KR102002457B1 - Automatic indoor water leak detecting method using statistical outlier detection - Google Patents

Abstract

에 의해 계산하는 단계; (f) 임계값 테이블(critical value table)에서 데이터의 개수에 따른 상위 5%의 유의수준(significance level)에 대응하는 임계값(CV)을 획득하는 단계; 및 (g) 상기 통계값(G)이 상기 임계값(CV) 보다 큰 경우, 옥내 누수로 결정하는 단계를 포함한다.An automatic indoor leak detection method using a statistical outlier detection method according to an embodiment of the present invention includes: (a) acquiring indoors water usage data measured at predetermined time intervals by a water meter for remote meter reading; (b) determining whether the indoor water usage pattern is abnormal based on the water usage data measured every predetermined time; (c) if the water usage pattern is abnormal, determining that the water usage pattern is an abnormal day as a day when the water leakage is suspected, and obtaining the daily usage amount of the water leakage day and the daily usage amount for the predetermined days; (d) calculating an average and standard deviation of the obtained daily use amount; (e) Statistical value (G) for water use (A) on the day when water leakage is suspected

≪ / RTI > (f) obtaining a threshold value (CV) corresponding to a significance level of the upper 5% according to the number of data in a critical value table; And (g) if the statistical value (G) is greater than the threshold value (CV), determining the indoor leak.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to an automatic indoor leak detection method using a statistical anomaly detection method,

The present invention relates to an automatic indoor leak detection method using a statistical outlier detection method.

The leakage in the indoor area includes leakage due to the breakdown of the water pipe installed in the house, leakage due to the faucet not being locked, or leakage from the toilet water in the toilet.

It is difficult for the user to recognize the leakage of the indoor water even if the indoor water use amount becomes extremely excessive due to various reasons. In some cases, if the metering probe detects a sudden increase in the volume while measuring the meter, the meter informs the user that the leak is indoors, but the user is indeed experiencing an indoor leak or suddenly increasing water usage. It is not easy to distinguish.

Accordingly, there is a need for an indoor leak detection method that can accurately detect whether indoor water leaks occur only by water usage data automatically without visiting a surveyor.

KR 10-1214986 B1

The object of the present invention is to provide an automatic indoor leak detection method using a statistical anomaly detection method that can accurately detect whether indoor water leaks occur automatically only by using water usage data without visiting a surveyor.

According to another aspect of the present invention, there is provided an automatic indoor leak detection method using a statistical outlier detection method,

(a) acquiring indoors water usage data measured by the remote meter reading water meter every predetermined time;

(b) determining whether the indoor water usage pattern is abnormal based on the water usage data measured every predetermined time;

(c) if the water usage pattern is abnormal, determining that the water usage pattern is an abnormal day as a day when the water leakage is suspected, and obtaining the daily usage amount of the water leakage day and the daily usage amount for the predetermined days;

(d) calculating an average and standard deviation of the obtained daily use amount;

에 의해 계산하는 단계;(e) Statistical value (G) for water use (A) on the day when water leakage is suspected

≪ / RTI >

(f) obtaining a threshold value (CV) corresponding to a significance level of the upper 5% according to the number of data in a critical value table; And

(g) if the statistical value (G) is greater than the threshold value (CV), determining as an indoor leak.

In the automatic indoor leak detection method using the statistical abnormal value search method according to an embodiment of the present invention, the step (b) may include a step of, if indoor water consumption data measured at the predetermined time is not zero, And determining that the water usage pattern is abnormal.

Also, in the automatic indoor leak detection method using the statistical outlier detection method according to an embodiment of the present invention, the threshold value table may include a threshold value table used in the Grubbs' method.

In the automatic indoor leak detection method using the statistical outlier detection method according to an embodiment of the present invention,

을 포함할 수 있다.

