An Anti-theft Method for Ultrasonic Water Meters Based on Multi-Sensor Fusion and Multi-Level Response

Abstract

This invention relates to the field of data fusion processing technology, specifically disclosing an ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response. This invention achieves comprehensive detection of various water theft behaviors such as magnetic interference, physical damage, and casing deformation through a sensor monitoring network; accurately identifies the type and severity of anomalies through multi-source information fusion and intelligent level judgment; securely protects communication data through encryption authentication and rolling code mechanisms; and implements differentiated handling from record keeping and remote alarm to valve shut-off through a hierarchical linkage response mechanism, effectively improving the comprehensiveness of ultrasonic water meter anti-theft monitoring and the adaptability of response handling.

Description

Technical Field

[0001] This invention relates to the field of data fusion processing technology, and more specifically, to an ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response. Background Technology

[0002] Ultrasonic water meters, as the core equipment for smart water metering, accurately measure water flow in pipelines using the ultrasonic time-difference method and are widely used in residential and industrial water metering. However, because water meters are directly related to water billing, some users resort to various means to steal water, causing serious economic losses to water supply companies. Current water theft mainly falls into three categories: First, magnetic interference theft, where thieves use powerful permanent magnets close to the water meter casing to interfere with internal electronic components or metering sensors, causing inaccurate metering data or malfunctioning metering functions; second, physical damage theft, where thieves damage the water meter's integrity by prying open the casing, disassembling parts, or using force to bypass the metering device and directly draw water; and third, data tampering theft, where thieves attack the water meter's communication link to intercept and tamper with uploaded metering data, or replay historical data to deceive the water management platform.

[0003] Existing water meter anti-theft technologies typically only protect against a single type of water theft, lacking comprehensive detection capabilities against multiple types of theft. Furthermore, their response methods after detecting anomalies are relatively simplistic, lacking a mechanism for tiered responses based on the severity of the anomaly. This can easily lead to false alarms causing water outages that affect normal users, or delayed responses that allow theft to continue. In addition, existing water meters have weak communication security, with data transmission often using plaintext or simple encryption methods, making them vulnerable to sophisticated communication attacks.

[0004] Therefore, it is necessary to provide an ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response to solve the above technical problems. In order to solve the above problems, a technical solution is provided. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of existing technologies, this invention provides an ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response. This method addresses the problem that existing ultrasonic water meter anti-theft technologies primarily target single water theft methods such as magnetic interference, physical damage, or data tampering, lacking the ability to comprehensively identify and collaboratively determine multiple types of water theft behaviors. Furthermore, these technologies suffer from limited anomaly handling methods, weak communication security protection, and difficulty in effectively curbing continuous or malicious water theft while preventing accidental water outages.

[0006] To achieve the above objectives, the present invention provides the following technical solution: An anti-theft method for ultrasonic water meters based on multi-sensor fusion and multi-level response includes the following steps: By constructing a sensor monitoring network, multi-source information data of water meters are monitored in real time. Based on the multi-source information data, multi-level anomaly detection is performed in real time and a first-level anomaly confirmation flag is output. The first-level anomaly confirmation flag includes magnetic field anomaly confirmation flag, physical anomaly confirmation flag, and deformation anomaly confirmation flag. An anomaly type code is generated by logical combination operation based on the first-level anomaly confirmation flag, and a quantitative indicator of the anomaly severity is obtained by setting an anomaly timer. A multi-source anomaly fusion judgment model is built based on the anomaly type code and the quantitative indicator of the anomaly severity to determine the anomaly level of the water meter. Obtain the metering data uploaded by the water meter to the water management platform, encrypt the metering data, and embed a rolling code mechanism to defend against replay attacks; Based on the level of water meter anomaly, corresponding response actions are executed to achieve hierarchical and coordinated control for handling water meter anomalies.

[0007] As a further aspect of the present invention, the multi-source information data includes magnetic field strength data, acceleration values, and voltage output values.

[0008] As a further aspect of the present invention, multi-layer anomaly detection is performed in real time based on multi-source information data and a first-level anomaly confirmation flag is output. The multi-layer anomaly detection includes a first-level anomaly confirmation detection, a second-level anomaly confirmation detection, and a third-level anomaly confirmation detection. The first-level anomaly confirmation detection is used to output a magnetic field anomaly confirmation flag, the second-level anomaly confirmation detection is used to output a physical anomaly confirmation flag, and the third-level anomaly confirmation detection is used to output a deformation anomaly confirmation flag.

[0009] As a further aspect of the present invention, a layer of anomaly confirmation detection is used to output a magnetic field anomaly confirmation flag, and the specific steps are as follows: By setting a magnetic field reference register in the main controller, the background magnetic field strength value under the normal operating environment of the water meter is stored. After each sampling, the absolute value of the difference between the current magnetic field strength data and the benchmark value is calculated, and the absolute value of the difference is compared with the preset magnetic field anomaly threshold. When the absolute value of the difference exceeds the magnetic field anomaly threshold, the current sampling result is marked as a magnetic field anomaly sample; when the absolute value of the difference does not exceed the magnetic field anomaly threshold, it is marked as a magnetic field normal sample. By setting a magnetic field anomaly counter, the number of magnetic field anomaly samples within a continuous sampling period is accumulated. When the accumulated number exceeds a preset threshold for the number of consecutive anomalies, a magnetic field anomaly confirmation flag is generated.

[0010] As a further aspect of the present invention, the two-layer anomaly confirmation detection is used to output physical anomaly confirmation flags, which include vibration anomaly confirmation flags and impact anomaly confirmation flags. The specific steps are as follows: By setting a sliding window buffer in the main controller, the acceleration data sequence within the most recent sampling periods can be stored; The standard deviation of acceleration data within the sampling period is calculated as the vibration intensity index, and the vibration intensity index is compared with the preset vibration anomaly threshold. When the vibration intensity index exceeds the preset vibration anomaly threshold, a vibration anomaly confirmation flag is generated; when the vibration intensity index does not exceed the preset vibration anomaly threshold, a normal vibration sample is generated. By calculating the rate of change between the current acceleration data and the previous acceleration data, an impact anomaly confirmation flag is generated when the rate of change exceeds a preset impact detection threshold; and an impact normal sample is generated when the rate of change does not exceed the preset impact detection threshold.

