Wireless digital pressure gauge performance data monitoring management method based on internet of things

Abstract

The application discloses a wireless digital pressure gauge performance data monitoring management method based on Internet of Things, relates to the technical field of pressure gauge technical performance data monitoring, and comprises the following steps: when a wireless digital pressure gauge is in an address abnormal overlap condition, a historical behavior model and a neighborhood behavior joint model are constructed, uploaded performance data are fitted respectively, a first residual error and a second residual error are obtained, and a residual error coupling degree and a coupling degree change trend for representing the transmission mapping confusion degree of the performance data are generated based on the first residual error and the second residual error; and whether the uploaded performance data exist a belonging error problem is judged according to the transmission mapping confusion degree corresponding to the residual error coupling degree and the coupling degree change trend. The application solves the data belonging error problem of the wireless digital pressure gauge of Internet of Things when the address is abnormally overlapped, and realizes intelligent identification, dynamic correction and trend modeling security protection of data belonging.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of pressure gauge performance data monitoring, and particularly relates to a wireless digital pressure gauge performance data monitoring and management method based on the Internet of Things. BACKGROUND

[0002] Wireless digital pressure gauge performance data refers to various dynamic performance index data related to the running state, measurement accuracy, response speed, signal stability, power state, environmental adaptability, etc. of the wireless digital pressure gauge generated during pressure measurement and transmitted wirelessly. These data not only reflect the real-time measurement capability of the pressure gauge, but also reveal the stability and reliability of the device itself in the actual working environment. Since traditional pressure gauges rely on manual inspection and local reading, they are not only inefficient and prone to errors, but also difficult to achieve centralized management and real-time response in large-scale industrial applications. Therefore, monitoring and managing the performance data of wireless digital pressure gauges based on the Internet of Things technology can achieve remote automatic data collection, real-time uploading and centralized analysis through a sensing network and a communication module, thereby greatly improving the transparency and controllability of device operation. The Internet of Things platform can uniformly manage, trend analyze, abnormally warn and maintain and dispatch the performance data of a large number of devices, effectively reducing operation and maintenance costs, improving system security and management efficiency, and being particularly suitable for industrial automation, energy management, smart city and other scenarios with high data monitoring requirements.

[0003] The existing wireless digital pressure gauge performance data monitoring and management technology based on the Internet of Things mainly integrates a pressure sensor, a data acquisition unit, a wireless communication module and a low-power processing chip in the wireless digital pressure gauge to realize real-time collection and local preprocessing of pressure data and device running state, and then upload the performance data to the cloud or edge server for centralized processing and management through Internet of Things communication methods (such as NB-IoT, LoRa, 4G / 5G or Wi-Fi). This technology usually includes several key links: first, the front-end sensing layer, composed of multiple wireless digital pressure gauges arranged at different measurement points, is used to collect pressure values and device running parameters in real time; second, the network transmission layer is responsible for transmitting the collected data to the background system through a stable and reliable Internet of Things communication link; third, the platform processing layer relies on the Internet of Things management platform for data reception, analysis, storage, statistical analysis and visualization display; fourth, the application management layer realizes intelligent monitoring and efficient management of device performance through threshold value judgment configuration, algorithm model analysis, trend warning and operation and maintenance scheduling. The entire process supports distributed device access, data encryption transmission, remote diagnosis and remote upgrading, forming an automated, highly real-time and scalable pressure gauge performance data management closed-loop system, effectively improving the overall reliability and intelligence level of the system.

[0004] The existing technology has the following deficiencies:

[0005] In the multi-point deployment process of the Internet of Things wireless digital pressure gauge system, if multiple wireless digital pressure gauges are installed in adjacent pressure measurement positions (such as between the main pipeline and the branch pipeline) in the same fluid network, and use a broadcast type communication protocol (such as LoRa) to cooperate with an automatic addressing mechanism for data reporting, it is possible that in the case that the devices experience re-online, address erasure or ID re-flashing, the device communication identifier (such as the logical ID or MAC address) will abnormally overlap, and then the background system will mistakenly attribute the performance data from different physical positions as the data source of the same logical measurement point; Since these devices are originally distributed in different pressure characteristic areas, their true pressure measurement behaviors differ, so after data merging, the platform will present "inconsistent with fluid physical laws" pressure fluctuations, such as frequent sudden rises and falls or periodic conflict waveforms, and the existing performance data monitoring and management technology based on the Internet of Things wireless digital pressure gauge cannot determine whether the current data has an attribution error problem according to the transmission mapping confusion degree of the performance data of the wireless digital pressure gauge in the device address abnormal overlap situation, resulting in the background continuing to use the error data for trend analysis, anomaly identification and strategy execution, which ultimately leads to misjudgment of device failure, false alarm of normal state, mislearning of trend model, and thus affects the decision accuracy and operation efficiency of the entire system.

[0006] The above information disclosed in the background section is only intended to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute the prior art known to those of ordinary skill in the art. SUMMARY

[0007] The purpose of the present application is to provide a performance data monitoring and management method for wireless digital pressure gauges based on the Internet of Things, to solve the problems in the background.

[0008] In order to achieve the above-mentioned purpose, the present application provides the following technical solution: a performance data monitoring and management method for wireless digital pressure gauges based on the Internet of Things, specifically comprising the following steps:

[0009] Obtain the pressure disturbance response data of the wireless digital pressure gauge in the initial static fluid environment, construct the behavior label based on the pressure disturbance response data, and bind the behavior label with the wireless digital pressure gauge;

[0010] After uploading the performance data on the wireless digital pressure gauge, extract the pressure disturbance response time sequence and compare it with the response time of the adjacent measurement point, and combine the physical topological relationship between the measurement points to determine whether the wireless digital pressure gauge is in the device address abnormal overlap situation;

[0011] When the wireless digital pressure gauge is in the device address abnormal overlap situation, a historical behavior model and a neighborhood behavior joint model are constructed, the uploaded performance data is fitted respectively to obtain a first residual and a second residual, and a residual coupling degree and a coupling degree change trend for representing the transmission mapping confusion degree of the performance data are generated based on the first residual and the second residual;

[0012] According to the transmission mapping confusion degree corresponding to the residual coupling degree and the coupling degree change trend, it is judged whether the uploaded performance data has a belonging error problem;

[0013] For the performance data with the belonging error problem, the performance data is divided into continuous time segments, the behavior characteristics of each time segment and the historical data of a preset number of measuring points are matched, and the target belonging measuring point of the performance data is determined through belonging voting according to the matching result, and the belonging adjustment operation is completed;

[0014] After the belonging adjustment is completed, the residual coupling degree, the coupling degree change trend, the judgment result and the target belonging measuring point form a belonging judgment track, and intervention processing is performed in the belonging error problem concentrated area, including reconstructing the behavior label, enabling the joint verification process of the behavior characteristics and the measuring point position, and limiting the uploaded data in this area to participate in the trend modeling.

[0015] Preferably, after the wireless digital pressure gauge uploads the performance data, the pressure disturbance response time sequence is extracted, and compared with the response time of the adjacent measuring point, and the physical topological relationship between the measuring points is combined to judge whether the wireless digital pressure gauge is in the device address abnormal overlap situation, specifically:

[0016] After the wireless digital pressure gauge uploads the performance data, the pressure disturbance response time sequence is extracted, which is composed of continuous pressure change data between the disturbance trigger starting point and the pressure value entering the stable interval;

[0017] The disturbance response time of the wireless digital pressure gauge is obtained by calculating the time interval between the time when the pressure change rate in the pressure disturbance response time sequence reaches the maximum value and the disturbance trigger starting point;

[0018] From the same data synchronization period, a preset number of adjacent measuring points are selected, and the pressure disturbance response time sequence of each measuring point is extracted, and the disturbance response time set of the adjacent measuring points is obtained in the same way;

[0019] According to the physical topological structure information between the wireless digital pressure gauge and each adjacent measuring point, the corresponding physical connection path and pipeline distance data are obtained, and the theoretical disturbance propagation time from each adjacent measuring point to the wireless digital pressure gauge is calculated combined with the preset fluid medium propagation speed;

[0020] Based on the difference between the response time of each measuring point in the set of disturbance response times of neighboring measuring points and its theoretical propagation time, the average propagation offset time of the disturbance under the current topology is calculated, and the theoretical response time delay range is constructed.

[0021] The disturbance response time of the wireless digital pressure gauge is compared with the theoretical response time delay range. When the response time exceeds the theoretical range boundary threshold, it is determined that the wireless digital pressure gauge is in a state of abnormal device address overlap.

[0022] Preferably, when the wireless digital pressure gauge is in a situation of abnormally overlapping device addresses, a historical behavior model is constructed based on the pressure disturbance response time series uploaded by the wireless digital pressure gauge in the historical period, and a neighborhood behavior joint model is constructed based on the pressure disturbance response time series of a preset number of neighboring measuring points with physical topological connections to the wireless digital pressure gauge. The uploaded performance data is used as input to fit the historical behavior model and the neighborhood behavior joint model to obtain the first residual and the second residual. By calculating the average difference and trend difference between the first residual and the second residual, the residual coupling degree and the coupling degree change trend are generated to characterize the degree of disorder in the transmission mapping of performance data.

