Water leakage early warning processing method, device and equipment for remote water meter

Abstract

The invention provides a water leakage early warning processing method, device and equipment for a remote water meter. The method comprises the following steps: acquiring real-time flow data of the remote water meter; preprocessing the real-time traffic data to obtain processed traffic data; monitoring the processed flow data in real time according to a preset monitoring rule to obtain abnormal condition information; wherein the preset monitoring rule comprises at least one of a data uploading rule, a negative water consumption rule and a threshold alarm rule; wherein the threshold alarm rule comprises at least one of a normal threshold rule, a dynamic threshold rule and a weight threshold rule; performing grading processing on the abnormal condition information to obtain alarm records of different grades; and alarm records of different levels are pushed. According to the scheme, different levels of alarms of various abnormal conditions can be achieved, the abnormal condition of water consumption of the remote water meter can be found in time, water waste is reduced, the problems of false alarm, alarm leakage, alarm lag and the like are avoided, and the abnormal condition of the remote water meter can be found and processed in time.

Description

TECHNICAL FIELD

[0001] The present application relates to the water meter leakage early warning technical field, in particular to a kind of remote water meter leakage early warning processing method, device and equipment. BACKGROUND

[0002] At present, remote water meter leakage monitoring is mainly realized by setting conventional threshold alarm to realize early warning and alarm reminder. However, this traditional method has many shortcomings: Lack of dynamic adjustment: cannot dynamically adjust the alarm threshold according to the actual water use of user, leading to false alarm, missed alarm and alarm lag, etc.; Limited monitoring range: only can monitor the threshold overrun of water consumption, cannot timely find other abnormalities except leakage, such as negative water consumption, remote water meter long-term non-uploaded data, etc.; Incomplete alarm processing: lack of grading processing and automatic alarm cancellation function of alarm record, cannot timely notify relevant personnel to handle abnormality, leading to water consumption loss and false charge, increasing user complaint rate. SUMMARY

[0003] The present application provides a kind of remote water meter leakage early warning processing method, device and equipment, can realize the alarm of different grades of multiple abnormal conditions, timely find remote water meter water abnormal condition, reduce water waste, avoid the occurrence of false alarm, missed alarm and alarm lag, etc. It is helpful to timely find and handle remote water meter abnormality, reduce false charge, reduce user complaint rate.

[0004] To solve the above technical problems, the technical scheme of the present application is as follows: A remote water meter leakage early warning processing method, comprising: obtaining real-time flow data of remote water meter; preprocessing the real-time flow data to obtain processed flow data; real-time monitoring the processed flow data according to preset monitoring rules to obtain abnormal condition information;Wherein, the preset monitoring rules include at least one of data uploading rule, negative water consumption rule and threshold alarm rule;Wherein, the threshold alarm rule includes at least one of normal threshold rule, dynamic threshold rule and weight threshold rule; grading processing the abnormal condition information to obtain alarm records of different levels;Push the alarm records of different levels to operation and maintenance management platform and user terminal equipment.

[0005] Optionally, real-time monitoring the processed flow data according to preset monitoring rules to obtain abnormal condition information, comprising: According to the data uploading rule, the processed flow data is monitored for data uploading, and if real-time flow data of the remote water meter is not acquired, first abnormal situation information is obtained; wherein the first abnormal situation information includes that flow data of the remote water meter on the day is not uploaded and flow data of the remote water meter is not uploaded for a first preset number of consecutive days.

[0006] Optionally, the processed flow data is monitored in real time according to a preset monitoring rule to obtain abnormal situation information, and the method further includes: According to the negative water use rule, the processed flow data is monitored for negative water use, and if the processed flow data of the remote water meter on the day minus the processed flow data of the last date with flow is negative, second abnormal situation information is obtained; wherein the second abnormal situation information includes that there is negative water use abnormality on the day, the number of times of negative water use abnormality occurring within a second preset number of days is greater than a preset number of times, and there is continuous negative water use abnormality for a third preset number of days.

[0007] Optionally, the processed flow data is monitored in real time according to a preset monitoring rule to obtain abnormal situation information, and the method further includes: According to the threshold value alarm rule, the processed flow data is monitored for a threshold value, and if a difference obtained by subtracting the processed flow data of the last date with flow from the processed flow data on the day is greater than a sum of daily threshold value alarm values of all dates within a date range of the day and the last date with flow, third abnormal situation information is obtained; wherein the third abnormal situation information includes that there is no threshold value alarm processing record when threshold value alarm occurs, there are a first preset number of threshold value alarm processing records when threshold value alarm occurs, and there are a second preset number of threshold value alarm processing records when threshold value alarm occurs; wherein the daily threshold value alarm value is at least one of a daily normal threshold value, a daily dynamic threshold value, and a daily weight threshold value.

[0008] Optionally, the abnormal situation information is processed according to a classification to obtain alarm records of different levels, including: According to the first abnormal situation information, a data uploading alarm record is generated; wherein the first abnormal situation information is that flow data of the remote water meter on the day is not uploaded, and a first-level data uploading alarm record is generated; the first abnormal situation information is that flow data of the remote water meter is not uploaded for a first preset number of consecutive days, and a second-level data uploading alarm record is generated.

[0009] Optionally, the abnormal situation information is processed according to a classification to obtain alarm records of different levels, and the method further includes: According to the second abnormal situation information, a negative water use alarm record is generated; wherein the second abnormal situation information is that there is negative water use abnormality on the day, and a first-level negative water use alarm record is generated; The second abnormal situation information is that the number of negative water usage anomalies within the second preset number of days is greater than the preset number, and a second-level negative water usage alarm record is generated; The second abnormal situation information is a continuous negative water usage abnormality for a third preset number of days, generating a third-level negative water usage alarm record.

[0010] Optionally, the process of classifying the abnormal situation information to obtain alarm records of different levels also includes: Based on the third abnormal situation information, a threshold alarm record is generated; The third abnormal situation information is that there is no threshold alarm processing record when a threshold alarm is triggered, and a first-level threshold alarm record is generated. The third abnormal situation information is that when a threshold alarm occurs, there is a first preset number of threshold alarm processing records, and a second level of threshold alarm records are generated. The third abnormal situation information is that when a threshold alarm occurs, there is a second preset number of threshold alarm processing records, generating a third level of threshold alarm records.

[0011] The present invention also provides a remote water meter leakage early warning and processing device, comprising: The acquisition module is used to acquire real-time flow data from the remote water meter. The processing module is used to preprocess the real-time traffic data to obtain processed traffic data; monitor the processed traffic data in real time according to preset monitoring rules to obtain abnormal situation information; wherein, the preset monitoring rules include at least one of: data upload rules, negative water usage rules, and threshold alarm rules; wherein, the threshold alarm rules include at least one of: normal threshold rules, dynamic threshold rules, and weighted threshold rules; classify the abnormal situation information to obtain alarm records of different levels; and push the alarm records of different levels to the operation and maintenance management platform and user terminal devices.

[0012] The present invention also provides a computing device, comprising: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above.

[0013] The present invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above.

[0014] The above-described solution of the present invention has at least the following beneficial effects: The above-mentioned solution of the present invention acquires real-time flow data from a remote water meter; preprocesses the real-time flow data to obtain processed flow data; monitors the processed flow data in real time according to preset monitoring rules to obtain abnormal situation information; wherein, the preset monitoring rules include at least one of data upload rules, negative water usage rules, and threshold alarm rules; wherein, the threshold alarm rules include at least one of normal threshold rules, dynamic threshold rules, and weighted threshold rules; the abnormal situation information is processed in a hierarchical manner to obtain alarm records of different levels; the alarm records of different levels are pushed to the operation and maintenance management platform and user terminal devices; it can realize different levels of alarms for various abnormal situations, promptly detect abnormal water usage of remote water meters, reduce water waste, avoid the occurrence of false alarms, missed alarms, and alarm lag, and help to promptly detect and handle abnormalities of remote water meters, reduce incorrect charges, and reduce user complaint rates. Attached Figure Description

[0015] Figure 1 This is a flowchart of a remote water meter leakage early warning processing method provided in an embodiment of the present invention; Figure 2 This is a module diagram of a remote water meter leakage early warning and processing device provided in an embodiment of the present invention. Detailed Implementation

[0016] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0017] like Figure 1 As shown, an embodiment of the present invention proposes a method for handling remote water meter leakage early warning, including: Step 11: Obtain real-time flow data from the remote water meter; Step 12: Preprocess the real-time traffic data to obtain processed traffic data; Step 13: Monitor the processed flow data in real time according to preset monitoring rules to obtain abnormal situation information; wherein, the preset monitoring rules include at least one of: data upload rules, negative water usage rules, and threshold alarm rules; wherein, the threshold alarm rules include at least one of: normal threshold rules, dynamic threshold rules, and weighted threshold rules. Step 14: Classify the abnormal situation information to obtain alarm records of different levels; push the alarm records of different levels to the operation and maintenance management platform and user terminal devices.