[0011] As a further aspect of the present invention, a three-layer anomaly confirmation detection is used to output a deformation anomaly confirmation flag, and the specific steps are as follows: The main controller reads the output voltage value of the photoelectric conversion circuit at a fixed sampling period. The output voltage value reflects the change in optical fiber transmission loss and indirectly characterizes the degree of deformation of the shell. A deformation reference register is set in the main controller to store the photoelectric conversion output reference value when the water meter casing is in a normal state after assembly. After each sampling, the relative rate of change between the current output voltage value and the photoelectric conversion output reference value is calculated, and the relative rate of change is compared with the preset deformation anomaly threshold. When the relative rate of change exceeds the deformation anomaly threshold, a deformation anomaly confirmation flag is generated; when the relative rate of change does not exceed the deformation anomaly threshold, a normal deformation sample is generated.

[0012] As a further aspect of the present invention, an exception type code is generated by performing logical combination operations based on the first-level exception confirmation flag, and the specific steps are as follows: An anomaly type encoding register is set to generate an anomaly type code based on the combined state of the first-level anomaly confirmation flags. The anomaly type codes include single magnetic interference anomaly, single magnetic interference anomaly, single shell anomaly, and compound anomaly types. Specifically: when only the magnetic field anomaly confirmation flag is valid, the code indicates a single magnetic interference anomaly; when only the physical anomaly confirmation flag is valid, the code indicates a single physical damage anomaly; when only the deformation anomaly confirmation flag is valid, the code indicates a single shell anomaly; when two or more anomaly confirmation flags are valid simultaneously, the code indicates a compound anomaly type. When any exception type is encoded, a comprehensive exception trigger signal is generated.

[0013] As a further aspect of the present invention, a quantitative indicator of the severity of the anomaly is obtained by setting an anomaly timer, and the specific steps are as follows: By setting an abnormal duration timer, the duration of the abnormal state is recorded during the validity period of the comprehensive abnormal trigger signal. By setting an anomaly accumulation counter, the effective number of comprehensive anomaly trigger signals within the monitoring time window is accumulated; The duration of abnormal states and the cumulative number of abnormalities are used as quantitative indicators of the severity of the abnormalities.

[0014] As a further aspect of the present invention, a multi-source anomaly fusion judgment model is constructed based on the anomaly type code and the quantitative index of anomaly severity to determine the anomaly level of the water meter. The specific steps for constructing the multi-source anomaly fusion judgment model are as follows: In the main controller, a level determination rule table is set up. The level determination rule table defines the mapping relationship between the abnormality type, the abnormality time threshold, the abnormal cumulative number threshold and the abnormality level. The duration of the anomaly is compared with the first anomaly time threshold in the level determination rule table. When the duration of the anomaly is less than the first anomaly time threshold, it is determined to be a level 1 anomaly; when the duration exceeds the first anomaly time threshold but is less than the second anomaly time threshold, it is determined to be a level 2 anomaly; when the duration exceeds the second anomaly time threshold, or when the anomaly type code indicates a composite anomaly, it is determined to be a level 3 anomaly. The cumulative number of anomalies is compared with the threshold number of cumulative anomalies in the level determination rule table. When the cumulative number of anomalies exceeds the threshold number, the current anomaly level is upgraded by one level, and a water meter anomaly level code is generated. The anomaly type code, water meter anomaly level code, and occurrence timestamp information are packaged to form an anomaly event record.

[0015] As a further aspect of the present invention, a corresponding response action is executed based on the water meter anomaly level to achieve hierarchical linkage control for water meter anomaly handling. The specific steps are as follows: Set up a response action control module in the main controller to trigger the response process according to the exception level code; When the exception level code indicates a level 1 exception, the response action control module executes a level 1 response, which specifically involves recording the response process locally. When the anomaly level code indicates a level 2 anomaly, the response action control module executes a level 2 response. Specifically, the level 2 response is an additional execution of the remote alarm response process on top of the level 1 response. When the anomaly level code indicates a level 3 anomaly, the response action control module executes a level 3 response. Specifically, the level 3 response involves executing a valve shut-off response procedure in addition to the level 1 and level 2 responses.

[0016] The technical effects and advantages of this invention, an ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response, are as follows: This invention utilizes a sensor monitoring network composed of magnetoresistive sensors, accelerometers, and fiber optic micro-bending sensors to comprehensively detect various water theft behaviors such as magnetic interference, physical damage, and casing deformation. This avoids misjudgments caused by single sensors or single detection methods, providing reliable and comprehensive basic data for subsequent anomaly analysis. Through multi-source information fusion and intelligent hierarchical judgment, it accurately identifies the type and severity of anomalies, effectively distinguishing between occasional interference and continuous, malicious water theft, avoiding the crude approach of "closing the valve at the first sign of an anomaly," and improving the accuracy and rationality of anomaly judgment. Based on encryption authentication and rolling code mechanisms, it achieves secure protection of communication data, integrating communication security into the overall anti-theft system. This effectively prevents metering data from being stolen, tampered with, or replayed during transmission, ensuring the authenticity and integrity of the data received by the water management platform, providing a reliable data foundation for anomaly handling and subsequent evidence collection. Through a hierarchical linkage response mechanism, it achieves differentiated handling from record keeping and remote alarm to valve closure, effectively improving the comprehensiveness of ultrasonic water meter anti-theft monitoring and the adaptability of response measures. This invention not only improves the comprehensiveness and accuracy of water meter anti-theft detection, but also significantly enhances the adaptability of abnormal handling and the overall reliability of the system operation. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response, provided as an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of this invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described technical solutions are only a part of this invention, and not all of it. All other technical solutions obtained by those skilled in the art based on the technical solutions of this invention without inventive effort are within the scope of protection of this invention.

[0019] like Figure 1 The diagram shown is a flowchart of an ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response provided by an embodiment of the present invention. Figure 1The execution entity of the method shown can be a software and / or hardware device. The execution entity of this application can include, but is not limited to, at least one of the following: user equipment, network equipment, etc. User equipment can include, but is not limited to, computers, smartphones, personal digital assistants (PDAs), and the aforementioned electronic devices. Network equipment can include, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers. Cloud computing is a type of distributed computing, consisting of a super virtual computer composed of a group of loosely coupled computers. This embodiment does not limit this. Steps S1 to S4 are detailed as follows: Step S1: By constructing a sensor monitoring network, multi-source information data of the water meter is monitored in real time. Multi-level anomaly detection is performed in real time based on the multi-source information data, and a first-level anomaly confirmation flag is output. The first-level anomaly confirmation flag includes magnetic field anomaly confirmation flag, physical anomaly confirmation flag, and deformation anomaly confirmation flag. Step S2: Based on the first-level anomaly confirmation flag, perform logical combination operations to generate anomaly type codes, and obtain quantitative indicators of anomaly severity by setting an anomaly timer. Based on the anomaly type codes and quantitative indicators of anomaly severity, build a multi-source anomaly fusion judgment model to determine the water meter anomaly level. Step S3: Obtain the metering data uploaded by the water meter to the water platform, encrypt the metering data, and embed a rolling code mechanism to defend against replay attacks; Step S4: Execute the corresponding response action based on the water meter anomaly level to achieve hierarchical linkage control for water meter anomaly handling.

