Blockchain-based charging pile multi-source data collaborative traceability and auditing method and system

Abstract

A blockchain-based charging pile multi-source data collaborative traceability and auditing method and system. The hardware gene identifier of the charging pile is obtained; and the multi-source detection data of the charging pile is collected; the time domain characteristic parameters and the frequency domain characteristic parameters are extracted based on the charging waveform data, the extracted characteristic parameters, the hardware gene identifier and the charging space-time information are nonlinearly coupled to generate dynamic data fingerprints; double-layer nested hash operation is performed on the dynamic data fingerprints and the multi-source detection data to generate evidence hash and digital signature to generate evidence data package; further, multi-dimensional cross verification is performed on the evidence data package in parallel; the evidence data package that passes the multi-dimensional cross verification is packed and chained for storage; based on the verification results obtained by the multi-dimensional cross verification and the historical evidence data of the charging pile, the dynamic credit score of the charging pile is calculated and updated; according to the hierarchical auditing strategy, the data generated subsequently by the charging pile is audited. The present application improves the efficiency and accuracy of charging pile detection.

Description

Technical Field

[0001] This invention belongs to the field of charging pile testing technology, specifically involving a blockchain-based method and system for collaborative traceability and auditing of multi-source data of charging piles. Background Technology

[0002] With the rapid development of the new energy vehicle industry, the operational safety and data reliability of charging piles, as critical infrastructure, are becoming increasingly prominent issues. Existing charging pile data management methods have obvious technical shortcomings, mainly reflected in the fact that traditional solutions focus on the storage of charging transaction data, lacking comprehensive collection and integration of multi-source detection data such as charging waveforms, equipment status, and environmental safety, making it difficult to truly reflect the equipment health and safety risks during the charging process.

[0003] Patent application CN120116783A, through multi-source data acquisition, dynamic fault threshold model construction, and multi-level fault detection and prediction mechanisms, achieves real-time and accurate monitoring of charging pile operation status and fault early warning. Patent application CN121524428A, through a three-dimensional data fusion architecture, edge computing fuzzy processing, digital twin modeling, and blockchain cross-chain storage, achieves multi-source data fusion, privacy protection, and trusted recommendations.

[0004] However, existing blockchain evidence storage technology can only guarantee the integrity of data after it is uploaded to the chain, and cannot verify the authenticity of data before it is uploaded. Furthermore, single data sources are vulnerable to tampering or cloning attacks. Current auditing methods are mostly passive, post-event tracing, applying a uniform auditing strategy to all transactions, resulting in wasted computing resources and low efficiency, and failing to achieve dynamic and accurate anomaly detection and accountability. Therefore, there is an urgent need for a blockchain-based collaborative traceability and auditing method for multi-source detection data in charging piles to address the data trust crisis and security supervision challenges in charging pile operation. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a blockchain-based method and system for collaborative traceability and auditing of multi-source data for charging piles.

[0006] The first aspect of this invention discloses a blockchain-based method for collaborative traceability and auditing of multi-source data for charging piles, employing the following technical solution: Obtain the hardware genetic identifier of the charging pile; and collect multi-source detection data of the charging pile, including charging waveform data, equipment operating status data, environmental safety data, and charging transaction data; Based on the charging waveform data, time-domain and frequency-domain feature parameters are extracted; and the hardware gene identifier, time-domain feature parameters, and frequency-domain feature parameters are nonlinearly coupled with the charging spatiotemporal information to generate a dynamic data fingerprint within the current charging cycle; the charging spatiotemporal information includes the charging start time, charging end time, and the geographical coordinates of the charging pile. Based on the dynamic data fingerprint and the multi-source detection data, a double-layer nested hash operation is performed to generate a storage hash and perform digital signature to generate a storage data package; the storage data package is uploaded to the blockchain network, and the smart contract of the blockchain network performs multi-dimensional cross-validation on the storage data package in parallel; if the multi-dimensional cross-validation passes, the data contained in the storage data package is packaged and stored on the blockchain; the multi-dimensional cross-validation includes electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification; Based on the verification results obtained from the multi-dimensional cross-validation and the historical evidence data of the charging pile, the dynamic credit score of the charging pile is calculated and updated; according to the dynamic credit score, a hierarchical audit strategy matching the dynamic credit score is determined to audit the data subsequently generated by the charging pile.

[0007] Furthermore, the nonlinear coupling of hardware gene identifiers, time-domain feature parameters, frequency-domain feature parameters, and charging spatiotemporal information includes: The hardware gene identifier, the time-domain feature parameters, the frequency-domain feature parameters, and the charging spatiotemporal information constitute a combined feature vector, and the combined feature vector is normalized. The hardware gene identifier is mapped to the initial value of a nonlinear recursive algorithm; Based on the initial value, the normalized combined feature vector is input as a parameter into the nonlinear recursive algorithm for iteration to generate a pseudo-random sequence. Each state value of the pseudo-random sequence is converted into a binary bit, and each component of the combined feature vector is converted into a binary bit. The two are then XORed bit by bit to generate the dynamic data fingerprint.

[0008] Furthermore, a double-layer nested hash operation is performed based on the dynamic data fingerprint and the multi-source detection data to generate a proof-of-existence hash, including: Perform a first hash operation on each type of data in the multi-source detection data to generate multiple corresponding first-level evidence storage hashes; Multiple first-level evidence hashes are concatenated and combined with the dynamic data fingerprint, and a second hash operation is performed on the concatenated and combined data to generate a second-level evidence hash.

[0009] Furthermore, the electrical data consistency verification includes: It obtains the power consumption from the power monitoring equipment at the charging pile and the charging power reported by the vehicle-side battery management system. The power consumption is compared with the charging power, and the electrical data is judged to be consistent based on the comparison result and the error threshold. The error threshold is calculated by weighting the base error threshold, the credit threshold narrowing rate, the environmental compensation coefficient, and the power adjustment error threshold. The credit threshold narrowing rate is calculated based on the credit score of the charging pile in the previous cycle within the current charging cycle.

