An electric vehicle charging safety early warning system based on vehicle-pile interaction data and VIN vehicle library fusion

Abstract

This invention discloses an electric vehicle charging safety early warning system based on the fusion of vehicle-to-charging pile interaction data and a VIN vehicle model database, belonging to the field of new energy vehicle technology. The system includes a charging pile intelligent data acquisition terminal, a data acquisition module, a VIN vehicle model database module, an early warning rule database module, a data processing and early warning engine, and an early warning response module. This invention uses the charging pile intelligent data acquisition terminal to collect CAN bus data from vehicle-to-charging pile interaction in real time and parse the VIN code. Combined with the VIN vehicle model database, it obtains the vehicle's static safety parameters and establishes personalized initial early warning thresholds. A dynamic threshold model is constructed, and dynamic learning and stable updating of the thresholds are achieved through anti-jitter calculations. Based on real-time comparison and analysis of multi-dimensional early warning rules, tiered early warnings are triggered, and corresponding response measures are executed. This invention overcomes the problems of poor adaptability and high false alarm / false negative rates associated with fixed thresholds, achieving accurate and adaptive early warning of charging safety risks, and significantly improving the safety and reliability of electric vehicle charging.

Description

Technical Field

[0001] This invention relates to the field of new energy vehicle technology, specifically to a safety early warning technology during the charging process of electric vehicles, and in particular to a system and method for personalized and adaptive dynamic safety early warning that integrates real-time vehicle-charging station interaction data with vehicle model information. Background Technology

[0002] With the widespread adoption of electric vehicles and the rapid development of high-power supercharging technology, charging safety has become a major concern in the industry. Batteries may experience thermal runaway during charging due to overcharging, overcurrent, or overheating, posing serious safety hazards.

[0003] In existing technologies, charging safety warnings mostly employ fixed threshold methods. For example, some solutions monitor battery parameters such as voltage, current, and temperature and compare them with preset fixed safety thresholds to trigger a warning. However, the rated parameters and safety boundaries vary significantly between different vehicle models and battery types. Using a "one-size-fits-all" fixed threshold cannot adapt to diverse vehicle models, resulting in high false alarm and false negative rates, severely impacting user experience and the effectiveness of the warnings.

[0004] Furthermore, some improved solutions attempt to incorporate historical data for learning. For example, patent CN114879049A discloses a method for evaluating battery consistency based on historical charging data fragments. However, this method focuses on assessing the long-term health of the battery and does not address the issue of how to "tailor-make" warning thresholds for a specific vehicle during a single real-time charging process. Another patent, CN202410560980A, discloses a method for evaluating battery performance using fast charging process messages, but its core lies in offline evaluation of battery performance, lacking the ability to monitor and intervene in the entire charging process in real time and dynamically.

[0005] Therefore, the existing technology has the following prominent problems:

[0006] 1. Lack of personalization: The warning thresholds are not associated with the specific vehicle model or battery characteristics, resulting in poor adaptability.

[0007] 2. Insufficient real-time performance and accuracy: The warning strategy is not dynamically adjusted in accordance with real-time charging conditions, and it is unable to cope with the impact of factors such as battery performance degradation and environmental changes.

[0008] 3. Limited data dimensions: Most data rely on single BMS data or pile-end data, failing to fully utilize the entire chain of information from vehicle-pile interaction for comprehensive analysis.

[0009] The present invention aims to solve one or more of the above-mentioned technical problems. Summary of the Invention

[0010] Purpose of the Invention The primary objective of this invention is to overcome the shortcomings of the prior art and provide an electric vehicle charging safety early warning system and method that can solve the problem of high false alarm and false alarm rates caused by fixed thresholds.

[0011] Another objective of this invention is to achieve personalized and precise safety warnings for different vehicle models by integrating static parameters provided by the VIN vehicle model database with real-time vehicle-pile interaction data.

[0012] Another objective of this invention is to introduce a data-driven dynamic threshold model, enabling the early warning system to adaptively learn and optimize early warning rules, thereby improving its adaptability to different vehicle states and charging scenarios.

[0013] To achieve the above objectives, the present invention adopts the following technical solution:

[0014] In a first aspect, the present invention provides an electric vehicle charging safety early warning system based on the fusion of collected vehicle-charging station interaction data and a VIN vehicle model database, characterized in that it includes:

[0015] • The charging pile intelligent data acquisition terminal is installed inside the charging pile and connected to the CAN data bus at the output end of the charging pile.

[0016] • Data acquisition module: During the charging process, the charging pile intelligent data acquisition terminal acquires the CAN bus data between the vehicle BMS and the charging pile in real time, and the standard it follows includes but is not limited to GB / T 27930.