[0023] Preferably, when the wireless digital pressure gauge is in a situation of abnormally overlapping device addresses, a historical behavior model is constructed based on the pressure disturbance response time series uploaded by the wireless digital pressure gauge in the historical period, specifically as follows:

[0024] Select a preset number of historical period data on continuous pressure changes from the start of the disturbance to when the pressure value enters a stable range.

[0025] Based on the continuous pressure change data, the pressure change rate per unit time is calculated to form a pressure change rate sequence. The pressure change rate sequence is clustered using a time series similarity algorithm, and a disturbance response behavior feature sequence containing pressure change rate samples with a preset time length is extracted.

[0026] A piecewise linear function model is constructed by fitting a sequence of perturbation response behavior characteristics. Combination weights are set according to the perturbation response amplitude and duration corresponding to each piecewise linear function model. The piecewise linear function models are then combined in a weighted manner to construct a historical behavior model that describes the perturbation response behavior of the wireless digital pressure gauge.

[0027] Preferably, a joint neighborhood behavior model is constructed based on the pressure disturbance response time series of a predetermined number of neighboring measuring points that have a physical topological connection with the wireless digital pressure gauge, specifically as follows:

[0028] Based on the physical pipeline connection diagram between the wireless digital pressure gauge and other measuring points, obtain the physical connection path and distance information, and filter out the nearby measuring points whose physical distance does not exceed the set distance threshold.

[0029] Continuous pressure change data of each neighboring measuring point from the start of disturbance to the stable pressure range within multiple historical periods are selected. Each data segment is normalized according to a uniform time sampling interval and length. The pressure change rate sequence is obtained by calculating the unit time difference of the continuous pressure data.

[0030] In each pressure change rate sequence, the maximum pressure change rate, the duration of the change rate, and the time location where the change rate occurs are calculated. A triplet containing the values ​​is constructed as the feature vector of the measuring point. The feature vectors of all neighboring measuring points are combined into a joint feature vector by weighted average according to distance. Based on the joint feature vector, a joint neighborhood behavior model is constructed through multi-segment linear fitting.

[0031] Preferably, the uploaded performance data is used as input to fit the historical behavior model and the neighborhood behavior joint model to obtain the first residual and the second residual. The average difference and trend difference between the first residual and the second residual are calculated to generate the residual coupling degree and the coupling degree change trend, which are used to characterize the degree of disorder in the transmission mapping of performance data. Specifically:

[0032] Using the pressure disturbance response time series from the uploaded performance data as input, and substituting them into the historical behavior model and the neighborhood behavior joint model corresponding to the wireless digital pressure gauge, respectively, at each unified sampling time point, the difference between the model prediction value and the actual pressure value is calculated to form the first residual sequence and the second residual sequence, respectively.

[0033] Based on the two residual sequences, the difference is calculated for the residual values ​​of each corresponding sampling point to form a residual difference sequence;

[0034] The average of all differences in the residual difference sequence is used to obtain the average difference that reflects the overall deviation of the fit.

[0035] Local linear regression is performed on the residual difference sequence using a sliding window with a fixed time length. The fitting slope within each window is extracted, and the standard deviation of all window slope values ​​is calculated to obtain the trend difference value used to characterize the trend of the fitting difference.

[0036] The average difference and trend difference are respectively subjected to min-max normalization to form residual coupling degree and coupling degree change trend, which are used to characterize the degree of transmission mapping disorder of the upload performance data under the condition of abnormal address overlap. Specifically: when both residual coupling degree and coupling degree change trend are in the preset high value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is high; when both residual coupling degree and coupling degree change trend are in the preset low value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is low; when neither residual coupling degree nor coupling degree change trend is in the preset high value range nor in the preset low value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is moderate.

[0037] Preferably, based on the degree of transmission mapping disorder corresponding to the residual coupling degree and the trend of coupling degree change, it is determined whether there is an attribution error in the uploaded performance data, specifically:

[0038] When the transmission mapping disorder level is high, it is determined that there is an attribution error in the current upload performance data;

[0039] When the degree of transmission mapping disorder is moderate, combined with the historical fitting error average level of the performance data in the current measurement point behavior model, if the first residual of the performance data at the current measurement point exceeds the preset tolerance range of the historical fitting error average level, it is determined that the performance data has an assignment error problem; otherwise, it is determined that the performance data does not have an assignment error problem.

[0040] When the level of transport mapping disorder is low, it is determined that there is no attribution error in the current upload performance data.

[0041] Preferably, for performance data with misclassification issues, the performance data is divided into continuous time segments according to a fixed time window length. The pressure disturbance response time series within each time segment is extracted, and its pressure change rate per unit time is calculated to form a pressure change rate sequence. This pressure change rate sequence is then matched with the historical pressure change rate sample sequence generated by the pressure disturbance response time series of a preset number of measuring points in the historical period of the same length. A similarity scoring algorithm based on dynamic time warping (DTW) is used to calculate the matching score for each measuring point. Based on the matching scores of all measuring points, a weighted vote is performed, and the measuring point with the highest number of votes is selected as the target measuring point for the current time segment. The frequency of the measuring points for all time segments is then statistically analyzed, and the measuring point with the highest frequency is finally determined as the target measuring point for the performance data, thus completing the classification adjustment operation.

[0042] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0043] 1. This invention, by constructing a closed-loop mechanism of "disturbance response behavior labeling + fitting residual coupling analysis + attribution behavior optimization," achieves for the first time automatic identification and correction of data attribution errors caused by abnormal overlap of device communication addresses in an IoT wireless digital pressure gauge system. Traditional IoT platforms often rely on communication identifiers (such as logical IDs and MAC addresses) as unique identifiers for measurement points, which are prone to conflicts during device re-connection or automatic addressing. This solution, however, uses "disturbance response behavior" as a dynamic identification criterion independent of communication identifiers, effectively solving the data mismatch problem caused by incorrect attribution of communication identifiers, and significantly improving the platform's data identification robustness and fault tolerance.

[0044] 2. This invention introduces a dual fitting mechanism of "historical behavior model and neighborhood behavior joint model". Through the difference sequence analysis and trend detection of the first residual and the second residual, a coupling degree index is constructed to quantify the "degree of disorder in performance data transmission mapping". This enables probabilistic early warning and quantitative identification of abnormal data attribution before the attribution error occurs, effectively avoiding the use of erroneous data for highly sensitive operations such as trend learning and fault judgment, improving the overall prediction accuracy and stability of the system, and possessing highly intelligent data governance capabilities.

[0045] 3. This invention employs a strategy of "attribution voting + behavioral feature matching + determination trajectory recording." After identifying data attribution errors, this solution automatically reassigns the target attribution of measurement points and executes intervention and optimization measures in areas with concentrated attribution errors, such as reconstructing behavioral labels, jointly verifying measurement point locations, and isolating data involved in modeling. This achieves closed-loop management of the entire process from identification and correction to self-optimization. Compared to traditional static configuration or manual intervention methods, this method not only responds more promptly and processes more accurately but also significantly improves the system's autonomous evolution capability and operational efficiency during long-term operation. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0047] Figure 1 This is a flowchart illustrating the IoT-based wireless digital pressure gauge performance data monitoring and management method of the present invention. Detailed Implementation

[0048] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0049] This invention provides, for example Figure 1 The IoT-based wireless digital pressure gauge performance data monitoring and management method shown includes the following steps:

[0050] Acquire pressure disturbance response data of wireless digital pressure gauges in static fluid environments during the initial deployment phase, construct behavioral tags based on this pressure disturbance response data, and bind the behavioral tags to the wireless digital pressure gauges;

[0051] After the wireless digital pressure gauge is physically installed, the platform sends a remote trigger command to induce a controllable, short-term disturbance in the fluid pipeline containing the pressure gauge. This can be achieved by slightly opening or closing a controllable valve or by applying a momentary load change, inducing a pressure disturbance response within the pipeline. During the disturbance, the system continuously collects pressure data from the wireless digital pressure gauge at a high sampling frequency, forming a complete disturbance response data sequence. To ensure that the environment is in a static fluid condition during data acquisition, the platform simultaneously monitors nearby measuring points. When the pressure change at other measuring points within a set time window is lower than a preset threshold, the current environment is considered to meet static conditions, and the disturbance response data is marked as initial deployment pressure disturbance response data.

[0052] After collecting pressure disturbance response data, the platform analyzes the data sequence using a feature extraction algorithm, extracting several key parameters including the starting position of the disturbance response, the maximum magnitude of pressure change, the duration of stabilization, the rate of change of waveform slope, and the attenuation trend. These parameters are then merged into multi-dimensional feature vectors. Subsequently, the platform calls a feature encoding function to encode and transform these feature vectors, generating structurally unique behavioral tags. These behavioral tags are then bound to the device identifier of the wireless digital pressure gauge. The binding operation is completed by writing the behavioral tag index into the device metadata database, ensuring that subsequently uploaded performance data can be compared with the tag's behavior, thereby establishing a behavioral reference relationship for data attribution.

[0053] Through the above operations, even in scenarios where device communication identifiers undergo abnormal changes or overlaps, data attribution can still be identified based on the behavioral characteristics of the wireless digital pressure gauge, thereby effectively improving the accuracy and robustness of data parsing. In IoT environments, phenomena such as address rescrambling, ID reuse, and communication conflicts may lead to performance data being incorrectly mapped to incorrect measurement points, thus interfering with the system's identification and decision-making regarding device status. By collecting disturbance response data from the initial deployment phase to establish behavioral tags, and completing tag binding and invocation in software, a behavioral identification mechanism independent of communication identifiers can be provided to the platform. This ensures a dynamic behavioral baseline for subsequent data attribution determination, facilitating accurate identification of attribution errors and the execution of attribution adjustment operations.