[0010] Furthermore, the device status baseline verification includes: Obtain the normal operation data of the charging pile within a preset historical period; Based on the normal operation data, a baseline model of the equipment health of the charging pile is constructed; Obtain device operating status data for the current charging cycle; Calculate the degree of deviation between the device operating status data of the current charging cycle and the device health baseline model; Based on the degree of deviation and the preset state deviation threshold, it is determined whether the equipment operating status of the charging pile is normal. The degree of deviation is the Mahalanobis distance between the device operating status data of the current charging cycle and the device health baseline model; The equipment health baseline model is a probability distribution model obtained by clustering the historical normal operation data of the charging pile based on a Gaussian mixture model.

[0011] Furthermore, the verification of the rationality of the spatiotemporal trajectory includes: Obtain the geographical location of the charging station; During the charging period, obtain the vehicle's geographical location; Calculate the spatial distance between the geographical location of the charging pile and the geographical location of the vehicle; Based on the spatial distance and the preset spatial error threshold, it is determined whether the vehicle position and the charging pile position meet the spatial consistency requirement.

[0012] Furthermore, based on the verification results and the historical evidence data of the charging pile, the dynamic credit score of the charging pile is calculated and updated, including: The historical pass rate of the consistency verification of the equipment operation status data, environmental safety data, and electrical data of the charging pile in the current charging cycle, as well as the authenticity evaluation coefficient obtained based on the last audit, are combined into an input feature vector; The input feature vector is fed into a trained credit risk prediction model to obtain the credit risk prediction result of the charging pile in the current charging cycle. The trained credit risk prediction model is a logistic regression model obtained through collaborative training based on a federated learning framework, used to predict the probability that the charging data is high-risk or low-risk. The dynamic credit score of the charging pile is updated based on the credit score of the previous cycle of the current charging cycle, the data quality contribution of the charging pile, and the audit score obtained based on the verification results. Based on the credit risk prediction results, adjust the calculation weight or value of the data quality contribution.

[0013] Furthermore, the data quality contribution of the charging pile includes: In the federated learning framework aggregation, the calculation includes charging pile nodes. Time performance ; Calculate the removal of the charging pile node Post-repolymerization performance ; The data quality contribution of the charging pile is, in terms of computing performance. With performance The difference is the value of the charging pile node. Data quality contribution .

[0014] Furthermore, the authenticity evaluation coefficient is obtained by weighting the scores obtained from electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification. The audit score obtained from the verification results is calculated by the average of the authenticity assessment coefficients of all charging transactions within the current charging cycle.

[0015] Further, determining a tiered audit strategy that matches the dynamic credit score includes: If the dynamic credit score is greater than or equal to the first preset credit score value, it is matched as a high credit level, and a sampling audit strategy is adopted for the data subsequently generated by the charging pile. If the dynamic credit score is greater than or equal to the second preset credit score value and less than the first preset credit score value, it is matched as a medium credit level, and a periodic audit strategy is adopted for the data subsequently generated by the charging pile. If the dynamic credit score is less than the second preset credit score value, it is matched as a low credit level, and a full audit strategy is adopted for the data subsequently generated by the charging pile.

[0016] Furthermore, including: When an anomaly is detected or an audit request is received, the full chain of evidence corresponding to the anomaly data is obtained from the blockchain. Based on the complete chain of evidence, the data tampering point or equipment failure point can be located; Based on the data tampering point or equipment failure point, and in conjunction with the preset responsibility determination rules, the party responsible for the data anomaly is determined, and the responsibility determination result is written into the blockchain.

[0017] The second aspect of this invention discloses a blockchain-based multi-source data collaborative traceability and auditing system for charging piles, which implements the technical solution of the multi-source data collaborative traceability and auditing method for charging piles as described in the first aspect of this invention. The system includes: Data acquisition module: acquires the hardware genetic identifier of the charging pile; and collects multi-source detection data of the charging pile, including charging waveform data, equipment operating status data, environmental safety data and charging transaction data; The fingerprint generation module extracts time-domain and frequency-domain feature parameters based on the charging waveform data; and nonlinearly couples the hardware gene identifier, time-domain feature parameters, and frequency-domain feature parameters with the charging spatiotemporal information to generate a dynamic data fingerprint for the current charging cycle; the charging spatiotemporal information includes the charging start time, charging end time, and the geographical coordinates of the charging pile. On-chain evidence storage module: Based on the dynamic data fingerprint and the multi-source detection data, a double-layer nested hash operation is performed to generate an evidence storage hash and perform digital signature to generate an evidence storage data package; the evidence storage data package is uploaded to the blockchain network, and the smart contract of the blockchain network performs multi-dimensional cross-validation on the evidence storage data package in parallel; if the multi-dimensional cross-validation passes, the data contained in the evidence storage data package is packaged and stored on the blockchain; the multi-dimensional cross-validation includes electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification; Credit assessment and tiered audit module: Based on the verification results obtained from the multi-dimensional cross-validation and the historical evidence data of the charging pile, calculate and update the dynamic credit score of the charging pile; according to the dynamic credit score, determine a tiered audit strategy matching the dynamic credit score to audit the data subsequently generated by the charging pile. The beneficial effect of this invention is that, compared with the prior art, (1) This invention extracts the output value of the physically unclonable function circuit on the charging pile control motherboard as a hardware gene identifier, and uniquely binds the dynamic data fingerprint to the specific physical device. This identifier is generated based on the inherent process deviation generated during chip manufacturing and has the characteristics of being unclonable and untamperable, which can effectively prevent device cloning attacks and motherboard replacement behavior. (2) The present invention uses a nonlinear parameter iteration sequence generation algorithm to nonlinearly weight couple hardware gene identifiers, multi-dimensional feature parameters and spatiotemporal information to generate dynamic data fingerprints. The sequence generation algorithm has a strong initial value dependency. Its output sequence will exhibit exponential and unpredictable differences with the slight differences in the initial input. This ensures that any slight change in the original data will be amplified into significant and irrelevant output results in the final generated fingerprint. (3) The present invention uses smart contracts to perform electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification in parallel, thus constructing a multi-dimensional evidence loop; at the same time, by dynamically adjusting the verification threshold, it avoids misjudgment caused by fixed thresholds and effectively identifies data fraud and equipment failure. (4) The present invention adopts a double-layer nested hash and digital signature evidence storage mechanism, which can ensure the authenticity of the on-chain data through the off-chain pre-verification mechanism, while protecting the privacy of the original data. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method for collaborative traceability and auditing of multi-source data in charging piles according to the present invention. Detailed Implementation

[0019] 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 of the embodiments of this invention. The embodiments described in this invention are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0020] As an embodiment of the present invention, a specific implementation method for a blockchain-based multi-source data collaborative traceability and auditing method for charging piles is disclosed. The execution flow of the method embodiment is as follows: Figure 1 .

