Charging pile on-line metering calibration verification method

Abstract

The invention relates to the technical field of charging pile on-line metering calibration, in particular to a charging pile on-line metering calibration verification method, which comprises the steps of constructing a charging pile network topology, dividing sub-networks, selecting reference equipment, generating a calibration correction coefficient, transmitting calibration parameters, carrying out localization processing and the like. And the consistency and long-term stability of the calibration state are ensured through a dynamic error transfer model and a distributed consensus algorithm. According to the method, efficient connection and data transmission of equipment in the region can be realized, calibration resource allocation is optimized, the accuracy and adaptability of calibration parameters are improved, the integrity and reliability of calibration information are ensured, meanwhile, high adaptation of the calibration parameters and equipment characteristics is realized, and the stability and reliability of the whole calibration system are ensured.

Description

Technical Field

[0001] This invention belongs to the field of electric vehicle charging technology and metrology testing technology, specifically a method for online metrology calibration and verification of charging piles. Background Art

[0002] With the rapid development of the new energy vehicle industry, charging piles, as a critical infrastructure, are being deployed on a large scale across the country. Their metrological accuracy directly impacts consumer rights protection and the fairness of electricity transactions, making charging pile metrological calibration and verification a core technical link in ensuring the healthy development of the industry. Currently, charging pile metrological calibration mainly relies on the traditional single-point, one-by-one verification model. This requires professional technicians to carry standard metrological equipment to independently calibrate each charging pile. This method not only consumes a large amount of manpower and resources but also results in a long verification cycle and high cost when dealing with a large and geographically distributed network of charging piles, making it difficult to adapt to the rapidly growing market demand.

[0003] In a large-scale charging pile network environment, the independent calibration mode of a single device fails to fully utilize the correlation between multiple devices within a region, resulting in low efficiency in calibration resource allocation. Once a charging pile in the area completes accurate calibration, its calibration data and parameters are difficult to effectively transmit to surrounding devices, leading to the isolation of standard metrological information. This information isolation further increases the technical difficulty of multi-device collaborative calibration, making it difficult to establish an effective calibration parameter sharing mechanism between adjacent charging piles, thus limiting the improvement of calibration efficiency for the entire charging pile group in the region. The lack of a unified networked calibration system makes it difficult to form distributed metrological standard transmission paths between charging pile devices, ultimately resulting in a persistently high overall verification workload. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned shortcomings and provide an online metering calibration and verification method for charging piles.

[0005] To achieve the above objectives, the specific solution of the present invention is as follows:

[0006] A method for online metering calibration and verification of charging piles may specifically include the following steps:

[0007] Construct a charging pile network topology and determine the optimal connection relationship based on device distance and communication capabilities;

[0008] Based on the charging pile network topology, a clustering algorithm is used to divide the charging pile sub-network, and the maximum topological diameter of the sub-network is limited to ensure the data interaction efficiency between devices within the sub-network.

[0009] Obtain the device's historical calibration records and select a reference device according to preset rules as the core node for subsequent calibration parameter transmission;

[0010] Based on the baseline device parameters, weighting coefficients are assigned according to the device runtime, and a weighted average algorithm is used to generate calibration correction coefficients for calibration adjustment of other devices in the sub-network.

[0011] The calibration parameters are transmitted wirelessly, and a retransmission mechanism is initiated when continuous transmission fails to ensure the integrity of the calibration information.

[0012] The received calibration parameters are localized by matching device features, and feature parameters such as temperature drift coefficient, voltage nonlinearity and clock synchronization deviation are parsed and extracted. The deviation value is calculated by error model, and the local calibration coefficient is iteratively optimized by gradient descent method.

[0013] A distributed consensus algorithm is used to synchronize the calibration status, and a dynamic error propagation model is established to evaluate the cumulative error. When accuracy degradation is detected, the parameter propagation depth is limited.

[0014] When the equipment error changes beyond the limit, recalibration is triggered, and the calibration deviation of the entire network is checked regularly to ensure the long-term stability and reliability of the overall calibration system.