[0054] After the wireless digital pressure gauge uploads performance data, the pressure disturbance response time series is extracted and compared with the response time of neighboring measuring points. Combined with the physical topology relationship between the measuring points, it is determined whether the wireless digital pressure gauge is in a situation of abnormal device address overlap.

[0055] In this embodiment, after the wireless digital pressure gauge uploads performance data, the pressure disturbance response time series is extracted and compared with the response time of neighboring measuring points. Based on the physical topology relationship between the measuring points, it is determined whether the wireless digital pressure gauge is experiencing abnormal address overlap. Specifically:

[0056] After the performance data is uploaded by the wireless digital pressure gauge, the pressure disturbance response time series is extracted. This pressure disturbance response time series consists of continuous pressure change data from the start of the disturbance to the point where the pressure value enters the stable range.

[0057] After the wireless digital pressure gauge uploads performance data, the disturbance trigger point is located in the uploaded data stream. The complete response process from the disturbance trigger point to the pressure value stabilizing is then extracted to construct a pressure disturbance response time series. This process first identifies the location in the time series where the pressure slope first continuously exceeds the threshold by setting a disturbance detection threshold (such as a pressure change rate threshold or a small abrupt change threshold). The system then continuously tracks the pressure curve changes after this point and determines whether the change amplitude within several consecutive time windows is lower than the set stability interval threshold. If the stability criteria are met within consecutive time windows, the pressure is considered to have entered a stable interval, and this time point is recorded as the end point of the sequence. Finally, the pressure sampling data from the start point to the end point is extracted in chronological order to form the pressure disturbance response time series. For example, if a wireless digital pressure gauge with a sampling frequency of 1Hz detects a rapid increase in pressure from 0.85 MPa to 0.96 MPa at the 120th second within a certain period, and then gradually approaches 0.92 MPa with a fluctuation amplitude of less than 0.002 MPa over the following 50 seconds, the system can determine that the 120th second is the start of the disturbance and the 170th second is the start of the stable interval. The pressure value sequence within this 50-second period can be extracted as a complete pressure disturbance response time series. This method allows for structured segmentation of the raw performance data at the software level, providing an accurate dynamic behavioral basis for subsequent response behavior analysis and attribution determination.

[0058] The disturbance response time of the wireless digital pressure gauge is obtained by calculating the time interval between the moment when the pressure change rate reaches its maximum value and the start of the disturbance in the pressure disturbance response time series.

[0059] After obtaining the pressure disturbance response time series, differential calculation and rate analysis are performed on the time series to extract the time point when the pressure change rate reaches its maximum value. Combined with the disturbance triggering point location, the disturbance response time of the wireless digital pressure gauge is calculated. Specifically, the system first performs first-order derivative calculation on the disturbance response time series, that is, differential processing is performed on adjacent pressure sampling points using a sliding time window to obtain the pressure change rate sequence per unit time. Next, the position with the largest absolute rate value is located in this rate sequence; the original time point corresponding to this position is the strongest response moment during the disturbance process. Then, the time difference between this time point and the initial disturbance triggering point is calculated to obtain the disturbance response time. For example, if the starting point of the disturbance response time series is 120 seconds, and after differential analysis it is found that the pressure change rate reaches its maximum value (e.g., 0.007 MPa / s) at 125 seconds, then 125 seconds can be determined as the maximum response moment. Finally, subtracting 120 from 125 yields a disturbance response time of 5 seconds. This method can be implemented in software using standard algorithms such as data sliding window, difference array, and extreme value extraction. It has the characteristics of versatility and high accuracy, and can provide accurate behavioral feature parameters for subsequent construction of topological response models and determination of attribution errors.

[0060] Within the same data synchronization period, a predetermined number of neighboring measuring points are selected, and the pressure disturbance response time series of each measuring point is extracted. The set of disturbance response times of neighboring measuring points is then calculated in the same way.

[0061] Within the same data synchronization period, a predetermined number of neighboring measuring points are selected and their pressure disturbance response time series are extracted. This can be achieved by setting physical proximity relationships and time synchronization windows in the IoT platform via software, combined with pressure change feature identification methods, to realize dynamic behavior extraction and response time calculation for multiple measuring points. Specifically, the platform first selects a predetermined number of measuring points as a set of neighboring measuring points from those directly connected to the target wireless digital pressure gauge via pipelines or within a physical distance threshold, based on the topology configuration table or deployment coordinate system. Then, within the synchronization time window (e.g., ±2 seconds) when the target device detects a disturbance event, continuous pressure data segments corresponding to the disturbance initiation point are extracted from the performance data of these measuring points. For each neighboring measuring point, the disturbance response time is calculated using the same processing method as the target device—identifying the disturbance trigger point, extracting the response time series, and calculating the time difference between the maximum pressure change rate and the trigger point. Finally, these disturbance response times are combined into a response time set. For example, if the target device exhibits a disturbance response at 120 seconds, the system will simultaneously extract data from its three adjacent measurement points within a window from 118 to 122 seconds. The same disturbance time extraction logic as for the target device will be applied to each measurement point, yielding response times such as 5.2 seconds, 4.8 seconds, and 5.5 seconds, respectively. These results will be combined to form a set of disturbance response times for neighboring measurement points. This set will be used for subsequent comparative analysis with the theoretical propagation model, helping to determine whether the target device has exhibited incorrect data attribution or abnormal response behavior.

[0062] Based on the physical topology information between the wireless digital pressure gauge and each adjacent measuring point, the corresponding physical connection path and pipeline distance data are obtained, and combined with the preset fluid medium propagation speed, the theoretical disturbance propagation time from each adjacent measuring point to the wireless digital pressure gauge is calculated.

[0063] Based on the physical topology information between the wireless digital pressure gauge and each neighboring measuring point, the physical connection path and pipeline distance data are obtained. Combined with the preset fluid medium propagation speed, the theoretical disturbance propagation time is calculated. This can be achieved by constructing a topology model of the fluid network in the IoT platform and introducing a path weight calculation method. Specifically, during the deployment or initialization phase, the physical location information and pipeline connection relationship of each measuring point are entered into the topology database, and a graphical connection structure between measuring points is established. Each side represents a pipeline segment, with the pipeline length included as a path weight. After the target wireless digital pressure gauge triggers a disturbance, the system performs a shortest path search (such as Dijkstra's algorithm or A* algorithm) on each neighboring measuring point in the topology map to obtain the physical connection path from that neighboring measuring point to the target measuring point and its cumulative pipeline length. Then, combined with the pre-configured fluid medium type (such as water or gas) and its propagation speed parameters in that environment (e.g., the shock wave speed of water in a steel pipe is approximately 1400 m / s), the theoretical disturbance propagation time from the neighboring measuring point to the target wireless digital pressure gauge is calculated using the formula "propagation time = total pipeline length / propagation speed". To illustrate with a concrete example, if the total length of the connection path between a nearby measuring point and the target device is 21 meters, the current medium is water, and the propagation speed is set to 1400 meters per second, then the theoretical propagation time is 0.015 seconds. Through this software modeling and path calculation method, the theoretical propagation time can be accurately estimated quickly and efficiently without additional hardware intervention, providing a foundation for subsequent response behavior comparisons and address anomaly detection.

[0064] Based on the difference between the response time of each measuring point in the set of disturbance response times of neighboring measuring points and its theoretical propagation time, the average propagation offset time of the disturbance under the current topology is calculated, and the theoretical response time delay range is constructed.

[0065] Based on the difference between the response time of each measuring point in the set of disturbance response times of neighboring measuring points and its theoretical propagation time, the average propagation offset time of the disturbance under the current topology is calculated, and the theoretical response time delay range is constructed. This can be achieved by statistically analyzing and modeling the response error of multiple measuring points using software. In specific operation, firstly, for each neighboring measuring point, the difference between its measured disturbance response time and its corresponding theoretical disturbance propagation time is calculated to obtain the propagation offset time of that measuring point; then, the propagation offset times of all neighboring measuring points are summarized, and the average propagation offset time under the current topology is obtained by using the mean or weighted mean calculation method. This value can be regarded as the overall dynamic offset characteristic of the disturbance response in the local fluid system; on this basis, the standard deviation or maximum fluctuation amplitude of all offset times is further calculated to characterize the degree of uncertainty of the propagation offset. Finally, the theoretical response time delay range is constructed using "average propagation offset time ± tolerance range" as a reference interval for judging the rationality of the target device's response time. For example, if three neighboring measurement points have response times of 5.3 seconds, 4.9 seconds, and 5.6 seconds, respectively, and their theoretical propagation times are 5.0 seconds, 4.8 seconds, and 5.2 seconds, then the offset times are 0.3 seconds, 0.1 seconds, and 0.4 seconds, respectively, with an average offset time of 0.266 seconds and a standard deviation of approximately 0.124 seconds. The theoretical response time delay range can be set as "theoretical propagation time + 0.266 seconds ± 0.15 seconds". This method can comprehensively consider neighborhood behavior offset characteristics to construct an adaptive judgment interval, improving the robustness and accuracy of address anomaly identification.