[0018] In this embodiment, real-time flow data from the remote water meter is acquired; the real-time flow data is preprocessed to obtain processed flow data; the processed flow data is monitored in real time according to preset monitoring rules to obtain abnormal situation information; wherein, the preset monitoring rules include at least one of data upload rules, negative water usage rules, and threshold alarm rules; wherein, the threshold alarm rules include at least one of normal threshold rules, dynamic threshold rules, and weighted threshold rules; the abnormal situation information is processed in a hierarchical manner to obtain alarm records of different levels; the alarm records of different levels are pushed to the operation and maintenance management platform and user terminal devices; by monitoring the processed flow data in real time through data upload rules, negative water usage rules, and threshold alarm rules, various abnormal situation information is obtained, and alarm records of different levels are obtained based on the various abnormal situation information, realizing different levels of alarms for various abnormal situations, timely detection of abnormal water usage in the remote water meter, reducing water waste, avoiding the occurrence of false alarms, missed alarms, and alarm lag, etc., which helps to timely detect and handle abnormalities in the remote water meter, reduce incorrect charges, and reduce user complaint rates.

[0019] In this embodiment, real-time flow data of the meter for only 30 days is obtained. If there are at least 15 records, the meter is identified as a remote water meter. The remote water meter has a built-in dual-mode communication module (NB-IoT+LoRa Mesh). The module firmware supports the MQTT5.0 protocol stack and QoS2 message confirmation mechanism, and has a Flash power-off resume cache (≥128KB) to cache flow data that was not successfully uploaded.

[0020] In this embodiment, the data upload rule is based on the MQTT5.0 dual-channel redundancy mechanism (NB-IoT (Narrowband Internet of Things) main channel reports every 15 minutes, LoRaMesh (wireless self-organizing network) backup channel reports every 1 hour). If any channel experiences 3 consecutive ACK (Acknowledgement) timeouts, a data upload anomaly is triggered. After negative water usage is detected, the negative water usage rule performs a secondary verification by comparing the mechanical reading of the electromechanical synchronous counter with the pressure change curve of the downstream pressure sensor. The downstream pressure sensor is a MEMS piezoresistive sensor with a range of 0-1.6MPa, an accuracy of ±0.25%FS, a sampling frequency of ≥1Hz, and communicates with the remote water meter via the Modbus-RTU protocol. The message header of the MQTT 5.0 protocol includes custom attribute fields: qos_level=2 (ensure one-time delivery); retain_flag=1 (retain the last message); user_property={"channel":"nbiot / lora","retry":3}.

[0021] In an optional embodiment of the present invention, step 12 includes: Step 121: Perform data cleaning on the real-time traffic data to obtain the first processed data; Step 122: Filter the first processed data to obtain the second processed data; Step 123: Perform format conversion on the second processed data to obtain processed traffic data.

[0022] In this embodiment, outlier removal and missing value imputation are performed on the real-time traffic data to obtain the first processed data; Specifically, the real-time traffic data is sorted to obtain a sorted sequence. ,in ; pass Find the position of the first quartile of the sorted sequence. Where n represents the total number of data in the sorted sequence; Indicates the position of the first quartile in the sorted sequence; if If it is an integer, then ;in, This represents the first quartile of the sorted sequence, i.e., the quartile of the sorted sequence. The value of each data point; if If it is a decimal, then ;in, express Integers; Indicates the first position in the sorted sequence The value of each data item; Indicates the first position in the sorted sequence The value of each data item; Indicates the first quartile of the sorted sequence; pass Find the position of the third quartile of the sorted sequence. ;like If it is an integer, then ;in, This represents the third quartile of the sorted sequence, i.e., the quartile of the sorted sequence. The value of each data point; if If it is a decimal, then ;in, express Integers; Indicates the first position in the sorted sequence The value of each data item; Indicates the first position in the sorted sequence The value of each data item; Represents the third quartile of the sorted sequence; pass The interquartile range is obtained; where, Indicates the interquartile range; Represents the third quartile of the sorted sequence; Indicates the first quartile of the sorted sequence; The upper and lower limits of the outlier threshold are determined using the interquartile range, the first quartile, and the third quartile. The upper limit is... The lower limit is The real-time traffic data that is greater than the upper limit or less than the lower limit is considered an outlier and will be removed. The removed data is linearly interpolated and adjusted for holidays to obtain the first processed data; specifically, the real-time traffic data sequence after outlier removal is as follows: ,pass This yields the interpolation for the missing time step; where t represents the missing time step. Indicates the moment preceding the missing moment; Indicates the moment immediately following the missing moment; Interpolation representing missing moments; express Real-time traffic data values ​​at any given moment; express Real-time traffic data values ​​at any given moment; For real-time traffic data that has undergone linear interpolation, arrange the data in chronological order and mark the positions of missing values; identify the types of holidays in the data; divide the data into "weekday group," "weekend group," and "specific holiday group" (such as National Day group, Spring Festival group), and statistically analyze the characteristics of each group (mean, median, and correlation with previous and subsequent days); perform preliminary linear interpolation calculations on missing values ​​to obtain preliminary interpolation values; extract traffic data for holidays in the past preset historical time period and calculate the difference patterns between them and the same period's weekdays; adjust the preliminary interpolation based on the difference patterns to obtain the final holiday interpolation values; obtain the first processed data; The first processed data is filtered to obtain the second processed data; specifically, the real-time traffic data is subjected to traffic pulse self-learning filtering. The filter coefficients are issued online by the cloud based on the residual sequence of the past 7 days. By filtering the first processed data, the device drift in the data can be deducted to ensure the accuracy of the data used subsequently. Convert the original units corresponding to the second processed data to a uniform value. The data is rounded to three decimal places to obtain the processed traffic data.

[0023] In an optional embodiment of the present invention, step 13 includes: Step 131: Monitor the data upload of the processed flow data according to the data upload rules. If the real-time flow data of the remote water meter is not obtained, obtain the first abnormal situation information. The first abnormal situation information includes the failure to upload the flow data of the remote water meter on the current day and the failure to upload the flow data of the remote water meter for a first preset number of consecutive days.

[0024] In this embodiment, if real-time flow data of the remote water meter is not obtained on a given day, an abnormal situation information indicating that the remote water meter flow data was not uploaded on that day is obtained, which is used as the first abnormal situation information. If real-time flow data of the remote water meter is not obtained for several consecutive days, and the number of days exceeds a first preset number of days, an abnormal situation information indicating that the remote water meter flow data has not been uploaded for a consecutive first preset number of days is obtained, which is used as the first abnormal situation information. The first preset number of days can be any integer greater than 1, for example, the first preset number of days can be 5. In specific applications, it can be adjusted according to the actual situation. Through the above process, abnormal situations such as the failure to upload remote water meter flow data can be detected in a timely manner, making it easier for relevant personnel to discover and respond promptly.

[0025] In an optional embodiment of the present invention, step 13 further includes: Step 1311: Based on the information of the first abnormal situation, perform root cause inference and automatic repair.

[0026] In this embodiment, when a data upload anomaly is triggered, i.e. when the first anomaly information is obtained, the Bayesian network is invoked to infer the root cause set {device power failure, SIM (subscriber identification card) failure, signal blind spot, platform anomaly}, and automatic repair operations are automatically performed: calling the operator API (application programming interface) base station signaling tracing; or, issuing the device AT (attention) command for remote wake-up; Based on the aforementioned root cause inference and automatic repair mechanism, the root cause of abnormal data uploads can be identified in a timely manner and automatically repaired, ensuring the normal operation of data uploads.