[0015] As a further aspect of the present invention, the construction of the charging pile network topology is specifically as follows: based on the location distribution of charging pile devices within a geographical area, a distance matrix is ​​generated by calculating the Euclidean distance between devices, an initial network topology map is constructed by combining the signal transmission capability threshold, the minimum spanning tree algorithm is used to determine the optimal connection relationship, and then the data transmission path is optimized by the shortest path algorithm to ensure the efficient transmission of calibration information between devices.

[0016] As a further aspect of the present invention, the process of limiting the maximum topological diameter of the subnetwork includes: calculating the shortest path length of all node pairs in the subnetwork topology graph; when the maximum path length exceeds a set threshold, adjusting by adding relay nodes or splitting the subnetwork to ensure that the maximum topological diameter of the subnetwork meets the preset requirements.

[0017] As a further aspect of the present invention, the process of obtaining historical calibration records of the equipment includes: when executed for the first time, manually calibrating at least 10% of the equipment in the sub-network through a mobile calibration terminal to generate an initial historical record; calculating the comprehensive score of the equipment based on the average measurement error and continuous running time in the historical record, and selecting the equipment with the highest comprehensive score as the benchmark equipment; if the highest-scoring equipment does not meet the preset accuracy requirements, then triggering supplementary calibration and updating the historical record.

[0018] As a further aspect of the present invention, the process of generating the calibration correction coefficient is as follows: weighting coefficients are assigned based on the equipment running time, and a weighted average algorithm is used to generate the calibration correction coefficient; when the deviation of the compensation parameter exceeds a preset threshold, the measurement data is re-extracted and iteratively calculated to ensure the accuracy of the calibration correction coefficient.

[0019] As a further aspect of the present invention, the process of transmitting calibration parameters for wireless communication includes: when the number of times no transmission confirmation signal is received within a preset time window exceeds a threshold, it is determined as a continuous transmission failure, and a retransmission mechanism is initiated to ensure the integrity of the calibration information.

[0020] As a further aspect of the present invention, the localization process for device feature matching includes: parsing a set of device feature parameters from calibration parameters, including at least temperature drift coefficient, voltage nonlinearity, and clock synchronization deviation; calculating the deviation value using a pre-built error model, with the formula: ,

[0021] Where k1, k2, and k3 are error coefficients, and ΔT, ΔV, and Δt represent the changes in temperature drift, voltage nonlinearity, and clock synchronization deviation, respectively.

[0022] When δ exceeds the set threshold, the local calibration coefficient is iteratively optimized using the gradient descent method to ensure a high degree of fit between the calibration parameters and the device characteristics.

[0023] As a further aspect of the present invention, the specific implementation of the distributed consensus algorithm includes: establishing a dynamic error propagation model. ,

[0024] in, Accumulate the error at the current moment; This is the new error introduced in this transmission; α and β are the historical attenuation coefficients;

[0025] When cumulative error At this time: terminate the current link transmission and start the reverse verification process to directly compare the parameters of the end device with the original parameters of the reference device;

[0026] If the reverse verification deviation exceeds the preset tolerance, the link is marked as a failed path and the network topology is reconstructed.

[0027] As a further aspect of the present invention, the reverse verification process includes: comparing the measurement data of the reference device and the end device at the same time using timestamp alignment technology, and calculating the absolute deviation value. ,

[0028] Among them, V b and I b V represents the voltage and current values ​​of the reference device, respectively. e and I e These represent the voltage and current values ​​of the terminal device, respectively. When If the result is greater than 0.5%, the verification is considered to have failed.

[0029] As a further aspect of the present invention, the distributed consensus algorithm employs an asynchronous Byzantine fault-tolerant algorithm to ensure the consistency and reliability of the calibration state in a complex network environment.

[0030] As a further aspect of the present invention, the process for determining when the equipment error change exceeds the limit includes: triggering a recalibration mechanism when the equipment error change exceeds a preset threshold; periodically verifying the calibration deviation of the entire network to ensure the long-term stability and reliability of the overall calibration system.

[0031] As a further aspect of the present invention, the triggering conditions of the recalibration mechanism include: automatically triggering the recalibration procedure and updating historical calibration records when the equipment error changes beyond a preset threshold; periodically verifying the calibration deviation of the entire network, combining historical data and real-time data, and quantitatively predicting the future trend of calibration parameter changes through statistical modeling to ensure the long-term stability of the overall calibration system.

