Intelligent gas meter remote reading method and system based on Internet of Things

Abstract

The invention provides an intelligent gas meter remote meter reading method and system based on the Internet of Things, and relates to the technical field of intelligent gas meters. The method comprises the following steps: collecting local metering data and Internet of Things transmission data; constructing a dual-source data multi-dimensional conflict identification system, and screening conflict data groups; performing accuracy verification on the conflict data group; an encrypted transmission channel is established, and the traceability of the data transmission process is realized; triggering a grading early warning mechanism for the abnormal usage data deviating from the baseline; and carrying out iterative updating on the meter reading data knowledge base, and carrying out remote calibration. According to the method, double-source data conflicts are identified through multiple dimensions, conflict data are corrected in combination with a hybrid model, metering accuracy is guaranteed, and disputes are reduced; through constructing a baseline model and a grading early warning mechanism, abnormal usage is accurately identified and disposed, hidden dangers are timely prevented, and service is excellent. Through dynamic on-demand remote calibration of equipment, the metering precision is guaranteed, the operation and maintenance cost is reduced, the user data transparency is improved, and intelligent gas management is assisted.

Description

Technical Field

[0001] This invention relates to the field of smart gas meter technology, and in particular to a remote meter reading method and system for smart gas meters based on the Internet of Things (IoT). Background Technology

[0002] With the acceleration of urbanization and the upgrading of demand for smart energy management, gas, as a key energy source for people's livelihood, is undergoing a transformation in its metering and reading methods from traditional manual meter reading to intelligent remote meter reading. The traditional manual meter reading method relies on staff to record gas meter data door-to-door. Limited by time and space, it is not only inefficient and costly in terms of labor, but also prone to inaccurate metering data due to human error, omissions, and incorrect readings, which can lead to billing disputes between users and gas companies. Furthermore, it cannot achieve real-time data collection, making it difficult for gas companies to keep abreast of fluctuations in users' gas consumption. This hinders their ability to quickly respond to safety hazards such as gas leaks and abnormal excessive gas consumption, and makes it difficult to optimize gas dispatching and supply planning based on data.

[0003] To address the drawbacks of manual meter reading, the industry has gradually introduced remote meter reading technology to achieve automatic collection and remote transmission of metering data. However, existing remote meter reading solutions rely on a single metering data source at the gas meter, which is prone to data distortion, leading to unreliable metering results. Furthermore, they cannot comprehensively consider multiple factors, making it difficult to accurately identify hidden data conflicts and affecting the accuracy of billing and safety warnings. In addition, as users demand greater transparency and security in gas consumption data, existing solutions are insufficient to meet the comprehensive needs of smart gas management for data accuracy, real-time performance, security, and intelligent services. Therefore, there is an urgent need for a remote meter reading technology solution that can achieve multi-source data fusion verification, intelligent conflict identification, efficient data processing, and dynamic optimization. Summary of the Invention

[0004] This invention provides a remote meter reading method and system for smart gas meters based on the Internet of Things, which addresses the shortcomings of existing technologies in meeting the comprehensive requirements of smart gas management for data accuracy, real-time performance, security, and intelligent services.

[0005] On one hand, the present invention provides a method for remote meter reading of smart gas meters based on the Internet of Things, including: Real-time collection of local metering data from smart gas meters and data transmitted via the Internet of Things.

[0006] Construct a multi-dimensional conflict identification system for dual-source data to screen conflicting data groups that pose a risk of data inconsistency.

[0007] A hybrid LSTM-Kalman filter model with an integrated attention mechanism is constructed to verify the accuracy of conflicting data sets and output corrected reliable measurement data.

[0008] Edge computing nodes are used to perform layered preprocessing of trusted metering data and an encrypted transmission channel is established to transmit the preprocessed trusted metering data to the remote meter reading cloud platform, thereby achieving traceability of the data transmission process.

[0009] Based on the preprocessed reliable metering data, a multi-dimensional user gas consumption report is generated, and a user gas consumption baseline model is constructed. An tiered early warning mechanism is triggered for abnormal consumption data that deviates from the baseline.

[0010] The meter reading data knowledge base is archived and stored according to a preset cycle, and the meter reading data knowledge base is iteratively updated using a federated learning algorithm. The smart gas meters are also remotely calibrated at a preset frequency.

[0011] According to the IoT-based remote meter reading method for smart gas meters provided by this invention, the local metering data includes the gas meter's instantaneous flow rate, cumulative gas consumption, metering timestamp, and temperature compensation coefficient. The IoT transmission data includes IoT instantaneous flow rate, IoT cumulative gas consumption, signal strength, and data transmission delay.

[0012] According to the IoT-based remote meter reading method for smart gas meters provided by the present invention, the process of filtering conflicting data groups with data consistency risks includes: Calculate the deviation between the local metering timestamp and the IoT transmission timestamp, set a time synchronization threshold, and when the deviation exceeds the time synchronization threshold, correct the timestamp using linear interpolation. If the deviation is still greater than the time synchronization threshold after correction, it is marked as a candidate group for time conflict.

