Water leakage detection method and system for closed-circuit cooling water system of wind turbine

Abstract

The present invention belongs to the field of water leakage detection technology, and relates to a water leakage detection method for a closed-circulation cooling water system of a wind turbine. Based on the pressure data of the water system, it can detect and warn of short-term rapid water leakage and long-term hidden water leakage in the pipeline; the core part of the short-term water leakage method includes two parts. One is to divide the pressure data into hours and then take the average value. This algorithm is to eliminate the pressure fluctuation factors under the start-stop state of the pump after taking the average value, so that the algorithm is more accurate. The other is that when the threshold is triggered, it is necessary to determine that M consecutive points meet the threshold before it is considered valid. The core part of the long-term water leakage algorithm includes two parts. One is to divide the pressure data into days and then take the median. This algorithm can eliminate the pressure fluctuation factors under the start-stop state of the pump, so that the algorithm is more accurate. The other is that when the threshold is triggered, it is necessary to determine that N consecutive points meet the threshold before it is considered valid.

Description

Technical Field

[0001] The present invention belongs to the technical field of water leakage detection, and in particular relates to a water leakage detection method and system for a closed-circulation cooling water system of a wind turbine generator. Background Art

[0002] At present, the cooling method for generator equipment, converter equipment, transformer and other equipment in wind power systems usually adopts the form of a closed cooling water system. Since water circulates in a closed pipe, there will be no evaporation loss caused by contact with air. Therefore, if there is no water leakage, except for slight fluctuations due to changes in ambient temperature, the pressure distribution in the loop system remains basically unchanged when the water pump is not turned on. The pressure of each part is determined by the pre-charge pressure of the original expansion tank and the most recent system pressure.

[0003] If a water leak occurs, the system pressure will gradually decrease until the cooling system fails to work properly, further affecting the normal operation of the wind turbine. Therefore, water leaks must be effectively identified and protected.

[0004] A novel search revealed patent document CN201710457244.6, which discloses a method for detecting water leakage in a converter valve cooling system. The method comprises the following steps: Step 1: Obtaining historical operating data of the converter valve cooling system from the converter station SCADA system, and establishing a corresponding relationship between the expansion tank pressure, the expansion tank liquid level, and the converter valve outlet water temperature; Step 2: When the converter valve cooling system is operating, calculating the predicted pressure of the expansion tank based on the expansion tank liquid level, the converter valve outlet water temperature, and the corresponding relationship; Step 3: Determining whether the converter valve cooling system is leaking based on the difference between the actual pressure and the predicted pressure. The detection method and detection system of the present invention fully utilize the characteristics of pressure sensitivity and the accuracy of big data analysis, achieve high prediction accuracy, and implement the functions of detecting minor leaks in the valve cooling system and continuously monitoring them.

[0005] The current defects are:

[0006] 1) Currently, leak detection primarily relies on setting a minimum pressure threshold to trigger an alarm. Leakage continues until the pressure drops below a certain threshold, and in more severe cases, the equipment is shut down. However, to prevent false alarms, the pressure threshold is often set too low. By then, the leak has already persisted for some time, preventing timely detection.

[0007] 2) The patented "A Method and System for Detecting Water Leakage in a Converter Valve Cooling System" uses a relatively scientific approach to calculate the total water volume of the entire water system in real time. When a decrease in total water volume is detected, the predicted system pressure trend is compared with the actual trend to issue an alarm. This method is a relatively scientific way to assess leaks, but accurate calculations require a large number of parameters, such as the water temperature within the water circuit, the liquid level in the water tank, and the volume of the expansion tank. These parameters are difficult to obtain, and the varying temperatures at various points in the cooling circuit can result in an excessive number of monitoring points, making implementation impossible. Summary of the Invention

[0008] The purpose of the present invention is to provide a method and system for detecting water leakage in a closed-circuit cooling water system of a wind turbine, which solves the problems in the prior art of not being able to detect water in a timely manner due to a relatively small pressure threshold setting and not being able to implement water leakage due to too many monitoring points.

