and circumferential displacement show relatively strong signals in the natural vibration frequency bands of 1.7 Hz (mode 1) and 6.7 Hz (mode 2), associated with the occurrence of water hammer pressure due to valve blockage.

Furthermore, as shown in Figures 15 and 17, the scalograms of water pressure fluctuations detect a decay in signal strength over time due to water hammer pressure. A rapid decay in the amplitude of water pressure fluctuations over time can be seen during leakage (Figures 11(a) and 13(a)), and similarly, in the time-frequency domain, a more rapid decay in signal strength during leakage is detected in the same frequency band as the natural vibration. Focusing on this trend, the scalograms of circumferential displacement, as shown in Figures 16 and 18, detect a decay in signal strength over time due to water hammer pressure in the frequency band of the natural vibration. This is particularly noticeable in Cases C and D, where the valve opening is large (Figure 16), and the rapid decay associated with leakage is also clearly detected.

In Cases E and F, where the valve opening was small, no attenuation of water hammer pressure was detected at the natural frequency of 6.7 Hz (mode 2) to the same extent as in Cases C and D.

These experimental results on a model pipeline confirmed that, in the time-frequency domain of circumferential displacement, it is possible to detect vibrations caused by water hammer pressure in the natural vibration frequency band, as well as differences in damping due to the presence or absence of a leak. This demonstrates that by using the circumferential displacement obtained from the image analysis results using the DIC method as an index, it is possible to detect and evaluate the actual state of a leak in a pipeline without contact.

(System configuration)
The configuration of the leakage phenomenon identification system 100 according to the embodiment of the present invention will be described below.

As shown in Figure 4, the leakage phenomenon identification system 100 in an embodiment of the present invention comprises at least an image analysis surface A installed on the surface of a pipeline pipe material, an imaging means (CCD cameras 11, 12) for imaging the image analysis surface A, a displacement amount measurement means 101 for measuring the circumferential displacement of the pipe material based on the image captured by the imaging means, and a leakage phenomenon determination means 103 for performing analysis based on the measured circumferential displacement amount and for determining at least the presence or absence of a fluid leakage phenomenon.

In this embodiment, as shown in Figure 4, at least two CCD cameras 11 and 12 are provided as the imaging means to capture images of image analysis surface A from different directions. However, the number of imaging cameras, including the CCD camera, may be one or more, as long as the imaging means is capable of capturing a three-dimensional image of image analysis surface A. The imaging means (CCD cameras 11 and 12) are connected to a PC incorporating the leakage phenomenon determination means 103, etc.

The leak phenomenon determination means 103 in the PC also includes an analysis means 104 and a learning means 105 that performs machine learning (learning step) on at least the amount of circumferential displacement of the pipe material, and analysis is performed based on a predetermined algorithm of the learning means 105. Then, based on the analysis results, identification information for the leak phenomenon is output to the leak information output unit 102.

In addition, the leakage phenomenon discrimination means 103 is connected to a measurement facility information input means 106, which makes it possible to input, for example, the diameter of the actually discovered leakage hole H, the distance from the image capture position to the leakage hole H, and so on, in addition to the diameter and material of the pipeline. By using this type of information and the circumferential displacement of the pipe material as learning data, it becomes possible to identify not only the presence or absence of a leakage hole H from the acquired 3D image, but also the location and scale of the leakage, further improving the accuracy of identifying leakage phenomena.

(Method for identifying water leakage phenomena)
Since most existing pipelines are buried underground, an image analysis surface A is set on the surface of the exposed pipe material of the pipeline inside a manhole where a water stop valve, etc. is located (analysis surface setting step). Then, photographing means (CCD cameras 11, 12) are installed, and the image analysis surface A is photographed by the photographing means (analysis surface photographing step).

Images captured by the imaging means are sent to a PC, which measures the circumferential displacement of the pipe material based on these images (displacement measurement step). Analysis is then performed based on the measured circumferential displacement of the pipe material, making it possible to determine whether or not a fluid leakage phenomenon has occurred (leakage phenomenon determination step). In other words, the deformation behavior of the pipe body is evaluated using the DIC method described above, the circumferential displacement in the cylindrical coordinate system ("θ" in Figure 5) is determined, and the attenuation behavior in both time and frequency (time-frequency domain) can be determined to identify leakage phenomena in the pipeline.