[0013] Extract the gas meter device number from the local metering data and the device identifier from the IoT transmission data. Verify the consistency through hash mapping. If they do not match, mark them as a device identifier conflict group.

[0014] Construct feature vectors from two sources and use the cosine similarity algorithm to calculate the correlation between the two feature vectors.

[0015] Set a correlation threshold. When the correlation is lower than the correlation threshold, it is marked as a candidate group for feature correlation conflict.

[0016] A dynamic threshold model is constructed based on historical conflict-free data to calculate the cumulative gas consumption deviation rate and instantaneous flow deviation.

[0017] A sliding window algorithm is used to calculate the dynamic threshold. When the cumulative gas consumption deviation rate is greater than the dynamic threshold of the cumulative gas consumption deviation rate and the instantaneous flow deviation is greater than the dynamic threshold of the instantaneous flow deviation, it is determined to be a conflicting data group, triggering the accuracy verification process.

[0018] According to the IoT-based remote meter reading method for smart gas meters provided by this invention, the process of constructing an LSTM-Kalman filter hybrid model with an attention mechanism to verify the accuracy of conflicting data sets includes: The historical time series data of the confirmed conflict data group is divided into time steps, and the weight of each feature in the time series prediction is calculated. The input features for each time step include instantaneous flow rate, cumulative gas consumption, signal strength, and transmission delay.

[0019] The LSTM network architecture is configured, including hidden layers and an output layer. The hidden layer contains multiple neurons and uses the ReLU activation function to model the weighted temporal features. The output layer is used to output the predicted trend values ​​of the data.

[0020] The predicted data trend is used as a priori estimate for Kalman filtering to establish the state equation and observation equation.

[0021] Historical conflict data and manually labeled reliable data were used as the training set, and the mean square error between the predicted and labeled values ​​was used as the loss function.

[0022] The model is iteratively trained using the Adam optimizer until the loss function converges. The trained hybrid model is then input into the conflict data set, and the model outputs corrected reliable econometric data. When the prediction confidence of the model output reaches the preset confidence threshold, the verification result is deemed reliable.

[0023] The remote meter reading method for smart gas meters based on the Internet of Things provided by the present invention includes a process of hierarchical preprocessing of trusted metering data using edge computing nodes, comprising: The wavelet packet transform algorithm is used to perform threshold quantization on the high-frequency coefficients based on Stein's unbiased risk estimation.

[0024] Define a standard data format, with fields including user ID, gas meter number, metering time, cumulative gas consumption, instantaneous flow rate, temperature compensation coefficient, and data status code, and encapsulate it in JSON format to obtain standardized data.

[0025] An isolated forest algorithm model is constructed to detect anomalies in standardized data. An anomaly score is calculated for each data point. When the anomaly score is greater than a preset score threshold, it is determined to be an anomaly, automatically removed, and marked as awaiting manual review. At the same time, the feature information of the anomaly data is recorded.

[0026] According to the IoT-based remote meter reading method for smart gas meters provided by the present invention, the process of establishing an encrypted transmission channel includes: The system employs a dual mechanism of username and password authentication and client certificate authentication. The client certificate includes the unique identifier of the gas meter device, its validity period, and public key information.

[0027] The ECC-256 algorithm is used to encrypt the Payload field of the MQTT protocol to generate encrypted data packets.

[0028] Calculate the data integrity check code and transmit the integrity check code along with the encrypted data packet.

[0029] Edge computing nodes package the metadata of transmitted data into blockchain transactions according to a preset period and submit them to the metrology agency. The metadata includes user ID, data transmission time, verification code and device identifier.

[0030] The metrology agency uses the PoS consensus mechanism, where each block contains the hash value of the previous block, ensuring that the data transmission process is immutable and traceable.

[0031] According to the IoT-based remote meter reading method for smart gas meters provided by the present invention, the process of triggering a tiered early warning mechanism for abnormal usage data deviating from the baseline includes: Collect users' historical normal gas consumption data and combine it with user profile information to construct a multiple linear regression baseline model.

[0032] The daily and weekly usage fluctuation thresholds for users are calculated based on a multiple linear regression baseline model.

[0033] Based on the calculation results, multi-dimensional usage reports are generated, including daily usage trend charts, weekly usage comparison tables, and monthly usage statistics. Warning levels are also classified according to the proportion of daily usage deviating from the baseline.

[0034] The remote meter reading method for smart gas meters based on the Internet of Things provided by this invention includes the following process for iteratively updating the meter reading data knowledge base using a federated learning algorithm: The meter reading data knowledge base of each regional edge computing node is used as a local dataset. The meter reading data knowledge base includes metering data, conflict handling records, and early warning information.

[0035] The initial conflict identification model and verification algorithm parameters are distributed to each edge computing node through the cloud platform.

[0036] The model is trained on each edge computing node using a local dataset, the gradient of the model parameter update is calculated, and the gradient of the model parameter update is encrypted using a homomorphic encryption algorithm before being uploaded to the cloud platform.