[0066] The disturbance response time of the wireless digital pressure gauge is compared with the theoretical response time delay range. When the response time exceeds the theoretical range boundary threshold, it is determined that the wireless digital pressure gauge is in a state of abnormal device address overlap.

[0067] The system compares the disturbance response time of a wireless digital pressure gauge with its theoretical response time delay range to determine whether it is experiencing abnormal address overlap. This comparison and anomaly identification can be achieved through an integrated interval judgment mechanism in the software. Specifically, after constructing the theoretical response time delay range, the system uses this range as a reference interval for the disturbance response behavior of the target measurement point. For example, it generates a dynamic judgment threshold interval in the form of "theoretical propagation time + average offset time ± tolerance". Subsequently, the system calls the previously calculated actual disturbance response time of the wireless digital pressure gauge and compares it with the upper and lower boundaries of this interval. If the response time is earlier than the lower boundary or later than the upper boundary, it is determined that the behavior response does not conform to the reasonable timing derived from the topology and propagation law, thus marking the device as having a risk of abnormal address overlap. For example, if the system constructs a theoretical response time delay range of "5.0 seconds to 5.4 seconds" based on the behavior of neighboring measuring points, while the disturbance response time of the target wireless digital pressure gauge is 4.6 seconds or 5.8 seconds, then it significantly deviates from this reasonable range. This indicates that the uploaded data did not arrive in the physical logical propagation order, and it is highly likely that data attribution errors have occurred due to overlapping device addresses. This judgment process can be embedded into the data pre-processing workflow of the IoT platform for automatic execution. It is not only real-time but also does not rely on manual review, making it suitable for dynamic anomaly identification and tag correction needs under large-scale deployment.

[0068] When the wireless digital pressure gauge is in a situation of abnormal device address overlap, a joint model of historical behavior and neighborhood behavior is constructed, and the uploaded performance data is fitted to obtain the first residual and the second residual. Based on the first residual and the second residual, residual coupling degree and coupling degree change trend are generated to characterize the degree of transmission mapping disorder of performance data.

[0069] In this embodiment, when the wireless digital pressure gauge is in a state of abnormal device address overlap, a historical behavior model is constructed based on the pressure disturbance response time series uploaded by the wireless digital pressure gauge in the historical period, and a neighborhood behavior joint model is constructed based on the pressure disturbance response time series of a preset number of neighboring measuring points with physical topological connections to the wireless digital pressure gauge. The uploaded performance data is used as input to fit the historical behavior model and the neighborhood behavior joint model to obtain the first residual and the second residual. By calculating the average difference and trend difference between the first residual and the second residual, the residual coupling degree and the coupling degree change trend are generated to characterize the degree of disorder in the transmission mapping of performance data.

[0070] In this embodiment, when the wireless digital pressure gauge is in a situation of abnormally overlapping device addresses, a historical behavior model is constructed based on the pressure disturbance response time series uploaded by the wireless digital pressure gauge in the historical period, specifically as follows:

[0071] Select a preset number of historical period data on continuous pressure changes from the start of the disturbance to when the pressure value enters a stable range.

[0072] To select a preset number of historical periods containing continuous pressure change data from the start of a disturbance to the point where the pressure value enters a stable range, the system can segment historical data records from a wireless digital pressure gauge into periods and identify disturbance events, then extract valid data segments using a stability assessment mechanism. First, the system identifies the start time of the disturbance in each historical period based on known disturbance triggering mechanisms or periodic external event markers (such as valve opening, pump start-up, and other control signals). Then, it scans the pressure data sequence backward from that time point, using set stability assessment rules (e.g., pressure fluctuations below a set threshold within several consecutive sampling points, and the gradient rate of change close to zero) to detect the starting point of the pressure stabilization range. Based on this, the system extracts the complete pressure change data from the disturbance start point to the stable start point as the disturbance response segment for the current period. During continuous operation, the system can select a preset number of periods that meet data integrity and accuracy requirements according to time windows or event labels for processing, and standardizes and manages the data segments corresponding to each period for subsequent behavioral modeling and analysis. In this way, the system can accurately extract structurally consistent, temporally clear, and physically complete disturbance response fragments from large-scale datasets, thereby enabling the preparation of basic data on the historical behavior of wireless digital pressure gauges.

[0073] Based on the continuous pressure change data, the pressure change rate per unit time is calculated to form a pressure change rate sequence. The pressure change rate sequence is clustered using a time series similarity algorithm, and a disturbance response behavior feature sequence containing pressure change rate samples with a preset time length is extracted.

[0074] The calculation of the pressure change rate per unit time based on continuous pressure variation data and the construction of a pressure change rate sequence can be achieved through differential calculation. Specifically, the system first performs sampling point time normalization on the disturbance response segment extracted in each cycle, then performs difference calculation on the pressure values ​​between adjacent sampling points and divides by the corresponding time interval to obtain the pressure change rate per unit time. This sequence reflects the response speed and trend during the disturbance process, and better reflects the essential characteristics of the disturbance. Subsequently, based on a sliding window mechanism, the system divides each pressure change rate sequence into a group of time segments of equal length as pressure change rate samples. To identify typical disturbance response patterns, the system uses a time series similarity algorithm to cluster all samples and extracts the samples that best represent the structural characteristics of each cluster as the disturbance response behavior feature sequence. For example, in historical data from a wireless digital pressure gauge containing 30 cycles, the system can generate 30 pressure change rate sequences. Each sequence is divided into several 1-second segments using a sliding window. After clustering, it is found that they can be classified into three main response modes: rapid rise-slow stability, slow rise-rapid stability, and linear smooth transition. A representative sample with the minimum distance center is extracted from each class, which constitutes the final behavioral feature sequence used for modeling.

[0075] Time series similarity algorithms are used to measure the structural similarity of two sequences along the time axis. Common methods include Dynamic Time Warping (DTW), Euclidean Distance, and Edit Distance on Real Sequence (EDR). DTW, in particular, is an algorithm that allows for non-linear alignment along the time dimension. It enables "flexible" comparisons of sequences with inconsistent lengths or time shifts, making it especially suitable for processing perturbation response data. In this scenario, DTW can automatically match pressure change rate sequences from two different periods along the time axis, allowing for the identification of structural similarity even when two perturbation responses have different response durations. For example, if a rapid pressure increase lasts only 0.8 seconds in one period and 1.2 seconds in another, DTW will automatically "align" their structural changes along the time dimension, thus more accurately assessing their behavioral similarity. This method provides a robust similarity metric foundation for identifying pressure perturbation patterns and is a crucial step in behavioral feature extraction.

[0076] A piecewise linear function model is constructed by fitting a sequence of perturbation response behavior characteristics. Combination weights are set according to the perturbation response amplitude and duration corresponding to each piecewise linear function model. The piecewise linear function models are then combined in a weighted manner to construct a historical behavior model that describes the perturbation response behavior of the wireless digital pressure gauge.

[0077] The process of constructing a piecewise linear function model based on the characteristic sequence of each disturbance response behavior and generating a historical behavior model accordingly can be accomplished by analyzing the changing trends and identifying key inflection points in the characteristic sequences. Specifically, the system first applies first-order derivative analysis to each disturbance response behavior characteristic sequence to identify significant inflection points in the rate of pressure change. Based on these inflection points, the entire sequence is divided into several intervals with a single trend, such as a continuously rising segment, a stable transition segment, or a slowly falling segment. The data for each interval is fitted to a linear function of the form P(t) = at + b, where a represents the rate of pressure change within that segment, t is time, and b is the initial pressure value. During the fitting process, the least squares method can be used to solve for the optimal linear parameters to minimize the fitting error. Next, the system calculates the contribution of each segment's pressure change amplitude and duration to the overall behavior pattern and sets normalized weights. All piecewise linear functions are weighted and combined according to their weights to form a set of continuous function sequences with different slope characteristics, which serve as the historical behavior model of the wireless digital pressure gauge. For example, if a certain feature sequence shows three response characteristics in succession: a rapid increase (0–0.4s), a slow increase (0.4–1.0s), and a stable level (1.0–2.0s), then each segment is fitted as a linear function and then weighted and combined to reflect the typical disturbance behavior process of the measuring point as a whole.

[0078] Piecewise linear function models are mathematical models that approximate nonlinear or multi-stage changes as a series of continuous linear segments. They are suitable for describing pressure response trajectories exhibiting different trends during disturbance responses. Each piecewise model approximates the actual data using a linear function within a local time interval, significantly reducing fitting errors and improving modeling accuracy. For example, in a pressure response sequence, the 0–0.3 second segment shows a rapid linear increase, which can be fitted as P1(t) = 2.5t + 0.8; the 0.3–0.9 second segment shows a slow pressure change, fitted as P2(t) = 0.6t + 1.4; and the 0.9–1.5 second segment tends to stabilize, fitted as an approximate horizontal line P3(t) = 1.8. This structured representation preserves the stages of the pressure response process and facilitates comparison and combination of models across different cycles, making it an important tool for achieving behavioral normalization modeling.

[0079] In this embodiment, a joint neighborhood behavior model is constructed based on the pressure disturbance response time series of a predetermined number of neighboring measuring points that have a physical topological connection with the wireless digital pressure gauge. Specifically:

[0080] Based on the physical pipeline connection diagram between the wireless digital pressure gauge and other measuring points, obtain the physical connection path and distance information, and filter out the nearby measuring points whose physical distance does not exceed the set distance threshold.