[0020] Preferably, the multi-source information data includes magnetic field strength data, acceleration values, and voltage output values.

[0021] It should be specifically explained that the steps for acquiring magnetic field strength data are as follows: Anomalies in the magnetic field around the water meter are detected using a magnetoresistive sensor to identify water theft caused by strong magnetic interference. A tunnel magnetoresistive sensor is installed on the circuit board inside the water meter. The sensitive axis of the tunnel magnetoresistive sensor is arranged perpendicular to the main plane of the water meter's outer casing to obtain optimal detection sensitivity to external magnets approaching. The magnetoresistive sensor is connected to the main controller via an analog-to-digital converter circuit. The main controller reads the output voltage value of the magnetoresistive sensor at a fixed sampling period and converts the voltage value into the corresponding magnetic field strength value. The steps for obtaining acceleration values ​​are as follows: an acceleration sensor detects abnormal vibration and impact of the water meter casing to identify water theft behavior caused by physical damage. A microelectromechanical system acceleration sensor is installed on the circuit board inside the water meter. This sensor has three mutually orthogonal sensitive axes and can detect acceleration changes in three-dimensional space. The acceleration sensor is connected to the main controller through a serial communication interface. The main controller reads the acceleration values ​​of the three axes at a high sampling frequency.

[0022] The steps for obtaining the voltage output value are as follows: An abnormal deformation of the water meter casing is detected using a fiber optic micro-bend sensor to identify prying or dismantling-type water theft. Fiber optic micro-bend sensors are installed at key structural locations on the water meter casing, with the bent sections of the sensor's fiber optic cable attached to the casing seams and stress-prone areas. The transmitting end of the fiber optic sensor is connected to a light source driving circuit, and the receiving end is connected to a photoelectric conversion circuit. The output of the photoelectric conversion circuit is connected to the analog-to-digital conversion channel of the main controller.

[0023] One embodiment of the present invention takes a residential ultrasonic water meter as an example. The water meter internally constructs a sensor monitoring network consisting of a magnetoresistive sensor, an accelerometer, and a fiber optic micro-bend sensor to collect multi-source information data in real time during the water meter's operation. Specifically, the multi-source information data includes magnetic field strength data, acceleration values, and voltage output values; these three types of data correspond to the physical characteristics of different types of water theft.

[0024] In real-world scenarios, when a user attempts to interfere with the water meter's meter readings using an external strong magnet, the magnet close to the meter's casing significantly alters the magnetic field distribution around the meter. At this point, the tunnel magnetoresistive sensor, mounted on the circuit board inside the water meter, comes into play. Because the sensitive axis of the tunnel magnetoresistive sensor is perpendicular to the main plane of the water meter's casing, it has high detection sensitivity to magnetic field disturbances generated by external magnets. The magnetoresistive sensor converts the sensed magnetic field changes into a voltage signal, which is then sent to the main controller via an analog-to-digital converter. The main controller continuously reads the voltage value according to a preset sampling period and converts it into magnetic field strength data, thereby enabling real-time monitoring of water theft attempts caused by strong magnetic interference.

[0025] In another type of real-world water theft scenario, users might damage the water meter's structure or affect its metering accuracy by striking, prying, or violently shaking the meter casing. To address this type of physical damage, a microelectromechanical system (MEMS) accelerometer sensor is installed inside the water meter to monitor casing vibration and impact. This accelerometer has three mutually orthogonal sensing axes, enabling it to simultaneously acquire changes in the water meter's acceleration in three-dimensional space. The accelerometer is connected to the main controller via a serial communication interface. The main controller continuously reads the acceleration values ​​along the three axes at a high sampling frequency, thus promptly acquiring corresponding acceleration data when the water meter experiences abnormal vibration or transient impact, providing a basis for identifying physically destructive water theft.

[0026] Furthermore, during the actual disassembly or prying of water meters, key structural parts of the water meter casing often undergo minute but continuous deformation. To accurately detect this behavior, fiber optic micro-bend sensors are installed at the seams and stress-prone locations of the water meter casing. The bent sections of the fiber optic micro-bend sensors are tightly attached to the casing structure; when the casing is pried or deformed, the fiber optic transmission loss changes. The transmitting end of the fiber optic sensor is connected to a light source driving circuit, and the receiving end is connected to a photoelectric conversion circuit. The photoelectric conversion circuit converts the changes in the optical signal into a voltage output value, which is then acquired through the analog-to-digital conversion channel of the main controller. The main controller periodically reads this voltage output value, which indirectly reflects the deformation state of the water meter casing, thereby effectively identifying water theft behaviors such as prying and disassembly.

[0027] By using magnetic field strength data, acceleration values, and voltage output values, the abnormal state of the water meter is characterized from three different physical levels: external intervention, implementation behavior, and structural results. Multi-source information data is collected uniformly within the same water meter and sent to the main controller, providing a real and reliable data foundation for subsequent anomaly fusion judgment and graded response control.

[0028] Preferably, multi-layer anomaly detection is performed in real time based on multi-source information data, and a first-level anomaly confirmation flag is output. The multi-layer anomaly detection includes a first-level anomaly confirmation detection, a second-level anomaly confirmation detection, and a third-level anomaly confirmation detection. The first-level anomaly confirmation detection is used to output a magnetic field anomaly confirmation flag, the second-level anomaly confirmation detection is used to output a physical anomaly confirmation flag, and the third-level anomaly confirmation detection is used to output a deformation anomaly confirmation flag.