[0037] The cloud platform aggregates the update gradients uploaded by edge computing nodes.

[0038] The aggregated gradients are updated to the global model, and the process of updating the gradients and aggregating the updated gradients is repeated until the model loss function converges. The updated global model parameters are then distributed to each edge computing node and the smart gas meter terminal.

[0039] According to the IoT-based remote meter reading method for smart gas meters provided by the present invention, the process of remotely calibrating the smart gas meter at a preset frequency includes: The gas meter sends a standard flow command at a preset frequency. After receiving the command, the gas meter operates at the standard flow rate and collects and feeds back the metering value.

[0040] The calibration deviation is calculated based on the standard flow rate and the feedback measurement value.

[0041] When the absolute value of the calibration deviation exceeds the preset deviation threshold, a parameter adjustment command is sent to the gas meter to correct the metering coefficient.

[0042] After adjustment, the gas meter is run again at the standard flow rate. New feedback metering values ​​are collected, and new deviations are calculated. When the absolute value of the new calibration deviation is less than the preset deviation threshold, the calibration is deemed qualified.

[0043] Generate a calibration report and upload it to the cloud platform for archiving. The calibration report includes calibration time, standard flow rate, feedback value, deviation rate, and adjustment parameters.

[0044] On the other hand, the present invention also provides a remote meter reading system for smart gas meters based on the Internet of Things, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, the processor including: The data acquisition module is used to collect local metering data and IoT transmission data from smart gas meters in real time.

[0045] The conflict screening module is used to build a multi-dimensional conflict identification system for dual-source data and screen conflicting data groups that pose a risk of data inconsistency.

[0046] The data correction module is used to construct an LSTM-Kalman filter hybrid model with an integrated attention mechanism, to verify the accuracy of conflicting data sets, and to output corrected reliable measurement data.

[0047] The data transmission module is used to perform layered preprocessing of trusted metering data using edge computing nodes and establish an encrypted transmission channel to transmit the preprocessed trusted metering data to the remote meter reading cloud platform, thereby achieving traceability of the data transmission process.

[0048] The anomaly warning module is used to generate multi-dimensional user gas consumption reports based on preprocessed reliable metering data, and to build a user gas consumption baseline model. It triggers a graded warning mechanism for abnormal consumption data that deviates from the baseline.

[0049] The update calibration module is used to archive and store the meter reading data knowledge base according to a preset cycle, iterate and update the meter reading data knowledge base using a federated learning algorithm, and remotely calibrate the smart gas meters at a preset frequency.

[0050] The remote meter reading method and system for smart gas meters based on the Internet of Things provided by this invention have the following beneficial effects: By constructing a multi-dimensional conflict identification system for dual-source data, it is possible to accurately identify data consistency risk groups from multiple dimensions such as time synchronization, device identification, and feature association, avoiding the problem of deviation from a single data source. Through an LSTM-Kalman filter hybrid model that integrates an attention mechanism, it can intelligently correct conflicting data by combining historical time-series data and real-time feature weights, outputting reliable measurement data, reducing disputes caused by inaccurate measurement, and ensuring the reliability of measurement results.

[0051] Edge computing nodes are used for layered preprocessing of trusted metrology data. Wavelet packet transform is used to compress the data, and the isolated forest algorithm is used to filter out outliers, reducing the amount of data uploaded to the cloud and alleviating network transmission and cloud storage pressure. An encrypted transmission channel based on the MQTT protocol and blockchain technology employs encryption and verification methods to ensure data transmission security. Blockchain-based evidence storage enables data traceability and prevents data tampering.

[0052] By constructing a multiple linear regression baseline model of user gas consumption and combining it with daily and weekly consumption fluctuation thresholds to divide a three-level early warning mechanism, different response measures can be dynamically triggered according to the degree of abnormality, so as to prevent safety hazards in a timely manner and improve the user service experience.

[0053] The federated learning algorithm is used to iteratively update the meter reading data knowledge base. Each edge node only uploads the model gradient instead of the original data. This protects user data privacy while achieving global model collaborative optimization, balancing data privacy protection and algorithm iteration efficiency.

[0054] The calibration frequency is dynamically adjusted based on historical data conflicts in the gas meter, and intermediate calibration is added for high-frequency conflict devices to achieve on-demand calibration. By remotely sending standard flow commands and dynamically correcting the metering coefficient, calibration efficiency is improved, long-term metering accuracy of the gas meter is guaranteed, and maintenance costs are reduced.

[0055] Multi-dimensional usage report generation, meter reading data knowledge base archiving, and model iteration optimization provide gas companies with data support such as user gas consumption trends and equipment operating status, helping them optimize gas scheduling planning and accurately predict gas demand. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0057] Figure 1This is a flowchart illustrating the remote meter reading method for smart gas meters based on the Internet of Things provided in this embodiment of the invention. Figure 2 This is a schematic diagram of the structure of the IoT-based smart gas meter remote meter reading system provided in an embodiment of the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0059] The following is combined with Figures 1-2 This invention describes a remote meter reading method and system for smart gas meters based on the Internet of Things (IoT).