[0081] The physical connection paths and distances between wireless digital pressure gauges and other measuring points can be obtained by constructing a topology graph database. Then, a graph theory path search algorithm is used to filter out neighboring measuring points whose physical distance does not exceed a set threshold. Specifically, the system first presets a complete pipeline topology graph. This graph uses nodes to represent wireless digital pressure gauges or other measuring points, edges to represent pipeline connections, and assigns a weight to each edge based on the actual pipeline length. The system loads this topology graph into a data structure in the form of an adjacency matrix or adjacency list. During actual operation, the system uses the current target wireless digital pressure gauge node as the starting point and employs a weighted shortest path algorithm (such as Dijkstra's algorithm or A* algorithm) to calculate the shortest physical path between it and all other measuring point nodes, accumulating the lengths of all pipe segments in the path to obtain the actual physical distance. The system then marks all measuring points whose distance does not exceed a set threshold (e.g., 50 meters or 100 meters) as neighboring measuring points and records the corresponding path structure for subsequent disturbance propagation time calculations. For example, if measuring point A is connected to target pressure gauge B via three pipe segments with a total length of 47 meters, and the path in the topology map is A→X→Y→B, then A will be included in the set of neighboring measuring points. In this way, the system can accurately and automatically identify the neighboring areas with physical topological associations to a certain measuring point in a complex large-scale pipeline network environment, providing accurate spatial boundaries for subsequent disturbance behavior correlation analysis.

[0082] Continuous pressure change data of each neighboring measuring point from the start of disturbance to the stable pressure range within multiple historical periods are selected. Each data segment is normalized according to a uniform time sampling interval and length. The pressure change rate sequence is obtained by calculating the unit time difference of the continuous pressure data.

[0083] To select disturbance response data from each neighboring measuring point across multiple historical periods and generate a pressure change rate sequence in a unified format, the system first establishes a historical data caching system at the measuring point level within the software. This, combined with external disturbance event logs or an internal pressure mutation detection algorithm, locates the disturbance trigger point within each historical period. Then, starting from the disturbance trigger point, the system continuously scans the pressure change data backward until the pressure curve enters a stable range (i.e., the numerical fluctuation is below a stable threshold within a specified window). The complete pressure change sequence within this range is then extracted as a single-period response data segment. To standardize data analysis, the system performs time alignment and length normalization on each extracted response data segment. This involves using linear interpolation or resampling algorithms to map data with different numbers of sampling points across different periods to a standard format with the same time length and sampling interval. Finally, the system performs unit-time difference operations on these standardized pressure sequences to calculate the pressure change rate between adjacent sampling points, forming a complete pressure change rate sequence. Taking a nearby measuring point as an example, its historical data contains 5 disturbance events. The system extracts the start and end pressure data segments of each disturbance and unifies them into a standard sequence of 100 time points through linear interpolation. Then, it generates 5 equal-length rate of change sequences through differential calculation for subsequent feature analysis and joint modeling.

[0084] The time period from the start of the disturbance to the pressure value entering the stable range refers to the period from when the system detects the first obvious pressure disturbance (such as an instantaneous increase / decrease) until the pressure fluctuation range stabilizes within a set threshold (such as ±0.05 MPa) for multiple consecutive time points. This is a key data window for characterizing the response of the measuring point to the disturbance. "Normalization processing" refers to unifying pressure data of different lengths or sampling densities into a standard structure with equal length and equal sampling intervals. A common method is linear resampling. For example, the data from 83 sampling points collected within 0.8 seconds is linearly interpolated and expanded to 100 points, giving it a basis for alignment with other periodic data. "Unit time difference calculation" is to perform a first-order difference operation on the normalized sequence, that is, the difference between the pressure values ​​of every two adjacent sampling points is divided by the sampling time interval (such as dividing the difference by 0.01s if sampling once every 10ms), to obtain the rate of change of pressure per unit time, thus forming a numerical time series describing the rate of change of the disturbance, which is convenient for subsequent trend identification and model fitting. For example, a normalized pressure sequence of [1.0, 1.2, 1.35, 1.4] MPa corresponds to a rate of change sequence of [(1.2–1.0) / 0.01, (1.35–1.2) / 0.01, (1.4–1.35) / 0.01], or [20, 15, 5] MPa / s. This sequence visually reveals the rate of change and stability during the disturbance response.

[0085] In each pressure change rate sequence, the maximum pressure change rate, the duration of the change rate, and the time location where the change rate occurs are calculated. A numerical triple containing these three parameters is constructed as the feature vector of the measuring point. The feature vectors of all neighboring measuring points are combined into a joint feature vector by weighted average according to distance. Based on the joint feature vector, a joint neighborhood behavior model is constructed through multi-segment linear fitting.

[0086] To achieve feature vector extraction and joint modeling of neighboring measurement points, each pressure change rate sequence is first scanned and statistically analyzed point by point. The system identifies local extreme points using a sliding window detection method, thereby extracting the maximum pressure change rate (i.e., the maximum positive or negative value in the change rate sequence) in each sequence. Subsequently, the system determines the minimum time period (e.g., ≥0.1 seconds) during which the maximum change rate continuously exceeds a certain set threshold, and records the length of this time period as the change rate duration. Then, the system calculates the sampling point index position where the maximum change rate first appears (e.g., the 30th point), and converts it into the occurrence time by combining it with the standard sampling period (e.g., every 10ms). Thus, the system forms a numerical triplet consisting of the maximum change rate value, the change rate duration, and the change rate occurrence time for each sequence, which serves as the disturbance response feature vector for that data. For example, if the maximum change rate of a sequence is 15 MPa / s, the duration is 0.12 seconds, and it first appears at the 25th point (i.e., 0.25 seconds), then its feature vector is (15, 0.12, 0.25). After performing the above operations sequentially on a preset number of neighboring measuring points, the system obtains a set of feature vectors. Combining the physical distance from each measuring point to the target wireless digital pressure gauge, the system applies an inverse weighting function (e.g., wi=1 / di) to weight and sum the vectors, resulting in a joint vector that incorporates neighborhood perturbation features. Finally, using the joint vector as input parameters, the system employs a multi-segment linear fitting method to reconstruct the neighborhood perturbation trend model, which serves as the expression for the joint neighborhood behavior model.

[0087] "Multi-segment linear fitting" refers to treating the entire neighborhood disturbance response process as a linear change phase composed of several time segments, fitting the trend of each segment separately, and representing these phases with linear expressions of different slopes and intercept parameters, thus piecing them together to form a complete model. It can capture multiple trend inflection points of pressure change during the disturbance process, reflecting the dynamic process of the disturbance from rapid response to mitigation and then to stability. For example, the joint vector reflecting the rapid rise in the early stage of the disturbance (0–0.3s), the gradual change in the middle stage (0.3–0.7s), and the gradual stabilization in the later stage (0.7–1.0s) can be fitted as three straight lines: y1=20t+1.0, y2=5t+1.5, and y3=1.8, respectively. The constructed "neighborhood behavior joint model" is a combination of these multi-segment functions, representing the typical change law of the disturbance response in the physical neighborhood. It is a key criterion for subsequent evaluation of whether the data uploaded by the current target measurement point conforms to the physical laws of the neighborhood, especially in scenarios with abnormal addresses or incorrect data attribution, serving as a benchmark for comparison.

[0088] In this embodiment, the uploaded performance data is used as input to fit the historical behavior model and the neighborhood behavior joint model to obtain the first residual and the second residual. The average difference and trend difference between the first residual and the second residual are calculated to generate the residual coupling degree and the coupling degree change trend, which are used to characterize the degree of disorder in the transmission mapping of performance data. Specifically:

[0089] Using the pressure disturbance response time series from the uploaded performance data as input, and substituting them into the historical behavior model and the neighborhood behavior joint model corresponding to the wireless digital pressure gauge, respectively, at each unified sampling time point, the difference between the model prediction value and the actual pressure value is calculated to form the first residual sequence and the second residual sequence, respectively.

[0090] To fit the pressure disturbance response time series from the uploaded performance data into the historical behavior model and the joint neighborhood behavior model of the wireless digital pressure gauge, a model function mapping relationship can be established. First, for the wireless digital pressure gauge, the system pre-constructs its historical behavior model based on multiple historical periods during the initialization phase, storing it as a model expression using piecewise linear or regression functions (e.g., slope and intercept for each time period). Simultaneously, disturbance response features are extracted and aggregated from a predetermined number of neighboring measuring points to construct a joint neighborhood behavior model, expressed as multiple linear or weighted curves. Then, during the actual monitoring period, the system receives the performance data uploaded by the wireless digital pressure gauge, extracts the pressure disturbance response time series for the current period, and standardizes the time axis to ensure the input is aligned with the time of the two models. Next, the system substitutes this time series point by point into the historical behavior model and the joint neighborhood behavior model, outputting predicted values ​​based on the fitting expression of each model, forming a one-to-one predicted result sequence with the original data.