Other Embodiments
The above describes an embodiment of the method and system for identifying leakage phenomena using three-dimensional image analysis according to the present invention, taking as an example a method for identifying leakage phenomena in a water transmission/distribution pipeline. However, the present invention is not necessarily limited to the configuration described above, and various modifications such as those described below are possible.

For example, the leakage phenomenon identification method and identification system of the present invention using 3D image analysis is not necessarily limited to water transmission and distribution pipelines, but can also be applied to pressurized pipelines that transport various fluids, such as fuel pipes, gas pipes, and steam pipes. By taking into account the characteristics of the fluid and the characteristics of the pipe body, it is possible to identify fluid leakage phenomena with high accuracy.

Furthermore, the magnitude of deformation of the pipe material depends on the elastic modulus and second moment of area of the pipe material. Therefore, the identification method and identification system of the present invention can be more suitably applied to polyvinyl chloride pipe material than to cast iron pipes and other steel pipes.

Although the embodiments of the present invention have been described above with reference to the drawings, the specific configuration is not limited to these embodiments. The scope of the present invention is defined by the claims, not by the description of the above embodiments, and includes all modifications equivalent to and within the scope of the claims. Furthermore, the specific materials, dimensions, shapes, etc. described in the above examples may be modified within the scope of resolving the problems of the present invention.

A: Image analysis surface H: Leak hole P: Water pressure gauge B: Valve 11: CCD camera 12: CCD camera 20: Lighting 100: Leak phenomenon identification system 101: Displacement amount measurement means 102: Leak information output means 103: Leak phenomenon discrimination means 104: Analysis means 105: Learning means 106: Measurement facility information input means

Claims (6)

1. A method for identifying a fluid leakage phenomenon in a pipeline, comprising:
an analysis surface setting step of setting an image analysis surface on a surface of the pipeline pipe material;
an analysis surface photographing step of photographing the image analysis surface with a photographing means;
a displacement amount measuring step of measuring a circumferential displacement amount of the pipe material based on the image captured by the imaging means;
a water pressure measuring step of measuring water pressure in the pipeline;
a damping behavior determination step of determining the damping behavior of the circumferential displacement amount in the time-frequency domain based on the frequency of the water hammer pressure measured in the water pressure measurement step and the circumferential displacement amount measured in the displacement amount measurement step;
a leakage phenomenon determination step that is capable of determining whether or not there is at least a leakage phenomenon of the fluid based on the determination result in the damping behavior determination step. The method for identifying a leakage phenomenon by three-dimensional image analysis according to claim 1, wherein the photographing means comprises at least two photographing cameras that photograph the image analysis plane from different directions. a learning step of machine learning at least the circumferential displacement amount,
The method for identifying a leakage phenomenon by three-dimensional image analysis according to claim 1 or 2, wherein in the leakage phenomenon determining step, the damping behavior determining step is performed based on a predetermined algorithm by machine learning. 1. A system for identifying a fluid leakage phenomenon in a pipeline, comprising:
an image analysis surface installed on a surface of the pipeline pipe;
an imaging means for imaging the image analysis surface;
a displacement amount measuring means for measuring a circumferential displacement amount of the pipe material based on the image captured by the imaging means;
a pressure sensor for measuring the water pressure in the pipeline;
An analysis means for determining the attenuation behavior of the circumferential displacement in the time-frequency domain based on the frequency of the water hammer pressure measured by the pressure sensor and the circumferential displacement measured by the displacement measurement means ;
and a leakage phenomenon determination means capable of determining whether or not a leakage phenomenon of the fluid has occurred based on the analysis result of the analysis means . 5. The system for identifying a leakage phenomenon by three-dimensional image analysis according to claim 4, wherein the photographing means comprises at least two photographing cameras that photograph the image analysis plane from different directions. a learning means for machine learning at least the circumferential displacement amount,
The system for identifying a leakage phenomenon using three-dimensional image analysis according to claim 4 or 5 , wherein the leakage phenomenon determination means determines the attenuation behavior of the circumferential displacement amount using the analysis means based on a predetermined algorithm of the learning means.