[0029] Preferably, a single layer of anomaly confirmation detection is used to output a magnetic field anomaly confirmation flag, and the specific steps are as follows: By setting a magnetic field reference register in the main controller, the background magnetic field strength value under the normal operating environment of the water meter is stored. After each sampling, the absolute value of the difference between the current magnetic field strength data and the benchmark value is calculated, and the absolute value of the difference is compared with the preset magnetic field anomaly threshold. When the absolute value of the difference exceeds the magnetic field anomaly threshold, the current sampling result is marked as a magnetic field anomaly sample; when the absolute value of the difference does not exceed the magnetic field anomaly threshold, it is marked as a magnetic field normal sample. By setting a magnetic field anomaly counter, the number of magnetic field anomaly samples within a continuous sampling period is accumulated. When the accumulated number exceeds a preset threshold for the number of consecutive anomalies, a magnetic field anomaly confirmation flag is generated.

[0030] Preferably, the two-layer anomaly confirmation detection is used to output physical anomaly confirmation flags, which include vibration anomaly confirmation flags and impact anomaly confirmation flags. The specific steps are as follows: By setting a sliding window buffer in the main controller, the acceleration data sequence within the most recent sampling periods can be stored; The standard deviation of acceleration data within the sampling period is calculated as the vibration intensity index, and the vibration intensity index is compared with the preset vibration anomaly threshold. When the vibration intensity index exceeds the preset vibration anomaly threshold, a vibration anomaly confirmation flag is generated; when the vibration intensity index does not exceed the preset vibration anomaly threshold, a normal vibration sample is generated. By calculating the rate of change between the current acceleration data and the previous acceleration data, an impact anomaly confirmation flag is generated when the rate of change exceeds a preset impact detection threshold; and an impact normal sample is generated when the rate of change does not exceed the preset impact detection threshold.

[0031] Preferably, the three-layer anomaly confirmation detection is used to output a deformation anomaly confirmation flag, and the specific steps are as follows: The main controller reads the output voltage value of the photoelectric conversion circuit at a fixed sampling period. The output voltage value reflects the change in optical fiber transmission loss and indirectly characterizes the degree of deformation of the shell. A deformation reference register is set in the main controller to store the photoelectric conversion output reference value when the water meter casing is in a normal state after assembly. After each sampling, the relative rate of change between the current output voltage value and the photoelectric conversion output reference value is calculated, and the relative rate of change is compared with the preset deformation anomaly threshold. When the relative rate of change exceeds the deformation anomaly threshold, a deformation anomaly confirmation flag is generated; when the relative rate of change does not exceed the deformation anomaly threshold, a normal deformation sample is generated.

[0032] One embodiment of the present invention takes an ultrasonic water meter installed in a residential water supply network as an example. The main controller, based on multi-source information data collected by a magnetoresistive sensor, an accelerometer, and a fiber optic micro-bending sensor, performs real-time anomaly detection on the water meter's operating status according to a hierarchical structure, and outputs corresponding first-level anomaly confirmation flags layer by layer. By dividing the anomaly detection process into first-level anomaly confirmation detection, second-level anomaly confirmation detection, and third-level anomaly confirmation detection, different types of water theft are independently identified at their respective physical levels, while maintaining a clear logical division of labor.

[0033] In the first-level anomaly detection process, the focus is on identifying water theft behaviors caused by strong magnetic interference. A pre-set magnetic field reference register in the main controller stores the background magnetic field strength values ​​collected by the water meter under normal operating conditions. During water meter operation, the main controller acquires the current magnetic field strength data according to a fixed sampling period and calculates the absolute value of the difference between this magnetic field strength data and the magnetic field reference value. Then, the absolute value of the difference is compared with a preset magnetic field anomaly threshold. When the absolute value of the difference exceeds the magnetic field anomaly threshold, the corresponding sampling result is marked as a magnetic field anomaly sample; otherwise, it is marked as a normal magnetic field sample. Simultaneously, the main controller accumulates the number of magnetic field anomaly samples within consecutive sampling periods using a magnetic field anomaly counter. When the accumulated number exceeds a preset threshold for consecutive anomalies, the magnetic field anomaly is determined to have persistent characteristics, thereby generating a magnetic field anomaly confirmation flag to avoid misjudgments caused by transient magnetic field fluctuations.

[0034] In the second-level anomaly confirmation detection process, detection is performed for water theft behaviors involving physical damage such as knocking, prying, or violent shaking of the water meter casing. The main controller first sets up a sliding window buffer to store the triaxial acceleration data sequence within the most recent sampling periods. Based on the acceleration data in the buffer, the main controller calculates the standard deviation of the acceleration data within the sampling period and uses it as a vibration intensity index. When the vibration intensity index exceeds a preset vibration anomaly threshold, a vibration anomaly confirmation flag is generated; when the vibration intensity index does not exceed the threshold, it is determined to be a normal vibration sample. Simultaneously, the main controller also detects transient impact behaviors by calculating the rate of change between the current acceleration data and the previous acceleration data. When the rate of change exceeds a preset impact detection threshold, an impact anomaly confirmation flag is generated; when the rate of change does not exceed the impact detection threshold, an impact normal sample is generated. By simultaneously introducing vibration anomaly confirmation flags and impact anomaly confirmation flags, continuous vibration behaviors and transient impact behaviors can be distinguished, improving the accuracy of identifying physically destructive water theft behaviors.

[0035] In the three-layer anomaly confirmation detection process, deformation anomalies caused by prying or disassembling the water meter casing are identified. The main controller reads the output voltage value of the photoelectric conversion circuit at a fixed sampling period. This output voltage value reflects the change in fiber optic transmission loss in the fiber optic micro-bend sensor, thus indirectly characterizing the degree of deformation of the water meter casing. A deformation reference register is set in the main controller to store the photoelectric conversion output reference value when the water meter is assembled and the casing is in a normal state. After each sampling, the main controller calculates the relative rate of change of the current output voltage value relative to the deformation reference value and compares this relative rate of change with a preset deformation anomaly threshold. When the relative rate of change exceeds the deformation anomaly threshold, a deformation anomaly confirmation flag is generated; when the relative rate of change does not exceed the deformation anomaly threshold, a normal deformation sample is generated. Through the above, structural stress or minor deformation anomalies can be detected in advance before the casing is obviously damaged.

[0036] The multi-layer anomaly confirmation detection identifies abnormal states of water meters from three different levels: magnetic interference, physical damage, and shell deformation. Each layer of detection outputs a clear first-level anomaly confirmation mark, providing a clear and reliable input basis for subsequent multi-source anomaly fusion judgment and anomaly level assessment.