[0091] Taking a specific calculation process as an example, assume the uploaded pressure disturbance response time series is a sequence of length 100, P={p1,p2,...,p100}, with each item corresponding to a time sampling point. The historical behavior model is a three-segment linear expression, such as: points 1 to 30 correspond to the function f1(t)=20t+1.0, points 31 to 70 to f2(t)=5t+1.5, and points 71 to 100 to f3(t)=1.8. The system compares the pi corresponding to each ti with the model's fj(ti) output value at each time point and calculates the residual. The first residual sequence R(1) = {r1(1),...,r100(1)} is formed. The neighborhood behavior joint model performs the same steps to form the second residual sequence R(2). The entire process is synchronized on the time axis, so that the subsequent residual difference sequence calculation has an accurate sampling point correspondence. This calculation method can efficiently identify model fitting bias and reveal whether the uploaded data is more consistent with the local historical pattern or the neighborhood linkage pattern, thus laying the foundation for the attribution judgment.

[0092] Based on the two residual sequences, the difference is calculated for the residual values ​​of each corresponding sampling point to form a residual difference sequence;

[0093] To calculate the residual difference for each corresponding sampling point based on two residual sequences, and thus form a residual difference sequence, the two residual sequences can be indexed and aligned at the data structure level for point-by-point calculation. Specifically, the system first ensures that the historical behavior model and the neighborhood behavior joint model use identical time sampling points when constructing the residual sequence, for example, 100 points each with a 10ms interval. Let the first residual sequence be R(1)={r1(1),r2(1),...,r100(1)}, and the second residual sequence be R(2)={r1(2),r2(2),...,r100(2)}. The system can construct a dual-pointer synchronous traversal structure in memory and perform point-by-point difference calculation: for each sampling point i, calculate... The difference Δri is stored in a new sequence ΔR={Δr1,Δr2,...,Δr100}, which is the residual difference sequence. For example, if the 25th sampling point is 0.15 in the first residual sequence and -0.05 in the second residual sequence, then the difference is... The entire residual difference sequence reflects the difference in response between the historical model fitting results and the neighborhood model fitting results at each time step, serving as an important foundation for subsequent evaluation of fit consistency and identification of data assignment errors. This sequence can also be used for further analytical tasks such as constructing statistical graphs, sliding trend windows, and similarity threshold identification.

[0094] The average of all differences in the residual difference sequence is used to obtain the average difference that reflects the overall deviation of the fit.

[0095] To quantify the overall deviation of the residual difference sequence from the model's fit, the "average difference" can be obtained by performing an arithmetic mean operation on all differences in the sequence. In this implementation, the system first ensures that the residual difference sequence is a complete floating-point array, for example, ΔR={Δr1,Δr2,...,Δrn}, where each item represents the difference between the two model residual values ​​at the i-th time sampling point. The system iterates through the entire array, accumulating the sum of all Δr1 values, and then divides this sum by the total number of sampling points n to calculate the average difference. For example, assuming there are 5 sampling points in the residual difference sequence with values ​​{0.02, 0.04, 0.01, 0.03, 0.05}, the sum is 0.15, and the average difference is 0.03. A higher value indicates a greater difference in the fitting results of the two models to the current uploaded performance data, suggesting that the data may not conform to the typical behavior of either model and has a higher risk of attribution confusion. Conversely, a lower average difference indicates that both models can interpret the data well, and the risk of attribution determination is lower. This processing method is not only simple to calculate and produces intuitive results, but also facilitates subsequent normalization and trend coupling processing, making it a crucial foundation for building an intelligent attribution identification mechanism.

[0096] Local linear regression is performed on the residual difference sequence using a sliding window with a fixed time length. The fitting slope within each window is extracted, and the standard deviation of all window slope values ​​is calculated to obtain the trend difference value used to characterize the trend of the fitting difference.

[0097] To quantitatively identify the changing trends of fitted differences in the residual difference sequence, a local linear regression analysis can be performed using a fixed-time-length sliding window approach. Specifically, the system first sets a time window length (e.g., 10 sampling points) and iteratively slides across the entire residual difference sequence with a sliding step size (e.g., 1 or 2), extracting continuous difference subsequences of equal length segment by segment. Linear regression is performed on the subsequence within each sliding window, using the time index within the window as the independent variable t={1,2,...,k} and the corresponding difference Δr={Δr1,...,Δrk} as the dependent variable, fitting a linear model y=at+b, where the slope a represents the changing trend of the difference within that time period. Linear regression is performed once for each window, and its slope is extracted, forming a sliding slope sequence. Assuming the entire residual difference sequence length is 100, the window length is 10, and the sliding step size is 1, then 91 windows are obtained, and 91 slope values ​​are calculated, forming a slope sequence A={a1,a2,...,a91}.

[0098] Calculating the standard deviation of the above slope sequence yields the trend difference, which characterizes the changing trend of fitting discrepancies. A larger standard deviation indicates greater slope fluctuations in each local window, suggesting a significant upward or downward trend in the residual difference over time. This indicates a structural shift in the fitting results of the current data in the two models, suggesting potential instability in the assignment judgment. Conversely, a smaller standard deviation indicates more stable fitting errors and higher stability in the assignment judgment. For example, if the slopes of the five windows are {0.01, 0.03, −0.02, 0.00, 0.02}, the mean is 0.008, and the standard deviation is approximately 0.017. The sliding window method provides the ability to capture local data fluctuations, while linear regression can extract trend signals. The aggregated slope standard deviation, used as the trend difference, effectively reflects the stability of the overall fitting state over time, providing a dynamic quantitative basis for assignment error risk analysis.

[0099] The average difference and trend difference are respectively subjected to min-max normalization to form residual coupling degree and coupling degree change trend, which are used to characterize the degree of transmission mapping disorder of the upload performance data under the condition of abnormal address overlap. Specifically: when both residual coupling degree and coupling degree change trend are in the preset high value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is high; when both residual coupling degree and coupling degree change trend are in the preset low value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is low; when neither residual coupling degree nor coupling degree change trend is in the preset high value range nor in the preset low value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is moderate.

[0100] To normalize the average difference and trend difference, and to construct two evaluation indicators—residual coupling degree and coupling degree change trend—min-max normalization can be used to standardize the original values. This method transforms values ​​of different dimensions or scales to a unified range by setting a normalization interval (usually [0,1]), facilitating subsequent joint evaluation. Specifically, the system first extracts the historical minimum value of the average difference from historical training data or long-term statistical samples. With the maximum value Similarly, the historical minimum and maximum values ​​of the trend difference are obtained. Let the currently calculated average difference be... Then its normalized value is calculated as follows: The trend difference is handled in the same way. For example, if the historical range of the average difference is [0.01, 0.10], and the current average difference is 0.07, then the normalized value is... This value is defined as the "residual coupling degree," used to measure the consistency of the overall deviation of the residual fitting between the two models. Similarly, the normalized trend difference is defined as the "coupling degree change trend," used to quantify the intensity of fluctuation in the fitting state over time. This normalization process not only improves the algorithm's versatility and comparability but also provides a unified and controllable evaluation dimension for subsequent attribution judgment and classification based on threshold intervals, ensuring that the system can stably measure the degree of confusion in the transmission mapping of performance data even in cases of abnormally overlapping device addresses.

[0101] In performance data attribution judgment based on residual comparison, residual coupling degree is used to reflect the overall fitting difference of the current upload performance data under the joint model of historical behavior and neighborhood behavior, while the coupling degree change trend is used to reflect the dynamic fluctuation characteristics of this difference over time. When both indicators are in the preset high value range, it indicates that the upload performance data deviates significantly from both models, and this deviation shows a high degree of fluctuation over time, reflecting extremely poor matching consistency with historical behavior patterns and neighborhood behavior structures, thus being judged as a high degree of transmission mapping chaos. Conversely, when both indicators are in the preset low value range, it indicates that the upload performance data shows stable and small fitting errors in both models, exhibiting high consistency, with almost no risk of attribution errors caused by abnormal address overlap, thus being judged as a low degree of chaos. When neither indicator is in the high value range nor in the low value range, i.e., they are in the intermediate ambiguity zone, it indicates that the current data may be partially distorted, but the overall perturbation behavior has not completely deviated from the model prediction, and there is a possibility of attribution anomalies, but they are not serious, thus being classified as a moderate degree.

[0102] Different levels of transmission mapping disorder correspond to different risk levels and handling strategies: High levels of disorder mean that the data segment is highly likely to originate from a non-target measurement point, and its attribution label is extremely misleading. If used for subsequent trend modeling or strategy execution, it will significantly reduce prediction accuracy and system stability. Therefore, data isolation and attribution reconstruction operations must be performed immediately. Moderate levels of disorder indicate a high probability of attribution anomalies but still some behavioral consistency. Such data requires triggering intermediate intervention measures such as behavioral feature backtracking and measurement point matching correction, and should be marked as pending review for participation in semi-supervised modeling. Low levels of disorder indicate that the data is basically reliable and can be used as normal samples for behavior learning and state inference without intervention. Through the above classification mechanism, the data attribution judgment of wireless digital pressure gauges in IoT systems can be more accurate and the hierarchical response more flexible in the case of abnormal address overlap, thereby effectively improving the intelligence and robustness of the monitoring platform.

[0103] Based on the degree of transmission mapping disorder corresponding to the residual coupling degree and the trend of coupling degree change, determine whether there is an attribution error in the uploaded performance data;

[0104] In this embodiment, the degree of transmission mapping disorder corresponding to the residual coupling degree and the trend of coupling degree change is used to determine whether the uploaded performance data has an attribution error. Specifically:

[0105] When the transmission mapping disorder level is high, it is determined that there is an attribution error in the current upload performance data;

[0106] When the degree of transmission mapping disorder is moderate, combined with the historical fitting error average level of the performance data in the current measurement point behavior model, if the first residual of the performance data at the current measurement point exceeds the preset tolerance range of the historical fitting error average level, it is determined that the performance data has an assignment error problem; otherwise, it is determined that the performance data does not have an assignment error problem.