. ≪ / RTI >

According to another aspect of the present invention, there is provided an automatic indoor leak detection method using a statistical outlier detection method,

(a) acquiring indoors water usage data measured by the remote meter reading water meter every predetermined time;

(b) determining whether the indoor water usage pattern is abnormal based on the water usage data measured every predetermined time;

(c) If the water usage pattern is abnormal, determine that the water usage pattern is abnormal on the day when the water leakage is suspected, and obtain the daily and weekly usage amount on the day when the water leakage is suspected and the daily and weekly usage amount for the previous predetermined days ;

(d) converting the two-dimensional data of the obtained daily and weekly usage amounts into one-dimensional data using principal component analysis (PCA);

(e) calculating an average and a standard deviation of the data converted into the one-dimensional data;

에 의해 계산하는 단계;(f) The statistical value (G) for the one-dimensional data (B)

≪ / RTI >

(g) obtaining a threshold value (CV) corresponding to a significance level of the upper 5% according to the number of data in the critical value table; And

(h) if the statistical value (G) is greater than the threshold value (CV), determining as indoor leakage.

In the indoor automatic water leakage detection method according to another embodiment of the present invention, in the step (b), when the water usage minimum value among the indoor water usage data measured at the predetermined time is not 0, And determining that the water usage pattern is abnormal.

In the automatic indoor leak detection method using the statistical outlier detection method according to another embodiment of the present invention, the threshold value table may include a threshold value table used in the Grubbs' method.

In the automatic indoor leak detection method using the statistical outlier detection method according to another embodiment of the present invention,

을 포함할 수 있다.

. ≪ / RTI >

In the automatic indoor leak detection method using the statistical outlier detection method according to another embodiment of the present invention, the step (d)

(d-1) determining an axis having a maximum variance as the principal component axis for the obtained two-dimensional data of daily and weekly usage amounts; And

(d-2) projecting the acquired two-dimensional data of the daily and weekly usage amounts onto the determined principal component axis, and converting the two-dimensional data of the obtained daily and weekly usage amounts into one-dimensional data .

According to the automatic indoor leak detection method using the statistical outlier detection method according to an embodiment of the present invention, since the indoor leak detection is detected using the statistical outlier detection method, It is possible to accurately detect the occurrence of leakage, thereby reducing the cost of detecting the leakage of indoor water and preventing the waste of water caused by indoor leakage, thereby protecting the water resources.

1 shows an exemplary ultrasonic water meter which is a water meter for remote meter reading.
FIG. 2 illustrates a system to which an automatic indoor leak detection method using a statistical outlier detection method according to an embodiment of the present invention is applied. FIG.
3 is a flowchart of an automatic indoor leak detection method using a statistical outlier detection method according to an embodiment of the present invention.
FIG. 4A is a graph showing leakage data of a customer who has caused a leakage of indoor water. FIG.
Figure 4b is a graph showing a normal water usage pattern.
Figure 4c is a graph showing an abnormal water usage pattern.
FIG. 5 illustrates exemplary daily water usage data; FIG.
6 shows a threshold table used in the Grubsc method;
FIG. 7 is a flowchart of an automatic indoor leak detection method using a statistical outlier detection method according to another embodiment of the present invention. FIG.
FIG. 8 illustrates exemplary daily and weekly water usage data; FIG.
9 is a diagram for explaining conversion of two-dimensional data of daily water use and weekly water usage into one-dimensional data as a main dimension by using PCA;
10 is a view showing two-dimensional data of daily water consumption and weekly water consumption and one-dimensional data converted by using PCA.

BRIEF DESCRIPTION OF THE DRAWINGS The objectives, specific advantages and novel features of the present invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which: FIG.

Prior to that, terms and words used in the present specification and claims should not be construed in a conventional and dictionary sense, and the inventor may properly define the concept of the term in order to best explain its invention Should be construed in accordance with the principles and the meanings and concepts consistent with the technical idea of the present invention.

It should be noted that, in the present specification, the reference numerals are added to the constituent elements of the drawings, and the same constituent elements are assigned the same number as much as possible even if they are displayed on different drawings.

Also, the terms "first", "second", "one side", "other side", etc. are used to distinguish one element from another, It is not.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following description of the present invention, a detailed description of known arts which may unnecessarily obscure the gist of the present invention will be omitted.