[0037] Preferably, the exception type code is generated by performing logical combination operations based on the first-level exception confirmation flag. The specific steps are as follows: An anomaly type encoding register is set to generate an anomaly type code based on the combined state of the first-level anomaly confirmation flags. The anomaly type codes include single magnetic interference anomaly, single magnetic interference anomaly, single shell anomaly, and compound anomaly types. Specifically: when only the magnetic field anomaly confirmation flag is valid, the code indicates a single magnetic interference anomaly; when only the physical anomaly confirmation flag is valid, the code indicates a single physical damage anomaly; when only the deformation anomaly confirmation flag is valid, the code indicates a single shell anomaly; when two or more anomaly confirmation flags are valid simultaneously, the code indicates a compound anomaly type. When any exception type is encoded, a comprehensive exception trigger signal is generated.

[0038] In one embodiment of the present invention, after the main controller completes the layered detection of magnetic field anomalies, physical anomalies, and deformation anomalies and obtains the corresponding first-level anomaly confirmation flags, it further performs logical combination operations on each anomaly confirmation flag to determine the anomaly type of the current water meter. To this end, an anomaly type encoding register is set in the main controller to generate the corresponding anomaly type code based on the valid status of each first-level anomaly confirmation flag.

[0039] Specifically, when the main controller detects that only the magnetic field anomaly confirmation flag is valid, while both the physical anomaly confirmation flag and the deformation anomaly confirmation flag are invalid, the anomaly type encoding register is assigned the value of "single magnetic interference anomaly," indicating that the water meter may be affected by external strong magnetic interference but has not yet shown obvious physical damage or casing deformation. When only the physical anomaly confirmation flag is valid, while both the magnetic field anomaly confirmation flag and the deformation anomaly confirmation flag are invalid, the anomaly type encoding register is assigned the value of "single physical damage anomaly," indicating that the water meter casing has been subjected to physical damage such as knocking, vibration, or impact. When only the deformation anomaly confirmation flag is valid, the anomaly type encoding register is assigned the value of "single casing anomaly," indicating that the water meter casing has been pried or deformed but no obvious magnetic interference or severe vibration has been detected.

[0040] Furthermore, when the main controller detects that two or more of the magnetic field anomaly confirmation flag, physical anomaly confirmation flag, and deformation anomaly confirmation flag are simultaneously in a valid state, the anomaly type encoding register is assigned a composite anomaly type. This is used to characterize the water meter's simultaneous presence of multiple abnormal behaviors such as external magnetic interference, physical damage, and / or casing deformation within the same time period, reflecting that the water theft behavior has a stronger continuity and malice.

[0041] After the anomaly type code is generated, the main controller generates a comprehensive anomaly trigger signal as long as any valid code exists in the anomaly type code register. This comprehensive anomaly trigger signal serves as a unified trigger condition for subsequent anomaly fusion judgment and anomaly level assessment. It enables timely entry into the anomaly handling process upon detection of any valid anomaly type, thereby achieving rapid response and effective handling of abnormal water meter conditions.

[0042] Preferably, a quantitative indicator of the severity of the anomaly is obtained by setting an anomaly timer. The specific steps are as follows: By setting an abnormal duration timer, the duration of the abnormal state is recorded during the validity period of the comprehensive abnormal trigger signal. By setting an anomaly accumulation counter, the effective number of comprehensive anomaly trigger signals within the monitoring time window is accumulated; The duration of abnormal states and the cumulative number of abnormalities are used as quantitative indicators of the severity of the abnormalities.

[0043] In one embodiment of the present invention, after the device generates a comprehensive anomaly trigger signal through multi-source sensing and logic combination during operation, the severity of the anomaly is further quantitatively assessed through an anomaly timing and counting mechanism. Specifically, an anomaly duration timer is set in the main controller. When the comprehensive anomaly trigger signal is valid, the anomaly duration timer automatically starts and continuously counts to record the duration of the current anomaly. For example, when the water meter is subjected to strong magnetic interference for a long time or its casing is continuously pried open, the anomaly duration timer can accurately reflect the persistence of the abnormal behavior, thereby distinguishing between momentary interference and long-term malicious damage.

[0044] Simultaneously, an anomaly accumulation counter is set up to cumulatively count each valid trigger of the comprehensive anomaly trigger signal within a preset monitoring time window. For example, if the device triggers multiple comprehensive anomaly signals within a daily or weekly monitoring period, even if each anomaly is short in duration, the anomaly accumulation counter can still reflect the frequency of the abnormal behavior. By using the duration of the abnormal state and the cumulative number of anomalies as quantitative indicators of anomaly severity, it is possible not only to identify whether an anomaly exists but also to comprehensively assess its persistence and repetitiveness. This provides a reliable basis for subsequent anomaly level classification, risk alarm strategy adjustment, and maintenance handling priority, making the anomaly judgment results closer to real-world usage scenarios and avoiding misjudgments based solely on a single anomaly trigger.

[0045] Preferably, a multi-source anomaly fusion judgment model is built based on the anomaly type code and the quantitative index of anomaly severity to determine the anomaly level of the water meter. The specific steps for building the multi-source anomaly fusion judgment model are as follows: In the main controller, a level determination rule table is set up. The level determination rule table defines the mapping relationship between the abnormality type, the abnormality time threshold, the abnormal cumulative number threshold and the abnormality level. The duration of the anomaly is compared with the first anomaly time threshold in the level determination rule table. When the duration of the anomaly is less than the first anomaly time threshold, it is determined to be a level 1 anomaly; when the duration exceeds the first anomaly time threshold but is less than the second anomaly time threshold, it is determined to be a level 2 anomaly; when the duration exceeds the second anomaly time threshold, or when the anomaly type code indicates a composite anomaly, it is determined to be a level 3 anomaly. The cumulative number of anomalies is compared with the threshold number of cumulative anomalies in the level determination rule table. When the cumulative number of anomalies exceeds the threshold number, the current anomaly level is upgraded by one level, and a water meter anomaly level code is generated. The anomaly type code, water meter anomaly level code, and occurrence timestamp information are packaged to form an anomaly event record.

[0046] In one embodiment of the present invention, after completing multi-source anomaly detection, anomaly type encoding generation, and anomaly severity quantification, the main controller further uses a multi-source anomaly fusion judgment model to uniformly determine the anomaly level of the water meter. Specifically, a level judgment rule table is pre-configured within the main controller. This rule table specifies, in the form of table entries, the anomaly duration threshold, the cumulative anomaly count threshold, and the corresponding anomaly level for different anomaly type codes, reflecting the risk differences of different anomaly behaviors in actual operation. For example, a relatively high time threshold can be set for a single magnetic interference anomaly, while a more stringent time and count threshold can be set for anomalies involving casing deformation.

