KR100420717B1 - Method for detecting leakage of service water tube and its system - Google Patents

Abstract

The leakage rate of tap water in Korea is about 25%, which is considerably higher than that in developed countries. Leakage generally refers to structural, as well as loss of water due to corrosion, holes, and incorrect connections. In case of leaks, foreign substances can invade and cause tap water to be contaminated. The present invention creates a leak pattern based on data such as flow rate, flow rate, hydraulic pressure at various points, and then learns it by using an artificial neural network program, and proposes a new method for detecting leaks in real time. Based on this, if a leak occurs in the water pipe, the structure of the system to detect it quickly is suggested.

Description

Leak detection method of water pipe and its system {METHOD FOR DETECTING LEAKAGE OF SERVICE WATER TUBE AND ITS SYSTEM}

The present invention relates to a leak detection method and a system implementing the same, and more specifically, to construct a pattern of a sensor value according to a leak location in a knowledge base DB, and then detect a leak location using a sensor value input from a field system. A leak detection method and a system implementing the same.

Currently, the old water pipes are not replaced in a timely manner, and there are a lot of poor constructions, so that many tap waters are leaked and the water budget is wasted, and there is a possibility of water pollution due to the inhalation of foreign substances from the leaked locations.

In order to reduce the leak rate, the leak location is repaired and the old pipe is replaced, but the conventional leak detection method is difficult to detect a small amount of leak, and a lot of funds are required to replace the old pipe. It is important to develop leak detection techniques to diagnose leaks on a regular basis.

The leakage rate of tap water in Korea is about 25%, which is considerably higher than that in developed countries. Leakage generally refers to structural, as well as loss of water due to corrosion, holes, and incorrect connections. In case of leaks, foreign substances can invade and cause tap water to be contaminated. Therefore, a technique for quickly detecting leaks is urgently needed. However, the current method is too wide a range to be detected in determining the leakage by measuring the noise from the leak and requires manual intervention. Conventional water pipe leak detection usually detects leaks from the ground with fresh food. When sound generated by water leaks is transmitted to the surface, the sound is detected and amplified from the surface and listened to the receiver or analyzed by meter. It was a way to detect.

By the way, the water supply pipeline is buried at a depth of 1.2M or less from the ground, so it is difficult to detect leaks in small quantities and it is difficult to accurately detect leak points. In case of leaks, traffic noise along with traffic was transmitted together, making it difficult or almost impossible to detect. Therefore, if it can be detected automatically using a computer and the leak area can be limited to some extent, it will greatly contribute to the prevention of leakage.

For this reason, technology development for leak detection has been studied for a long time. As a general method, there is a sonic leak detection method using a sound amplifier to detect leaks. There is a Water Audit based on Mass Balance Analysis. Water audit is a leak detection method that measures the flow rate by dividing zones or districts by using valves in the lumen, and in terms of accuracy of data when there are leaks or cross-connections between regions. There is a disadvantage of low reliability. The acoustic leak detection method can detect leaks in water supply or valves well, and in contrast, the water audit has a characteristic of detecting leaks in the drainage main pipe and the water supply pipe. The former has the advantage of being economical while the latter has the advantage of being expensive but finding a large leak. New leak detection techniques often use electronic equipment or trace gases and dyes.

The most widely used leak detection technique is the leak correlators, which utilizes the principle that two channel microprocessors calculate a leak point by using a time difference of sound generated at the leak point. However, this method can lead to errors if the distance between the two transducers is incorrectly entered, or if the material is changed and proper contact with the transducer is not achieved.

The present invention assumes a case where water leak occurs at each point of the water supply pipe, and then makes a database of sensor values installed in each water supply pipe in each case, and uses the same to easily detect a leak point remotely. An object of the present invention is to provide a leak inspection method and a system for enabling the same.