[0107] When the level of transport mapping disorder is low, it is determined that there is no attribution error in the current upload performance data.

[0108] In practical implementation, the degree of transmission mapping disorder of performance data under abnormal address overlap can be divided into three categories by using preset thresholds for high, medium, and low intervals of residual coupling degree and coupling degree change trend, and corresponding data attribution judgment logic can be formulated accordingly. When the disorder level is high, it indicates that the fitting deviation between the historical behavior model and the neighborhood behavior joint model for the current data is very significant, indicating that the behavior pattern of the data cannot be reasonably explained in the local measurement point and its neighborhood. Therefore, it is directly judged that the performance data has an attribution error problem, thus avoiding the inclusion of obviously abnormal data in the subsequent analysis process. When the disorder level is low, it indicates that the performance data has high consistency and small difference between the local model and the neighborhood model, and the data behavior is stable and predictable. Therefore, it is judged that the data does not have an attribution error problem. For the case of medium disorder level, since the fitting deviation has not yet reached the abnormal level, but there is still a potential risk of misattribution, it is necessary to compare it with the average level of the first residual in the historical behavior model over a recent period. If the first residual of the current data exceeds the preset tolerance range of the historical average residual, it indicates that its fitting error is significantly large and tends to be regarded as abnormal attribution; otherwise, it is regarded as normal data. This multi-level judgment mechanism not only enables timely identification of high-risk misclassified data, but also preserves the intelligent fault-tolerant processing capability for boundary data, thereby improving the accuracy and robustness of data attribution judgment, avoiding misjudgments caused by error fluctuations, and ensuring the stable operation of the monitoring and management system in complex field environments.

[0109] For performance data with misclassification issues, the performance data is divided into continuous time segments. Each time segment is matched with the historical data of a preset number of measurement points for behavioral features. Based on the matching results, the target measurement point of the performance data is determined through classification voting, thus completing the classification adjustment operation.

[0110] In this embodiment, for performance data with misattribution issues, the performance data is divided into continuous time segments according to a fixed time window length. The pressure disturbance response time series within each time segment is extracted, and its pressure change rate per unit time is calculated to form a pressure change rate sequence. This pressure change rate sequence is then matched with the historical pressure change rate sample sequence generated by the pressure disturbance response time series of the same length from a preset number of measuring points in the historical period. A similarity scoring algorithm based on Dynamic Time Warping (DTW) is used to calculate the matching score for each measuring point. Based on the matching scores of all measuring points, a weighted vote is performed, and the measuring point with the highest number of votes is selected as the target measuring point for the current time segment. The frequency of the measuring points for all time segments is then statistically analyzed, and the measuring point with the highest frequency is finally determined as the target measuring point for the performance data, thus completing the attribution adjustment operation.

[0111] In the implementation process, the performance data with attribution errors is first divided into multiple continuous, non-overlapping time segments according to a fixed time window length set by the system (e.g., 10 seconds, 30 seconds, or 1 minute). Then, a disturbance response time series is extracted from each time segment. This time series consists of pressure data recorded from the start to the end of the time segment. Assuming the data is recorded at a fixed sampling rate, the pressure difference between adjacent time points in the sequence is calculated and divided by the sampling interval to obtain the pressure change rate sequence per unit time. For example, if pressure data is collected once per second in a time segment and continuously recorded as [P1, P2, P3, ..., Pn], then the change rate sequence is [(P2-P1) / 1, (P3-P2) / 1, ..., (Pn-Pn-1) / 1], which represents the pressure change rate per second. This change rate sequence can effectively extract the disturbance behavior characteristics within the current segment for subsequent dynamic matching analysis with historical measurement data. This process can be completed in real time after data upload via programming, without manual intervention, and is suitable for large-scale concurrent device data processing.

[0112] In the implementation process, the first step is to obtain the pressure disturbance response time series of each measuring point from a preset number of candidate measuring points, representing a historical period of equal length to the current time segment. The pressure change rate per unit time is then calculated for these series to form a historical pressure change rate sample series. Next, the Dynamic Time Warping (DTW) algorithm is used to calculate the similarity score between the pressure change rate series of the current time segment and each historical sample series. DTW can effectively align fluctuation features with temporal misalignment, improving matching robustness. Based on the DTW matching score of each measuring point, it is standardized and converted into matching weights; the lower the score, the higher the similarity, and the greater the weight. Subsequently, a weighted voting mechanism is used to include the matching weights of all measuring points in the total voting score. The measuring point with the highest number of votes is designated as the target measuring point for that time segment. Frequency statistics are performed on the target measuring points for all time segments, i.e., the number of times each measuring point is selected as the target, and the measuring point with the highest frequency is selected as the final target measuring point for the entire performance data. Based on this, an assignment adjustment operation is performed to ensure the rationality of data assignment. For example, if performance data is divided into 5 time segments, with 3 segments voting to belong to measurement point A and 2 segments voting to belong to measurement point B, then the data will ultimately be assigned to measurement point A.

[0113] Similarity matching is the process of evaluating the similarity between two data sequences in a feature space, commonly used in pattern recognition and anomaly detection scenarios. The Time-to-Wave (DTW) algorithm is a classic time series alignment algorithm capable of handling data streams with inconsistent rates. It uses dynamic programming to find the shortest "alignment path" between two sequences to minimize the overall distance. For example, comparing a response sequence with a pressure jump to a historical sample sequence that occurred at slightly different times but with a consistent trend will result in a lower DTW score. Weighted voting is an aggregation decision-making mechanism. In this scheme, after converting DTW similarity into voting weights, a comprehensive decision is made on the attribution tendency of multiple measurement points, avoiding occasional mismatches in a single time slice. Frequency statistics, through global statistical analysis, extracts the consistent attribution trend of most time slices, thereby improving the stability and accuracy of the final judgment. For example, if measurement point A receives 3 votes and measurement point B receives 2 votes in 5 segments, then measurement point A is the final target attribution. The combined use of these methods can effectively achieve intelligent judgment and automatic correction of misattributed data.

[0114] After the attribution adjustment is completed, the residual coupling degree, the trend of coupling degree change, the judgment result and the target attribution measurement point are recorded to form an attribution judgment trajectory. Intervention processing is performed in the area with concentrated attribution errors, including reconstructing behavior labels, enabling the joint verification process of behavior features and measurement point location, and restricting the data uploaded in this area from participating in trend modeling.

[0115] To enable traceability and optimization of the attribution adjustment process, key data from the attribution determination process can be structured and recorded to form an "attribution determination trajectory." Specifically, after each attribution adjustment, the system automatically extracts and records parameters related to the attribution decision, including the normalized values ​​of residual coupling degree and coupling degree change trends, the attribution determination conclusion based on these values, the identification information of the target attribution measurement point, and its corresponding matching confidence level. This data is bound and stored in time-series format in the logical record corresponding to the wireless digital pressure gauge, and can be used for subsequent attribution determination result comparison, pattern tracing, and strategy adjustment. This method, through the combination of a judgment module and storage mechanism in the software logic, achieves traceability and interpretability of the attribution determination process without human intervention.

[0116] To further improve the accuracy of attribution determination and the system's adaptability, an intervention process can be triggered when the system detects frequent and concentrated attribution errors within a specific physical area. The first step of the intervention is to reconstruct the behavioral tags of the wireless digital pressure gauges within that area. This involves re-collecting their pressure disturbance response data under static fluid conditions and generating and binding new behavioral features through a unified process. Secondly, the system initiates a joint verification process based on behavioral features and the spatial location of the measuring points. This means that in addition to relying on the behavioral model to determine data attribution, the relative position of the device within the topology is also included in the comparison parameters to ensure that the behavioral logic is reasonable and the spatial attribution is correct. This joint verification can be achieved by setting data consistency indicators and distance confidence weights within the spatial neighborhood.

[0117] Furthermore, to prevent erroneous data from interfering with the trend modeling process, the system will temporarily restrict the direct participation of uploaded raw data within this physical area in the modeling operation. Before the data attribution stability is restored, this data will only be used for auxiliary comparison and will not be used as training input for trend prediction model updates. This strategy is implemented through software-level modeling permission control and labeling filtering mechanisms, ensuring both the accuracy of model training data and preventing misattributed data from misleading the overall monitoring decision-making chain, thereby improving the stability and robustness of the entire IoT-based wireless digital pressure gauge performance data monitoring and management method.

[0118] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0119] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means (e.g., infrared, wireless, microwave, etc.). A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0120] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0121] 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 implementation should not be considered beyond the scope of this application.

[0122] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the 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 an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0123] 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.

[0124] In addition, the functional units in the various embodiments of this application 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.