[0021] S1: As one implementation method, the physical unclonable function output value of the charging pile control motherboard is obtained as the hardware gene identifier of the charging pile; and multi-source detection data during the charging process of the charging pile is collected, including charging waveform data, equipment operating status data, environmental safety data and charging transaction data.

[0022] Specifically, the charging pile control motherboard is equipped with a physically unclonable function (PUF) circuit to generate a globally unique hardware identifier. This identifier is generated based on inherent process variations in chip manufacturing and possesses physical unclonability. It includes two optional PUF implementation circuits: SRAM PUF circuit: When the SRAM storage cell on the charging pile control motherboard is powered on and initialized, the corresponding initial state value is read. Taking advantage of the uncertainty of the SRAM storage cell's power-on state caused by random deviations in the transistor threshold voltage during the manufacturing process, the initial state value is read and processed (such as error correction and hash operation) to extract and generate a 256-bit hardware gene identifier. This extraction and comparison verification process is automatically executed every time the charging pile is powered on. By comparing the generated identifier with the reference identifier pre-stored in the secure storage area (such as the security chip), it is possible to detect whether the control motherboard has been replaced or physically tampered with.

[0023] Ring Oscillator PUF Circuit: At least two identical ring oscillators are integrated on the charging pile control motherboard. The frequency difference between the different ring oscillators, caused by manufacturing process variations, is extracted using a counter and comparator as a hardware identifier. The oscillation frequency of the ring oscillator is affected by temperature and voltage. Environmental parameters are collected by built-in temperature and voltage sensors, and temperature and voltage compensations are applied to the frequency deviation value to ensure the stability of the hardware identifier under different environmental conditions.

[0024] Furthermore, during the operation of the charging pile, especially during the charging process, the system simultaneously collects multi-dimensional detection data, forming a comprehensive status monitoring surface, including: The charging waveform data is continuously acquired at a sampling frequency of not less than 1 kHz, including charging current and charging voltage waveforms, and the acquired current sequence is used as the basis for the data acquisition. and voltage sequence Construct a matrix, where, This represents the number of sampling points in a single charging cycle. The current sequence and voltage sequence have the same number of elements, and both the current sequence and voltage sequence have a number of elements equal to the number of sampling points in a single charging cycle.

[0025] The device operating status data is a vector composed of module temperature, insulation resistance, and controller operating log feature values ​​(the feature value corresponding to the operating status is 1, and vice versa, the corresponding feature value is 0); the environmental safety data of the multi-source detection data is a vector composed of ambient temperature, humidity, and smoke concentration; the charging transaction data of the multi-source detection data is a vector composed of charging start time, end time, BMS reported power consumption, metered power consumption, and average power.

[0026] S2: As one implementation method of the embodiment, the charging waveform data is decomposed into multiple scales to extract time-domain feature parameters and frequency-domain feature parameters; the hardware gene identifier, time-domain feature parameters, frequency-domain feature parameters and charging spatiotemporal information are nonlinearly coupled to generate the dynamic data fingerprint of the charging pile in the current charging cycle.

[0027] Specifically, a unique, unpredictable, and tamper-sensitive dynamic data fingerprint is generated for each charging pile in each charging cycle. The process mainly includes two stages: feature extraction and nonlinear coupling. The steps are as follows: S2.1: Extract frequency domain feature parameters and time domain feature parameters; S2.1.1: Perform wavelet packet decomposition on the charging current waveform and charging voltage waveform, and extract the energy features under different frequency bands as frequency domain feature parameters; Furthermore, the matrix corresponding to the charging waveform data is subjected to wavelet packet multi-scale decomposition using discrete wavelet packet transform, with the decomposition level being [number missing]. The wavelet packet coefficient set is obtained. ,in, Index for the number of decomposition levels; For frequency band indexing; For the first The layer and frequency band index are wavelet packet coefficients; Based on the set of wavelet packet coefficients Calculate the first Layer Energy of each frequency band , That is, the sum of the squares of all wavelet packet coefficients within that frequency band. Energy It characterizes the intensity of the original signal at that specific frequency component.

[0028] All calculated energy In a certain order (such as by) From 1 to Within each floor Arranged from smallest to largest, they form the frequency domain feature vector. = .

[0029] S2.1.2: Calculate statistical and shape features from the charging waveform data to construct a time-domain feature vector, including: Peak current The maximum current within a single charging cycle; Current valley The minimum current during a single charging cycle; Average current The average current during a single charging cycle; Current variance The degree of dispersion of current within a single charging cycle; Phase difference during charging cycle The phase difference between the voltage and current waveforms, calculated through cross-correlation analysis, is used to characterize the power factor characteristics.

[0030] This constitutes the time-domain feature vector. = .

[0031] S2.2: Nonlinear coupling generates dynamic data fingerprints.

[0032] Specifically, the hardware's unique identifier, time-domain / frequency-domain characteristic parameters representing the charging process, and charging spatiotemporal information are transformed and mixed using a nonlinear recursive algorithm to generate a final 256-bit dynamic data fingerprint. The charging spatiotemporal information includes the charging start time, charging end time, and the charging station's geographical coordinates. This includes: S2.2.1: First, construct a combined feature vector consisting of hardware gene identifier, time-domain feature parameters, frequency-domain feature parameters, and charging spatiotemporal information, denoted as... ;Will The dimension is denoted as , ,in, A function that represents the number of data dimensions to retrieve; S2.2.2: Combining eigenvectors Each component in the process is normalized. Due to the combined feature vectors The components in the equation have different dimensions (such as binary bits, current, time, etc.), and need to be uniformly mapped to the interval [0,1]. Each component Using minimum-maximum normalization, we obtain The calculation formula is: ; in, and The first The minimum and maximum values ​​of each component within historical data or a preset range.