The above object of the present invention is provided with a mapping table for storing a correlation between a leak database corresponding to each predicted sensor value of a leak database and a leak database for storing each predicted sensor value expected to be detected by each sensor when a leak occurs. A method of detecting a leak using a system, the method comprising: receiving a sensor value of a pipe to be leak checked; Analyzing a correspondence between the input sensor value and each expected sensor value of the leak database; Selecting an expected sensor value having the highest agreement among each expected sensor value and inferring a leak position corresponding to the selected expected sensor value using a mapping table; And it can be achieved by the leak detection method characterized in that it comprises the step of displaying the inferred leak location. Still another object of the present invention is to remotely receive the sensor value from the sensor installed in the water pipes A leak location detecting system for detecting a leak location using an input data storage unit for storing a sensor value transmitted remotely from a sensor; A leak database for dividing and storing each expected sensor value expected to be detected by each sensor when a leak occurs; A mapping table for recording correlations of leak positions corresponding to respective expected sensor values of the leak database; A processor configured to compare the sensor values of the input data storage unit with each of the predicted sensor values in the leak database, search for the most likely predicted sensor values, and infer a leak position corresponding to the detected predicted sensor values using a mapping table; And a display unit for displaying the inferred leak position.

In addition, another object of the present invention, a leak position detection system for detecting a leak position by receiving a sensor value remotely from the sensor installed in the water pipe, the input data storage unit for storing the sensor value; A knowledge base-based leak database created by creating a leak pattern by using sensor values of flow rate, flow rate, and hydraulic pressure at each point according to leaks generated at various points of a pipe; An inference engine for inferring a correspondence between the sensor value of the input data storage unit and the pattern of the leak database; And a display unit for displaying the result inferred from the inference engine. The present invention also provides a leak position detection system. In addition, another object of the present invention is to provide a plurality of pipes and the hydraulic pipe, An on-site system having a sensor for detecting a flow rate and a transmitter for transmitting a measurement of the sensor to a remote location; A relay system for amplifying and relaying sensor values received from the field system; And an input data storage unit for storing sensor values received from the relay system, a leak database storing sensor values to be detected by each sensor when a leak occurs, and a leak position corresponding to a sensor value of the leak database. A processing unit for searching for a sensor value having the highest degree of agreement by comparing a position mapping unit to be defined, a sensor value of the input data storage unit, and a leak database sensor value, and outputting a leak position corresponding to the detected sensor value using the location mapping unit; It can also be achieved by the leak detection system, characterized in that it consists of a leak server system having an output for outputting the leak location.

1 is an embodiment of a system diagram of the present invention.

2 is an embodiment of an expert system configuration used in the present invention.

3 is a flow chart illustrating a water pipe leak detection algorithm of the present invention.

4A is a piping configuration diagram of a single pipe, and FIG. 4B is an explanatory diagram for explaining a method of detecting a leak position using a knowledge base DB including a sensor value of the pipe shown in FIG. 4A.

5 is a piping configuration diagram of a two-pronged pipe.

6 is an explanatory diagram for explaining a method of grouping a leak detection zone;

Advantages, features and preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is an embodiment of the overall system configuration of the present invention. The entire system of the present invention is composed of a field system 10, a relay server 20 and a leak detection server 30. The field system 10 includes a water supply pipe 11, a sensing sensor 12 attached thereto, and a telemetry having a sensor information transmitting means 13 for wirelessly transmitting an output value of the sensing sensor. The relay server 20 includes a local relay 21 for relaying output signals of the telemetry, a communication server 23 for processing communication signals transmitted from the plurality of local relays 21, and a security means 22. The local repeater 21 is preferably installed at a geographical distance adjacent to the water supply pipe 11. The leak detection server 30 is composed of a security means 31, expert system 32, knowledge base DB 33 and the display unit 34.

The field system 10 measures the information on the hydraulic pressure, flow rate, and flow rate of the plurality of water supply pipes 11 using the sensor 12, and then relays the server 20 using the sensor information transmission means 13. To send. The local repeater 21 functions to relay sensor information. The local repeater 21 is installed at a distance geographically adjacent to the sensor 12 to amplify the sensor value and transmits the sensor value to the communication server 23. The communication server 23 encodes the sensor measurement value and is transmitted to the leak detection server 30 using the wired / wireless communication network 40 via the security means 22. The leak detection server 30 decodes a plurality of sensor input values input through the security means 31, infers a leak location using the knowledge base DB 33, and then displays the inferred location. Will be displayed on the screen.