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

FIG. 1 is a view showing an exemplary ultrasonic water meter as a water meter for remote meter reading. FIG. 2 is a view showing a system to which an automatic indoor water leak detection method using a statistical outlier detection method according to an embodiment of the present invention is applied to be.

 In the automatic indoor leak detection method using the statistical outlier detection method according to an embodiment of the present invention, the exemplary ultrasonic water meter 200, which is a water meter for remote meter reading shown in FIGS. 1 and 2, Automatically measures the indoor water usage every hour, and transmits the measured value once a day to the database server 201 via the communication network. Since the battery of the water meter 200 is consumed quickly when the ultrasonic water meter 200 transmits the measurement value every time it is measured, it is preferable that the measured value is transmitted to the database server 201 once a day.

Referring to FIG. 2, an automatic indoor leak detection method using a statistical outlier detection method according to an exemplary embodiment of the present invention is performed in a server 202, specifically, a control unit (not shown) of a server 202.

The server 202 judges the indoor leakage based on the indoor water usage measured by the water meter 200 according to the statistical outliers searching method according to an embodiment of the present invention, And notifies the corresponding customer through the Web server 204.

FIG. 3 is a flowchart of an automatic indoor leak detection method using a statistical outlier detection method according to an embodiment of the present invention. FIG. 3 is a flowchart illustrating an automatic indoor leak detection method using a statistical outlier detection method according to an embodiment of the present invention shown in FIG. Detects indoor leaks based on daily water usage data.

Referring to Fig. 3, in step S300, the server 202 obtains indoor water usage data measured every hour by the water meter 200 for remote meter reading. In the automatic indoor leak detection method using the statistical outlier detection method according to an embodiment of the present invention, the water usage measurement time interval is one hour, but the present invention is not limited to this, and the server 202 may be shorter The water usage data may be obtained from the water meter 200 at every long measurement time interval.

In step S302, the server 202 determines whether the indoor water usage pattern is abnormal based on indoor indoor water usage data measured every hour.

That is, the server 202 determines that the indoor water usage pattern of the corresponding cafeteria is abnormal if the minimum value of the water usage data among the indoor water usage data measured every hour is not zero.

FIG. 4A shows leak data of an 8-day leaked water of a customer who has leaked indoors on the 6th day. Reference numeral 400 denotes an enlarged normal pattern of one day. Reference numeral 402 denotes an enlarged leak pattern of one day.

FIG. 4B shows a normal water usage pattern. When the water usage pattern is normal, it is determined that the water usage pattern is normal because the water usage minimum value is zero.

FIG. 4C shows an abnormal water usage pattern. When the water usage pattern is abnormal, since the minimum value of the water usage is not 0, it is doubtful as the indoor water leakage and the indoor water leakage is determined using the outlier detection algorithm.

That is, in the automatic indoor leak detection method using the statistical outliers search method according to an embodiment of the present invention, when the water meter 200 detects the indoor water leak, In order to reduce the number of indoor leaks, indoor leaks are detected only in case of suspected indoor leaks.

In step S304, when the water usage pattern is abnormal, the server 202 determines that the water usage pattern is an abnormal day as a day on which the water leakage is suspected. If the water usage pattern is abnormal on the day when the water leakage is suspected (500 in FIG. 5) , For example, 20 days of daily use (502 in FIG. 5).

For reference, it is assumed that the water usage data periodically measured by the water meter 200 is stored in the database server 201.

In step S306, the server 202 calculates an average and a standard deviation of the obtained total 21 daily use amounts.

Illustratively, for a total of 21 daily usage amounts shown in FIG. 5, the average is 19.97 and the standard deviation is 11.23.

In step S308, the server 202 calculates the statistical value G for the water use amount (A = 64.57) on the day when the leakage is suspected according to the equation (1).

In the present embodiment, G = | 64.57-19.97 | /11.23=3.97.

In step S310, the server 202 sets a threshold value corresponding to the significance level of the upper 5% according to the number of data (21 in this embodiment) in the critical value table shown in Fig. 6 To obtain a value (CV).