[0060] Figure 1 This is a flowchart illustrating the remote meter reading method for smart gas meters based on the Internet of Things provided in this embodiment of the invention.

[0061] like Figure 1 As shown in the embodiments of the present invention, the remote meter reading method and system for smart gas meters based on the Internet of Things (IoT) can be executed by a remote meter reading method for smart gas meters based on the IoT, and the method includes: Real-time collection of local metering data from smart gas meters and data transmitted via the Internet of Things.

[0062] Local metering data includes instantaneous flow rate, cumulative gas consumption, metering timestamp, and temperature compensation coefficient collected by the Hall sensor built into the gas meter. IoT transmission data includes IoT instantaneous flow rate, IoT cumulative gas consumption, signal strength, and data transmission latency transmitted via the LoRaWAN protocol.

[0063] Constructing a multi-dimensional conflict identification system for dual-source data and screening conflicting data groups with potential data consistency risks includes the following process: Calculate the deviation between the local metering timestamp and the IoT transmission timestamp, set a time synchronization threshold, and when the deviation exceeds the time synchronization threshold, correct the timestamp using linear interpolation. If the deviation is still greater than the time synchronization threshold after correction, it is marked as a candidate group for time conflict.

[0064] Extract the gas meter device number from the local metering data and the device identifier from the IoT transmission data. Verify the consistency through hash mapping. If they do not match, mark them as a device identifier conflict group.

[0065] Constructing dual-source data feature vectors, local feature vectors ,in, Indicates local instantaneous traffic. This indicates the local cumulative gas consumption. Indicates the temperature compensation coefficient. Represents the local metering timestamp. IoT feature vector. ,in, Indicates instantaneous traffic flow in the Internet of Things (IoT) This indicates the cumulative gas consumption via the Internet of Things (IoT), RSSI indicates the received signal strength indicator, and delay indicates the data transmission delay.

[0066] The cosine similarity algorithm is used to calculate the correlation between two feature vectors, and the formula is expressed as:

[0067] In the formula, Sim represents the degree of correlation. Represents the local feature vector. Represents the feature vector of the Internet of Things. Indicates the magnitude of the local feature vector. This represents the magnitude of the feature vector of the Internet of Things. This represents the dot product of vectors.

[0068] Set a correlation threshold. When the correlation is lower than the correlation threshold, it is marked as a candidate group for feature correlation conflict.

[0069] A dynamic threshold model is constructed based on historical conflict-free data to calculate the cumulative gas consumption deviation rate and instantaneous flow deviation, expressed by the following formula:

[0070]

[0071] In the formula, This indicates the cumulative gas consumption deviation rate. This indicates the local cumulative gas consumption. This indicates the cumulative gas consumption via the Internet of Things (IoT). Indicates instantaneous flow deviation. Indicates local instantaneous traffic. This represents the instantaneous traffic flow in the Internet of Things (IoT).

[0072] Calculate the dynamic threshold using the sliding window algorithm and The formula is expressed as:

[0073]

[0074] In the formula, This represents the dynamic threshold for the cumulative gas consumption deviation rate. This represents the average of the historical cumulative gas consumption deviation rate. The standard deviation of the historical cumulative gas consumption deviation rate. This represents the dynamic threshold for instantaneous flow deviation. This represents the mean of historical instantaneous flow deviations. The standard deviation represents the historical instantaneous flow rate deviation.

[0075] when and When a conflicting data group is identified, the accuracy verification process is triggered.

[0076] A hybrid LSTM-Kalman filter model incorporating an attention mechanism is constructed to verify the accuracy of conflicting data sets and output corrected reliable measurement data. The process includes: The historical time-series data of the confirmed conflict data group is divided into time steps. The input features for each time step include instantaneous flow rate, cumulative gas consumption, signal strength, and transmission delay, for a total of 6 dimensions.

[0077] The weight of each feature in the temporal prediction is calculated using the following formula:

[0078]

[0079] In the formula, This represents the attention weight, where n represents the total number of input features. This represents the intermediate calculated value obtained after performing a linear transformation on the i-th dimension of the input feature. Represents the weight matrix. Represents the i-th dimension of the input feature. This indicates the bias term.

[0080] The LSTM network architecture is configured, including hidden layers and an output layer. The hidden layer contains multiple neurons and uses the ReLU activation function to model the weighted temporal features. The output layer is used to output the predicted trend values ​​of the data.

[0081] Using the predicted data trend as a priori estimate for Kalman filtering, we establish the state equation and the observation equation. The state equation is expressed as:

[0082]

[0083]

[0084]

[0085] In the formula, This represents the state vector at time k. Let A represent the state vector at time k-1, A represent the state transition matrix, and B represent the control input matrix. Indicates the rate of change of flow rate. Indicates process noise. This represents the predicted cumulative gas consumption at time k. This represents the predicted instantaneous flow rate at time k. Indicates the sampling time interval.