2 illustrates one embodiment of an expert system configuration used in the present invention. The expert system 32 includes a data receiver 32-1 in charge of receiving data, an inference engine 32-3 based on the knowledge base DB 33 that defines a correlation between sensor values for each leak location, and a leak. The position display part 34 is comprised. The leak detection server 30 receiving each sensor output value stores it in the data receiver 32-1. The sensor value stored in the data receiver 32-1 undergoes a preprocessing process for comparison with the knowledge base DB 33, and then the most suitable pattern is obtained through the inference engine 32-3 using the knowledge base DB 33. After detecting, the leak position corresponding to the detected pattern is identified and displayed on the display unit 34.

In the present invention, a multi-layer perceptron was applied as an expert system. That is, when various data from the sensor 12 are allocated to the input node, the value of the output node is determined through internal calculation, which indicates the position where the node with the highest value has leaked. Detailed configuration or learning method and reasoning method of the expert system by the multilayer perceptron is beyond the scope of the present invention, and since a large number of known materials exist, detailed description thereof will be omitted. Representative examples of the known data of the multilayer perceptron include Korean Patent Publication No. 1996-012131 for the error signal generation method and Publication No. 1997-17000 for the learning method.

3 is a flow chart illustrating the water pipe leak detection algorithm of the present invention. The sensor value is input from the field system 10 (step S1). The input data is subjected to a preprocessing process to infer a correlation with the knowledge base DB. (Step S2) During the preprocessing process, the input data is a constant of some form. Include whether the information is input from the tube. For example, it may include information on whether the sensor value measured in a pipe array composed of a single water pipe is input or data received from a branched water pipe.

The input data after preprocessing is analyzed for correlation with the knowledge base DB (step S3). In the case of leakage at various leak points, the knowledge base DB stores sensor values of the corresponding points in each case. By extracting the pattern with the highest correlation (step S4), the leak position can be identified (step S5).

One of the most important things in expert systems is how to implement knowledge base DB. A method of constructing such a knowledge base DB is described below with reference to FIG. 4.

I. Single piping

4a shows that one main water supply pipe is provided and a flow meter is installed at points A and B of the main water supply pipe. It is shown that the main water pipe is provided with a branched lower water pipe 50 for supplying water to each household. Even when no leakage occurs between A and B, the amount of tap water used in each household changes according to the sensor value measured at points A and B, which makes it difficult to accurately diagnose leakage. However, in case of leakage between points A and B, the sensor values measured at points A and B have a constant pattern despite the change in tap water usage in each household. In addition, this pattern shows a difference between a case where a leak occurrence point occurs near a point A and a point that occurs far away. That is, if the same amount of water leakage occurs, if the leakage occurs near the point A, the flow rate at point B will be relatively low, and if the leak occurs near the point B, the flow rate at the point B will be relatively high. do.

This can be calculated with the formula of physics, but it takes considerable time in the example with a large number of nodes, such as water pipes. Therefore, in the present invention, a method of constructing a knowledge base DB using data obtained by performing experiments as learning data is not used to calculate this using physics. As the leakage point is changed from the area 1 to the area n of FIG. 4A, data on the flow rate, flow rate, hydraulic pressure, pipe diameter, and fluid type at the point A, and the flow rate, flow rate, oil pressure, and pipe The data on the diameter and the type of fluid are patterned and entered into the knowledge base DB. We learn using the suggested pattern and infer the input data and output the leak point.

(LQ) (V _A1 , P _A1 , Q _A1 , R _A , M _A ) (V _B1 , P _B1 , Q _B1 , R _B , M _B ),

(LQ) (V _A2 , P _A2 , Q _A2 , R _A , M _A ) (V _B2 , P _B2 , Q _B2 , R _B , M _B ),

(LQ) (V _A3 , P _A3 , Q _A3 , R _A , M _A ) (V _B3 , P _B3 , Q _B3 , R _B , M _B ),

...