The threshold value table shown in FIG. 6 is a part of a threshold value table used in the Grubbs' method.

It is assumed that the threshold value table is stored in a memory (not shown) of the server 202. The server 202 does not necessarily have to store the entire threshold table in the memory, but it does not necessarily have to store the threshold value CV of 2,580 (CV) corresponding to the significance level of the top 5% Or a threshold value CV corresponding to the significance level of the upper 5% according to the number of 31 data.

Since the number of data is 21 in this embodiment, the threshold value CV corresponding to the significance level of the upper 5% in the threshold value table of FIG. 6 is 2.580.

In step S312, the server 202 determines whether the statistical value G is greater than the threshold value CV.

In this embodiment, the statistical value G is 3.97 and the threshold value CV is 2.759, so that the statistical value G is larger than the threshold value CV.

Therefore, in step S314, the server 202 determines that the indoor leak is occurring, and notifies the customer that the indoor leak has occurred.

7 is a flowchart of an automatic indoor leak detection method using a statistical outlier detection method according to another embodiment of the present invention. In FIG. 7, in order to more accurately detect indoor leakage, indoor leak detection is performed based on daily and weekly water usage data. do.

Referring to Fig. 7, in step S700, the server 202 obtains the indoor water usage data measured every hour by the water meter 200 for remote meter reading. In the statistical outlier detection method according to the embodiment of the present invention, the water usage measurement time interval is one hour, but the present invention is not limited to this, and the server 202 may store the water usage data May be obtained from the water meter (200).

In step S702, the server 202 determines whether the water usage pattern is abnormal based on the indoor water usage data measured every hour.

That is, the server 202 determines that the indoor water usage pattern of the corresponding cafeteria is abnormal when the minimum value of the water use amount among the indoor water usage data measured every hour is not zero.

FIG. 4A shows leak data of an 8-day leaked water of a customer who has leaked indoors on the 6th day. Reference numeral 400 denotes an enlarged normal pattern of one day. Reference numeral 402 denotes an enlarged leak pattern of one day.

FIG. 4B shows a normal water usage pattern. When the water usage pattern is normal, it is determined that the water usage pattern is normal because the water usage minimum value is zero.

FIG. 4C shows an abnormal water usage pattern. When the water usage pattern is abnormal, since the minimum value of the water usage is not 0, it is doubtful as the indoor water leakage and the indoor water leakage is determined using the outlier detection algorithm.

That is, in the automatic indoor leak detection method using the statistical outliers search method according to an embodiment of the present invention, when the water meter 200 detects the indoor water leak, In order to reduce the number of indoor leaks, indoor leaks are detected only in case of suspected indoor leaks.

In step S704, when the water usage pattern is abnormal, the server 202 determines that the water usage pattern is abnormal on the day when the water leakage is suspected, and the daily usage amount (800 in FIG. 8) and the week usage amount 8 802), and the daily usage amount (804 in FIG. 8) and the weekly usage amount (806 in FIG. 8) between a predetermined day, for example, 20 days.

For reference, it is assumed that the water usage data periodically measured by the water meter 200 is stored in the database server 201.

In step S706, the server 202 converts the obtained two-dimensional data of the daily use amount and the weekly usage amount into one-dimensional data by using Principal Component Analysis (PCA).

Specifically, as shown in FIG. 9, the server 202 determines an axis having the maximum variance with respect to the obtained two-dimensional data of the daily use amount and the weekly usage amount as the principal component axis 900, Dimensional data of the weekly usage amount onto the determined main component axis 900 and converts the obtained two-dimensional data of daily use amount and weekly usage amount into one-dimensional data.

10 shows the conversion of the two-dimensional data 1000 of the daily use amount and the weekly usage amount into the one-dimensional data 1002 using the PCA.

In step S708, the server 202 calculates an average and a standard deviation of a total of 21 data converted into the one-dimensional data.

Illustratively, for a total of 21 data shown in FIG. 10, the mean is zero and the standard deviation is 19.37.