[0027] In an optional embodiment of the present invention, step 13 further includes: Step 132: Monitor the negative water usage data according to the negative water usage rule. If the negative water usage data of the remote water meter on the current day minus the negative water usage data of the previous day with water usage is a negative number, obtain the second abnormal situation information. The second abnormal situation information includes negative water usage abnormality on the current day, negative water usage abnormality occurring more than a preset number of times within a second preset number of days, and continuous negative water usage abnormality for a third preset number of days.

[0028] In this embodiment, if the processed flow rate data of the remote water meter on the current day minus the processed flow rate data of the previous day with flow is negative, it indicates that there is an abnormal situation of negative water usage. If there is negative water usage anomaly on a given day, obtain the anomaly information for negative water usage anomaly on that day, and use it as the second anomaly information; If multiple negative water usage anomalies occur within the second preset number of days, and the number of anomalies exceeds the preset number, then information indicating that the number of negative water usage anomalies within the second preset number of days exceeds the preset number is obtained and used as the second abnormal situation information. If negative water usage occurs for several consecutive days and the number of consecutive days exceeds the third preset number of days, the abnormal situation information of the third preset number of consecutive negative water usage is obtained as the second abnormal situation information. The second preset number of days, the preset number of times, and the third preset number of days can all be integers greater than 1. The second preset number of days is greater than the third preset number of days, and the preset number of times is less than the second preset number of days. For example, the second preset number of days can be 10; the preset number of times can be 3; and the third preset number of days can be 5. These can be adjusted according to the actual situation in specific applications. The above process can promptly detect abnormal water usage, facilitating timely detection and response by relevant personnel.

[0029] In an optional embodiment of the present invention, step 13 further includes: Step 1321: When the second abnormal situation information is obtained, the mechanical reading of the electromechanical synchronization counter of the remote water meter and the pressure change curve of the downstream pressure sensor during the negative water usage period are obtained. Step 1322: Perform bidirectional verification based on the mechanical reading of the electromechanical synchronization counter of the remote water meter and the pressure change curve of the downstream pressure sensor during the negative water usage period.

[0030] In this embodiment, when the second abnormal situation information is obtained, a two-way verification is performed based on the mechanical reading data of the electromechanical synchronization counter of the remote water meter and the pressure change curve data of the downstream pressure sensor during the negative water usage period. Specifically, a verification is performed when the absolute value of the difference between the mechanical and electronic readings of the electromechanical synchronous counter exceeds a first preset difference. A second verification is performed if the sudden change in the downstream pressure sensor exceeds the second preset difference. If both verifications pass, i.e., the absolute value of the difference between the mechanical and electronic readings of the electromechanical synchronous counter is greater than the first preset difference, and the sudden change amplitude of the downstream pressure sensor is greater than the second preset difference, then both verifications point to a true reversal, and the second abnormal situation information is determined, generating a negative water usage alarm record; and calling the DMA (regional metering area) leakage model to mark the negative water meter as a "suspected upstream node of pipe burst"; If the verification fails, the anomaly will be marked as "data drift" and zero-point self-calibration will be triggered; The downstream pressure sensor is a MEMS piezoresistive sensor with a range of 0-1.6MPa, an accuracy of ±0.25%FS, a sampling frequency of ≥1Hz, and communicates with the remote water meter via the Modbus-RTU protocol.

[0031] By implementing a two-way verification mechanism, negative water usage anomalies (secondary anomaly information) can be verified a second time, avoiding false alarms about negative water usage.

[0032] In an optional embodiment of the present invention, step 13 further includes: Step 133: Perform threshold monitoring on the processed traffic data according to the threshold alarm rules. If the difference between the processed traffic data of the current day and the processed traffic data of the previous day with traffic is greater than the sum of the daily threshold alarm values ​​of all days within the range of the current day and the previous day with traffic, a third abnormal situation information is obtained. The third abnormal situation information includes no threshold alarm processing record when a threshold alarm occurs, a first preset number of threshold alarm processing records when a threshold alarm occurs, and a second preset number of threshold alarm processing records when a threshold alarm occurs. The daily threshold alarm value is at least one of the daily normal threshold, the daily dynamic threshold, and the daily weighted threshold.

[0033] In this embodiment, if the difference between the processed traffic data of the current day and the processed traffic data of the previous day with traffic is greater than the sum of the daily threshold alarm values ​​of all days within the range of the current day and the previous day with traffic, it indicates that there is an abnormal situation of threshold alarm; wherein, the sum of the daily threshold alarm values ​​is at least one of the sum of the daily normal threshold, the sum of the daily dynamic threshold, and the sum of the daily weighted threshold. If the abnormal situation of the threshold alarm is the first occurrence, and there is no abnormal situation information in the threshold alarm processing record when the threshold alarm is obtained, it is regarded as the third abnormal situation information. If an abnormal situation occurs when a threshold alarm occurs, and there are already a first preset number of abnormal situations for threshold alarms, obtain the abnormal situation information of the first preset number of threshold alarm processing records at the time of the threshold alarm, and use it as the third abnormal situation information. If an abnormal situation occurs when a threshold alarm occurs, and there are already a second preset number of abnormal situations for threshold alarms, obtain the abnormal situation information of the second preset number of threshold alarm processing records at the time of the threshold alarm, and use it as the third abnormal situation information. The first and second preset quantities are both integers and gradually increase; for example, the first preset quantity can be 1 and the second preset quantity can be 2; they can be adjusted according to the actual situation in specific applications.

[0034] The daily threshold alarm value is at least one of the daily normal threshold, the daily dynamic threshold, and the daily weighted threshold; The daily normal threshold can include hourly normal thresholds and daily normal thresholds. The specific confirmation process is as follows: Hourly Normal Threshold: Set the hourly normal threshold range to a preset multiple of the average hourly water consumption; for example, if the average hourly water consumption is 2... / hour, the normal threshold can be set to an upper limit of 3 / hour (1.5 times the average value) and a lower limit of 1 / hour (0.5 times the average value); the specific preset multiple can be adjusted according to the specific situation; Daily normal threshold: The daily normal threshold range is set to the first preset percentile to the second preset percentile of daily water consumption; wherein, the first preset percentile can be the 10th percentile; the second preset percentile can be the 90th percentile; for example, if the 10th percentile is 5... If the 90th percentile is 15 / day, then the daily normal threshold range is 5 to 15. / sky; In practical applications, the daily normal threshold is updated periodically. Specifically, through the above process, the first and second preset percentiles of the average hourly and daily water consumption are recalculated periodically (e.g., monthly or quarterly), thereby adjusting the hourly and daily normal thresholds accordingly to adapt to long-term changes in water consumption patterns. By periodically adjusting the hourly and daily normal thresholds based on the user's actual water consumption, dynamic adjustment of the hourly and daily normal thresholds can be achieved, enabling the daily normal threshold to be updated periodically. Ultimately, this achieves dynamic adjustment of the daily threshold alarm value, avoiding problems such as false alarms, missed alarms, and alarm lag.

[0035] The process for confirming the daily dynamic threshold is as follows: Acquire historical flow data and related historical feature information from remote water meters; The historical traffic data is preprocessed to obtain historical processed traffic data; The historical processing traffic data is input into the first processing module of the processing layer of the dynamic threshold confirmation model to obtain the first processing result. The historical processing traffic data and the historical related feature information are input into the second module of the processing layer of the dynamic threshold confirmation model to obtain the second processing result; The first processing result and the second processing result are input into the output layer of the dynamic threshold confirmation model to obtain the dynamic threshold prediction value, which is used as the dynamic threshold for the day. The first processing module of the processing layer of the dynamic threshold confirmation model is trained based on the first preset network model; the second processing module of the processing layer of the dynamic threshold confirmation model is trained based on the second preset network model; and the output layer of the dynamic threshold confirmation model is trained based on the third preset network model. Historical flow data can be the daily water consumption of remote water meters over the past 90 days; historical related characteristics can be: time characteristics, such as day of the week, whether it is a holiday, lunar phase, and season; statistical characteristics, such as rolling 7 / 14 / 30-day mean, variance, skewness, and kurtosis; and anomaly markers, such as historical abnormal events (emergency repairs, water outages) during the same period. Training process of dynamic threshold confirmation model: Acquire training flow data and training-related feature information from remote water meters; The training traffic data is preprocessed to obtain training processed traffic data; according to The training data is decomposed to obtain trend components, seasonal components, and residual components; wherein... This represents the training processing traffic data at time t; Indicates trend components; Indicates seasonal quantity; Represents the residual components; The trend component and the seasonal component are input into a first preset network model. The trend component is processed to obtain a horizontal component and a trend component, where the horizontal component represents the base value of the trend component (such as the current average value); the trend component represents the rate of change of the trend component (such as the amount of increase / decrease). Update the horizontal components; where, Represents the horizontal component at time t; Indicates the horizontal smoothing coefficient. ; Represents the horizontal component at time t-1; represents the trend component at time t-1; m represents the length of the seasonal cycle; Indicates the seasonal component at time tm; This represents the training processing traffic data at time t. (Through...) Update the trend components; among them, Represents the trend component at time t; Indicates the trend smoothing coefficient. ; Represents the horizontal component at time t; Represents the horizontal component at time t-1; This represents the trend component at time t-1. (Through...) Update the seasonal quantities; among them, Indicates the seasonal component at time t; Indicates the seasonal smoothing coefficient. ; This represents the training processing traffic data at time t; Represents the horizontal component at time t-1; This represents the trend component at time t-1; This represents the seasonal component at time tm. (Through...) The first predicted value is obtained; where, The first predicted value is represented by h; the prediction step size is represented by h. Represents the horizontal component at time t; Represents the trend component at time t; Represent the seasonal component at time t+hm. α is searched using a grid. ,γ ∈ [0.1,0.9], prediction step size 0.1, until Minimize and save the current optimal model; where, express and Symmetric mean absolute percentage error; This represents the training processing traffic data at time t; This represents the first predicted value.