[0086] The observation equation is expressed as:

[0087]

[0088]

[0089] In the formula, Let H represent the observation vector at time k, and let H represent the observation matrix. Indicates observation noise. This represents the cumulative gas consumption measured locally at time k. This represents the instantaneous flow rate measured locally at time k. This represents the instantaneous traffic volume transmitted by the Internet of Things at time k.

[0090] Historical conflict data and manually annotated reliable data are used as the training set, and the mean squared error between the predicted and annotated values ​​is used as the loss function, expressed by the formula:

[0091] In the formula, N represents the sample size. Indicates the confidence value of the annotation. This represents the model's output value.

[0092] The model is iteratively trained using the Adam optimizer until the loss function converges. The trained hybrid model is then input into the conflict data set, and the model outputs corrected reliable econometric data. When the prediction confidence of the model output reaches the preset confidence threshold, the verification result is deemed reliable.

[0093] Edge computing nodes are used to perform layered preprocessing of trusted metering data and an encrypted transmission channel is established to transmit the preprocessed trusted metering data to the remote meter reading cloud platform, thereby achieving traceability of the data transmission process.

[0094] The process of using edge computing nodes to perform hierarchical preprocessing on trusted metrology data includes: The wavelet packet transform algorithm was adopted, the db5 wavelet basis function was selected, the number of decomposition layers was set to 4, and the high-frequency coefficients were subjected to threshold quantization based on Stein unbiased risk estimation.

[0095] Define a standard data format, with fields including user ID, gas meter number, metering time, cumulative gas consumption, instantaneous flow rate, temperature compensation coefficient, and data status code, and encapsulate it in JSON format to obtain standardized data.

[0096] Construct an isolation forest algorithm model, perform anomaly detection on standardized data, and calculate the anomaly score S for each data point, expressed by the formula:

[0097] In the formula, This represents the average path length of data point x in the forest. This represents the average path length when the sample size is n. When the abnormal score is greater than the preset score threshold, it is judged as an outlier, automatically removed and marked as awaiting manual review, and the feature information of the abnormal data is recorded.

[0098] The process of establishing an encrypted transmission channel includes: The system employs a dual mechanism of username and password authentication and client certificate authentication. The client certificate includes the unique identifier of the gas meter device, its validity period, and public key information.

[0099] The ECC-256 algorithm is used to encrypt the Payload field of the MQTT protocol to generate encrypted data packets.

[0100] Calculate the data integrity check code and transmit the integrity check code along with the encrypted data packet.

[0101] Edge computing nodes package the metadata of transmitted data into blockchain transactions according to a preset period and submit them to the metrology agency. The metadata includes user ID, data transmission time, verification code and device identifier.

[0102] The metrology institution adopts the PoS (Proof of Stake) consensus mechanism, where each block contains the hash value of the previous block, ensuring the immutability and traceability of data transmission.

[0103] Based on the preprocessed reliable metering data, a multi-dimensional user gas consumption report is generated, and a user gas consumption baseline model is constructed. A tiered early warning mechanism is triggered for abnormal consumption data deviating from the baseline. The process includes: Collect users' historical normal gas consumption data and combine it with user profile information to construct a multiple linear regression baseline model, expressed by the following formula:

[0104] In the formula, This represents the baseline value for a user's gas consumption, where S is the apartment size, N is the number of family members, and T is the number of gas appliances. is the monthly average temperature compensation coefficient, and a, b, c, and d are model coefficients, obtained by fitting historical data using the least squares method.

[0105] The daily and weekly usage fluctuation thresholds for users are calculated based on a multiple linear regression baseline model.

[0106] Based on the calculation results, multi-dimensional usage reports are generated, including daily usage trend charts, weekly usage comparison tables, and monthly usage statistics. Warning levels are also assigned based on the percentage of daily usage deviating from the baseline, including: Level 1 alerts are set at 10%-20% deviation from baseline in daily usage, sending notifications via the app without manual intervention. Level 2 alerts are set at 20%-30% deviation in daily usage or 15%-25% deviation in weekly usage, triggering customer service to verify via phone and generating an anomaly analysis report (including possible causes such as minor equipment malfunctions or temporary changes in user gas consumption). Level 3 alerts are set at more than 30% deviation in daily usage or more than 25% deviation in weekly usage for more than 3 consecutive days, immediately dispatching maintenance personnel for on-site investigation and suspending remote billing to prevent user losses.

[0107] The meter reading data knowledge base is archived and stored according to a preset cycle, and the meter reading data knowledge base is iteratively updated using a federated learning algorithm. The smart gas meters are also remotely calibrated at a preset frequency.

[0108] The process of iteratively updating the meter reading data knowledge base using a federated learning algorithm includes: The meter reading data knowledge base of each regional edge computing node is used as a local dataset. The meter reading data knowledge base includes metering data, conflict handling records, and early warning information.

[0109] The initial conflict identification model and verification algorithm parameters are distributed to each edge computing node through the cloud platform.