In step S710, the server 202 calculates the statistical value G for the one-dimensional data (B = 79.76) of the day when the leak is suspected according to the equation (2).

In the present embodiment, G = | 79.76-0 | /19.37=4.12.

In step S712, the server 202 sets a threshold value corresponding to the significance level of the upper 5% according to the number of data (21 in this embodiment) in the critical value table shown in Fig. 6 To obtain a value (CV).

The threshold value table shown in FIG. 6 is a part of a threshold value table used in the Grubbs' method.

It is assumed that the threshold value table is stored in a memory (not shown) of the server 202. The server 202 does not necessarily have to store the entire threshold table in the memory, but it does not necessarily have to store the threshold value CV of 2,580 (CV) corresponding to the significance level of the top 5% Or a threshold value CV corresponding to the significance level of the upper 5% according to the number of 31 data.

Since the number of data is 21 in this embodiment, the threshold value CV corresponding to the significance level of the upper 5% in the threshold value table of FIG. 6 is 2.580.

In step S714, the server 202 determines whether the statistical value G is larger than the threshold value CV.

In this embodiment, the statistical value G is 4.12, and the threshold value CV is 2.580, so that the statistical value G is larger than the threshold value.

Accordingly, in step S716, the server 202 determines the indoor leakage, and notifies the corresponding customer that the indoor leakage has occurred.

According to the automatic indoor leak detection method using the statistical outlier detection method according to an embodiment of the present invention, since the indoor leak detection is detected using the statistical outlier detection method, It is possible to accurately detect the occurrence of leakage, thereby reducing the cost of detecting the leakage of indoor water and preventing the waste of water caused by indoor leakage, thereby protecting the water resources.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. It is clear that the present invention can be modified or improved.

It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

200: Ultrasonic water meter 201: Database server
202: server 204: web server
400: one day normal pattern 402: one day leakage pattern
500: Daily usage data for a suspected leaky day
502: daily usage data for 20 days
800: daily usage data for suspected leak
802: Weekly usage data for suspected leak
804: daily usage data for 20 days
806: 20 day weekly usage data
900: main component axis
1000: Two-dimensional data of daily usage and weekly usage
1002: One-dimensional data

Claims (9)

delete delete delete delete (a) acquiring indoors water usage data measured by the remote meter reading water meter every predetermined time;
(b) determining whether the indoor water usage pattern is abnormal based on the water usage data measured every predetermined time;
(c) If the water usage pattern is abnormal, the water usage pattern is determined to be the day when the water leakage is suspected to be an abnormal day, and if the water usage pattern is suspected, the daily and weekly usage amount and the previous Obtaining daily and weekly usage amounts;
(d) converting the two-dimensional data of the obtained daily use amount and the weekly usage amount into one-dimensional data by using principal component analysis (PCA);
(e) calculating an average and a standard deviation of the data converted into the one-dimensional data;
(f) The statistical value (G) for the one-dimensional data (B)

≪ / RTI >
(g) obtaining a threshold value (CV) corresponding to a significance level of the upper 5% according to the number of data in the critical value table; And
(h) if the statistical value (G) is greater than the threshold value (CV), determining as an indoor leak,
The step (d)
(d-1) determining, as a principal component axis, an axis having a maximum dispersion with respect to two-dimensional data having the daily use amount as a first axis and the weekly usage amount as a second axis; And
(d-2) projecting two-dimensional data of the daily use amount and the weekly usage amount onto the principal component axis, and converting the two-dimensional data of the daily usage amount and the weekly usage amount into one- Automatic Indoor Leak Detection Using Statistical Outlier Detection Method. The method of claim 5,
The step (b)
And determining that the indoor water usage pattern is abnormal when the minimum water use amount among the indoor water usage data measured at the predetermined time is not zero. The method of claim 5,
Wherein the threshold table includes a threshold table used in a Grubbs' method. ≪ Desc / Clms Page number 21 > The method of claim 7,
Wherein the threshold value table comprises:

The method comprising the steps of: delete

Images