[0125] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations 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. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for monitoring and managing the performance data of wireless digital pressure gauges based on the Internet of Things, characterized in that, Specifically, the following steps are included: Acquire pressure disturbance response data of wireless digital pressure gauges in static fluid environments during the initial deployment phase, construct behavioral tags based on this pressure disturbance response data, and bind the behavioral tags to the wireless digital pressure gauges; After the wireless digital pressure gauge uploads performance data, the pressure disturbance response time series is extracted and compared with the response times of neighboring measurement points. Combined with the physical topology between measurement points, it is determined whether the wireless digital pressure gauge is experiencing abnormal address overlap. When the wireless digital pressure gauge is experiencing abnormal address overlap, a joint model of historical behavior and neighborhood behavior is constructed and fitted to the uploaded performance data to obtain a first residual and a second residual. Based on the first and second residuals, residual coupling degree and coupling degree change trend are generated to characterize the degree of transmission mapping disorder in the performance data. Based on the degree of transmission mapping disorder corresponding to the residual coupling degree and coupling degree change trend, it is determined whether the uploaded performance data has an incorrect attribution problem. For performance data with misclassification issues, the performance data is divided into continuous time segments. Each time segment is matched with the historical data of a preset number of measurement points for behavioral features. Based on the matching results, the target measurement point of the performance data is determined through classification voting, thus completing the classification adjustment operation. After the attribution adjustment is completed, the residual coupling degree, the trend of coupling degree change, the judgment result and the target attribution measurement point are recorded to form an attribution judgment trajectory. Intervention processing is performed in the area with concentrated attribution errors, including reconstructing behavior labels, enabling the joint verification process of behavior features and measurement point location, and restricting the data uploaded in this area from participating in trend modeling.

2. The method for monitoring and managing the performance data of a wireless digital pressure gauge based on the Internet of Things according to claim 1, characterized in that, After the wireless digital pressure gauge uploads performance data, the pressure disturbance response time series is extracted and compared with the response times of neighboring measuring points. Based on the physical topology relationship between the measuring points, it is determined whether the wireless digital pressure gauge is experiencing abnormal address overlap. Specifically: After the performance data is uploaded by the wireless digital pressure gauge, the pressure disturbance response time series is extracted. This pressure disturbance response time series consists of continuous pressure change data from the start of the disturbance to the point where the pressure value enters the stable range. The disturbance response time of the wireless digital pressure gauge is obtained by calculating the time interval between the moment when the pressure change rate reaches its maximum value and the start of the disturbance in the pressure disturbance response time series. Within the same data synchronization period, a predetermined number of neighboring measuring points are selected, and the pressure disturbance response time series of each measuring point is extracted. The set of disturbance response times of neighboring measuring points is then calculated in the same way. Based on the physical topology information between the wireless digital pressure gauge and each adjacent measuring point, the corresponding physical connection path and pipeline distance data are obtained, and combined with the preset fluid medium propagation speed, the theoretical disturbance propagation time from each adjacent measuring point to the wireless digital pressure gauge is calculated. Based on the difference between the response time of each measuring point in the set of disturbance response times of neighboring measuring points and its theoretical propagation time, the average propagation offset time of the disturbance under the current topology is calculated, and the theoretical response time delay range is constructed. The disturbance response time of the wireless digital pressure gauge is compared with the theoretical response time delay range. When the response time exceeds the theoretical range boundary threshold, it is determined that the wireless digital pressure gauge is in a state of abnormal device address overlap.

3. The method for monitoring and managing the performance data of a wireless digital pressure gauge based on the Internet of Things according to claim 2, characterized in that, When a wireless digital pressure gauge experiences abnormal address overlap, a historical behavior model is constructed based on the pressure disturbance response time series uploaded by the wireless digital pressure gauge in historical periods. A neighborhood behavior joint model is constructed based on the pressure disturbance response time series of a preset number of neighboring measuring points with physical topological connections to the wireless digital pressure gauge. The uploaded performance data is used as input to fit the historical behavior model and the neighborhood behavior joint model to obtain the first residual and the second residual. By calculating the average difference and trend difference between the first residual and the second residual, the residual coupling degree and the coupling degree change trend are generated to characterize the degree of disorder in the transmission mapping of performance data.

4. The method for monitoring and managing the performance data of a wireless digital pressure gauge based on the Internet of Things according to claim 3, characterized in that, When a wireless digital pressure gauge experiences abnormally overlapping device addresses, a historical behavior model is constructed based on the pressure disturbance response time series uploaded by the wireless digital pressure gauge in historical periods. Specifically: Select a preset number of historical period data on continuous pressure changes from the start of the disturbance to when the pressure value enters a stable range. Based on the continuous pressure change data, the pressure change rate per unit time is calculated to form a pressure change rate sequence. The pressure change rate sequence is clustered using a time series similarity algorithm, and a disturbance response behavior feature sequence containing pressure change rate samples with a preset time length is extracted. A piecewise linear function model is constructed by fitting a sequence of perturbation response behavior characteristics. Combination weights are set according to the perturbation response amplitude and duration corresponding to each piecewise linear function model. The piecewise linear function models are then combined in a weighted manner to construct a historical behavior model that describes the perturbation response behavior of the wireless digital pressure gauge.

5. The method for monitoring and managing the performance data of a wireless digital pressure gauge based on the Internet of Things according to claim 4, characterized in that, A joint neighborhood behavior model is constructed based on the pressure disturbance response time series of a predetermined number of neighboring measuring points with physical topological connections to the wireless digital pressure gauge. Specifically: Based on the physical pipeline connection diagram between the wireless digital pressure gauge and other measuring points, obtain the physical connection path and distance information, and filter out the nearby measuring points whose physical distance does not exceed the set distance threshold. Continuous pressure change data of each neighboring measuring point from the start of disturbance to the stable pressure range within multiple historical periods are selected. Each data segment is normalized according to a uniform time sampling interval and length. The pressure change rate sequence is obtained by calculating the unit time difference of the continuous pressure data. In each pressure change rate sequence, the maximum pressure change rate, the duration of the change rate, and the time location where the change rate occurs are calculated. A triplet containing the values ​​is constructed as the feature vector of the measuring point. The feature vectors of all neighboring measuring points are combined into a joint feature vector by weighted average according to distance. Based on the joint feature vector, a joint neighborhood behavior model is constructed through multi-segment linear fitting.

6. The method for monitoring and managing performance data of a wireless digital pressure gauge based on the Internet of Things according to claim 5, characterized in that, Using uploaded performance data as input, the model is fitted to both the historical behavior model and the neighborhood behavior joint model to obtain the first and second residuals. The average difference and trend difference between the first and second residuals are then calculated to generate the residual coupling degree and its changing trend, which characterize the degree of disorder in the transmission mapping of performance data. Specifically: Using the pressure disturbance response time series from the uploaded performance data as input, and substituting them into the historical behavior model and the neighborhood behavior joint model corresponding to the wireless digital pressure gauge, respectively, at each unified sampling time point, the difference between the model prediction value and the actual pressure value is calculated to form the first residual sequence and the second residual sequence, respectively. Based on the two residual sequences, the difference is calculated for the residual values ​​of each corresponding sampling point to form a residual difference sequence; The average of all differences in the residual difference sequence is used to obtain the average difference that reflects the overall deviation of the fit. Local linear regression is performed on the residual difference sequence using a sliding window with a fixed time length. The fitting slope within each window is extracted, and the standard deviation of all window slope values ​​is calculated to obtain the trend difference value used to characterize the trend of the fitting difference. The average difference and trend difference are respectively processed by min-max normalization and the residual coupling degree and the coupling degree change trend are formed to characterize the degree of transmission mapping disorder of the upload performance data under the condition of abnormal address overlap. Specifically, when the residual coupling degree and the coupling degree change trend are both in the preset high value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is high. When both the residual coupling degree and the coupling degree change trend are in the preset low value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is low. When the residual coupling degree and the trend of coupling degree change are neither in the preset high value range nor in the preset low value range, the degree of transmission mapping disorder of the current upload performance data under the condition of abnormal address overlap is moderate.

7. The method for monitoring and managing the performance data of a wireless digital pressure gauge based on the Internet of Things according to claim 6, characterized in that, Based on the degree of transmission mapping disorder corresponding to the residual coupling degree and the trend of coupling degree change, it is determined whether there is an attribution error in the uploaded performance data. Specifically: When the transmission mapping disorder level is high, it is determined that there is an attribution error in the current upload performance data; When the degree of transmission mapping disorder is moderate, combined with the historical fitting error average level of the performance data in the current measurement point behavior model, if the first residual of the performance data at the current measurement point exceeds the preset tolerance range of the historical fitting error average level, it is determined that the performance data has an assignment error problem; otherwise, it is determined that the performance data does not have an assignment error problem. When the level of transport mapping disorder is low, it is determined that there is no attribution error in the current upload performance data.

8. The method for monitoring and managing the performance data of a wireless digital pressure gauge based on the Internet of Things according to claim 7, characterized in that, For performance data with misclassification issues, the performance data is divided into continuous time segments according to a fixed time window length. The pressure disturbance response time series within each time segment is extracted, and the pressure change rate per unit time is calculated to form a pressure change rate series. This pressure change rate series is then matched with the historical pressure change rate sample series generated by the pressure disturbance response time series of the same length in the historical period for a preset number of measuring points. A similarity scoring algorithm based on Dynamic Time Warping (DTW) is used to calculate the matching score for each measuring point. Based on the matching scores of all measuring points, a weighted vote is performed, and the measuring point with the highest number of votes is selected as the target measuring point for the current time segment. The frequency of the measuring points for all time segments is then statistically analyzed, and the measuring point with the highest frequency is finally determined as the target measuring point for the performance data, completing the classification adjustment operation.

Images