[0017] • VIN Vehicle Model Database Module: Builds and maintains a VIN vehicle model database. By parsing the vehicle's VIN code, it associates and stores key static parameters such as the vehicle's brand, model, battery type, rated capacity, and rated voltage.

[0018] • Early warning rule library module: Stores multi-dimensional early warning rules based on national standards, manufacturer data and industry experience. These rules cover more than ten dimensions, including electrical safety (such as voltage and current), battery temperature (such as maximum temperature and temperature rise rate), and status monitoring (such as SOC jump and insulation resistance).

[0019] • Data Processing and Early Warning Engine: As the core of the system, this engine is configured as follows:

[0020] o receives real-time CAN bus data from the data acquisition layer and accurately parses the vehicle's VIN code from it.

[0021] Based on the parsed VIN code, query the VIN vehicle model database module to obtain the corresponding static safety parameters for that vehicle model (e.g., the maximum allowable charging voltage and maximum allowable temperature of the battery for that model).

[0022] Based on the obtained static safety parameters and the warning rule base, the initial threshold values ​​for various warnings applicable to this vehicle model are determined. For example, for model A, the upper limit for voltage warnings is set to 4.2V, while for model B, it may be set to 4.35V.

[0023] For early warning rules suitable for dynamic optimization (such as temperature rise rate), the engine is trained using historical and real-time data to dynamically learn and adapt to the optimal early warning rule, forming a dynamic threshold model. The engine smooths the threshold update process through anti-jitter calculations (such as lag comparison and moving average filtering), avoiding frequent false alarms caused by instantaneous data fluctuations, thereby achieving stable and reliable updates of the dynamic threshold.

[0024] • Early warning response layer module: responsible for delivering early warning information to users and the operation management platform through multiple channels such as APP push, SMS, and platform alarms. In high-risk situations, it can also coordinate with the charging pile to perform operations such as reducing power or urgently stopping charging.

[0025] Secondly, the present invention provides a method for early warning of electric vehicle charging safety based on the above-mentioned system, characterized by comprising the following steps:

[0026] S1: Real-time data acquisition. During the charging process, the CAN message data exchanged between the vehicle BMS and the charging pile is acquired in real time.

[0027] S2: VIN code parsing, which parses the vehicle's VIN code from the collected CAN message data.

[0028] S3: Static parameter query. Based on the parsed VIN code, query the VIN vehicle model database to obtain the static safety parameters of the vehicle model.

[0029] S4: Warning Threshold Determination. Based on the acquired static safety parameters, determine the various warning thresholds applicable to this vehicle model. This step includes:

[0030] For fixed rules (such as absolute upper limits), the initial threshold determined based on static parameters is directly adopted.

[0031] For dynamic optimization rules (such as relative rate of change), the dynamic threshold model is invoked to generate or update the currently applicable dynamic threshold based on historical and real-time data.

[0032] S5: Dynamic analysis and comparison, real-time parsing of dynamic parameters (such as real-time voltage, current, and temperature) in CAN messages, and real-time comparison and analysis with the threshold determined in step 4.

[0033] S6: Tiered Early Warning and Response. Based on the comparison results, tiered early warnings are triggered. For example, a Level 1 warning is an alert level, notifying the user via the app; a Level 2 warning is an operational level, where the system automatically links with the charging station to reduce power or stop charging, and simultaneously sends an alarm SMS.

[0034] Compared with the prior art, the present invention has the following significant beneficial effects:

[0035] 1. Significantly Improved Accuracy: By using a VIN vehicle database to "tailor-make" warning thresholds for different vehicle models, the drawbacks of a fixed threshold with a "one-size-fits-all" approach are fundamentally overcome. Experiments show that, compared to traditional methods, this invention can significantly reduce the false alarm rate to below 14.67%, and the false negative rate is close to zero.

[0036] 2. Strong Adaptability: The core of the system possesses "data-driven" characteristics. The dynamic threshold model can learn and optimize itself based on the vehicle's historical charging behavior and real-time status, enabling the warning system to adapt to long-term factors such as battery aging and seasonal changes, and possessing good scalability and adaptability.

[0037] 3. Comprehensive coverage with no blind spots: Data analysis is performed based on the complete vehicle-charging station interaction protocol, covering the entire process from charging handshake, parameter configuration, energy transfer to the end of charging. With up to 18 warning dimensions, it achieves comprehensive and thorough monitoring of charging safety risks.

[0038] 4. Timely response and effective handling: Through a tiered early warning mechanism and coordinated control capabilities, differentiated response measures can be taken for risks of different degrees. This ensures necessary user reminders while also automatically taking hard protection measures such as power outages in emergency situations, effectively preventing the occurrence and escalation of accidents.

[0039] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0040] Figure 1 This is a block diagram of the overall architecture of the system described in this embodiment of the invention.

[0041] Figure 2 This is a flowchart of the early warning method described in the embodiments of the present invention.