[0032] The technical advantages of the online metering calibration and verification method for charging piles of the present invention include: achieving efficient connection and data transmission between devices within the region by constructing a charging pile network topology; optimizing the allocation efficiency of calibration resources by dividing the network into sub-networks through clustering algorithms; improving the accuracy and adaptability of calibration parameters by selecting reference devices and generating calibration correction coefficients; ensuring the integrity and reliability of calibration information through wireless communication transmission and retransmission mechanisms; achieving a high degree of adaptation between calibration parameters and device characteristics through localized processing of device feature matching; and ensuring the consistency and long-term stability of calibration status through distributed consensus algorithms and dynamic error propagation models. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating the online metering calibration and verification method for charging piles in an embodiment of the present invention, showing the overall process from network topology construction to calibration parameter transmission and error verification. Detailed Implementation

[0034] This invention provides an online metering calibration and verification method for charging piles, the specific implementation of which is combined with... Figure 1 The flowchart shown below provides a detailed explanation. In practical applications, this method achieves online metrological calibration and verification of charging pile equipment through a series of steps, ensuring the efficiency and reliability of the entire calibration system.

[0035] First, in constructing the charging pile network topology, it is necessary to clarify the location distribution relationships between the various charging pile devices. This is done by obtaining the coordinate information of all charging pile devices within a geographical area and calculating the Euclidean distances between the devices to generate a distance matrix. The distance matrix between devices is then calculated as follows: ,

[0036] Among them, (x i ,y i Let (x) be the geographic coordinates of device i, and (x) be the geographic coordinates of device i. j ,y j Let be the geographical coordinates of device j. Each element in the distance matrix represents the straight-line distance between two devices; this distance value is the foundational data for subsequent network topology construction. Subsequently, the communication capabilities between devices are evaluated based on signal transmission capacity thresholds, and devices meeting the communication conditions are connected to form an initial network topology. Nodes in the initial network topology represent charging pile devices, and edges represent the communication connections between devices. To determine the optimal connection relationships, the minimum spanning tree algorithm is used to optimize the initial network topology, generating a minimum spanning tree. The minimum spanning tree ensures that the total length of the communication paths between all devices is minimized, thereby improving data transmission efficiency. Furthermore, the paths in the minimum spanning tree are optimized using a shortest path algorithm to ensure that calibration information can be efficiently transmitted between devices. At this point, the construction of the charging pile network topology is complete.

[0037] Next, based on the constructed charging pile network topology, a clustering algorithm is used to divide the charging pile devices into sub-networks and limit the maximum topological diameter of each sub-network. Specifically, the initial sub-network division uses the Agglomerative Hierarchical Clustering (AGNES) algorithm, generating initial sub-networks based on the Euclidean distance between devices. This includes: initializing each device as an independent cluster; iteratively merging the two closest clusters, with the inter-cluster distance calculated using a fully connected method. ,

[0038] Among them, among them, C represents the inter-cluster distance function. P Let C represent the p-th sub-network cluster. q Let i represent the q-th sub-network cluster, and let i represent the cluster C. P Within a cluster (charging pile device), j represents the cluster C. q Internal nodes (charging pile equipment);

[0039] When any inter-cluster distance > preset radius threshold R th The merging process stops when the clustering results are reached; the clustering results are output as the initial subnetwork division. The specific implementation process for limiting the maximum topological diameter of a subnetwork includes calculating the shortest path length for all node pairs in the subnetwork topology graph, and determining whether the subnetwork structure needs adjustment based on whether the maximum path length exceeds a set threshold. If the maximum path length of a subnetwork exceeds the preset requirement, adjustments are made by adding relay nodes or splitting the subnetwork to ensure that the maximum topological diameter within the subnetwork meets the design requirements. After the subnetwork division is completed, the number and distribution range of devices within each subnetwork are effectively controlled, thereby improving the efficiency of calibration resource allocation.