(LQ) (V _An , P _An , Q _An , R _A , M _A ) (V _Bn , P _Bn , Q _Bn , R _B , M _B )

LQ is the amount of leakage,

V _Ai and V _Bi are the flow rates at the A and B points when a leak occurs at the i point,

P _Ai and P _Bi are the hydraulic pressures at A and B points when a leak occurs at point i,

Q _Ai , Q _Bi are the flow rates at A and B points when water leaks at point i,

R _A , R _B is the diameter of the A, B pipe,

M _A, M _B is the type of fluid at points A and B

For example, as shown in FIG. 4A, when a leak occurs at a first point between A and B points, a pattern of sensor values measured at A and B points is (LQ) (V _A1 , P _A1 , Q _A1). , R _A , M _A ) (V _B1 , P _B1 , Q _B1 , R _B , M _B ), and if a leak occurs at the second point, the pattern of the sensor value measured at points A and B is (LQ) (V _A2 , P _A2 , Q _A2 , R _A , M _A ) (V _B2 , P _B2 , Q _B2 , R _B , M _B ), and if a leak occurs at point n in this way (LQ) It can be patterned as (V _An , P _An , Q _An , R _A , M _A ) (V _Bn , P _Bn , Q _Bn , R _B , M _B ). The sensor value of each point according to the leaking point can be measured by using a real water pipe or by using experimental data after making similar experimental equipment.

In this way, the knowledge base DB is constructed using the measurements in each case patterned. As shown in FIG. 4B, the pattern of each case is stored and constructed in the knowledge base DB, and the leakage position is inferred by inferring the most suitable pattern of the knowledge base DB using the input pattern transmitted from the field system 10. You can do it.

II. 2 branch piping

In general, it is economically difficult to install a sensor at each branch point in the water pipe installed in each home from the water treatment plant as shown in FIG. 4. FIG. 5 shows an application that is more suitable for the case of water supply piping than that of FIG. 4.

In the case where three branched conduits are arranged as shown in FIG. 5, assuming that the sensors are installed at points A, B, and C, a leak location is detected by measuring sensor values at points A, B, and C. can do.

(LQ) (V _A1 , P _A1 , Q _A1 , R _A , M _A ) (V _B1 , P _B1 , Q _B1 , R _B , M _B ) (V _C1 , P _C1 , Q _C1 , R _C , M _C ),

(LQ) (V _A2 , P _A2 , Q _A2 , R _A , M _A ) (V _B2 , P _B2 , Q _B2 , R _B , M _B ) (V _C2 , P _C2 , Q _C2 , R _C , M _C ),

(LQ) (V _A3 , P _A3 , Q _A3 , R _A , M _A ) (V _B3 , P _B3 , Q _B3 , R _B , M _B ) (V _C3 , P _C3 , Q _C3 , R _C , M _C ),

...

(LQ) (V _An , P _An , Q _An , R _A , M _A ) (V _Bn , P _Bn , Q _Bn , R _B , M _B ) (V _Cn , P _Cn , Q _Cn , R _C , M _C )

LQ is the amount of leakage,

V _Ai , V _Bi , and V _Ci are the flow rates at A, B, and C when leaking at point i,

P _Ai , P _Bi , P _Ci are the hydraulic pressures at A, B, and C,

Q _Ai , Q _Bi , and Q _Ci are the flow rates at A, B, and C when water leaks at point i,

R _A , R _B , R _C are the diameters of A, B, C pipes,

M _A, M _B and M _C are the types of fluids at points A, B and C

For example, as shown in FIG. 5, when a leak occurs at a first point between points A, B, and C, the pattern of sensor values measured at points A, B, and C is (LQ) (V _A1 , P _A1 , Q _A1 , R _A , M _A ) (V _B1 , P _B1 , Q _B1 , R _B , M _B ) (V _C1 , P _C1 , Q _C1 , R _C , M _C ) If leakage occurs at, the pattern of sensor values measured at points A, B, and C is (L Q) (V _A2 , P _A2 , Q _A2 , R _A , M _A ) (V _B2 , P _B2 , Q _B2 , R _B , M _B ) (V _C2 , P _C2 , Q _C2 , R _C , M _C ), and in this way, if a leak occurs at the nth point, (LQ) (V _An , P _An , Q _An , The pattern may be defined as R _A , M _A ) (V _Bn , P _Bn , Q _Bn , R _B , M _B ) (V _Cn , P _Cn , Q _Cn , R _C , M _C ). These patterns can be obtained using similar experimental equipment. After storing the patterns in each of these cases in the knowledge base DB, the sensor can be inferred by inferring the most appropriate pattern by sensing each sensor value transmitted from the field system 10 at regular intervals. The additional use of sound wave detection equipment enables more accurate leak location detection.