[0009] The present invention is achieved through the following technical solutions:

[0010] A method for detecting water leakage in a closed-circuit cooling water system of a wind turbine generator comprises the following steps:

[0011] S1. Obtaining the operating data of the closed-loop cooling water system of the wind turbines in the target wind farm and performing data cleaning. The cleaned data includes the pressure data of the cooling water system of each unit;

[0012] S2. Calculate the design value of the cooling system pressure;

[0013] S3. Input the pressure data into the first pressure processing model to obtain the average pressure data of i intervals of each unit on that day;

[0014] According to the pressure average data of the i intervals of each unit on that day and the design value of the cooling system pressure, the pressure average change rate of two adjacent intervals is calculated to obtain i-1 pressure average change rate data;

[0015] S4. Input the pressure data into the second pressure processing model, calculate the median pressure of the unit on the current day and the median pressure of the historical days, and calculate the change rate based on the design value of the cooling system pressure to obtain the daily median pressure change rate;

[0016] S5. Send the i-1 pressure average change rate data to the short-term water leakage warning model. The short-term water leakage warning model stores the hourly trend pressure change threshold. If M points in the i-1 pressure average change rate series data exceed the hourly trend pressure change threshold, an abnormality is marked and a short-term water leakage warning is triggered.

[0017] The daily median pressure change rate data is sent to the long-term water leakage warning model, which stores the daily trend pressure change threshold. If N points continuously exceed the daily trend pressure change threshold, an abnormality mark will be issued and a long-term water leakage warning will be triggered.

[0018] Furthermore, in S3, the processing of the first pressure processing model specifically includes the following steps:

[0019] 3.1. Segment the pressure data into i intervals at intervals of L hours, where L >= 2.

[0020] Calculate the average values of each interval Pavg1, Pavg2, Pavg3, ... Pavgi;

[0021] 3.2. Calculate the change rate of the average pressure between two adjacent intervals:

[0022] Kph i-1=(Pavg i-Pavg i-1) / Pdesign;

[0023] Then we get the data set {Kph 1, Kph 2...Kph i-1};

[0024] Where Pdesign is the design value of the cooling system pressure.

[0025] Furthermore, in S4, the processing of the second pressure processing model specifically includes the following steps:

[0026] 4.1. Select the historical data of a single unit to calculate the pressure change trend. The time range is N natural days before the calculation day, and calculate the median of daily pressure P. 50 i, where i represents the number of days, and a set of P 50 1. P 50 2. P 50 3. P 50 4. P 50 5. P 50 6. P 50 7......P 50 N+1;

[0027] 4.2. Calculate the median daily pressure change rate:

[0028] Kpd1=(P 50 1-P 50 2) / Pdesign;

[0029] Kpd2=(P 50 2-P 50 3) / Pdesign;

[0030] …

[0031] KpdN=(P 50 NP 50 N+1) / Pdesign;

[0032] A total of N Kpd values are obtained, where Pdesign is the design value of the cooling system pressure.

[0033] Furthermore, in S5, the hourly trend pressure change threshold is recorded as Kh0, which is divided into three levels: Kh1, Kh2 and Kh3, among which Kh1 <Kh2<Kh3:

[0034] When Kh0 selects Kh1, the short-term water leakage warning model will sound a warning;

[0035] Kh0 selects Kh2, and the short-term water leakage warning model issues an alarm;

[0036] Kh0 selects Kh3, and the short-term water leakage warning model reports an emergency.

[0037] Furthermore, in S5, the daily trend pressure change threshold is recorded as Kd0, which is divided into three levels: Kd1, Kd2 and Kd3, among which Kd1 <Kd2<Kd3:

[0038] When Kd0 selects Kd1, the long-term water leakage warning model will sound a warning;

[0039] Kd0 selects Kd2, and the long-term water leakage warning model issues an alarm;

[0040] Kd0 selects Kd3, and the long-term water leakage warning model reports an emergency.

[0041] Furthermore, in S5, M≥3, N≥3.

[0042] Furthermore, in S1, data cleaning specifically includes: deleting duplicate values, supplementing missing values, or eliminating interrupted data.

[0043] Furthermore, in S1, the operating data of the closed-loop cooling water system of the wind turbine generator is required to be no less than one month's data.