[0040] Based on the subnetwork partitioning, a reference device needs to be selected as the core node for calibration parameter transmission. The selection of the reference device is based on its historical calibration records. When the calibration process is executed for the first time, at least 10% of the devices within the subnetwork are manually calibrated using a mobile calibration terminal, generating initial historical records. Subsequently, a comprehensive score for each device is calculated based on the average measurement error and continuous operating time in the historical records, and the device with the highest comprehensive score is selected as the reference device. If the device with the highest score does not meet the preset accuracy requirements, supplementary calibration is triggered and the historical records are updated until a suitable reference device is selected. The reference device plays a crucial role throughout the calibration process, and its calibration parameters will serve as a reference for the calibration and adjustment of other devices.

[0041] Based on the parameters of the reference device, a weighted average algorithm is used to generate calibration correction coefficients. The specific implementation process of the weighted average algorithm is as follows: First, weight coefficients are assigned according to the device's runtime; devices with longer runtimes have higher weights, weakening the impact of the instability of new devices and strengthening the decision-making weight of reliable devices. Then, the calibration parameters of the reference device are combined with the operating data of other devices in the sub-network, and the calibration correction coefficients are calculated using the weighted average algorithm. When the deviation of the compensation parameters exceeds a preset threshold, the measurement data is re-extracted and iteratively calculated to ensure the accuracy of the calibration correction coefficients. After the calibration correction coefficients are generated, they are transmitted to other devices in the sub-network via a wireless communication module. During transmission, if the number of times a transmission acknowledgment signal is not received within a preset time window exceeds a threshold, the wireless communication module determines it as a continuous transmission failure and initiates a retransmission mechanism to ensure the integrity of the calibration information.

[0042] The device receiving the calibration correction coefficients needs to undergo localization processing to adapt them to its own characteristic parameters. The localization processing module is responsible for parsing the set of device characteristic parameters in the calibration parameters, which includes at least key parameters such as temperature drift coefficient, voltage nonlinearity, and clock synchronization deviation.

[0043] The deviation value is calculated using a pre-built error model. : ,

[0044] Where k1, k2, and k3 are error coefficients, and ΔT, ΔV, and Δt represent the changes in temperature drift, voltage nonlinearity, and clock synchronization deviation, respectively. When the deviation value δ exceeds a set threshold, the local calibration coefficients are iteratively optimized using the gradient descent method to ensure a high degree of fit between the calibration parameters and the device characteristics. The output of the localization processing module is directly applied to the calibration adjustment of the device, thereby improving calibration accuracy.

[0045] To ensure the consistency and long-term stability of the calibration status, a distributed consensus algorithm is used to synchronize the calibration status. The specific implementation of the distributed consensus algorithm includes establishing a dynamic error propagation model, the mathematical expression of which is: ,

[0046] in, Accumulate the error at the current moment; This introduces a new error in this transmission; α and β are historical attenuation coefficients, with values ​​ranging from 0.8 ≤ α ≤ 0.95 and 0.05 ≤ β ≤ 0.2, respectively.

[0047] When the cumulative error E t Exceeding the preset threshold E th When this happens, the current link transmission is terminated, and the reverse verification process is initiated. The reverse verification process compares the metering data of the reference device and the end device at the same time using timestamp alignment technology to calculate the absolute deviation value. ,

[0048] Among them, V b and I b V represents the voltage and current values ​​of the reference device, respectively. e and I e These represent the voltage and current values ​​of the terminal device, respectively. When If the error rate is greater than 0.5%, the verification is deemed to have failed, and the link is marked as a failed path. Subsequently, the network topology is reconstructed to restore the normal operation of the calibration system.

[0049] Furthermore, when equipment error changes exceed a preset threshold, a recalibration mechanism is automatically triggered. The recalibration process includes recalibrating equipment parameters and updating historical calibration records. Regularly verifying network-wide calibration deviations and combining historical and real-time data, statistical modeling is used to quantitatively predict future trends in calibration parameters, thereby ensuring the long-term stability and reliability of the overall calibration system.

[0050] In the above implementation, the collaborative relationships between the steps are as follows: Charging pile devices form a network topology through communication links; the initial network topology graph and the minimum spanning tree jointly determine the optimal connection relationship between devices; sub-network partitioning uses a clustering algorithm to group devices, ensuring the reasonable allocation of calibration resources; the selection of reference devices relies on historical calibration records, and their calibration parameters are transmitted to other devices via a wireless communication module; the localization processing module parses and optimizes the received calibration parameters, ensuring that the parameters match the device characteristics; the dynamic error propagation model and the reverse verification process jointly guarantee the consistency and reliability of the calibration status. The collaboration between these steps realizes the overall process from network construction to calibration parameter transmission and error verification, ensuring the efficiency and accuracy of the online metering calibration and verification method for charging piles.