[0036] The training processing traffic data and the training-related feature information are input into a second preset network model. The training processing traffic data and the training-related feature information are classified into dynamic variables, static variables, and temporal features. The static variables are processed by the static variable encoder of the second preset network model to generate a static context vector. The temporal features are processed by the temporal variable encoder of the second preset network model to generate a temporal feature vector. The dynamic vector, static context vector, and temporal feature vector are input into the variable selection network of the second preset network model for processing to filter key features. The key features are input into the encoder (which can be an LSTM encoder) and decoder (which can be a Transformer decoder) of the second preset network model. The encoder processes the key features to obtain short-term dependency features, and the decoder processes the key features to obtain long-term dependency features. The short-term dependency features and long-term dependency features are input into the fusion gate of the second preset network model for processing to obtain fused features. The fused features are input into the output layer of the second preset network model to obtain a second predicted value. When the validation set reaches P90... If the width calibration error does not decrease (i.e., does not improve) for 5 consecutive epochs (one cycle of traversing a complete training set during training), stop training and save the current best model; The first and second predicted values ​​are input into the third preset network model, and then... The training prediction values ​​are obtained; where, Indicates the training prediction value; Indicates the first predicted value; Indicates the weight of the first predicted value; Indicates the second predicted value; The weights represent the second predicted value. A grid search optimization algorithm is used to perform a weighted average of the first and second predicted values. By traversing different weight combinations, the weight combination that minimizes the root mean square error is selected as the final model weights. Training is then completed, and the current optimal model is saved.

[0037] In practical applications, the daily dynamic threshold is updated periodically. The training process of the model is confirmed by the above dynamic threshold, and the training traffic data is updated periodically to retrain the model, thereby updating the daily dynamic threshold. Specifically, new data is collected every 24 hours, and lightweight training is performed every 7 days. If the symmetrical average absolute percentage error between the actual value and the predicted value is greater than 15% for 3 consecutive days, full retraining is triggered. By periodically updating the daily dynamic threshold, the daily threshold alarm value is finally dynamically adjusted to avoid problems such as false alarms, missed alarms, and alarm lag.

[0038] The process for confirming the daily weight threshold is as follows: The system acquires historical flow data and historical alarm tag information from the remote water meter. The historical flow data may include: daily flow data for the past 24 months, non-zero water consumption flow data for the past 12 days, and daily peak flow data. The historical alarm tag information may include: holiday or emergency repair event tags marked by water industry experts. The historical traffic data is preprocessed to obtain historical processed traffic data; The historical processing traffic data is input into the first processing module of the weighted threshold confirmation model to obtain the first confirmation result; The historical alarm tag information is input into the second processing module of the processing layer of the weighted threshold confirmation model to obtain the second confirmation result; The first confirmation result and the second confirmation result are sent to the output layer of the weight threshold confirmation model to obtain the predicted value of the weight threshold, which is used as the weight threshold for the day. The first processing module of the weight threshold confirmation model is trained based on the fourth preset network model; the second processing module of the weight threshold confirmation model is trained based on the fifth preset network model; and the output layer of the weight threshold confirmation model is trained based on the sixth preset network model.

[0039] Weight thresholds confirm the model's training process: Training process of dynamic threshold confirmation model: Obtain training flow data and training alarm tag information from the remote water meter; The training traffic data is preprocessed to obtain training processed traffic data; The training traffic data is input into the fourth preset network model for judgment matrix processing to obtain the judgment matrix; The judgment matrix is ​​as follows:

[0040] Wherein, F1 represents the training processing traffic data for the past 24 months; F2 represents the training processing traffic data for the past 12 days; and F3 represents the daily peak training processing traffic data. The judgment matrix A is:

[0041] pass and This yields the largest eigenvalue and its corresponding eigenvector; where A represents the judgment matrix. Represents the eigenvector; Represents the largest eigenvalue; calculated as follows That is, the weights of F1 The weight of F2 is 0.558. The weight of F3 is 0.320. It is 0.122; Perform consistency verification, pass RI=0.58, thus obtaining the consistency ratio; where CR represents the consistency ratio; CI represents the consistency index; and RI represents the average random consistency index. Represents the largest eigenvalue; because If CR = 0.046, then the consistency check passes; the weights calculated above are solidified using the version number and hash value, and remain valid in the long term.

[0042] The training alarm label information is input into the fifth preset network model for weighted processing to obtain the loss curve; wherein, the training alarm label information includes the false positive rate and the false negative rate; specifically, through... The loss curve is obtained; where FAR represents the false alarm rate. Indicates the false negative rate; Represents the weighted loss function; The loss curve is processed by Gaussian process and expectation improvement to obtain the initial weight parameter α_star; with α∈[1.05,1.50], iterate 30 times; convergence is reached when the difference between two consecutive α values ​​is <0.02 or after 30 iterations. Update via The weight parameters are updated to obtain the target weight parameters; where, This indicates the target weight parameter being updated. This represents the target weight parameter from the last update.

[0043] pass ,in, F1 represents the weight threshold for the current day; F2 represents the training processing traffic data for the past 24 months; F3 represents the training processing traffic data for the past 12 days; and F3 represents the peak training processing traffic data for the day. Indicates the weights of F1; Indicates the weight of F2; Indicates the weight of F3; This indicates the target weight parameter being updated.

[0044] In practical applications, the daily weight threshold is updated periodically. The target weight parameter is updated periodically through the above process, thereby updating the daily weight threshold and ultimately realizing the dynamic adjustment of the daily threshold alarm value, avoiding problems such as false alarms, missed alarms, and alarm lag.

[0045] In this embodiment, at least one of the daily normal threshold, daily dynamic threshold, and daily weighted threshold is used as the daily threshold alarm value, and the daily normal threshold, daily dynamic threshold, and daily weighted threshold are updated periodically to achieve dynamic adjustment of the daily threshold alarm value, thereby avoiding problems such as false alarms, missed alarms, and alarm lag.

[0046] In an optional embodiment of the present invention, step 14 includes: Step 141: Generate a data upload alarm record based on the first abnormal situation information; Among them, the first abnormal situation information is that the remote water meter flow data was not uploaded on that day, generating a first-level data upload alarm record; The first abnormal situation information is that the remote water meter flow data has not been uploaded for a first preset number of consecutive days, and a second-level data upload alarm record is generated.

[0047] In this embodiment, when the remote water meter flow data for the day is not uploaded, an alarm history record is generated as a first-level data upload alarm record; for example, an alarm history record is generated with the alarm type "data not uploaded", the alarm handling level is "warning", and the alarm reason is "data not uploaded for the day". If the remote water meter flow data is not uploaded for the first preset number of consecutive days, an alarm processing record will be generated as a second-level data upload alarm record. For example, if the remote water meter flow data is not uploaded for 5 consecutive days, an alarm processing record will be generated with the alarm type "data not uploaded", the alarm processing level "alarm", and the alarm reason "data not uploaded for 5 consecutive days". Through the above process, different levels of data upload alarm records are generated based on different data upload anomaly information. This facilitates timely handling of anomaly information such as the failure to upload remote water meter flow data, ensuring the normal operation of remote water meters, reducing erroneous charges, and lowering the user complaint rate.