[0047] During the anomaly level determination process, the main controller first compares the currently detected anomaly duration with the first anomaly time threshold in the level determination rule table. When the anomaly duration is less than the first anomaly time threshold, the anomaly is determined to be a Level 1 anomaly with a relatively small impact on the metering results. When the anomaly duration exceeds the first anomaly time threshold but does not exceed the second anomaly time threshold, it is determined to be a Level 2 anomaly with a higher risk. When the anomaly duration exceeds the second anomaly time threshold, or the anomaly type code indicates a composite anomaly type containing at least two of the following: magnetic interference, physical damage, or casing deformation, it is directly determined to be a Level 3 anomaly with the highest risk, reflecting its significant threat to the safety and reliability of the water meter.

[0048] Based on this, the main controller further compares the cumulative number of anomalies with the preset threshold for cumulative anomalies in the level determination rule table. When the cumulative number of anomalies exceeds the corresponding threshold within the monitoring time window, the currently determined anomaly level is automatically upgraded by one level, thus reflecting the cumulative impact of the repetition of abnormal behavior on the risk level. Finally, the main controller generates the corresponding water meter anomaly level code and uniformly encapsulates the anomaly type code, water meter anomaly level code, and timestamp information of the anomaly occurrence to form a complete anomaly event record. This anomaly event record can be used for subsequent remote uploading, operation and maintenance analysis, or law enforcement evidence collection, ensuring that the anomaly determination results have clear logical basis and traceability in real-world applications.

[0049] Preferably, the metering data uploaded by the water meter to the water management platform is obtained, and the metering data is encrypted. The specific steps are as follows: Metering data uploaded by water meters to the water management platform is encrypted to prevent theft or tampering during transmission, ensuring the confidentiality and integrity of the metering data. A symmetric encryption algorithm module is integrated into the main controller, employing a combination of block encryption and authentication encryption to encrypt the communication payload. Before sending the metering data, the main controller first reads a preset encryption key from a secure storage area and divides the metering data to be sent into fixed-length data blocks. Each data block is then encrypted sequentially to generate a ciphertext block, and an authentication tag is simultaneously generated and appended to the ciphertext. All ciphertext blocks and authentication tags are combined to form the encrypted communication payload, which is then transmitted to the water management platform via the wireless communication module.

[0050] After receiving encrypted data, the water platform uses the same key and algorithm to decrypt and verify the authentication tag. Only data that passes authentication is accepted and processed, while data that fails authentication is judged to have been tampered with and discarded.

[0051] Preferably, a rolling code mechanism is embedded to defend against replay attacks, and the specific steps are as follows: By embedding a rolling code mechanism in the communication frames, replay attacks by interceptors can be prevented. A rolling code counter is set in the main controller, which automatically increments after each communication frame is sent. The value of the rolling code counter is embedded as the rolling code of the current frame in the header field of the communication frame.

[0052] When assembling a communication frame, the current value of the rolling code counter is filled into a designated position in the frame header. The frame header is then combined with the payload encrypted in step six to form a complete communication frame. The rolling code and payload participate in the encryption operation together, ensuring that the value of the rolling code is also encrypted and cannot be tampered with by an attacker.

[0053] The water management platform maintains a rolling code expectation value register corresponding to each water meter. Upon receiving a communication frame, the platform decrypts and extracts the rolling code from the frame, comparing it with the expectation value. If the received rolling code is less than or equal to the expectation value, the frame is determined to be a replay attack frame and discarded; if the received rolling code is greater than the expectation value and within the allowed window range, the frame is accepted and the expectation value register is updated.

[0054] To prevent metering data from being illegally intercepted, tampered with, or forged during wireless transmission, this invention implements encryption protection for the metering data during the data transmission phase. Specifically, before the water meter prepares to upload metering data, the main controller calls its internally integrated symmetric encryption algorithm module to read a preset communication key from the secure storage area and divides the metering data to be uploaded into several data packets according to a preset length. Subsequently, each data packet undergoes a packet encryption operation in sequence, and a corresponding authentication tag is generated synchronously during the encryption process for subsequent integrity and authenticity verification. All encrypted packets and authentication tags are combined to form an encrypted communication payload, which is sent to the water management platform via the wireless communication module. Upon receiving the data, the water management platform decrypts the communication payload using the same key and algorithm as the water meter and verifies the authentication tag. Only when authentication is successful is the metering data written into the business system; if authentication fails, the data is discarded, thereby effectively ensuring the confidentiality and integrity of the metering data during transmission.

[0055] Building upon this foundation, to further defend against attackers intercepting legitimate communication data and launching replay attacks, a rolling code mechanism is embedded in the communication process. Specifically, a rolling code counter is set up in the main controller. This counter automatically increments after each successful transmission of a communication frame, and its current value is written as the rolling code into a designated field in the communication frame header. During the communication frame assembly process, the rolling code and the encrypted communication payload participate in the encryption operation together, ensuring that the rolling code itself is also encrypted and protected, preventing independent tampering. The water platform maintains a corresponding rolling code expectation value register for each water meter. After receiving and decrypting a communication frame, it extracts the rolling code from the frame header and compares it with the expectation value. If the rolling code is less than or equal to the expectation value, the communication frame is determined to be a replay attack and discarded. If the rolling code is greater than the expectation value and within the allowed window range, it is determined to be a legitimate communication frame, and the expectation value register is updated. Through this mechanism, a collaborative design for encryption protection and replay attack defense in the metering data transmission process is achieved, further enhancing the overall security of smart water meter communication.