[0110] The model is trained on each edge computing node using a local dataset, the gradient of the model parameter update is calculated, and the gradient of the model parameter update is encrypted using a homomorphic encryption algorithm before being uploaded to the cloud platform.

[0111] The cloud platform aggregates the update gradients uploaded by edge computing nodes, and the aggregation formula is expressed as:

[0112] In the formula, This represents the global gradient update amount obtained after aggregation, where M is the number of edge nodes. Let be the sample size of the i-th node. This represents the gradient update of the model parameters calculated by the i-th edge computing node based on its local samples.

[0113] The aggregated gradients are updated to the global model, and the above process is repeated until the model loss function converges. The updated global model parameters are then sent to each edge computing node and smart gas meter terminal.

[0114] The process of remotely calibrating a smart gas meter at a preset frequency includes: The gas meter sends a standard flow command at a preset frequency. After receiving the command, the gas meter operates at the standard flow rate and collects and feeds back the metering value.

[0115] The calibration deviation is calculated using the following formula:

[0116] In the formula, e represents the calibration deviation. This indicates the feedback measurement value. This indicates the standard flow command, and t represents the time during which the gas meter operates at the standard flow rate.

[0117] When the absolute value of the calibration deviation exceeds the preset deviation threshold, a parameter adjustment command is sent to the gas meter to correct the metering coefficient. The adjustment formula is as follows:

[0118] In the formula, This represents the adjusted measurement coefficient. This indicates the measurement coefficient before adjustment.

[0119] After adjustment, the gas meter is run again at the standard flow rate. New feedback metering values ​​are collected, and new deviations are calculated. When the absolute value of the new calibration deviation is less than the preset deviation threshold, the calibration is deemed qualified, a calibration report is generated, and it is uploaded to the cloud platform for archiving. The calibration report includes calibration time, standard flow rate, feedback value, deviation rate, and adjustment parameters.

[0120] In summary, this embodiment provides a remote meter reading method for smart gas meters based on the Internet of Things, with the following beneficial effects: By constructing a dual-source data multi-dimensional conflict identification system, it is possible to accurately identify data consistency risk groups from multiple dimensions such as time synchronization, device identification, and feature association, avoiding the problem of deviation from a single data source. The LSTM-Kalman filter hybrid model, which integrates the attention mechanism, can intelligently correct conflict data by combining historical time-series data and real-time feature weights, outputting reliable measurement data, reducing disputes caused by inaccurate measurement, and ensuring the reliability of measurement results.

[0121] Edge computing nodes are used to perform layered preprocessing of trusted metrological data. Wavelet packet transform is used to compress data and the isolated forest algorithm is used to filter out outliers, reducing the amount of data uploaded to the cloud and thus reducing network transmission and cloud storage pressure. An encrypted transmission channel based on the MQTT protocol and blockchain technology is used to ensure data transmission security through encryption and verification methods. Blockchain evidence storage enables data traceability and prevents data tampering.

[0122] By constructing a multiple linear regression baseline model of user gas consumption and combining it with daily and weekly consumption fluctuation thresholds to divide a three-level early warning mechanism, different response measures can be dynamically triggered according to the degree of abnormality, so as to prevent safety hazards in a timely manner and improve the user service experience.

[0123] The federated learning algorithm is used to iteratively update the meter reading data knowledge base. Each edge node only uploads the model gradient instead of the original data. This protects user data privacy while achieving global model collaborative optimization, balancing data privacy protection and algorithm iteration efficiency.

[0124] The calibration frequency is dynamically adjusted based on historical data conflicts in the gas meter, and intermediate calibration is added for high-frequency conflict devices to achieve on-demand calibration. By remotely sending standard flow commands and dynamically correcting the metering coefficient, device calibration can be completed without manual on-site visits, improving calibration efficiency, ensuring long-term metering accuracy of the gas meter, and reducing operation and maintenance costs.

[0125] Multi-dimensional usage report generation, meter reading data knowledge base archiving, and model iteration optimization provide gas companies with data support such as user gas consumption trends and equipment operating status, helping them optimize gas scheduling planning and accurately predict gas demand; users can query usage data and calibration records in real time, improving data transparency and promoting the transformation of gas services towards proactive prediction.

[0126] Based on the same general inventive concept, this invention also protects an IoT-based smart gas meter remote meter reading system. The IoT-based smart gas meter remote meter reading system provided by this invention will be described below. The IoT-based smart gas meter remote meter reading system described below can be referred to in correspondence with the IoT-based smart gas meter remote meter reading method described above.

[0127] Figure 2 This is a schematic diagram of the structure of the IoT-based smart gas meter remote meter reading system provided in an embodiment of the present invention.

[0128] like Figure 2 As shown, the IoT-based smart gas meter remote meter reading system includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The processor includes a data acquisition module, a conflict screening module, a data correction module, a data transmission module, an anomaly warning module, and an update calibration module.

[0129] The data acquisition module is used to collect local metering data and IoT transmission data from smart gas meters in real time.

[0130] The conflict screening module is used to build a multi-dimensional conflict identification system for dual-source data and to screen conflicting data groups that pose a risk of data inconsistency.