[0048] In practical applications, data upload alarm records can be cleared. Specifically, if the remote water meter flow data is uploaded on the same day, the alarm clearing process for data not uploaded will be performed to clear all previous data upload alarm records of all levels. Specifically, the alarm clearing process for data not uploaded can be automatic or manual.

[0049] In an optional embodiment of the present invention, step 14 further includes: Step 142: Generate a negative water usage alarm record based on the second abnormal situation information; The second abnormal situation information is that there is a negative water usage abnormality on that day, and a first-level negative water usage alarm record is generated. The second abnormal situation information is that the number of negative water usage anomalies within the second preset number of days is greater than the preset number, and a second-level negative water usage alarm record is generated; The second abnormal situation information is a continuous negative water usage abnormality for a third preset number of days, generating a third-level negative water usage alarm record.

[0050] In this embodiment, when there is an abnormal negative water usage on a given day, an alarm history record is generated as a first-level negative water usage alarm record; for example, an alarm history record is generated with the alarm type being "negative water usage", the alarm handling level being "early warning", and the alarm reason being "negative water usage on that day". If the number of negative water usage anomalies exceeds the preset number within the second preset number of days, an alarm processing record will be generated as a second-level negative water usage alarm record. For example, if more than 3 negative water usage incidents occur within 10 days, an alarm processing record will be generated with the alarm type being "negative water usage", the alarm processing level being "alarm", and the alarm reason being "3 negative water usage incidents within 10 days". When there are consecutive negative water usage anomalies for the third preset number of days, the alarm handling record will be used as the third level of negative water usage alarm record; for example, if there are 5 consecutive days of negative water usage anomalies, an alarm handling record will be generated, with the alarm type being "negative water usage", the alarm handling level being "alarm", and the alarm reason being "continuous negative water usage". Through the above process, different levels of negative water usage alarm records are generated based on different situations of negative water usage anomalies. This facilitates timely handling of abnormal situations related to negative water usage by relevant personnel, ensuring the normal operation of remote water meters, reducing erroneous charges, and lowering the user complaint rate.

[0051] In practical applications, negative water usage alarm records can be cleared. Specifically, if the number of negative water usage anomalies within the second preset number of days up to the current day is less than the preset number, multiple negative water usage alarm records will be cleared. If there are no negative water usage anomalies on the current day, continuous negative water usage alarm records will be cleared. Specifically, the clearing of multiple negative water usage alarm records and the clearing of continuous negative water usage alarm records can be done automatically or manually.

[0052] In an optional embodiment of the present invention, step 14 further includes: Step 143: Generate a threshold alarm record based on the third abnormal situation information; The third abnormal situation information is that there is no threshold alarm processing record when a threshold alarm is triggered, and a first-level threshold alarm record is generated. The third abnormal situation information is that when a threshold alarm occurs, there is a first preset number of threshold alarm processing records, and a second level of threshold alarm records are generated. The third abnormal situation information is that when a threshold alarm occurs, there is a second preset number of threshold alarm processing records, generating a third level of threshold alarm records.

[0053] In this embodiment, when there is no threshold alarm processing record when a threshold alarm occurs, an alarm processing record is generated as the first-level threshold alarm record; for example, an alarm history record is generated, the alarm type is "threshold alarm", the alarm level is "continuous observation", and the alarm reason is "the actual water consumption on a specific date triggers a specific threshold alarm rule"; wherein, the specific threshold alarm rule can be at least one of normal threshold rule, dynamic threshold rule and weighted threshold rule; When a threshold alarm occurs and there are a first preset number of threshold alarm processing records, an alarm processing record is generated as a second-level threshold alarm record. For example, if there is one threshold alarm processing record, that is, if there is an unprocessed alarm processing record with the alarm type "threshold alarm" and the alarm level "continuous observation", then an alarm processing record is generated with the alarm type "threshold alarm", the alarm level "early warning", and the alarm reason "details of two specific dates' actual water consumption triggering specific threshold rules". When a threshold alarm occurs and there are a second preset number of threshold alarm processing records, an alarm processing record is generated as a third-level threshold alarm record. For example, if there are 2 threshold alarm processing records, that is, if there are unprocessed alarm processing records with alarm type "threshold alarm" and alarm level "warning", then an alarm processing record is generated with alarm type "threshold alarm", alarm level "alarm", and alarm reason "details of 3 specific dates' actual water consumption triggering specific threshold rules". Through the above process, different levels of threshold alarm records are generated based on different threshold alarm anomalies. This facilitates timely handling of abnormal threshold alarm information by relevant personnel, enabling timely detection and resolution of water leaks and reducing water loss.

[0054] In practical applications, threshold alarm records can be cleared; specifically, if the water consumption on the day no longer triggers the threshold alarm rule, the threshold alarm record clearing process will be performed to clear all current threshold alarm records; specifically, the threshold alarm record clearing process can be automatic or manual.

[0055] The alarm records of different levels can be pushed to the operation and maintenance management platform and user terminal devices. The alarm records of different levels can be sent to operation and maintenance personnel and managers through push methods such as SMS, email, and APP.

[0056] In an optional embodiment of the present invention, the remote water meter leakage early warning processing method further includes: Step 144: After the alarm records of different levels are pushed, if the user fills in the work order repair time, a preset observation period is started. If the same type of abnormality is not triggered again during the observation period, the alarm is automatically cleared. If it is triggered again, the alarm clearing is marked as "invalid" and the weight threshold of the remote water meter is increased by 10% as negative feedback.

[0057] In this embodiment, after different levels of alarm records are pushed, if the user fills in the work order repair time, a preset observation period is started. If the same type of abnormality is not triggered again during the observation period, the alarm is automatically cleared. If it is triggered again, the clearing of the alarm is marked as "invalid", and the weight threshold of the remote water meter is increased by 10% as negative feedback. Specifically, the weight threshold is calculated as follows: new weight threshold = original weight threshold × (1 + 0.1 × number of recurrences), where the number of recurrences is the cumulative number of times the same abnormality is marked within 30 days. The preset observation period can be 24 hours.

[0058] In an optional embodiment of the present invention, the remote water meter leakage early warning processing method further includes: Step 151: Obtain the number of days of data upload interruption, the frequency and consecutive days of negative water usage rules, and the deviation value of threshold alarm rules. Step 152: Based on the number of days of data transmission interruption in the data upload rule, the frequency and consecutive days of negative water usage rule, and the deviation value of the threshold alarm rule, obtain the cross-rule composite alarm and push it.

[0059] In this embodiment, the confidence level is obtained by using Confidence = α'·D + β'·(F+C) + γ'·ΔQ; if Confidence > θ, a cross-rule composite alarm is triggered; where Confidence represents the confidence level; θ represents the dynamic threshold; D represents the number of days of data transmission interruption; F represents the frequency of negative water usage; C represents the number of consecutive days of negative water usage; ΔQ represents the deviation value; α', β', and γ' represent adaptive weights, which can be trained based on historical false alarm data; and the cross-rule composite alarm can be "suspected equipment failure".

[0060] The cross-rule composite alarm obtained through the above process enables maintenance personnel to promptly detect "suspected equipment failures" such as the depletion of batteries in remote water meters, ensuring normal operation.

[0061] Let's take a user's actual data as an example: The user has not uploaded any data for 3 consecutive days, and the system has recorded this. The number of days the transmission was interrupted was D=3; The frequency of negative water usage is F=2, and the number of consecutive days with negative water usage is C=2; The deviation value ΔQ output by the threshold alarm rule is 0.3.

[0062] The weights obtained by training with historical false positive data are α=0.4, β=0.3, and γ=0.3. Confidence = 0.4 × 3 + 0.3 × (2 + 2) + 0.3 × 0.3 = 1.2 + 1.2 + 0.09 = 2.49; Since 2.49 > the trigger threshold θ (θ = 2.0), the system generates a "suspected equipment failure" composite alarm and pushes it to maintenance personnel via SMS and work order. On-site verification confirmed that the alarm was due to the water meter battery being depleted, verifying the effectiveness of the collaborative decision-making mechanism.