[0051] To enable those skilled in the art to fully understand and implement this invention, the specific implementation principle of this invention is further explained below in conjunction with a specific application scenario.

[0052] In a city's new energy vehicle charging network, hundreds of charging stations are distributed across an area, interconnected via wireless communication modules. First, technicians obtain the coordinates of all charging stations using a geographic information system (GIS) and calculate the Euclidean distances between them to generate a distance matrix. Then, the communication capabilities between the stations are evaluated based on signal transmission capacity thresholds, and stations meeting the communication requirements are connected to form an initial network topology. During this process, a minimum spanning tree algorithm is used to optimize the initial network topology, generating a minimum spanning tree to ensure the shortest total length of communication paths between all stations. The paths in the minimum spanning tree are further optimized using a shortest path algorithm, thereby improving the efficiency of calibration information transmission between stations. At this point, the charging station network topology is complete.

[0053] Next, based on the constructed charging pile network topology, the charging pile devices are divided into sub-networks using a clustering algorithm, and the maximum topological diameter of each sub-network is limited. First, a preliminary network partitioning of the charging pile devices is performed using the Aggregate Hierarchical Clustering (AGNES) algorithm. This initial sub-network is generated based on the Euclidean distance between devices, specifically including: initializing each device as an independent cluster; iteratively merging the two closest clusters; and calculating the inter-cluster distance using a fully connected method. ,

[0054] in, C represents the inter-cluster distance function. P Let C represent the p-th sub-network cluster. q Let i represent the q-th sub-network cluster, and let i represent the cluster C. P Within a cluster (charging pile device), j represents the cluster C. q Internal nodes (charging pile equipment);

[0055] When the distance between any clusters is greater than the preset radius threshold R th The merging process stops when the time is right; the clustering results are output as the initial subnetwork division; the shortest path length of all node pairs in the subnetwork topology graph is calculated. If the maximum path length of a subnetwork exceeds the preset requirement, adjustments are made by adding relay nodes or splitting the subnetwork to ensure that the maximum topological diameter within the subnetwork meets the design requirements. For example, if the maximum path length in a subnetwork exceeds 500 meters, a relay node is added to shorten the path length, thereby improving data exchange efficiency. After the subnetwork division is completed, the number and distribution range of devices within each subnetwork are effectively controlled, improving the allocation efficiency of calibration resources.

[0056] Based on the sub-network division, a reference device is selected as the core node for calibration parameter transmission. During the initial calibration, at least 10% of the devices within the sub-network are manually calibrated using a mobile calibration terminal, generating initial historical records. Subsequently, a comprehensive score for each device is calculated based on the average measurement error and continuous operating time in the historical records, and the device with the highest comprehensive score is selected as the reference device. For example, if device A in a sub-network has a comprehensive score of 95, higher than other devices, then device A is selected as the reference device. If the highest-scoring device does not meet the preset accuracy requirements, supplementary calibration is triggered and the historical records are updated until a suitable reference device is selected. The calibration parameters of the reference device will serve as a reference for the calibration and adjustment of other devices.

[0057] Based on the parameters of the reference equipment, a weighted average algorithm is used to generate calibration correction coefficients. In practice, weighting coefficients are assigned according to the equipment's runtime, with longer runtime equipment receiving higher weights. For example, if equipment B has a runtime of 1000 hours and equipment C has a runtime of 500 hours, then equipment B's weighting coefficient is higher than equipment C's. The calibration parameters of the reference equipment are combined with the runtime data of other equipment within the subnetwork, and the calibration correction coefficients are calculated using a weighted average algorithm. When the deviation of the compensation parameters exceeds a preset threshold, the measurement data is re-extracted and iteratively calculated to ensure the accuracy of the calibration correction coefficients. After the calibration correction coefficients are generated, they are transmitted to other equipment within the subnetwork via a wireless communication module. If the number of times a transmission acknowledgment signal is not received within a preset time window exceeds a threshold, it is considered a continuous transmission failure, and a retransmission mechanism is initiated to ensure the integrity of the calibration information.