[0033] S2.2.3: Map the 256-bit hardware gene identifier to the initial value of the non-linear recursive algorithm. This is to ensure that the starting points of different charging piles are differentiated during the dynamic generation process. The calculation formula is: ; in, Indicates the first in hardware gene identifier The value of the bit, and The value of the denominator is 0 or 1; As a normalization factor, ensure This conversion method ensures that even if the hardware fingerprints of different charging stations differ by only one bit, it will result in a corresponding... Significant differences are generated, thus utilizing the initial value sensitivity of nonlinear recursive algorithms to generate different fingerprint sequences.

[0034] S2.2.4: Construct a nonlinear recursive algorithm that can be perturbed by external data. This involves normalizing the combined feature vectors. As an additional parameter, it affects the computation of the core state transition function in each iteration, thereby achieving deep integration of data and the generation process. The iterative equation is as follows: ; in, The preset disturbance coefficient is a positive decimal much smaller than 1; For control parameters, For the pseudo-random sequence, the first... The state value of the next iteration. ; For the pseudo-random sequence, the first... The state value of the next iteration; For the number of iterations, .

[0035] S2.2.5: with Starting from this point, of Each component is sequentially used as a disturbance input for... After several iterations, a length of [length missing] is obtained. pseudo-random sequence This sequence has been deeply coupled with information from all input features.

[0036] S2.2.6: Convert the pseudo-random sequence into a binary fingerprint to obtain a fixed-length (256-bit) binary fingerprint, and fully utilize the high-precision randomness of the iterative state values, converting each state value... Convert to a single binary digit. The conversion algorithm is as follows: ; in, For precision parameters, such as =32; Indicates will It is mapped to a 32-bit integer, preserving high-precision information of the state value; for Operations corresponding to the binary bit conversion results; Pick The least significant bit is used as the final binary bit; since the distribution of state values ​​is uniform and unpredictable, this least significant bit has excellent randomness; from the pseudo-random sequence Extract binary bits .

[0037] S2.2.7: The obtained pseudo-random binary bit sequence With the original combined feature vector The binary representation is XORed bit by bit to generate the final dynamic data fingerprint.

[0038] Specifically, for Each component in Convert it to a binary bit. .like If it is a binary number, it is used directly; if it is a consecutive value, the same method is used. The method takes the least significant bit.

[0039] The final dynamic data fingerprint Bit The calculation method is as follows: ; in, ; For sequence ;like If the value is greater than 256, then take the first 256 bits. to As the final dynamic data fingerprint; if If the value is less than 256, the data can be expanded to 256 bits by repeatedly performing nonlinear iterations and sampling its output sequence.

[0040] In summary, by integrating unique hardware identifiers and charging data from specific times and spaces, the fingerprints of different charging stations and different charging sessions are almost guaranteed to be different. This is because the nonlinear recursive algorithm requires specific initial values... The extreme sensitivity of the input data (derived from the hardware ID) and parameters (derived from the feature data) means that even a tiny alteration to the input data (such as a current sampling point or a timestamp) can lead to huge and unpredictable changes in the final fingerprint. Sensitive information such as the original hardware ID and waveform features cannot be deduced from the final 256-bit fingerprint, thus protecting privacy. High-dimensional, multi-source data (waveform, state, spatiotemporal information) is compressed into a fixed-length binary string, facilitating storage, transmission, and comparison.

[0041] It should be noted that the dynamic data fingerprint produces an exponentially amplified deviation from any modification to the original data, making it impossible to generate a matching fingerprint from the tampered data.

[0042] S3: As one implementation method of the embodiment, a double-layer nested hash operation is performed on the dynamic data fingerprint of the charging pile in the current charging cycle and the multi-source detection data at the edge computing node to generate a first-level evidence storage hash and a second-level evidence storage hash, and the operation result is digitally signed with national cryptographic standards; the signed data is uploaded to the blockchain network.

[0043] S3.1: Double-layer nested hash operation; The double-layer nested hash operation adopts a hierarchical processing strategy. First, it independently digests various types of raw data, and then it aggregates the global digest and dynamic data fingerprints. This ensures data privacy (without uploading all raw data) while establishing a tight data association chain.

[0044] S3.1.1: First-level evidence storage hash; Perform an independent hash operation on each type of data in the multi-source detection data to generate its unique content digest.

[0045] SHA-3 hash operations are performed on the matrices corresponding to the charging waveform data, the device operating status data vector, the environmental safety data vector, and the charging transaction data vector, respectively, to obtain four fixed-length hash values, denoted as . , , , This is the first-level evidence storage hash set. Each hash value uniquely represents the complete content of the corresponding category of data, and any slight change in the data will result in a completely different hash value.

[0046] To further enhance security and prevent attackers from replaying valid historical data packets, timeliness and randomness factors are introduced when generating the first-level hash. Specifically, before calculating the SHA-3 hash of each data type, the current timestamp (accurate to milliseconds) and a 32-bit cryptographic random number generated in real-time by a hardware security chip (such as an SE or TEE) are concatenated with the original data for that type. The concatenated data is then hashed. This method ensures that even if the original data is the same, the hash value generated at different times will necessarily be different, effectively resisting replay attacks.

[0047] S3.1.2: Second-level evidence hash; aggregate the summaries of each item generated in the first level with the dynamic data fingerprint representing the uniqueness of this charging to generate a global, nested final hash value.

[0048] The 256-bit dynamic data fingerprint generated in step S2 is compared with the four first-level hash values ​​mentioned above. , , , The data is concatenated according to a preset sorting rule (such as lexicographical order of data category names) to form a longer data string.

[0049] Perform a Keccak-256 hash operation on the concatenated data string, and the output is a 256-bit (32-byte) hash value, which is the second-level evidence storage hash.