[0056] Preferably, a corresponding response action is executed based on the water meter anomaly level to achieve hierarchical linkage control for water meter anomaly handling. The specific steps are as follows: Set up a response action control module in the main controller to trigger the response process according to the exception level code; When the anomaly level code indicates a Level 1 anomaly, the response action control module executes a Level 1 response, specifically by recording the response process locally. The generated anomaly event record is written to the evidence storage area of ​​the non-volatile memory. The record includes information such as the anomaly type, occurrence time, duration, and raw sensor data. Storage uses a circular overwrite method; when the storage area is full, the oldest record is overwritten by the newest record. When the anomaly level code indicates a level 2 anomaly, the response action control module executes a level 2 response. Specifically, the level 2 response, in addition to executing the level 1 response, involves performing an additional remote alarm response procedure. This includes assembling an alarm communication frame containing anomaly event information, using data encryption and rolling code mechanisms for security processing, and then sending the alarm frame to the water management platform via the wireless communication module. Upon receiving the alarm, the water management platform can conduct manual verification or dispatch on-site handling. When the anomaly level code indicates a Level 3 anomaly, the response action control module executes a Level 3 response. Specifically, the Level 3 response, in addition to executing Level 1 and Level 2 responses, additionally performs a valve shut-off response procedure. A shut-off command is sent to the electric valve drive circuit built into the water meter. The drive circuit controls the electric valve to rotate to the closed position, cutting off the water supply pipeline and directly blocking water theft. After the valve is shut off, a valve status lock flag is set to prevent automatic restoration of water supply without authorization. Restoration requires an unlock command issued by the water management platform or manual on-site operation.

[0057] In one embodiment of the present invention, after the main controller completes the anomaly level determination and generates the corresponding anomaly level code, the response action control module intervenes to execute the graded linkage handling process. The control module is pre-embedded in the water meter main controller and is used to automatically trigger response actions of different intensities according to the anomaly level code, thereby ensuring metering safety while avoiding excessive intervention and balancing safety and the user's normal water use needs.

[0058] When a water meter is identified as having a Level 1 anomaly, such as a short-term magnetic field fluctuation or slight vibration, the response action control module only executes the local recording response procedure. At this time, the generated anomaly event record is written to the non-volatile memory evidence storage area inside the water meter. The record includes the anomaly type, the time of occurrence, the duration, and the corresponding sensor's original sampling data. The evidence storage area operates in a cyclic overwrite mode, automatically overwriting the oldest historical record when the storage space is full. This ensures that the water meter continuously retains the latest and most valuable anomaly evidence during long-term operation without frequent manual intervention.

[0059] When a water meter is identified as having a Level 2 anomaly, if the anomaly persists for an extended period or the abnormal behavior is repeated multiple times, a remote alarm response process is triggered in addition to the aforementioned local recording. The main controller assembles the anomaly event information into an alarm communication frame and securely processes the alarm frame using data encryption and rolling code mechanisms to prevent data eavesdropping or replay. Subsequently, the alarm information is uploaded to the water management platform in real time via the wireless communication module. The platform can then combine historical water usage data and regional risk conditions to arrange manual verification or dispatch maintenance personnel for on-site inspections, thereby achieving proactive and controllable escalation of anomaly handling.

[0060] When an anomaly is classified as Level 3, such as a combination of magnetic interference, physical damage, and casing deformation, or when the anomaly duration significantly exceeds the limit, the response action control module executes the highest-level handling procedure. Simultaneously with local recording and remote alarm activation, the main controller sends a shut-off command to the water meter's built-in electric valve drive circuit, driving the electric valve to the closed position and directly cutting off the water supply, physically preventing continuous water theft or malicious damage. After the valve is shut off, a valve status lock flag is simultaneously set, preventing the water meter from automatically restoring water supply without authorization. Restoration requires a remote unlocking command from the water management platform or on-site handling by maintenance personnel, ensuring a clear management loop and secure controllability throughout the anomaly handling process.

[0061] Through the above embodiments, this invention utilizes a sensor monitoring network comprised of magnetoresistive sensors, accelerometers, and fiber optic micro-bending sensors to comprehensively detect various water theft behaviors, such as magnetic interference, physical damage, and casing deformation. This avoids misjudgments caused by single sensors or detection methods, providing reliable and comprehensive foundational data for subsequent anomaly analysis. By fusing multi-source information and using intelligent hierarchical judgment, it accurately identifies the type and severity of anomalies, effectively distinguishing between occasional interference and continuous, malicious water theft, avoiding the crude approach of "closing the valve at the first sign of an anomaly," and improving the accuracy and rationality of anomaly judgment. Based on encryption authentication and rolling code mechanisms, it achieves secure protection of communication data, integrating communication security into the overall anti-theft system. This effectively prevents metering data from being stolen, tampered with, or replayed during transmission, ensuring the authenticity and integrity of the data received by the water management platform, providing a reliable data foundation for anomaly handling and subsequent evidence collection. Through a tiered linkage response mechanism, it achieves differentiated handling from record keeping and remote alarm to valve shut-off, effectively improving the comprehensiveness of ultrasonic water meter anti-theft monitoring and the adaptability of response measures. This invention not only improves the comprehensiveness and accuracy of water meter anti-theft detection, but also significantly enhances the adaptability of abnormal handling and the overall reliability of the system operation.

[0062] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

[0063] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for preventing theft of ultrasonic water meters based on multi-sensor fusion and multi-level response, characterized in that, Includes the following steps: By constructing a sensor monitoring network, multi-source information data of water meters are monitored in real time. Based on the multi-source information data, multi-level anomaly detection is performed in real time and a first-level anomaly confirmation flag is output. The first-level anomaly confirmation flag includes magnetic field anomaly confirmation flag, physical anomaly confirmation flag, and deformation anomaly confirmation flag. An anomaly type code is generated by logical combination operation based on the first-level anomaly confirmation flag, and a quantitative indicator of the anomaly severity is obtained by setting an anomaly timer. A multi-source anomaly fusion judgment model is built based on the anomaly type code and the quantitative indicator of the anomaly severity to determine the anomaly level of the water meter. Obtain the metering data uploaded by the water meter to the water management platform, encrypt the metering data, and embed a rolling code mechanism to defend against replay attacks; Based on the level of water meter anomaly, corresponding response actions are executed to achieve hierarchical and coordinated control for handling water meter anomalies.

2. The ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response according to claim 1, characterized in that, Multi-source information data includes magnetic field strength data, acceleration values, and voltage output values.

3. The ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response according to claim 2, characterized in that, Multi-level anomaly detection is performed in real time based on multi-source information data, and a first-level anomaly confirmation flag is output. Multi-level anomaly detection includes first-level anomaly confirmation detection, second-level anomaly confirmation detection, and third-level anomaly confirmation detection. The first-level anomaly confirmation detection is used to output a magnetic field anomaly confirmation flag, the second-level anomaly confirmation detection is used to output a physical anomaly confirmation flag, and the third-level anomaly confirmation detection is used to output a deformation anomaly confirmation flag.