[0058] The device receiving the calibration correction coefficients needs to undergo localization processing to adapt them to its own characteristic parameters. The localization processing module is responsible for parsing the device characteristic parameter set from the calibration parameters, including key parameters such as temperature drift coefficient, voltage nonlinearity, and clock synchronization deviation. The deviation value δ is calculated using a pre-built error model, with the formula: Where k1, k2, and k3 are error coefficients, and ΔT, ΔV, and Δt represent the changes in temperature drift, voltage nonlinearity, and clock synchronization deviation, respectively. For example, if the temperature drift change ΔT of a device is 0.5℃, the voltage nonlinearity change ΔV is 0.1V, and the clock synchronization deviation change Δt is 0.01 seconds, then the deviation value δ can be calculated using the above formula. When the deviation value δ exceeds a set threshold, the gradient descent method is used to iteratively optimize the local calibration coefficients to ensure a high degree of fit between the calibration parameters and the device characteristics. The output of the localization processing module is directly applied to the calibration adjustment of the device, thereby improving calibration accuracy.

[0059] To ensure the consistency and long-term stability of the calibration status, a distributed consensus algorithm is used to synchronize the calibration status. The mathematical expression of the dynamic error propagation model is as follows: ,in, Accumulate the error at the current moment; This represents the new error added in this transmission; α and β are historical attenuation coefficients, with values ​​ranging from 0.8 ≤ α ≤ 0.95 and 0.05 ≤ β ≤ 0.2, respectively; for example, the cumulative error E at the current moment... t The new error δ is 0.3%. t If the cumulative error is 0.05%, then the cumulative error at the next time step can be calculated using the formula above. When the cumulative error E t Exceeding the preset threshold E th When this occurs, the current link transmission is terminated, and the reverse verification process is initiated. The reverse verification process compares the metering data of the reference device and the end device at the same time using timestamp alignment technology to calculate the absolute deviation value. ,

[0060] Where V b and I b V represents the voltage and current values ​​of the reference device, respectively. e and I e These represent the voltage and current values ​​of the terminal device, respectively. For example, the voltage value V of the reference device. b The voltage value V of the terminal device is 220V. e If the voltage is 218V, then the voltage deviation is |(220-218) / 220|×100%=0.91%. When Δ>0.5%, the calibration is considered to have failed, and the link is marked as a failed path. Subsequently, the network topology is reconstructed to restore the normal operation of the calibration system.

[0061] Furthermore, a recalibration mechanism is automatically triggered when equipment error changes exceed a preset threshold. For example, if the error change of a device exceeds 1%, a recalibration procedure is triggered to update historical calibration records. The entire network's calibration deviation is periodically checked, and by combining historical and real-time data, statistical modeling is used to quantitatively predict future trends in calibration parameters, thereby ensuring the long-term stability and reliability of the overall calibration system.

[0062] In the above implementation, the collaborative relationships between the steps are as follows: Charging pile devices form a network topology through communication links; the initial network topology graph and the minimum spanning tree jointly determine the optimal connection relationship between devices; sub-network partitioning uses a clustering algorithm to group devices, ensuring the reasonable allocation of calibration resources; the selection of reference devices relies on historical calibration records, and their calibration parameters are transmitted to other devices via a wireless communication module; the localization processing module parses and optimizes the received calibration parameters, ensuring that the parameters match the device characteristics; the dynamic error propagation model and the reverse verification process jointly guarantee the consistency and reliability of the calibration status. The collaboration between these steps realizes the overall process from network construction to calibration parameter transmission and error verification, ensuring the efficiency and accuracy of the online metering calibration and verification method for charging piles.

[0063] The above description is only a preferred embodiment of the present invention. Therefore, any equivalent changes or modifications made to the structure, features and principles described in the claims of this patent application are included within the protection scope of this patent application.