In the above, the case of a single water pipe and a branched water pipe has been described, but various classifications may be possible depending on the number of pipes in the pipe. The reason why such a classification is necessary will be described, for example, when the water pipes installed in "Yeoksam-dong" are grouped into a plurality of small water pipes having a single water pipe or a bifurcated water pipe, and the grouped water pipes as one input terminal. You can easily apply it. 6 shows a zone divided into three groups (first group, third group and fifth group) having a single water pipe and five groups having two branched water pipes (second group and fourth group). Shows. The leak detection server 30 periodically receives a sensor value from the first group to the fifth group and processes whether there is a leak. The input data is the data of which group through the data preprocessing process and how many points the sensor value is composed.

While the preferred embodiments of the present invention have been described using specific terms, such descriptions are for illustrative purposes only, and it is understood that various changes and modifications may be made without departing from the spirit and scope of the following claims. Should be done.

For example, the present invention has been described using an expert system, but a database for storing a plurality of sensor values corresponding to each leak point in one pattern and a leak corresponding to each pattern in association with the database. Obviously, it can also be implemented by a system with a mapping table that describes the association of locations.

The advantage of using artificial neural networks is that they are easier to generalize than using physics formulas. In other words, even if it is not learned, the appropriate value is output. Artificial neural networks also take a long time to learn, but once learned, they calculate the output very quickly. On the other hand, when using the laws of physics, it takes considerable computation time.

According to the present invention, it is assumed that water leakage occurs at each point of the water supply pipe, and by using a sensor sensor value installed in the water supply pipe generated at this time, it is possible to easily detect the water leakage point using this.

Claims (6)

In the leak position detection system for receiving the sensor value from the sensor installed in the water supply pipe remotely and detecting the leak position using the sensor value, An input data storage unit for storing the sensor value transmitted remotely from the sensor; A leak database for dividing and storing each expected sensor value expected to be detected by each sensor when a leak occurs; A mapping table for recording correlations of leak positions corresponding to the expected sensor values of the leak database; Comparing the sensor values of the input data storage unit with each of the predicted sensor values of the leak database to search for a predicted sensor value having the highest match, and inferring a leak position corresponding to the detected predicted sensor value using the mapping table. Processing unit; And And a display unit for displaying the inferred leak position. The method of claim 1, And the sensor value and the expected sensor value include oil pressure, flow rate, and flow rate. In the leak position detection system for receiving the sensor value from the sensor installed in the water supply pipe remotely and detecting the leak position using the sensor value, An input data storage unit for storing the sensor value; A knowledge base-based leak database created by creating a leak pattern by using a sensor value of flow rate, flow rate, and hydraulic pressure at each point according to leaks generated at various points of the pipe; An inference engine that infers a degree of agreement between a sensor value of the input data storage unit and a pattern of the leak database; And And a display unit displaying a result inferred from the inference engine. The method of claim 3, wherein Leakage position detection system, characterized in that the system is implemented with a multi-layer perceptron. A field system having a plurality of pipes and a sensor installed in the pipe, the sensor detecting a hydraulic pressure and a flow rate, and a transmitter for transmitting a measurement value of the sensor to a remote location; A relay system for amplifying and relaying sensor values received from the field system; And An input data storage unit for storing a sensor value received from the relay system; A leak database storing sensor values to be detected at each sensor when a leak occurs; A position mapping unit defining an association relationship between a leak position corresponding to a sensor value of the leak database; A processor configured to compare the sensor value of the input data storage unit with the leak database sensor value, search for a sensor value having the highest degree of agreement, and output a leak position corresponding to the detected sensor value using the location mapping unit; And The leak detection system, characterized in that configured as a leak server system having an output unit for outputting the leak position. By using a system having a leakage table for storing each expected sensor value expected to be detected by each sensor when a leak occurs, and a mapping table for recording the correlation between the leak location corresponding to each expected sensor value of the leak database. As a method of detecting leaks, Receiving a sensor value of a pipe to check for leaks; Analyzing a degree of agreement between the input sensor value and each expected sensor value of the leak database; Selecting an expected sensor value having the highest agreement among each of the expected sensor values, and inferring a leak position corresponding to the selected expected sensor value using the mapping table; And And displaying the inferred leak position.