[0044] Furthermore, in S2, the design pressure value of the cooling system is Pdesign. When the Pdesign value cannot be obtained, all pressure data P of all organic groups in the entire site within one month are taken and sorted in descending order. The data ranked 5%-25% are taken for average calculation, and the final result is Pdesign.

[0045] The present invention also discloses a water leakage detection system for a closed-circulation cooling water system of a wind turbine generator, comprising:

[0046] A pressure data acquisition module is used to obtain the operating data of the closed-loop cooling water system of the wind turbines in the target wind farm;

[0047] The data cleaning module is used to clean the operating data of the closed-loop cooling water system of the wind turbine. The cleaned data includes the pressure data of the cooling water system of each unit;

[0048] A cooling system pressure design value obtaining module is used to obtain the cooling system pressure design value;

[0049] The first pressure processing model is used to process the pressure data to obtain the average pressure data of each unit in i intervals on the same day;

[0050] According to the pressure average data of the i intervals of each unit on that day and the design value of the cooling system pressure, the pressure average change rate of two adjacent intervals is calculated to obtain i-1 pressure average change rate data;

[0051] The second pressure processing model is used to process the pressure data, obtain the median pressure of the unit on the day and the median pressure of the historical days, and solve the change rate in combination with the design value of the cooling system pressure to obtain the daily median pressure change rate;

[0052] The short-term water leakage warning model stores an hourly trend pressure change threshold, which is used to compare the i-1 pressure average change rate data with the hourly trend pressure change threshold. If M points in the i-1 pressure average change rate series data exceed the hourly trend pressure change threshold, an abnormality mark is issued, triggering a short-term water leakage warning;

[0053] The long-term water leakage warning model stores a daily trend pressure change threshold, which is used to compare the daily median pressure change rate data with the daily trend pressure change threshold. If N points continuously exceed the daily trend pressure change threshold, an abnormality mark will be issued to trigger a long-term water leakage warning.

[0054] Compared with the prior art, the present invention has the following beneficial technical effects:

[0055] The present invention discloses a method for detecting water leakage in a closed-circuit cooling water system of a wind turbine. Based on the water system pressure data in SCADA historical data, the method adopts an early warning algorithm combining long-term water leakage detection and short-term water leakage, and can detect and warn short-term rapid water leakage and long-term hidden water leakage in the pipeline. The core parts of the short-term water leakage method include two parts: one is to divide the pressure data into hours and then take the average value. This algorithm is to eliminate the pressure fluctuation factors under the pump start-stop state after taking the average value, making the algorithm more accurate; the other is that when the threshold is triggered, it is necessary to determine whether M consecutive points meet the threshold before it is considered valid. The core parts of the long-term water leakage algorithm include two parts: one is to divide the pressure data into days and then take the median. This algorithm can eliminate the pressure fluctuation factors under the pump start-stop state after taking the average value, making the algorithm more accurate; the other is that when the threshold is triggered, it is necessary to determine whether N consecutive points meet the threshold before it is considered valid.

[0056] Furthermore, in order to compare with the design value of the cooling system pressure and make the result more accurate, a method for obtaining the design value is specially introduced. By calculating the data of all organic groups in the entire site and sorting them, the high and low values are removed, and the data in the range of 5%-25% are selected for average value processing to avoid the problem of inaccurate threshold setting due to the missing design value. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 The present invention is a flow chart of a method for detecting water leakage in a closed-circulation cooling water system of a wind turbine generator. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solutions and advantages of the present invention more clear, the following is a further detailed description with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. That is, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments.

[0059] The components described and illustrated in the drawings and embodiments of the present invention may be arranged and designed in a variety of different configurations. Therefore, the detailed description of the embodiments of the present invention provided in the following drawings is not intended to limit the scope of the claimed invention, but merely represents a selected embodiment of the present invention. All other embodiments derived by those skilled in the art based on the drawings and embodiments of the present invention without inventive effort shall fall within the scope of protection of the present invention.

[0060] The features and performance of the present invention are further described in detail below with reference to the embodiments.