[0063] In an optional embodiment of the present invention, the remote water meter leakage early warning processing method further includes: Step 161: Obtain the markings of the alarm results by the operations and maintenance personnel; Step 162: Update the relevant parameters of the dynamic threshold confirmation model according to the marking of the alarm results by the operation and maintenance personnel.

[0064] In this embodiment, after the alarm records of different levels are pushed to the operation and maintenance management platform, the operation and maintenance personnel will mark the alarm results. The specific marking content may be false alarm or missed alarm. If a false alarm is detected, the parameter γ of the dynamic threshold confirmation model is lowered by γ_new = max(γ_min, γ×0.8); where γ represents the seasonal smoothing coefficient in the dynamic threshold confirmation model; γ_new represents the updated seasonal smoothing coefficient; and γ_min represents the configurable lower limit of the seasonal smoothing coefficient, γ_min=0.1. If a false negative is detected, the parameter β of the dynamic threshold confirmation model is updated by β_new = β_old × 1.2; where β represents the trend smoothing coefficient in the dynamic threshold confirmation model; β_new represents the updated trend smoothing coefficient; and β_old represents the original trend smoothing coefficient. Updating the relevant parameters of the dynamic threshold confirmation model based on the marking of alarm results by the maintenance personnel can ensure the monitoring accuracy of the dynamic threshold rules.

[0065] In an optional embodiment of the present invention, the remote water meter leakage early warning processing method further includes: Step 171: Obtain the coordinate information of the remote water meters within the preset area; Step 172: Generate a spatial cluster of remote water meters within the preset area based on the coordinate information of the remote water meters within the preset area. Step 173: If ≥30% of the remote water meters in the spatial cluster of the preset area trigger the threshold alarm at the same time, an "area leakage event" alarm is generated; if only a single remote water meter triggers the threshold alarm and the data of the surrounding remote water meters is normal, it is marked as "independent device failure".

[0066] In this embodiment, by monitoring remote water meters within a preset range, if ≥30% of the remote water meters in the spatial cluster of remote water meters within the preset area simultaneously trigger threshold alarms, an "area leakage event" alarm is generated, which can promptly detect area leakage events.

[0067] In an optional embodiment of the present invention, the remote water meter leakage early warning processing method further includes: Step 181: Obtain the daily water consumption peak-to-valley ratio, water consumption period entropy value, and holiday water consumption fluctuation rate from historical flow data; Step 182: Generate user profile tags based on the daily peak-to-valley ratio of water usage, entropy value of water usage period, and water usage fluctuation rate during holidays in the historical flow data; wherein the user profile tags can be at least one of the following: morning peak type, nighttime continuous type, and holiday fluctuation type; Step 183: Dynamically adjust the daily threshold based on the user profile tags; Step 184: Obtain fault information based on the second abnormal situation information and the user profile tag.

[0068] In this embodiment, user profile tags are generated based on the daily peak-to-valley ratio of water usage, the entropy value of water usage periods, and the fluctuation rate of water usage during holidays from historical flow data. These user profile tags can be at least one of the following: morning peak type, nighttime continuous type, and holiday fluctuating type. Based on the user profile tags, the daily threshold is dynamically adjusted. For example, if the user profile tag is nighttime continuous type, the corresponding nighttime threshold is set to be 20% higher. Based on the second abnormal situation information and the user profile tags, fault information is obtained. Specifically, if the user type corresponding to the user profile tag experiences negative water usage during off-peak hours (second abnormal situation information), the fault information can be "high probability equipment failure." For example, if a morning peak user experiences negative water usage during off-peak hours, it is marked as "high probability equipment failure." Through the above process, corresponding threshold adjustments can be made according to different user types, thereby improving the monitoring accuracy for different user types.

[0069] In an optional embodiment of the present invention, the remote water meter leakage early warning processing method further includes: Step 191: Obtain meteorological early warning information; wherein, the meteorological early warning information may be abnormal meteorological warnings that affect water use, such as rainstorm warnings and freezing warnings. Step 192: Based on the meteorological early warning information, run the disaster relief mode.

[0070] In this embodiment, meteorological warning information is obtained by accessing the National Meteorological Administration API (weather data interface) via HTTPS (HTTP channel for security purposes). The disaster mitigation mode may include: monitoring only negative water usage and equipment offline; multiplying the dynamic threshold by a factor of 1.5; delaying alarms until the abnormality lasts for 3 hours, etc. By setting the disaster mitigation mode, false alarms caused by instantaneous fluctuations can be avoided, and the monitoring accuracy of remote water meters can be guaranteed under abnormal meteorological warnings that affect water use, such as rainstorm warnings and freezing warnings.

[0071] The remote water meter leakage early warning processing method provided in the above embodiments of the present invention is implemented based on a remote water meter leakage early warning processing system, which includes: The data acquisition module is used to collect real-time flow data from the remote water meter; specifically, it collects real-time flow data from the remote water meter periodically and uploads it to the server through IoT devices and communication networks. The data processing module is used to preprocess the collected real-time traffic data, including data cleaning and format conversion. Specifically, it uses data processing algorithms to remove noisy data, convert data formats, and ensure the accuracy and consistency of the data. The monitoring module is used to monitor water meter data in real time according to the set rules and determine whether there are any abnormalities. Specifically, by setting various preset monitoring rules (such as data upload rules, negative water usage rules, threshold alarm rules, etc.), the module analyzes the data in real time and generates abnormal situation information. The alarm processing module is used to generate alarm records based on abnormal situation information and to classify and process alarms according to their severity, thereby generating alarm records of different levels. Specifically, it generates alarm records of different levels according to alarm type and severity, and provides automatic alarm cancellation and manual alarm cancellation functions. The notification module is used to promptly notify relevant personnel of alarm records of different levels so that abnormal situations can be handled in a timely manner; specifically, alarm information is sent to operation and maintenance personnel and administrators through push methods such as SMS, email, and APP. The user interface module provides a user interaction interface, making it convenient for users to view alarm records, configure thresholds, etc. The anomaly confidence fusion module is used to receive the number of days of data transmission interruption D, the frequency F and the number of consecutive days of negative water use rules, and the deviation value ΔQ of threshold alarm rules. The confidence level is calculated using the formula Confidence=α'·D+β'·(F+C)+γ'·ΔQ; when Confidence>θ, a cross-rule composite alarm is triggered, where θ is a dynamic threshold. The alarm closed-loop feedback module is used to receive the marking of alarm results by operation and maintenance personnel {false alarm, leak confirmation}; based on the marking results, the threshold model parameters are automatically adjusted: if marked as a false alarm, γ is reduced by γ_new = max(γ_min,γ×0.8), where γ_min=0.1 is a configurable lower limit; if marked as a missed alarm, β←β×1.2; the updated parameters are written to the model configuration file; The spatiotemporal correlation analysis engine is used to construct a water meter topology network based on GIS coordinates and identify physically adjacent water meter clusters. If ≥30% of the water meters in the cluster trigger threshold alarms at the same time, an "area leakage event" alarm is generated. If only a single water meter alarms and the data of the surrounding water meters is normal, it is marked as "independent device failure". The water usage behavior profiling module is used to extract the daily water usage peak-to-valley ratio, water usage period entropy, and holiday water usage volatility based on historical data, and generate user profile tags {morning peak type, nighttime continuous type, and holiday volatility type}; the threshold is dynamically adjusted according to the profile tags: the nighttime threshold for nighttime continuous type users is increased by 20%; if a morning peak type user has negative water usage during off-peak hours, it is marked as "high probability equipment failure"; The disaster mitigation mode adaptive module is used to connect to the meteorological early warning API and automatically trigger the disaster mitigation mode when a rainstorm or freezing warning is issued. In the disaster mitigation mode: only negative water usage and offline equipment are monitored; the dynamic threshold is multiplied by a coefficient of 1.5; and the alarm is delayed until 3 hours after the abnormality continues.

[0072] The remote water meter leakage early warning processing method provided in the above embodiments of the present invention is implemented based on a lightweight edge computing architecture; wherein, the lightweight edge computing architecture is layered as follows: Edge node: ARM Cortex-M33 MCU + 256KB RAM, running a lightweight rule engine, responsible for data uploading, negative water usage monitoring and 7-day rolling caching; Cloud-based: Performs training and global threshold updates for complex models such as TFT and Bayesian methods. Offline mode: Edge nodes continue to run the rule engine based on local cache during network outages; Once the network is restored, the offline results are uploaded in one go via MQTT, and the data is merged and the history is completed in the cloud.