[0042] Figure 3 This is a schematic diagram of the dynamic threshold model training and update process in one embodiment of the present invention. Detailed Implementation

[0043] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments, but the scope of protection of the present invention is not limited thereto.

[0044] Example 1: System Architecture Reference Figure 1 The system in this embodiment includes:

[0045] • The data acquisition layer is deployed inside the charging pile and captures communication messages conforming to the GB / T 27930 protocol through a CAN interface card.

[0046] The VIN vehicle model database module, as an independent database service, stores static parameters for over 1,000 common electric vehicle models, which can be quickly queried using the VIN code.

[0047] The early warning rule base module exists in the form of a configuration file, which defines rules for 18 monitoring dimensions, including overvoltage / undervoltage, overcurrent, overheating, rapid temperature rise, abnormal SOC, and poor voltage consistency.

[0048] The data processing and early warning engine is a software module running on a cloud server, employing a microservice architecture. It receives data streams from a large number of charging piles and performs VIN parsing, data matching, threshold calculation, and risk assessment. The dynamic threshold model is implemented using online machine learning algorithms (such as regression models based on random forests or gradient boosting trees).

[0049] The early warning response layer integrates message queues and API interfaces, which can distribute early warning events to user mobile apps and operation and maintenance management dashboards, and can send control commands to charging piles through standard protocols (such as OCPP).

[0050] Example 2: Early Warning Method Flow

[0051] refer to Figure 2 A complete charging warning process is as follows:

[0052] S1. Install the charging pile intelligent data acquisition terminal inside the charging pile and connect it to the CAN data bus at the output end of the charging pile.

[0053] S2. Real-time data acquisition: When the vehicle is plugged into the charging gun to start charging, the data acquisition layer starts and continuously acquires CAN messages.

[0054] S3. VIN code parsing: The engine parses the VIN code (such as "LSVNV133XXXXXXXX") from the message and initiates a query to the VIN model database.

[0055] S4. Static Parameter Query: The vehicle database returns the static parameters of the corresponding vehicle model (e.g., "Model S 2023 of a certain brand") for this VIN: rated battery capacity 85kWh, rated voltage 350V, maximum allowable charging voltage 4.2V / cell, maximum allowable temperature 55℃, etc.

[0056] S5. Warning Threshold Determination: For the "Individual Cell Voltage Too High" rule, the engine load warning rule directly uses 4.2V as the initial threshold. For the "Temperature Rise Rate Too Fast" rule, a dynamic threshold model is invoked. Based on data from the vehicle's past 30 charging cycles, this model calculates that, at the current ambient temperature (25℃) and initial SOC (20%), the safe threshold for the temperature rise rate should be dynamically adjusted to ≤1.5℃ / min.

[0057] S6. Dynamic analysis and comparison: During the charging process, the engine monitors the voltage and temperature changes of each individual cell in real time. If a cell's voltage reaches 4.19V at a certain moment, the engine will trigger a level one warning, sending a notification to the user via the app: "Charging voltage is approaching its upper limit, please pay attention."

[0058] S7. Tiered Early Warning and Response: The engine detected that the average temperature rise rate of the battery pack reached 1.6℃ / min within one minute, exceeding the dynamic threshold of 1.5℃ / min. The engine immediately triggered a level two early warning, sending an emergency alarm to the user and the operation and maintenance platform stating "Abnormal temperature rise, system intervention is imminent," and simultaneously reducing the charging power to 50% through the charging pile control system.

[0059] Example 3: Implementation Reference of Dynamic Threshold Model Figure 3 The training and update process of the dynamic threshold model is as follows:

[0060] 1. Feature Engineering and Data Preprocessing

[0061] Before model training, feature extraction and standardization of the raw charging data are required:

[0062] • Feature data extraction: Based on the collected data and threshold rules and related parameters, feature data is extracted;

[0063] • Feature standardization: Z-score standardization is performed on all continuous features to ensure comparability of features with different dimensions.

[0064]

[0065] in:

[0066] oμ is the characteristic mean;

[0067] oσ is the characteristic standard deviation.

[0068] • Time series feature extraction: For dynamic parameters such as temperature rise rate, a sliding window calculation is used.

[0069]

[0070] in:

[0071] o Indicates the current temperature;

[0072] o This represents the temperature n sampling periods ago;

[0073] oτ is the sampling interval time.