[0061] like Figure 1As shown, this invention analyzes operational data from closed-loop water systems to develop a closed-loop waterway system leakage detection method. This method can be divided into long-term and short-term leakage detection. The algorithm can be run in real time on a wind farm-level server. When an anomaly is detected, an alarm is triggered, prompting maintenance personnel to perform maintenance. The invention consists of three parts: data preparation, model calculation, early warning judgment, and operation and maintenance prompts.

[0062] 1. Data preparation

[0063] 1) Collect and store the operating data of the closed-loop cooling water system of the wind turbines in the target wind farm. The data should be no less than one month, which can fully reflect various operating conditions and external wind conditions to make the data more effective.

[0064] 2) The returned data is placed on the field server according to the specified location and number for storage;

[0065] 3) Data cleaning work, including deleting duplicate values, supplementing missing values, and removing interrupted data.

[0066] 2. Model calculation

[0067] 2.1 Obtaining the design value of the cooling system pressure

[0068] Assume that Pdesign is the design value of the cooling system pressure, which is the pressure value when the system is operating normally. If this value is difficult to obtain, the calculation can be performed based on the unit water cooling system pressure data P. Take all the pressure data P of all units in the site within one month and sort them in descending order (from large to small). Take the data ranked 5%-25% and calculate the average value to obtain the final result Pdegisn.

[0069] For example, if the pressure data of all units in a month is 10,000 points, they are sorted from large to small, and the values of 2,000 points between the 501st value and the 2,500th value are averaged.

[0070] 2.2 Calculate short-term water leakage

[0071] Perform early warning calculations on a single unit, calculate the 2-hour average pressure of the unit on that day, and solve the rate of change (short-term water leakage):

[0072] 1) Calculate the average value Pavgh1, Pavgh2, Pavgh3, ... of the pressure data at intervals of L hours (L>=2). For example, if L=2, all data are sorted by time and the average value is calculated every 2 hours. For example, Pavgh1 is the average value of the pressure data from 00:00 to 02:00, and Pavgh2 is the average value of the pressure data from 02:00 to 04:00.

[0073] 2) Calculate the 2-hour average pressure change rate of two adjacent points:

[0074] Kph1=(Pavgh1-Pavgh2) / Pdesign;

[0075] Kph2=(Pavgh2-Pavgh3) / Pdesign;

[0076] …

[0077] Kph11=(Pavgh11-Pavgh12) / Pdesign;

[0078] A total of 11 Kph values (hourly pressure average change rate) were obtained.

[0079] 2.3 Calculate long-term water leakage

[0080] Perform early warning calculations on a single unit, obtain the median pressure of the unit on that day and the median pressure of historical days, and calculate the rate of change (long-term water leakage and seepage)

[0081] 1) Select the historical data of a single unit to calculate the pressure change trend. The time range is usually N natural days before the calculation day. For example, if the calculation day is January 20, select the data from January 13 to January 20, where N is 7 (usually a natural number N ≥ 7). For the data from January 13 to January 20, calculate the median of the daily pressure (50% ranking number): P 50 i (where i represents the number of days), and obtain a set of P 50 1. P 50 2. P 50 3. P 50 4. P 50 5. P 50 6. P 50 7. P 50 8. Corresponding to the data from January 13th to January 20th.

[0082] 2) Calculate the median daily pressure change rate

[0083] Kpd1=(P 50 1-P 50 2) / Pdesign;

[0084] Kpd2=(P 50 2-P 50 3) / Pdesign;

[0085] …

[0086] Kpd7=(P 50 7-P 508) / Pdesign;

[0087] A total of 7 Kph values (daily median pressure change rate) were obtained.

[0088] 3. Early warning judgment

[0089] 1) Set the hourly trend pressure change threshold Kh0 and the daily trend pressure change threshold Kd0;

[0090] 2) When M points appear consecutively (such as the average pressure change rate for M consecutive hours) exceeding the hourly trend pressure change threshold, an abnormality is marked and a short-term water leakage warning is triggered.

[0091] Kphi,Kphi+1,Kphi+M>Kh0(M>=3);

[0092] 3) When N consecutive points (such as the median pressure change rate for N consecutive days) exceed the daily trend pressure change threshold, an abnormality mark is issued and a long-term water leakage warning is triggered.