[0131] The data correction module is used to construct an LSTM-Kalman filter hybrid model with an integrated attention mechanism, to verify the accuracy of conflicting data sets, and to output corrected reliable measurement data.

[0132] The data transmission module is used to perform layered preprocessing of trusted metering data using edge computing nodes and establish an encrypted transmission channel to transmit the preprocessed trusted metering data to the remote meter reading cloud platform, thereby achieving traceability of the data transmission process.

[0133] The anomaly warning module is used to generate multi-dimensional user gas consumption reports based on preprocessed reliable metering data, and to build a user gas consumption baseline model. It triggers a graded warning mechanism for abnormal consumption data that deviates from the baseline.

[0134] The update calibration module is used to archive and store the meter reading data knowledge base according to a preset cycle, it uses a federated learning algorithm to iteratively update the meter reading data knowledge base, and remotely calibrates the smart gas meters at a preset frequency.

[0135] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A remote meter reading method for an intelligent gas meter based on the Internet of Things, characterized in that, The application relates to a method for realizing accurate and reliable data transmission of a smart gas meter. The method comprises the following steps: real-time acquisition of local metering data and Internet of Things transmission data of a smart gas meter; construction of a multi-dimensional conflict identification system for double-source data to screen conflict data groups with data consistency risks; construction of an LSTM-Kalman filter hybrid model with an attention mechanism to accurately verify the conflict data groups and output corrected reliable metering data; layered preprocessing of the reliable metering data by an edge computing node, establishment of an encrypted transmission channel, transmission of the preprocessed reliable metering data to a remote meter reading cloud platform, and realization of traceability of the data transmission process; generation of multi-dimensional user gas consumption reports according to the preprocessed reliable metering data, construction of a user gas consumption baseline model, triggering of a hierarchical early warning mechanism for abnormal consumption data deviating from the baseline; 2. The IoT-based smart gas meter remote reading method according to claim 1, characterized in that, archival storage of the meter reading data knowledge base according to a preset period, iterative updating of the meter reading data knowledge base by a federated learning algorithm, and remote calibration of the smart gas meter according to a preset frequency. 3.The IoT-based smart gas meter remote reading method according to claim 1, wherein, The local metering data comprises instantaneous flow, cumulative gas consumption, metering time stamp and temperature compensation coefficient of the gas meter; the Internet of Things transmission data comprises Internet of Things instantaneous flow, Internet of Things cumulative gas consumption, signal strength and data transmission time delay. The process of screening conflict data groups with data consistency risks comprises the following steps: calculation of the deviation of the local metering time stamp and the Internet of Things transmission time stamp, setting of a time synchronization threshold value, correction of the time stamp by a linear interpolation method when the deviation exceeds the time synchronization threshold value, and marking as a time conflict candidate group if the deviation after correction is still greater than the time synchronization threshold value; extraction of the gas meter equipment number in the local metering data and the equipment identifier in the Internet of Things transmission data, verification of consistency by a hash mapping, and direct marking as an equipment identifier conflict group if the consistency is not met; construction of double-source data feature vectors, calculation of the correlation degree of the two feature vectors by a cosine similarity algorithm; setting of a correlation degree threshold value, marking as a feature correlation conflict candidate group when the correlation degree is lower than the correlation degree threshold value; construction of a dynamic threshold value model based on historical non-conflict data, calculation of the cumulative gas consumption deviation rate and the instantaneous flow deviation; 4. The IoT based smart gas meter remote reading method as claimed in claim 1 wherein, calculation of the dynamic threshold value by a sliding window algorithm, determination of a confirmed conflict data group when the cumulative gas consumption deviation rate is greater than the cumulative gas consumption deviation dynamic threshold value and the instantaneous flow deviation is greater than the instantaneous flow deviation dynamic threshold value, and triggering of an accuracy verification process. The process of accurately verifying the conflict data groups by the LSTM-Kalman filter hybrid model with the attention mechanism comprises the following steps: division of the historical time series data of the confirmed conflict data group according to time steps, calculation of the weight of each feature in time series prediction, and input of each time step feature including instantaneous flow, cumulative gas consumption, signal strength and transmission time delay; setting of an LSTM network architecture, including a hidden layer and an output layer; the hidden layer comprises a plurality of neurons, and a ReLU activation function is used to model the weighted time series features; the output layer is used to output data trend prediction values; the data trend prediction values are used as prior estimates of Kalman filtering, and state equations and observation equations are established. The historical conflict data and the artificial labeled reliable data are used as a training set, and the mean square error of the predicted value and the labeled value is used as a loss function; The loss function is iteratively trained by an Adam optimizer until convergence, the mixed model is input into the conflict data set, and the corrected reliable metering data is output. When the prediction confidence of the model output reaches the preset confidence threshold, the verification result is determined to be reliable. 5.The IoT-based smart gas meter remote reading method according to claim 1, wherein, The process of hierarchical preprocessing of reliable metering data by edge computing nodes includes: The wavelet packet transform algorithm is used to quantize the high-frequency coefficients based on Stein unbiased risk estimation threshold; A standard data format is defined, including user ID, gas meter number, metering time, cumulative gas consumption, instantaneous flow, temperature compensation coefficient and data status code, and is packaged in JSON format to obtain standardized data; An isolation forest algorithm model is constructed to detect anomalies in the standardized data, calculate the anomaly score of each data point, and when the anomaly score is greater than a preset score threshold, determine it as an outlier, automatically exclude and mark it for manual review, and record the feature information of the abnormal data. 6.The IoT-based smart gas meter remote reading method according to claim 1, wherein, The process of establishing an encrypted transmission channel includes: A dual mechanism of username and password authentication and client certificate authentication is used, and the client certificate includes a unique gas meter device identifier, a validity period and public key information; The Payload field of the MQTT protocol is encrypted using the ECC-256 algorithm to generate an encrypted data packet; Calculate the data integrity check code and transmit it with the encrypted data packet; The edge computing node packages the meta information of the transmitted data into a blockchain transaction at a preset period, and submits it to the metering agency. The meta information includes user ID, data transmission time, check code and device identifier; The metering agency uses the PoS consensus mechanism, and each block contains the hash value of the previous block, realizing the non-tamperability and traceability of the data transmission process. 7.The IoT-based smart gas meter remote reading method according to claim 1, wherein, The process of triggering a hierarchical early warning mechanism for abnormal baseline deviation includes: Collect historical normal gas consumption data of users, combine user profile information, and construct a multivariate linear regression baseline model; Calculate the user's daily consumption fluctuation threshold and weekly consumption fluctuation threshold based on the multivariate linear regression baseline model; Generate a multi-dimensional consumption report based on the calculation results, including daily consumption trend chart, weekly consumption comparison table and monthly consumption statistics, and divide the early warning level according to the proportion of daily consumption deviation from the baseline. 8.The IoT-based smart gas meter remote reading method according to claim 1, wherein, The process of iteratively updating the meter reading data knowledge base using federated learning algorithm includes: The meter reading data knowledge base of each regional edge computing node is used as a local data set, and the meter reading data knowledge base includes metering data, conflict processing records and warning information; The cloud platform sends the initial conflict identification model and verification algorithm parameters to each edge computing node; Use the local data set to train the model of each edge computing node, calculate the model parameter update gradient, and upload it to the cloud platform after encryption by the homomorphic encryption algorithm; The cloud platform aggregates the update gradient uploaded by the edge computing nodes; The aggregated gradient is updated to the global model, and the process of calculating the model parameter update gradient and updating the gradient aggregation is repeated until the model loss function converges, and the updated global model parameters are distributed to each edge computing node and the intelligent gas meter terminal. 9.The IoT-based smart gas meter remote reading method according to claim 1, wherein, The process of remotely calibrating the intelligent gas meter at a preset frequency includes: sending a standard flow instruction to the gas meter at a preset frequency, the gas meter running according to the standard flow after receiving the instruction, and collecting feedback measurement values; calculating the calibration deviation according to the standard flow and the feedback measurement values; when the absolute value of the calibration deviation is greater than a preset deviation threshold, sending a parameter adjustment instruction to the gas meter to correct the measurement coefficient; running the adjusted gas meter again according to the standard flow, collecting new feedback measurement values, calculating a new deviation, and determining that the calibration is qualified when the absolute value of the new calibration deviation is less than the preset deviation threshold; generating a calibration report and uploading it to the cloud platform for archiving, the calibration report including calibration time, standard flow, feedback value, deviation rate and adjustment parameter.