[0073] The lightweight edge computing architecture described above enables resource control: edge node duty cycle is less than 5%, and battery life is extended by 30%. Compared to a pure cloud architecture, system availability is improved to 99.5% in extreme weather or communication outage scenarios.

[0074] The process for handling remote water meter leakage early warning is as follows: Obtain real-time flow data from remote water meters; The real-time traffic data is preprocessed to obtain processed traffic data; specifically, outlier removal and missing value imputation are performed on the real-time traffic data to obtain first processed data; the first processed data is then converted to obtain processed traffic data. The processed traffic data is monitored in real time according to preset monitoring rules to obtain abnormal situation information; the abnormal situation information is then classified and processed to obtain alarm records of different levels; specifically: If real-time flow data from the remote water meter is not obtained on a given day, an alarm history record will be generated indicating that the remote water meter flow data for that day has not been uploaded. The alarm type will be "data not uploaded", the alarm handling level will be "early warning", and the alarm reason will be "data not uploaded on that day". If real-time flow data from the remote water meter is not obtained for several consecutive days, and the number of days exceeds 5, an abnormal situation information of "data not uploaded for 5 consecutive days" is obtained, and an alarm processing record is generated. The alarm type is "data not uploaded", the alarm processing level is "alarm", and the alarm reason is "data not uploaded for 5 consecutive days". If there is negative water usage abnormality on the day, obtain the abnormal situation information of negative water usage abnormality on the day, generate alarm history record, alarm type is "negative water usage", alarm handling level is "early warning", alarm reason is "negative water usage on the day"; If multiple negative water usage anomalies occur within 10 days, and the number of anomalies is greater than 3, an anomaly information of more than 3 negative water usage anomalies within 10 days will be obtained, an alarm handling record will be generated, the alarm type will be "negative water usage", the alarm handling level will be "alarm", and the alarm reason will be "3 negative water usage incidents within 10 days". If negative water usage occurs for several consecutive days and the number of consecutive days is greater than 5 days, obtain the abnormal situation information of 5 consecutive days of negative water usage, generate an alarm handling record, the alarm type is "negative water usage", the alarm handling level is "alarm", and the alarm reason is "continuous negative water usage". If the threshold alarm is the first occurrence of an anomaly, and there is no threshold alarm handling record when the threshold alarm is obtained, an alarm history record is generated, the alarm type is "threshold alarm", the alarm level is "continuous observation", and the alarm reason is "the actual water consumption on a specific date triggers the specific threshold alarm rule". If an abnormal situation occurs when a threshold alarm occurs, there is already one abnormal situation for a threshold alarm. If the abnormal situation information of one threshold alarm processing record is obtained when the threshold alarm occurs, an alarm processing record is generated. The alarm type is "threshold alarm", the alarm level is "early warning", and the alarm reason is "details of two specific dates when actual water consumption triggers specific threshold rules". If an abnormal situation occurs when a threshold alarm occurs, and there are already 2 abnormal situations for threshold alarms, obtain the abnormal situation information that there are 2 threshold alarm processing records when the threshold alarm occurs, generate an alarm processing record, the alarm type is "threshold alarm", the alarm level is "alarm", and the alarm reason is "details of 3 specific dates' actual water consumption triggering specific threshold rules". The alarm records of different levels are pushed to the operation and maintenance management platform and user terminal devices. Specifically, the alarm records of different levels can be sent to operation and maintenance personnel and management personnel through push methods such as SMS, email, and APP. Alarm clearing, specifically: If the remote water meter flow data is uploaded on the same day, the alarm will be cleared if the data is not uploaded, in order to eliminate all previous data upload alarm records. If the number of abnormal water usage incidents within 10 days up to that day is less than 3, multiple negative water usage alarm records will be processed to clear the alarms. If there is no abnormal negative water usage on that day, record the continuous negative water usage alarm and clear the alarm. If the water consumption on that day no longer triggers the threshold alarm rule, the threshold alarm record will be cleared to eliminate all current threshold alarm records.

[0075] The above process supports various alarm methods, including data upload failure, negative water usage, continuous negative water usage, and threshold alarms. Thresholds can be configured in multiple ways, such as dynamic thresholds, normal thresholds, and weighted thresholds. Monitoring across multiple dimensions, including normal, dynamic, and weighted thresholds, allows for timely detection of various anomalies in remote water meters, especially leaks, reducing leakage and minimizing economic losses. Alarm records are generated based on severity, distinguishing alarm levels as continuous observation, early warning, and alarm. Alarm handling supports both automatic and manual alarm cancellation and can be sent to relevant personnel according to configuration. This enables timely monitoring of remote water meter anomalies and facilitates communication with residents through flexible notification methods, allowing for timely repair of on-site remote water meters, minimizing inconvenience caused by anomalies, and improving user service quality. This process also enables timely detection of abnormal water usage in remote water meters, reducing water waste, promptly identifying and handling on-site equipment malfunctions, reducing incorrect billing, lowering user complaint rates, and sending anomaly information to relevant maintenance personnel immediately, shortening the problem-solving cycle and improving efficiency.

[0076] The remote water meter leakage early warning processing method provided by the above embodiments of the present invention can support multiple rule settings such as normal threshold, dynamic threshold, and weighted threshold, and can cover different scenarios, while existing technologies can usually only set a single fixed threshold. Among them, the weighted threshold is data-driven, using AHP to solidify weights (compared to the average daily water consumption data of the same month, the average of the last 12 non-zero water consumption data, and the daily peak water consumption) + Bayesian monthly fine-tuning (setting initial values ​​according to water consumption patterns and scenario requirements, adjusting based on historical data statistical analysis, and dynamically optimizing); the dynamic threshold is fused through STL-ES + TFT dual-model balancing speed and accuracy. It can monitor multiple anomaly types, including data not uploaded, negative water consumption, continuous negative water consumption, and threshold alarms, while existing technologies can usually only monitor simple threshold over-limit alarms. It can continuously observe, provide early warning, and issue alarms at three levels, allowing users to take corresponding processing measures according to the processing level, while existing technologies often cannot classify alarms. Alarms support both automatic and manual alarm cancellation, while existing technologies usually only have one alarm cancellation method. It can be configured to send SMS notifications to relevant personnel, so that alarm information can be sent to relevant personnel as soon as possible, while existing technologies are not flexible enough in terms of alarm notification.

[0077] like Figure 2 As shown, an embodiment of the present invention also provides a remote water meter leakage early warning and processing device 20, comprising: The acquisition module 21 is used to acquire real-time flow data from the remote water meter; The processing module 22 is used to preprocess the real-time traffic data to obtain processed traffic data; to monitor the processed traffic data in real time according to preset monitoring rules to obtain abnormal situation information; wherein, the preset monitoring rules include at least one of: data upload rules, negative water usage rules, and threshold alarm rules; wherein, the threshold alarm rules include at least one of: normal threshold rules, dynamic threshold rules, and weighted threshold rules; to perform hierarchical processing on the abnormal situation information to obtain alarm records of different levels; and to push the alarm records of different levels to the operation and maintenance management platform and user terminal devices.

[0078] Optionally, the processed traffic data can be monitored in real time according to preset monitoring rules to obtain abnormal situation information, including: According to the data upload rules, the processed flow data is monitored for data upload. If the real-time flow data of the remote water meter is not obtained, the first abnormal situation information is obtained. The first abnormal situation information includes the failure to upload the flow data of the remote water meter on the current day and the failure to upload the flow data of the remote water meter for a first preset number of consecutive days.

[0079] Optionally, the process further includes real-time monitoring of the processed traffic data according to preset monitoring rules to obtain abnormal situation information: According to the negative water usage rule, the processed flow data is monitored for negative water usage. If the processed flow data of the remote water meter on the current day minus the processed flow data of the previous day with flow is negative, a second abnormal situation information is obtained. The second abnormal situation information includes negative water usage abnormality on the current day, negative water usage abnormality occurring more than a preset number of times within a second preset number of days, and continuous negative water usage abnormality for a third preset number of days.