Claims

1. A method for online metering calibration and verification of charging piles, characterized in that, Includes the following steps: Construct a charging pile network topology and determine the optimal connection relationship based on device distance and communication capabilities; The charging pile sub-network is divided using a clustering algorithm, and the maximum topological diameter of the sub-network is limited. Obtain the device's historical calibration records and select the reference device according to preset rules; Calculate the calibration correction coefficients for devices within the sub-network based on the reference device parameters; The calibration parameters are transmitted wirelessly, and a retransmission mechanism is activated when continuous transmission fails. The received calibration parameters are localized by matching device characteristics. A distributed consensus algorithm is used to synchronize the calibration status, and the parameter propagation depth is limited when accuracy degradation is detected. When the equipment error changes beyond the limit, a recalibration is triggered, and the calibration deviation of the entire network is checked periodically.

2. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The construction of the charging pile network topology is specifically as follows: Based on the location distribution of charging piles within a geographical area, a distance matrix is ​​generated by calculating the Euclidean distance between devices. An initial network topology is constructed by combining the signal transmission capacity threshold. The minimum spanning tree algorithm is used to determine the optimal connection relationship. Subsequently, the shortest path algorithm is used to optimize the data transmission path to ensure the efficient transmission of calibration information between devices.

3. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The process of limiting the maximum topological diameter of the subnetwork includes: Calculate the shortest path length for all node pairs in the subnetwork topology. When the maximum path length exceeds a set threshold, adjust by adding relay nodes or splitting the subnetwork to ensure that the maximum topological diameter of the subnetwork meets the preset requirements.

4. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The process of obtaining historical calibration records of the device includes: Upon initial execution, at least 10% of the devices within the subnetwork are manually calibrated using a mobile calibration terminal to generate an initial historical record. The comprehensive score of the equipment is calculated based on the average measurement error and continuous operating time in historical records, and the equipment with the highest comprehensive score is selected as the benchmark equipment. If the highest-scoring device does not meet the preset accuracy requirements, a supplementary calibration will be triggered and the historical records will be updated.

5. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The specific process for generating the calibration correction coefficient is as follows: Weighting coefficients are assigned based on equipment runtime, and a weighted average algorithm is used to generate calibration correction coefficients. When the deviation of the compensation parameter exceeds the preset threshold, the measurement data is re-extracted and iteratively calculated to ensure the accuracy of the calibration correction coefficients.

6. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The process of calibrating the wireless communication transmission parameters includes: If the number of times no transmission confirmation signal is received within the preset time window exceeds a threshold, it is determined as a continuous transmission failure, and a retransmission mechanism is initiated to ensure the integrity of the calibration information.

7. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The localization process for device feature matching includes: The set of characteristic parameters of the equipment is extracted from the calibration parameters, including at least the temperature drift coefficient, voltage nonlinearity, and clock synchronization deviation; The deviation value is calculated using a pre-built error model, using the following formula: ， Where k1, k2, and k3 are error coefficients, and ΔT, ΔV, and Δt represent the changes in temperature drift, voltage nonlinearity, and clock synchronization deviation, respectively. When the deviation value δ exceeds the set threshold, the local calibration coefficient is iteratively optimized using the gradient descent method.

8. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The specific implementation of the distributed consensus algorithm includes: A dynamic error propagation model is established, and its mathematical expression is as follows: ， in, Accumulate the error at the current moment; This is the new error introduced in this transmission; α and β are the historical attenuation coefficients; When cumulative error When: Terminate the current link transmission and start the reverse verification process.

9. The online metering calibration and verification method for charging piles according to claim 8, characterized in that, The reverse verification process includes: The absolute deviation value is calculated by comparing the measurement data of the reference device and the end device at the same time using timestamp alignment technology. The formula is as follows: ， Among them, V b and I b V represents the voltage and current values ​​of the reference device, respectively. e and I e These represent the voltage and current values ​​of the terminal device, respectively. The verification is considered to have failed when Δ is greater than 0.5%.

10. The online metering calibration and verification method for charging piles according to claim 1, characterized in that, The triggering conditions for the recalibration mechanism include: When the equipment error changes beyond a preset threshold, a recalibration procedure is automatically triggered, and historical calibration records are updated. Regularly verify the calibration deviation across the entire network, and combine historical and real-time data to quantitatively predict future trends in calibration parameters through statistical modeling, thereby ensuring the long-term stability of the overall calibration system.

Images