[0050] Furthermore, when generating the second-level hash, a unique device certificate serial number of the edge computing node can be appended to the end of the concatenated string. This further binds the final evidence hash to a specific hardware device, enhancing the traceability of the data source.

[0051] S3.2: After generating the final nested hash value, digitally sign it using the national cryptographic algorithm to prove that the hash value was generated by a legitimate edge computing node at a specific time and has not been tampered with since signing.

[0052] The final nested hash value and the current timestamp are digitally signed using the national cryptographic standard SM2 through a hardware security chip to generate a signature data packet (the signature data packet consists of the final nested hash value, the current timestamp, the corresponding national cryptographic standard SM2 digital signature, and the device certificate serial number of the corresponding edge computing node); the resulting signature data packet is uploaded to the transaction pool of the blockchain network to await consensus.

[0053] S4: As one implementation method of the embodiment, the smart contract in the blockchain network performs multi-dimensional cross-validation on the uploaded data in parallel. The multi-dimensional cross-validation includes electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification. Only when all verifications pass, the dynamic data fingerprint and the multi-source detection data are packaged and uploaded to the blockchain for evidence storage.

[0054] S4.1: Electrical data consistency verification compares the charging power reported by the vehicle-side BMS (Battery Management System) with the power consumption obtained from the charging station's power monitoring equipment to determine whether the difference between the two is within the preset error threshold range.

[0055] Furthermore, to adapt to different credit levels, environmental conditions, and equipment statuses, and to avoid misjudgments caused by environmental changes or normal equipment fluctuations, this embodiment introduces a dynamic error threshold adjustment mechanism. The error threshold range for the electrical data consistency verification is dynamically adjusted to avoid misjudgments caused by environmental changes, resulting in a final error threshold. The calculation formula is: ; in, The basic error threshold; This refers to the credit threshold narrowing rate. This is the environmental compensation coefficient; This is the power adjustment error threshold.

[0056] The basic error threshold is set based on the accuracy level and calibration records of the charging pile, with a default value of ±3.0%. Based on the current charging cycle Within this range, the credit threshold narrowing rate is based on the charging pile node. In the previous credit assessment cycle The dynamic credit score at the end of the period is adjusted by a factor. ,in, For the previous period's credit score, This means that for every 10-point increase in credit score, the error threshold narrowing rate increases by 0.2 percentage points, under the initial conditions. (New equipment is shipped with a default credit score of 50). The environmental compensation coefficient These are empirical coefficients pre-calibrated based on historical operating environment data (such as temperature and humidity) of the charging piles through statistical regression analysis. The power adjustment error threshold It is the theoretical maximum error estimated based on the hardware precision level and control parameters of the charging pile.

[0057] Specifically, during the system initialization phase, an error benchmark comparison table containing different environmental operating conditions and power error combinations is pre-constructed, and... and As an index parameter, this lookup table is queried to obtain the initial error threshold base. .

[0058] For example, a credit score of 85, in a high temperature and high humidity environment ( =0.05) charging at 20% of rated power ( For charging piles with a failure rate of 0.5%, the final error threshold is: =3.0%×(1-0.017)×(1+0.05)+0.5%=3.0%×0.983×1.05+0.5%≈3.095%+0.5%≈3.595%.

[0059] This means that under this complex operating condition, the allowable error range has been dynamically adjusted from the basic ±3% to approximately ±3.6%, thereby reducing false alarms while ensuring safety.

[0060] S4.2: Perform equipment status baseline verification by comparing the equipment operating status data within the current charging cycle with the historical health baseline of the charging pile (using a multi-dimensional feature vector constructed from the charging pile's normal operating data over the past 30 days read from the blockchain, calculating the Mahalanobis distance between the constructed multi-dimensional feature vector and the charging pile's historical health baseline, and comparing the obtained Mahalanobis distance with a preset threshold; the charging pile's historical health baseline is obtained by querying a preset form in the database) to determine if there are any parameter mutations or abnormal deviations. Specifically, this includes: S4.2.1: Read the normal operation data of the charging pile over the past 30 days from the blockchain, and extract the equipment operation status data from each record. This data is a feature vector containing multiple status parameters, which is a multi-dimensional feature vector space. S4.2.2: Use Gaussian Mixture Model (GMM model) to perform cluster analysis on normal operation data to generate a probability distribution model of equipment health baseline; The probability density function of the Gaussian mixture model is ; in, For a single sample; The number of Gaussian components; The mixing coefficient; It is the mean vector; It is the covariance matrix; For For the mean, Let be the probability density function of the multivariate Gaussian distribution of the covariance matrix, representing . In the Likelihood under each component; =1,2,…, .

[0061] The GMM model parameters were optimized using the expectation-maximization algorithm. , , The process involves iterative optimization until convergence. The trained GMM model then becomes the baseline probability distribution model for the device health of the charging station.

[0062] S4.2.3: Calculate the Mahalanobis distance between the device operating status data of the current charging cycle and the health baseline; For accurate measurement Calculate the Mahalanobis distance to the nearest Gaussian component from the distance to the entire health data distribution. First, find the distance to the nearest Gaussian component. The Gaussian component with the highest Euclidean distance or probability membership is denoted by its index. The corresponding parameters are Then the Mahalanobis distance The calculation formula is as follows: ; in, Nearest Neighbor Gaussian Components The mean vector; Gaussian components The inverse of the covariance matrix; The device operating state vector for the current charging cycle.

[0063] The calculated Mahalanobis distance Deviation from preset state threshold (Usually, 3 times the standard deviation is used for comparison.) If this occurs, it is determined that the equipment is in an abnormal state, triggering the early warning mechanism.

[0064] It should be noted that the health baseline is updated regularly to accommodate performance drift during the normal aging process of equipment.

[0065] S4.3: Spatial-temporal trajectory rationality verification. The geofence information of the vehicle is obtained from a third-party map service through an off-chain oracle to verify the spatial consistency between the vehicle location and the charging pile location (if the distance between the two is less than the spatial error threshold, it is determined that the spatial consistency is satisfied; otherwise, it is determined that the spatial consistency is not satisfied).