[0093] Kpdi,Kpdi+1,Kpdi+N>Kd0(N>=3).

[0094] 4. Operation and maintenance tips

[0095] 1) The threshold Kh0 is divided into 3 levels, Kh1, Kh2 and Kh3. The above 3 numbers are all positive numbers, among which Kh1 <Kh2<Kh3:

[0096] When Kh0 is selected as Kh1, it will report attention; when Kh0 is selected as Kh2, it will report warning; when Kh0 is selected as Kh3, it will report emergency;

[0097] 2) The threshold value Kd0 is divided into 3 levels, Kd1, Kd2 and Kd3. The above 3 numbers are all positive numbers, among which Kd1 <Kd2<Kd3:

[0098] When Kd0 selects Kd1, it reports attention; when Kd0 selects Kd2, it reports warning; when Kd0 selects Kd3, it reports emergency.

[0099] The present invention also discloses a water leakage detection system for a closed-circulation cooling water system of a wind turbine generator, comprising:

[0100] A pressure data acquisition module is used to obtain the operating data of the closed-loop cooling water system of the wind turbines in the target wind farm;

[0101] The data cleaning module is used to clean the operating data of the closed-loop cooling water system of the wind turbine. The cleaned data includes the pressure data of the cooling water system of each unit;

[0102] A cooling system pressure design value obtaining module is used to obtain the cooling system pressure design value;

[0103] The first pressure processing model is used to process the pressure data to obtain the average pressure data of each unit in i intervals on the same day;

[0104] According to the pressure average data of the i intervals of each unit on that day and the design value of the cooling system pressure, the pressure average change rate of two adjacent intervals is calculated to obtain i-1 pressure average change rate data;

[0105] The second pressure processing model is used to process the pressure data, obtain the median pressure of the unit on the day and the median pressure of the historical days, and solve the change rate in combination with the design value of the cooling system pressure to obtain the daily median pressure change rate;

[0106] The short-term water leakage warning model stores an hourly trend pressure change threshold, which is used to compare the i-1 pressure average change rate data with the hourly trend pressure change threshold. If M points in the i-1 pressure average change rate series data exceed the hourly trend pressure change threshold, an abnormality mark is issued, triggering a short-term water leakage warning;

[0107] The long-term water leakage warning model stores a daily trend pressure change threshold, which is used to compare the daily median pressure change rate data with the daily trend pressure change threshold. If N points continuously exceed the daily trend pressure change threshold, an abnormality mark will be issued to trigger a long-term water leakage warning.

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for detecting water leakage in a closed-circuit cooling water system of a wind turbine generator, characterized in that: The following steps are involved: S1. Obtaining the operating data of the closed-loop cooling water system of the wind turbines in the target wind farm and performing data cleaning. The cleaned data includes the pressure data of the cooling water system of each unit; S2. Calculate the design value of the cooling system pressure; S3. Input the pressure data into the first pressure processing model to obtain the average pressure data of i intervals of each unit on that day; According to the pressure average data of the i intervals of each unit on that day and the design value of the cooling system pressure, the pressure average change rate of two adjacent intervals is calculated to obtain i-1 pressure average change rate data; S4. Input the pressure data into the second pressure processing model, calculate the median pressure of the unit on the current day and the median pressure of the historical days, and calculate the change rate based on the design value of the cooling system pressure to obtain the daily median pressure change rate; S5. Send the i-1 pressure average change rate data to the short-term water leakage warning model. The short-term water leakage warning model stores the hourly trend pressure change threshold. If M points in the i-1 pressure average change rate series data exceed the hourly trend pressure change threshold, an abnormality is marked and a short-term water leakage warning is triggered. The daily median pressure change rate data is sent to the long-term water leakage warning model, which stores the daily trend pressure change threshold. If N points continuously exceed the daily trend pressure change threshold, an abnormality mark will be issued and a long-term water leakage warning will be triggered.

2. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S3, the processing of the first pressure processing model specifically includes the following steps: 3.