[0080] Optionally, the process further includes real-time monitoring of the processed traffic data according to preset monitoring rules to obtain abnormal situation information: According to the threshold alarm rules, the processed traffic data is monitored for thresholds. If the difference between the processed traffic data of the current day and the processed traffic data of the previous day with traffic is greater than the sum of the daily threshold alarm values ​​of all days within the range of the current day and the previous day with traffic, a third abnormal situation information is obtained. The third abnormal situation information includes no threshold alarm processing record when a threshold alarm occurs, a first preset number of threshold alarm processing records when a threshold alarm occurs, and a second preset number of threshold alarm processing records when a threshold alarm occurs. The daily threshold alarm value is at least one of the daily normal threshold, the daily dynamic threshold, and the daily weighted threshold.

[0081] Optionally, the abnormal situation information is processed in a hierarchical manner to obtain alarm records of different levels, including: Based on the first abnormal situation information, a data upload alarm record is generated; Among them, the first abnormal situation information is that the remote water meter flow data was not uploaded on that day, generating a first-level data upload alarm record; The first abnormal situation information is that the remote water meter flow data has not been uploaded for a first preset number of consecutive days, and a second-level data upload alarm record is generated.

[0082] Optionally, the process of classifying the abnormal situation information to obtain alarm records of different levels also includes: Based on the second abnormal situation information, a negative water usage alarm record is generated; The second abnormal situation information is that there is a negative water usage abnormality on that day, and a first-level negative water usage alarm record is generated. The second abnormal situation information is that the number of negative water usage anomalies within the second preset number of days is greater than the preset number, and a second-level negative water usage alarm record is generated; The second abnormal situation information is a continuous negative water usage abnormality for a third preset number of days, generating a third-level negative water usage alarm record.

[0083] Optionally, the process of classifying the abnormal situation information to obtain alarm records of different levels also includes: Based on the third abnormal situation information, a threshold alarm record is generated; The third abnormal situation information is that there is no threshold alarm processing record when a threshold alarm is triggered, and a first-level threshold alarm record is generated. The third abnormal situation information is that when a threshold alarm occurs, there is a first preset number of threshold alarm processing records, and a second level of threshold alarm records are generated. The third abnormal situation information is that when a threshold alarm occurs, there is a second preset number of threshold alarm processing records, generating a third level of threshold alarm records.

[0084] It should be noted that this device is the same as the method described above. All implementations in the above method embodiments are applicable to the embodiments of this device and can achieve the same technical effect.

[0085] An embodiment of the present invention also provides a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described in the above embodiments. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0086] In this embodiment of the invention, a computer-readable storage medium is also provided, storing instructions that, when executed on a computer, cause the computer to perform the method described in the above embodiments. All implementations of the methods described in the above embodiments are applicable to this embodiment and can achieve the same technical effect.

[0087] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0088] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0089] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0092] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0093] Furthermore, it should be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of the present invention. Moreover, the steps performing the above series of processes can naturally be executed in the order described, but are not necessarily required to be executed in chronological order; some steps can be executed in parallel or independently of each other. Those skilled in the art will understand that all or any step or component of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or network of computing devices, in hardware, firmware, software, or a combination thereof. This is something that those skilled in the art can achieve by using their basic programming skills after reading the description of the present invention.

[0094] Therefore, the object of the present invention can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose device. Therefore, the object of the present invention can also be achieved simply by providing a program product containing program code implementing the method or apparatus. That is, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent to the present invention. Furthermore, the steps performing the above series of processes can naturally be performed in the order described, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

[0095] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for handling leakage early warning of remote water meters, characterized in that, include: Obtain real-time flow data from remote water meters; The real-time traffic data is preprocessed to obtain processed traffic data; The processed flow data is monitored in real time according to preset monitoring rules to obtain abnormal situation information; wherein, the preset monitoring rules include at least one of the following: data upload rules, negative water usage rules, and threshold alarm rules; wherein, the threshold alarm rules include at least one of the following: normal threshold rules, dynamic threshold rules, and weighted threshold rules. The abnormal situation information is processed in a hierarchical manner to obtain alarm records of different levels; the alarm records of different levels are pushed to the operation and maintenance management platform and user terminal devices.

2. The remote water meter leakage early warning and processing method according to claim 1, characterized in that, The processed traffic data is monitored in real time according to preset monitoring rules to obtain abnormal situation information, including: According to the data upload rules, the processed flow data is monitored for data upload. If the real-time flow data of the remote water meter is not obtained, the first abnormal situation information is obtained. The first abnormal situation information includes the failure to upload the flow data of the remote water meter on the current day and the failure to upload the flow data of the remote water meter for a first preset number of consecutive days.

3. The remote water meter leakage early warning and processing method according to claim 1, characterized in that, The system monitors the processed traffic data in real time according to preset monitoring rules to obtain information on abnormal situations, and also includes: According to the negative water usage rule, the processed flow data is monitored for negative water usage. If the processed flow data of the remote water meter on the current day minus the processed flow data of the previous day with flow is negative, a second abnormal situation information is obtained. The second abnormal situation information includes negative water usage abnormality on the current day, negative water usage abnormality occurring more than a preset number of times within a second preset number of days, and continuous negative water usage abnormality for a third preset number of days.

4. The remote water meter leakage early warning and processing method according to claim 1, characterized in that, The system monitors the processed traffic data in real time according to preset monitoring rules to obtain information on abnormal situations, and also includes: According to the threshold alarm rules, the processed traffic data is monitored for thresholds. If the difference between the processed traffic data of the current day and the processed traffic data of the previous day with traffic is greater than the sum of the daily threshold alarm values ​​of all days within the range of the current day and the previous day with traffic, a third abnormal situation information is obtained. The third abnormal situation information includes no threshold alarm processing record when a threshold alarm occurs, a first preset number of threshold alarm processing records when a threshold alarm occurs, and a second preset number of threshold alarm processing records when a threshold alarm occurs. The daily threshold alarm value is at least one of the daily normal threshold, the daily dynamic threshold, and the daily weighted threshold.

5. The remote water meter leakage early warning and processing method according to claim 2, characterized in that, Based on the abnormal situation information, alarm records of different levels are obtained through hierarchical processing, including: Based on the first abnormal situation information, a data upload alarm record is generated; Among them, the first abnormal situation information is that the remote water meter flow data was not uploaded on that day, generating a first-level data upload alarm record; The first abnormal situation information is that the remote water meter flow data has not been uploaded for a first preset number of consecutive days, and a second-level data upload alarm record is generated.

6. The remote water meter leakage early warning and processing method according to claim 3, characterized in that, Based on the abnormal situation information, alarm records of different levels are obtained through hierarchical processing, including: Based on the second abnormal situation information, a negative water usage alarm record is generated; The second abnormal situation information is that there is a negative water usage abnormality on that day, and a first-level negative water usage alarm record is generated. The second abnormal situation information is that the number of negative water usage anomalies within the second preset number of days is greater than the preset number, and a second-level negative water usage alarm record is generated; The second abnormal situation information is a continuous negative water usage abnormality for a third preset number of days, generating a third-level negative water usage alarm record.

7. The remote water meter leakage early warning and processing method according to claim 4, characterized in that, Based on the abnormal situation information, alarm records of different levels are obtained through hierarchical processing, including: Based on the third abnormal situation information, a threshold alarm record is generated; The third abnormal situation information is that there is no threshold alarm processing record when a threshold alarm is triggered, and a first-level threshold alarm record is generated. The third abnormal situation information is that when a threshold alarm occurs, there is a first preset number of threshold alarm processing records, and a second level of threshold alarm records are generated. The third abnormal situation information is that when a threshold alarm occurs, there is a second preset number of threshold alarm processing records, generating a third level of threshold alarm records.

8. A remote water meter leakage early warning and processing device, characterized in that, include: The acquisition module is used to acquire real-time flow data from the remote water meter. The processing module is used to preprocess the real-time traffic data to obtain processed traffic data; The processed traffic data is monitored in real time according to preset monitoring rules to obtain abnormal situation information; wherein, the preset monitoring rules include at least one of: data upload rules, negative water usage rules, and threshold alarm rules; wherein, the threshold alarm rules include at least one of: normal threshold rules, dynamic threshold rules, and weighted threshold rules; the abnormal situation information is processed in a hierarchical manner to obtain alarm records of different levels; the alarm records of different levels are pushed to the operation and maintenance management platform and user terminal devices.

9. A computing device, characterized in that, include: A processor, a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 7.

Images