[0066] The geographical coordinates of the charging station were directly parsed from the data packet uploaded to the blockchain and denoted as follows: ;in, Longitude Latitude; Furthermore, the smart contract initiates a request to a trusted third-party map location service provider (such as Amap or Baidu Maps API) by invoking an off-chain oracle. The request parameters include the identifier of the vehicle to be verified (such as license plate number or VIN code) and the time window for this charging (start time). Until the end time The oracle returns the geofence information recorded by the vehicle during that time period or the last reported latitude and longitude coordinates, denoted as... ;in, The longitude coordinates of the vehicle. This refers to the latitude coordinates of the vehicle.

[0067] Furthermore, to calculate the true distance between two points on the Earth's surface, the Haversine formula is used to calculate the great circle distance between the charging station location and the vehicle location. ; ; in, The average radius of the Earth The unit is radians. The unit is radians.

[0068] The calculated spatial distance With a preset spatial error threshold A comparison is made. This threshold takes into account factors such as positioning error and charging station coverage; for example, it can be set to 50 meters or 100 meters. If the vehicle's location and the charging station's location are found to be spatially consistent, the verification is successful; otherwise, if they are not, the verification fails, indicating that there may be a falsification of the charging station's geographical location.

[0069] In summary, if all three items in the multi-dimensional cross-validation pass, the smart contract will package the dynamic data fingerprint of the charging pile in the current charging cycle and the multi-source detection data into a new block and store it on the blockchain; if any of the three validation items fails, the smart contract will trigger an early warning mechanism and store the abnormal event separately.

[0070] S5: As one implementation of the embodiment, at the end of each preset credit assessment period (e.g., 24 hours), the dynamic credit score of the charging pile is calculated and updated based on the historical evidence data and cross-validation results of the charging pile within this period; according to the level to which the dynamic credit score belongs, the corresponding hierarchical audit strategy is matched to conduct sampling audit, periodic audit or full audit on the data subsequently generated by the charging pile. In the process of S500 calculating and updating the dynamic credit score of the charging pile, based on the historical evidence data of the charging pile and the results of each cross-validation, the local model parameters of multiple charging pile operation nodes are aggregated using a federated learning framework, and the data quality contribution of each charging pile is quantified by the Shapley value algorithm to generate a dynamic credit score.

[0071] The calculation of dynamic credit scores specifically includes the collaborative training and contribution quantification of the credit assessment model: S5.1: Train a credit risk prediction model locally for each charging pile operation node.

[0072] The credit risk prediction model is a binary logistic regression model. The input feature vector is composed of the historical pass rate of the current charging cycle's equipment operating status data, environmental safety data, electrical data consistency verification, and the authenticity assessment coefficient of the last audit. The input features are normalized before model training. The model uses cross-entropy as the loss function and optimizes the model weights through gradient descent during training.

[0073] Each charging transaction undergoes multi-dimensional verification. The authenticity evaluation coefficient is obtained based on the results of multi-dimensional cross-validation, including weighted averages of scores obtained from electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification, to obtain the authenticity score of a single charging transaction.

[0074] The training labels for the model are binary values. If all the charging data passes the cross-validation of S400, the label is 0 (low risk); if any one fails, the label is 1 (high risk).

[0075] Furthermore, each node uses local historical data to train a credit risk prediction model and generate model weight parameters. The data is then uploaded to the federated aggregation server. The server aggregates the model weight parameters using a weighted average method based on the data volume of each node. And it was issued. After multiple rounds of iteration, a global model was obtained that can comprehensively assess the risks of submitting charging pile data.

[0076] S5.2: Contribution quantification for evaluating charging station nodes The contribution to the global model is evaluated using the leave-one-out method. This is calculated when charging pile nodes are included. Performance of the global model With the removal of charging station nodes Global model performance after re-aggregation Charging pile nodes Data quality contribution Defined as: = - ; The training set, which includes various normal and abnormal charging samples collected from the central server, is used to measure performance, and the scores obtained from testing on the validation set are used to measure performance. , The difference in scores reflects the marginal value of the charging pile node data in improving the accuracy of the global model.

[0077] The performance This refers to the classification accuracy of the federated learning model on the preset test set.

[0078] Specifically, before calculating the contribution, a test set is first constructed using historical evidence data; then, after local training is completed and the global model is aggregated, the global model is used to make predictions on the test set, and the classification accuracy at this point is recorded as [the result]. Subsequently, assuming that data from one participant is removed, a new model is retrained, and its accuracy is calculated. The aforementioned contribution This is the difference between the global model accuracy and the model accuracy after removing the participant.

[0079] S5.3: Charging Pile Node In the Periodic dynamic credit scoring The score is determined by the individual's basic credit score, contribution level, and audit results within the current period. The calculation formula is as follows: ; in, Credit score for the previous period; The charging pile nodes will be completed within this cycle. Contribution Values ​​normalized to the [0,1] interval; Charging pile nodes within this cycle The audit score (obtained by averaging the authenticity scores of all charging transactions within this period); , , These are the corresponding weights, and ,like =0.6, =0.2, =0.2.

[0080] S5.4: Based on the level to which the dynamic credit score belongs, match the corresponding tiered audit strategy, specifically including: The dynamic credit score that is greater than or equal to the first preset credit score value is recorded as a high credit level. A sampling audit strategy (such as a preset ratio of 10%) is adopted to randomly select a preset ratio of on-chain transactions for hash value comparison and verification. Dynamic credit scores that are greater than or equal to the second preset credit score and less than the first preset credit score are classified as medium credit scores. A periodic audit strategy is adopted, and a full data verification is triggered once every preset number of transactions. The dynamic credit score that is less than the second preset credit score is recorded as a low credit score. A full audit strategy is adopted to trigger real-time data backtracking verification for each transaction and extend the storage time of the associated video surveillance data; the first preset credit score is greater than the second preset credit score. After each audit, an authenticity assessment coefficient is calculated, and the dynamic credit score is dynamically updated based on the authenticity assessment coefficient.

[0081] S6: As one implementation of the embodiment, when data anomaly is detected or an audit request is received, the tampering point or fault point is automatically located based on the full chain of evidence stored on the blockchain, and the responsible party is identified according to the cause of the anomaly, and the responsibility determination result is written into the blockchain.