10. An intelligent gas meter remote reading system based on Internet of Things, comprising a memory, a processor and a computer program stored on the memory and capable of running on the processor, characterized in that, The processor implements the remote meter reading method of the intelligent gas meter based on the Internet of Things as claimed in any one of claims 1 to 9 when executing the program, and the processor includes: a data acquisition module for acquiring local measurement data and Internet of Things transmission data of the intelligent gas meter in real time; a conflict screening module for constructing a multi-dimensional conflict identification system of double-source data to screen conflict data groups with data consistency risks; a data correction module for constructing an LSTM-Kalman filter hybrid model with a fusion attention mechanism to accurately verify the conflict data groups and output corrected reliable measurement data; a data transmission module for using edge computing nodes to perform hierarchical preprocessing on the reliable measurement data, establishing an encrypted transmission channel, transmitting the preprocessed reliable measurement data to a remote meter reading cloud platform, and realizing traceability of the data transmission process; an abnormality early warning module for generating a multi-dimensional user gas consumption report according to the preprocessed reliable measurement data, constructing a user gas consumption baseline model, and triggering a hierarchical early warning mechanism for abnormal consumption data deviating from the baseline; an update and calibration module for archiving and storing the meter reading data knowledge base at a preset period, iteratively updating the meter reading data knowledge base using a federated learning algorithm, and remotely calibrating the intelligent gas meter at a preset frequency.

Images