4. The anti-theft method for ultrasonic water meters based on multi-sensor fusion and multi-level response according to claim 3, characterized in that, The first-layer anomaly confirmation detection is used to output a magnetic field anomaly confirmation flag. The specific steps are as follows: By setting a magnetic field reference register in the main controller, the background magnetic field strength value under the normal operating environment of the water meter is stored. After each sampling, the absolute value of the difference between the current magnetic field strength data and the benchmark value is calculated, and the absolute value of the difference is compared with the preset magnetic field anomaly threshold. When the absolute value of the difference exceeds the magnetic field anomaly threshold, the current sampling result is marked as a magnetic field anomaly sample; When the absolute value of the difference does not exceed the magnetic field anomaly threshold, it is marked as a sample with normal magnetic field. By setting a magnetic field anomaly counter, the number of magnetic field anomaly samples within a continuous sampling period is accumulated. When the accumulated number exceeds a preset threshold for the number of consecutive anomalies, a magnetic field anomaly confirmation flag is generated.

5. The anti-theft method for ultrasonic water meters based on multi-sensor fusion and multi-level response according to claim 3, characterized in that, The second-layer anomaly confirmation detection is used to output physical anomaly confirmation flags, which include vibration anomaly confirmation flags and impact anomaly confirmation flags. The specific steps are as follows: By setting a sliding window buffer in the main controller, the acceleration data sequence within the most recent sampling periods can be stored; The standard deviation of acceleration data within the sampling period is calculated as the vibration intensity index, and the vibration intensity index is compared with the preset vibration anomaly threshold. When the vibration intensity index exceeds the preset vibration anomaly threshold, a vibration anomaly confirmation flag is generated; when the vibration intensity index does not exceed the preset vibration anomaly threshold, a normal vibration sample is generated. By calculating the rate of change between the current acceleration data and the previous acceleration data, an impact anomaly confirmation flag is generated when the rate of change exceeds a preset impact detection threshold. When the rate of change does not exceed the preset impact detection threshold, a normal impact sample is generated.

6. The anti-theft method for ultrasonic water meters based on multi-sensor fusion and multi-level response according to claim 3, characterized in that, The three-layer anomaly confirmation detection is used to output deformation anomaly confirmation flags. The specific steps are as follows: The main controller reads the output voltage value of the photoelectric conversion circuit at a fixed sampling period. The output voltage value reflects the change in optical fiber transmission loss and indirectly characterizes the degree of deformation of the shell. A deformation reference register is set in the main controller to store the photoelectric conversion output reference value when the water meter casing is in a normal state after assembly. After each sampling, the relative rate of change between the current output voltage value and the photoelectric conversion output reference value is calculated, and the relative rate of change is compared with the preset deformation anomaly threshold. When the relative rate of change exceeds the deformation anomaly threshold, a deformation anomaly confirmation flag is generated; when the relative rate of change does not exceed the deformation anomaly threshold, a normal deformation sample is generated.

7. The anti-theft method for ultrasonic water meters based on multi-sensor fusion and multi-level response according to claim 1, characterized in that, The exception type code is generated by performing logical combination operations based on the first-level exception confirmation flag. The specific steps are as follows: An anomaly type encoding register is set to generate an anomaly type code based on the combined state of the first-level anomaly confirmation flags. The anomaly type codes include single magnetic interference anomaly, single magnetic interference anomaly, single shell anomaly, and compound anomaly types. Specifically: when only the magnetic field anomaly confirmation flag is valid, the code indicates a single magnetic interference anomaly; when only the physical anomaly confirmation flag is valid, the code indicates a single physical damage anomaly; when only the deformation anomaly confirmation flag is valid, the code indicates a single shell anomaly; when two or more anomaly confirmation flags are valid simultaneously, the code indicates a compound anomaly type. When any exception type is encoded, a comprehensive exception trigger signal is generated.

8. The ultrasonic water meter anti-theft method based on multi-sensor fusion and multi-level response according to claim 1, characterized in that, The following steps illustrate how to obtain quantitative indicators of anomaly severity by setting an anomaly timer: By setting an abnormal duration timer, the duration of the abnormal state is recorded during the validity period of the comprehensive abnormal trigger signal. By setting an anomaly accumulation counter, the effective number of comprehensive anomaly trigger signals within the monitoring time window is accumulated; The duration of abnormal states and the cumulative number of abnormalities are used as quantitative indicators of the severity of the abnormalities.

9. The anti-theft method for ultrasonic water meters based on multi-sensor fusion and multi-level response according to claim 1, characterized in that, A multi-source anomaly fusion judgment model is built based on anomaly type coding and quantitative indicators of anomaly severity to determine the anomaly level of water meters. The specific steps for building the multi-source anomaly fusion judgment model are as follows: In the main controller, a level determination rule table is set up. The level determination rule table defines the mapping relationship between the abnormality type, the abnormality time threshold, the abnormal cumulative number threshold and the abnormality level. The duration of the anomaly is compared with the first anomaly time threshold in the level determination rule table. When the duration of the anomaly is less than the first anomaly time threshold, it is determined to be a level 1 anomaly; when the duration exceeds the first anomaly time threshold but is less than the second anomaly time threshold, it is determined to be a level 2 anomaly. When the duration exceeds the second abnormal time threshold, or when the abnormal type code indicates a compound abnormality, it is judged as a level three abnormality; The cumulative number of anomalies is compared with the threshold number of cumulative anomalies in the level determination rule table. When the cumulative number of anomalies exceeds the threshold number, the current anomaly level is upgraded by one level, and a water meter anomaly level code is generated. The anomaly type code, water meter anomaly level code, and occurrence timestamp information are packaged to form an anomaly event record.

10. The anti-theft method for ultrasonic water meters based on multi-sensor fusion and multi-level response according to claim 1, characterized in that, Based on the level of water meter anomaly, corresponding response actions are executed to achieve hierarchical linkage control for handling water meter anomalies. The specific steps are as follows: Set up a response action control module in the main controller to trigger the response process according to the exception level code; When the exception level code indicates a level 1 exception, the response action control module executes a level 1 response, which specifically involves recording the response process locally. When the anomaly level code indicates a level 2 anomaly, the response action control module executes a level 2 response. Specifically, the level 2 response is an additional execution of the remote alarm response process on top of the level 1 response. When the anomaly level code indicates a level 3 anomaly, the response action control module executes a level 3 response. Specifically, the level 3 response involves executing a valve shut-off response procedure in addition to the level 1 and level 2 responses.

Images