1. Segment the pressure data into i intervals at intervals of L hours, where L >= 2. Calculate the average values of each interval Pavg1, Pavg2, Pavg3, ... Pavgi; 3.

2. Calculate the change rate of the average pressure between two adjacent intervals: Kph i-1=(Pavg i-Pavg i-1) / Pdesign; Then we get the data set {Kph 1, Kph 2...Kph i-1}; Where Pdesign is the design value of the cooling system pressure.

3. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S4, the processing of the second pressure processing model specifically includes the following steps: 4.

1. Select the historical data of a single unit to calculate the pressure change trend. The time range is N natural days before the calculation day, and calculate the median of daily pressure P. 50 i, where i represents the number of days, to obtain a set of P 50 1. P 50 2. P 50 3. P 50 4. P 50 5. P 50 6. P 50 7......P 50 N+1; 4.

2. Calculate the median daily pressure change rate: Kpd1=(P 50 1-P 50 2) / Pdesign; Kpd2=(P 50 2-P 50 3) / Pdesign; … KpdN=(P 50 N-P 50 N+1) / Pdesign; A total of N Kpd values are obtained, where Pdesign is the design value of the cooling system pressure.

4. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S5, the hourly trend pressure change threshold is recorded as Kh0, which is divided into three levels: Kh1, Kh2 and Kh3. <Kh2<Kh3: When Kh0 selects Kh1, the short-term water leakage warning model will sound a warning; Kh0 selects Kh2, and the short-term water leakage warning model issues an alarm; Kh0 selects Kh3, and the short-term water leakage warning model reports an emergency.

5. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S5, the daily trend pressure change threshold is recorded as Kd0, which is divided into three levels: Kd1, Kd2 and Kd3. <Kd2<Kd3: When Kd0 selects Kd1, the long-term water leakage warning model will sound a warning; Kd0 selects Kd2, and the long-term water leakage warning model issues an alarm; Kd0 selects Kd3, and the long-term water leakage warning model reports an emergency.

6. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S5, M≥3, N≥3.

7. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S1, data cleaning specifically includes: deleting duplicate values, supplementing missing values, or eliminating interrupted data.

8. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S1, the operating data of the closed-loop cooling water system of the wind turbine generator must be no less than one month's data.

9. A method for detecting water leakage in a closed-circulation cooling water system of a wind turbine according to claim 1, characterized in that: In S2, the design pressure value of the cooling system is Pdesign. When the Pdesign value cannot be obtained, all pressure data P of the organic groups in the entire site within one month are taken and sorted in descending order. The data ranked 5%-25% are taken for average calculation, and the final result is Pdesign.

10. A water leakage detection system for a closed-circuit cooling water system of a wind turbine generator, characterized in that: include: A pressure data acquisition module is used to obtain the operating data of the closed-loop cooling water system of the wind turbines in the target wind farm; The data cleaning module is used to clean the operating data of the closed-loop cooling water system of the wind turbine. The cleaned data includes the pressure data of the cooling water system of each unit; A cooling system pressure design value obtaining module is used to obtain the cooling system pressure design value; The first pressure processing model is used to process the pressure data to obtain the average pressure data of each unit in i intervals on the same day; According to the pressure average data of the i intervals of each unit on that day and the design value of the cooling system pressure, the pressure average change rate of two adjacent intervals is calculated to obtain i-1 pressure average change rate data; The second pressure processing model is used to process the pressure data, obtain the median pressure of the unit on the day and the median pressure of the historical days, and solve the change rate in combination with the design value of the cooling system pressure to obtain the daily median pressure change rate; The short-term water leakage warning model stores an hourly trend pressure change threshold, which is used to compare the i-1 pressure average change rate data with the hourly trend pressure change threshold. If M points in the i-1 pressure average change rate series data exceed the hourly trend pressure change threshold, an abnormality mark is issued, triggering a short-term water leakage warning; The long-term water leakage warning model stores a daily trend pressure change threshold, which is used to compare the daily median pressure change rate data with the daily trend pressure change threshold. If N points continuously exceed the daily trend pressure change threshold, an abnormality mark will be issued to trigger a long-term water leakage warning.

Images