[0082] The automatic location of tampering points or fault points based on the full evidence chain stored on the blockchain specifically includes: when an audit discovers data anomalies or receives an audit request initiated by a regulator, retrieving the full evidence chain data corresponding to the abnormal transaction from the blockchain. The full evidence chain data includes hardware genetic identifiers, dynamic data fingerprints, multi-source detection data, nested hash values, digital signatures, and timestamps; recalculating the hash value of the original data stored off-chain and comparing it step by step with the hash value stored on-chain to locate the specific data category and time point of tampering; based on the location results and combined with the change trend of the dynamic credit score before and after the tampering, identifying the responsible party as the user end, charging pile end, network transmission link, or third-party attack through a preset responsibility determination rule base.

[0083] The preset liability determination rule base includes: User-side liability determination rules: If the vehicle-side BMS data is inconsistent with the charging pile-side data, and there is a spatial deviation between the vehicle positioning trajectory and the charging pile location, it is determined that the user-side data is falsified or the equipment is faulty. The spatial deviation refers to the verification of the spatial consistency between the vehicle location and the charging pile location in S4.3. Charging pile terminal responsibility determination rules: If the charging pile terminal sensor data is abnormal but the vehicle terminal data is normal, and the hardware gene identification verification is passed, it is determined to be a charging pile sensor failure or calibration deviation. Network transmission liability determination rules: If the hash values ​​of off-chain data and on-chain data are inconsistent, but the edge computing node logs show that the data was sent completely, it is determined that the data was lost or hijacked during network transmission; Third-party attack liability determination rules: If a large number of abnormal transactions, abnormal charging outside of working hours, or repeated failures to verify hardware genetic identifiers are detected within a short period of time, it is determined to be a malicious attack by a third party.

[0084] As an embodiment of the present invention, a blockchain-based multi-source data collaborative traceability and auditing system for charging piles is disclosed. Employing the specific implementation method described above for the blockchain-based multi-source data collaborative traceability and auditing method for charging piles, the system includes: Data acquisition module: acquires the hardware genetic identifier of the charging pile; and collects multi-source detection data of the charging pile, including charging waveform data, equipment operating status data, environmental safety data and charging transaction data; The fingerprint generation module extracts time-domain and frequency-domain feature parameters based on the charging waveform data; and nonlinearly couples the hardware gene identifier, time-domain feature parameters, and frequency-domain feature parameters with the charging spatiotemporal information to generate a dynamic data fingerprint for the current charging cycle; the charging spatiotemporal information includes the charging start time, charging end time, and the geographical coordinates of the charging pile. On-chain evidence storage module: Based on the dynamic data fingerprint and the multi-source detection data, a double-layer nested hash operation is performed to generate an evidence storage hash and perform digital signature to generate an evidence storage data package; the evidence storage data package is uploaded to the blockchain network, and the smart contract of the blockchain network performs multi-dimensional cross-validation on the evidence storage data package in parallel; if the multi-dimensional cross-validation passes, the data contained in the evidence storage data package is packaged and stored on the blockchain; the multi-dimensional cross-validation includes electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification; Credit assessment and tiered audit module: Based on the verification results obtained from the multi-dimensional cross-validation and the historical evidence data of the charging pile, calculate and update the dynamic credit score of the charging pile; determine a tiered audit strategy that matches the dynamic credit score to audit the data subsequently generated by the charging pile.

[0085] As an embodiment of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded into the processor, it adopts the specific implementation method described above for the multi-source data collaborative traceability and auditing method for charging piles.

[0086] As an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it adopts the specific implementation method described above for the multi-source data collaborative traceability and auditing method for charging piles.

[0087] 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 it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A blockchain-based method for collaborative traceability and auditing of multi-source data for charging piles, characterized in that: include: Obtain the hardware genetic identifier of the charging pile; It also collects multi-source detection data of the charging pile, including charging waveform data, equipment operating status data, environmental safety data, and charging transaction data; Time-domain and frequency-domain feature parameters are extracted based on the charging waveform data; The hardware gene identifier, time-domain feature parameters, and frequency-domain feature parameters are nonlinearly coupled with the charging spatiotemporal information to generate a dynamic data fingerprint for the current charging cycle. The charging spatiotemporal information includes the charging start time, the charging end time, and the geographical coordinates of the charging pile. Based on the dynamic data fingerprint and the multi-source detection data, a double-layer nested hash operation is performed to generate a storage hash and perform digital signature to generate a storage data package; the storage data package is uploaded to the blockchain network, and the smart contract of the blockchain network performs multi-dimensional cross-validation on the storage data package in parallel; if the multi-dimensional cross-validation passes, the data contained in the storage data package is packaged and stored on the blockchain; the multi-dimensional cross-validation includes electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification; Based on the verification results obtained from the multi-dimensional cross-validation and the historical evidence data of the charging pile, the dynamic credit score of the charging pile is calculated and updated; according to the dynamic credit score, a hierarchical audit strategy matching the dynamic credit score is determined to audit the data subsequently generated by the charging pile.

2. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, The nonlinear coupling of hardware gene identifiers, time-domain feature parameters, frequency-domain feature parameters, and charging spatiotemporal information includes: The hardware gene identifier, the time-domain feature parameters, the frequency-domain feature parameters, and the charging spatiotemporal information constitute a combined feature vector, and the combined feature vector is normalized. The hardware gene identifier is mapped to the initial value of a nonlinear recursive algorithm; Based on the initial value, the normalized combined feature vector is input as a parameter into the nonlinear recursive algorithm for iteration to generate a pseudo-random sequence. Each state value of the pseudo-random sequence is converted into a binary bit, and each component of the combined feature vector is converted into a binary bit. The two are then XORed bit by bit to generate the dynamic data fingerprint.

3. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, Based on the dynamic data fingerprint and the multi-source detection data, a double-layer nested hash operation is performed to generate a proof-of-existence hash, including: Perform a first hash operation on each type of data in the multi-source detection data to generate multiple corresponding first-level evidence storage hashes; Multiple first-level evidence hashes are concatenated and combined with the dynamic data fingerprint, and a second hash operation is performed on the concatenated and combined data to generate a second-level evidence hash.

4. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, The electrical data consistency verification includes: It obtains the power consumption from the power monitoring equipment at the charging pile and the charging power reported by the vehicle-side battery management system. The power consumption is compared with the charging power, and the electrical data is judged to be consistent based on the comparison result and the error threshold. The error threshold is calculated by weighting the base error threshold, the credit threshold narrowing rate, the environmental compensation coefficient, and the power adjustment error threshold. The credit threshold narrowing rate is calculated based on the credit score of the charging pile in the previous cycle within the current charging cycle.

5. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, The device status baseline verification includes: Obtain the normal operation data of the charging pile within a preset historical period; Based on the normal operation data, a baseline model of the equipment health of the charging pile is constructed; Obtain device operating status data for the current charging cycle; Calculate the degree of deviation between the device operating status data of the current charging cycle and the device health baseline model; Based on the degree of deviation and the preset state deviation threshold, it is determined whether the equipment operating status of the charging pile is normal. The degree of deviation is the Mahalanobis distance between the device operating status data of the current charging cycle and the device health baseline model; The equipment health baseline model is a probability distribution model obtained by clustering the historical normal operation data of the charging pile based on a Gaussian mixture model.

6. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, The verification of the rationality of the spatiotemporal trajectory includes: Obtain the geographical location of the charging station; During the charging period, obtain the vehicle's geographical location; Calculate the spatial distance between the geographical location of the charging pile and the geographical location of the vehicle; Based on the spatial distance and the preset spatial error threshold, it is determined whether the vehicle position and the charging pile position meet the spatial consistency requirement.

7. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, Based on the verification results and the historical evidence data of the charging pile, the dynamic credit score of the charging pile is calculated and updated, including: The historical pass rate of the consistency verification of the equipment operation status data, environmental safety data, and electrical data of the charging pile in the current charging cycle, as well as the authenticity evaluation coefficient obtained based on the last audit, are combined into an input feature vector; The input feature vector is fed into a trained credit risk prediction model to obtain the credit risk prediction result of the charging pile in the current charging cycle. The trained credit risk prediction model is a logistic regression model obtained through collaborative training based on a federated learning framework, used to predict the probability that the charging data is high-risk or low-risk. The dynamic credit score of the charging pile is updated based on the credit score of the previous cycle of the current charging cycle, the data quality contribution of the charging pile, and the audit score obtained based on the verification results. Based on the credit risk prediction results, adjust the calculation weight or value of the data quality contribution.

8. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 7, characterized in that, The data quality contribution of the charging pile includes: In the federated learning framework aggregation, the calculation includes charging pile nodes. Time performance ; Calculate the removal of the charging pile node Post-repolymerization performance ; The data quality contribution of the charging pile is, in terms of computing performance. With performance The difference is the value of the charging pile node. Data quality contribution .

9. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 7, characterized in that, The authenticity evaluation coefficient is obtained by weighting the scores obtained from electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification. The audit score obtained from the verification results is calculated by the average of the authenticity assessment coefficients of all charging transactions within the current charging cycle.

10. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, Determining a tiered audit strategy that matches the dynamic credit score includes: If the dynamic credit score is greater than or equal to the first preset credit score value, it is matched as a high credit level, and a sampling audit strategy is adopted for the data subsequently generated by the charging pile. If the dynamic credit score is greater than or equal to the second preset credit score value and less than the first preset credit score value, it is matched as a medium credit level, and a periodic audit strategy is adopted for the data subsequently generated by the charging pile. If the dynamic credit score is less than the second preset credit score value, it is matched as a low credit level, and a full audit strategy is adopted for the data subsequently generated by the charging pile.

11. The blockchain-based multi-source data collaborative traceability and auditing method for charging piles according to claim 1, characterized in that, include: When an anomaly is detected or an audit request is received, the full chain of evidence corresponding to the anomaly data is obtained from the blockchain. Based on the complete chain of evidence, the data tampering point or equipment failure point can be located; Based on the data tampering point or equipment failure point, and in conjunction with the preset responsibility determination rules, the party responsible for the data anomaly is determined, and the responsibility determination result is written into the blockchain.

12. A blockchain-based multi-source data collaborative traceability and auditing system for charging piles, executing the multi-source data collaborative traceability and auditing method for charging piles as described in any one of claims 1-11, characterized in that, The system includes: Data acquisition module: acquires the hardware genetic identifier of the charging pile; and collects multi-source detection data of the charging pile, including charging waveform data, equipment operating status data, environmental safety data and charging transaction data; The fingerprint generation module extracts time-domain and frequency-domain feature parameters based on the charging waveform data; and nonlinearly couples the hardware gene identifier, time-domain feature parameters, and frequency-domain feature parameters with the charging spatiotemporal information to generate a dynamic data fingerprint for the current charging cycle; the charging spatiotemporal information includes the charging start time, charging end time, and the geographical coordinates of the charging pile. On-chain evidence storage module: Based on the dynamic data fingerprint and the multi-source detection data, a double-layer nested hash operation is performed to generate an evidence storage hash and perform digital signature to generate an evidence storage data package; the evidence storage data package is uploaded to the blockchain network, and the smart contract of the blockchain network performs multi-dimensional cross-validation on the evidence storage data package in parallel; if the multi-dimensional cross-validation passes, the data contained in the evidence storage data package is packaged and stored on the blockchain; the multi-dimensional cross-validation includes electrical data consistency verification, equipment status baseline verification, and spatiotemporal trajectory rationality verification; Credit assessment and tiered audit module: Based on the verification results obtained from the multi-dimensional cross-validation and the historical evidence data of the charging pile, calculate and update the dynamic credit score of the charging pile; determine a tiered audit strategy that matches the dynamic credit score to audit the data subsequently generated by the charging pile.

13. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is loaded into the processor, it implements the multi-source data collaborative traceability and auditing method for charging piles according to any one of claims 1-11.

14. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the multi-source data collaborative traceability and auditing method for charging piles according to any one of claims 1-11.

Images