[0074] 2. Threshold learning

[0075] • Calculate the temperature difference using Euclidean distance:

[0076]

[0077] in:

[0078] o Two eigenvectors

[0079] o Feature Dimension

[0080] o The i-th eigenvalue

[0081] • Calculation of outlier scores

[0082] K-nearest neighbor average distance

[0083]

[0084] in:

[0085] o The i-th nearest neighbor of sample x

[0086] o Number of neighbors

[0087] o Distance function

[0088] Threshold setting formula

[0089]

[0090] in:

[0091] o : p-th percentile

[0092] o Expected abnormality rate

[0093] o Normal sample size

[0094] 1) Anti-shake filtering algorithm

[0095] To prevent frequent fluctuations in the threshold, an improved anti-shake algorithm combining hysteresis comparison and moving average filtering is used for the calculated threshold.

[0096] • Moving average filtering:

[0097]

[0098] in:

[0099] oλ is the filter coefficient (0 < λ < 1), set according to system response requirements.

[0100] • Lag comparison update conditions:

[0101] Update conditions:

[0102]

[0103] in:

[0104] oδ is the fixed dead zone threshold.

[0105] oκ is the sensitivity coefficient (usually taken as 2-3).

[0106] o The standard deviation of recent threshold changes reflects system stability.

[0107] • Incremental update strategy: When the update conditions are met, an incremental update is adopted.

[0108]

[0109] in:

[0110] o Maximum update step size in a single operation

[0111] oη is the update rate coefficient.

Claims

1. An electric vehicle charging safety early warning system based on the fusion of vehicle-charging station interaction data and VIN vehicle model database, characterized in that, include: • Data acquisition layer, used to acquire CAN bus data exchanged between the vehicle BMS and the charging pile in real time during the charging process; • VIN vehicle model database module, used to store and maintain a VIN vehicle model database, which associates the vehicle's brand, model, battery type, rated capacity and rated voltage static parameters by parsing the vehicle's VIN code; • The early warning rule base module is used to store multi-dimensional early warning rules based on national standards, manufacturer data, and industry data; The data processing and early warning engine is configured as follows: o receives real-time CAN bus data from the data acquisition layer and parses out the vehicle VIN code from it; Based on the parsed VIN code, the static safety parameters of the corresponding vehicle model are retrieved from the VIN vehicle model database module; Based on the static safety parameters, determine the initial warning threshold applicable to this vehicle model; o For early warning rules suitable for dynamic optimization, training is performed based on historical and real-time vehicle-pile interaction data to dynamically learn and adapt the optimal early warning rules, forming a dynamic threshold model, and updating the dynamic threshold through anti-shake calculation. • The early warning response layer is used to trigger tiered early warnings based on the data processing and early warning engine outputs, and push the early warning information to users and management platforms through at least one communication channel.

2. The system according to claim 1, characterized in that, When forming a dynamic threshold model, the data processing and early warning engine is further configured to use a machine learning algorithm, taking the dynamic parameters in historical vehicle-charging station interaction data and their corresponding charging safety status as training samples, to iteratively optimize the threshold in the early warning rules, and then smoothing the optimized threshold through a de-shake filter to suppress noise interference and achieve stable updates of the dynamic threshold.

3. The system according to claim 1 or 2, characterized in that, The warning rules stored in the warning rule base module cover at least ten dimensions in electrical safety, battery temperature, and battery status monitoring.

4. The system according to claim 1, characterized in that, The early warning response layer can be linked to control the charging pile to reduce power or stop charging.

5. A method for early warning of electric vehicle charging safety based on the fusion of vehicle-charging station interaction data and VIN vehicle model database, characterized in that, Includes the following steps: Real-time acquisition of CAN message data between vehicle and charging station during the charging process; The vehicle VIN code is parsed from the CAN message data; Based on the VIN code, query the VIN vehicle model database to obtain the static safety parameters of the vehicle model; based on the obtained static safety parameters, determine the various warning thresholds applicable to the vehicle model; The system analyzes the dynamic parameters in the CAN message in real time and compares them with the warning threshold. Based on the comparison results, it triggers a graded warning and executes corresponding response measures.

6. The method according to claim 5, characterized in that, The steps of "determining the various warning thresholds applicable to this vehicle model" include: For fixed rules, the initial thresholds determined based on the static safety parameters are directly adopted; For dynamic optimization rules, the currently applicable dynamic thresholds are generated and updated by training a dynamic threshold model based on historical and real-time vehicle-pile interaction data.

7. The method according to claim 6, characterized in that, The training and updating process of the dynamic threshold model includes: collecting dynamic parameter sequences and corresponding charging safety status tags from historical charging processes; using machine learning algorithms to optimize the warning threshold parameters with the goal of minimizing false alarms and false alarms; performing anti-shake calculations on the newly optimized threshold and the current threshold, and retaining the current threshold if the change is within a preset tolerance range, otherwise updating the threshold by a predetermined step size.

8. The method according to claim 5, characterized in that, The tiered warning system includes at least two levels: Level 1 warning is an alert level, which notifies the user via APP or SMS; Level 2 warning is an operational level, in which the system automatically links with the charging pile to reduce power or stop charging.

Images