Battery intelligent monitoring and fault diagnosis method based on multi-source data fusion

Abstract

The invention relates to the technical field of new energy power batteries, in particular to an intelligent battery monitoring and fault diagnosis method based on multi-source data fusion, which comprises the following steps: step S10, performing battery appearance detection, if the detection is qualified, performing the next step, and if the detection is not qualified, performing fault analysis on the battery appearance, and then performing the next step; step S20, detecting the interior of the battery, if the detection is qualified, carrying out the next step, and if the detection is not qualified, carrying out fault analysis on the interior of the battery, and then carrying out the next step; step S30, classifying the batteries according to different faults detected by the batteries; step S40, uploading to a cloud system; according to the fault diagnosis method combining multi-source data fusion and environment dynamic calibration, the limitation of single parameter detection is broken through; a three-level decision-making tree and an intelligent sorting system are innovatively designed, and automatic grading processing of faults is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy power batteries, and in particular to a battery intelligent monitoring and fault diagnosis method based on multi-source data fusion. Background Art

[0002] New energy, also known as unconventional energy, refers to various forms of energy other than traditional energy. It refers to energy that has just begun to be developed and utilized or is being actively researched and is yet to be promoted, such as solar energy, geothermal energy, wind energy, ocean energy, biomass energy and nuclear fusion energy.

[0003] With the widespread application of new energy vehicles, a series of battery-related problems have also followed. When faulty packs on the market need to be returned to after-sales service, if the battery packs are disassembled blindly without being classified according to the faults, it will not only delay work hours but also cause unnecessary damage to the components inside the battery packs.

[0004] For internal faults in battery packs, existing technologies mostly rely on single parameter threshold comparisons, without considering the dynamic impact of environmental factors such as temperature and humidity on battery parameters (such as voltage and internal resistance), which can easily lead to misjudgments. In addition, the monitoring data of a single vehicle lacks multi-vehicle collaborative analysis. The cloud system only implements data storage and does not fully utilize the correlation of cluster data. At the same time, historical fault diagnosis relies on static models, which are difficult to adapt to complex scenarios such as battery aging and low-frequency faults.

[0005] In summary, the applicant has proposed a battery intelligent monitoring and fault diagnosis method based on multi-source data fusion. Summary of the Invention

[0006] The purpose of the present invention is to provide a battery intelligent monitoring and fault diagnosis method based on multi-source data fusion to save labor costs and improve network security. To achieve the above technical objectives, the technical solution of the present invention is as follows:

[0007] A battery intelligent monitoring and fault diagnosis method based on multi-source data fusion includes the following steps:

[0008] Step S10, perform a battery appearance inspection. If the inspection passes, proceed to the next step. If the inspection fails, perform a fault analysis on the battery appearance and then proceed to the next step.

[0009] Step S20, perform internal battery testing. If the test is qualified, proceed to the next step. If the test is unqualified, perform internal battery fault analysis and then proceed to the next step.

[0010] Step S30, classifying the batteries according to different faults detected;

[0011] Step S40: Upload to the cloud system.

[0012] Furthermore, the battery appearance inspection specifically includes the following steps:

[0013] Step S11: Check the lug system. Use a three-dimensional coordinate measuring instrument to detect the installation position deviation of the lugs (required to be ≤ ±0.5 mm), and check the depth of the threaded holes;

[0014] Step S12: Fastener inspection. Use a digital display torque wrench to randomly inspect the bolts outside the housing, and focus on checking the integrity of the sealing rubber ring of the waterproof screw of the maintenance hole cover;

[0015] Step S13: Evaluate the damage of the housing. Use a 3D profile scanner to detect the flatness of the outer shell, and use an infrared thermal imager to detect hidden cracks;

[0016] Step S14: Waterproof performance test. Conduct IP68 level verification on a rain test bench;

[0017] Step S15: Corrosion detection. Use an XRF spectrometer to analyze the composition of the oxide layer on the surface of the housing and determine the corrosion grade;

[0018] Step S16: Verify the identification system. Use a barcode scanner to check the identity code of the battery pack and check the integrity of the safety warning label.

[0019] Furthermore, the battery internal inspection specifically includes the following steps:

[0020] Step S21: Establish a data connection with the battery pack using a sensing device, and obtain the battery operation status parameters in real time and present them visually through a human-machine interface;

[0021] Step S22: After completing the acquisition of basic parameters, the system synchronously starts the environmental monitoring unit to dynamically capture environmental variables such as temperature, humidity, and air pressure at the work site, and generates a multi-dimensional environmental data map on the monitoring interface;

[0022] Step S23: By constructing a correlation model between environmental factors and battery performance parameters, the system can accurately quantify the influence weight of the external environment on the battery working characteristics, and then perform dynamic compensation and parameter calibration on the original detection data;

[0023] Step S24: The battery core parameters corrected by environmental factors will enter a dual diagnosis process. First, primary fault screening is performed based on the threshold comparison of environment-independent parameters, and then in-depth fault diagnosis is performed through the analysis of the time-series changes of the compensated parameters. Finally, a graded battery health status assessment report is formed.

[0024] Furthermore, the specific steps of step S30 specifically include the following steps:

[0025] Step S31, insulation failure. If the insulation resistance < 500 kΩ (the lowest national standard threshold), immediately trigger the highest-priority sorting and transfer it to the "insulation failure area".

[0026] Step S32, relay adhesion. If the relay cannot disconnect after being energized (the current continuously > 1 A for more than 10 s), it is determined as a relay failure and sorted to the "BMS failure area".

[0027] Step S33, abnormal internal resistance. If the internal resistance value exceeds the nominal value by 20% (e.g., the nominal internal resistance is 1 mΩ, and the measured value ≥ 1.2 mΩ), it is determined that the battery cell is aging or the connector is corroded, and sorted to the "battery cell aging area" or the "connector failure area".

[0028] Step S34, voltage imbalance. If ΔV > 50 mV and it does not recover after 3 consecutive charge-discharge cycles, it is determined that the battery cell consistency deteriorates and sorted to the "battery cell equalization and repair area".

[0029] Step S35, multiple faults occurring simultaneously. If insulation failure and abnormal internal resistance are detected simultaneously, give priority to handling according to the insulation fault.

[0030] Step S36, unknown fault. If the data does not meet the preset threshold, start the manual re-inspection mode, sort it to the "to-be-confirmed area" and trigger an alarm signal.

[0031] After improvement, the present invention further has the following beneficial effects:

[0032] 1. The present invention adopts a comprehensive detection strategy from the outside to the inside of the battery, greatly improving the battery detection efficiency. After adopting this method, the overall battery repair efficiency is increased by about 30% - 50% compared with the traditional method, and the residence time of the faulty battery pack in the after-sales link is greatly shortened.

[0033] 2. The present invention uses an advanced data processing algorithm to preprocess the collected battery data; by establishing a mathematical model of the relationship between environmental factors and battery parameter changes, it automatically eliminates the influence of external environmental factors such as temperature, humidity, and altitude on the battery pack parameters, so as to obtain more accurate battery performance data; through experimental verification, under complex and changeable environmental conditions, this method can improve the detection accuracy from about 70% - 80% of the traditional method to more than 90%, effectively avoiding misjudgment caused by environmental factor interference.

[0034] 3. With the aid of modern communication technologies, the present invention uploads the monitoring data of the battery to the cloud system in real time. The cloud system has powerful data storage and analysis capabilities and can continuously and dynamically analyze the battery data. Regardless of the operating state of the battery, users and vehicle manufacturers can obtain the latest monitoring information of the battery at any time through the network. This not only realizes the continuous monitoring of the repaired battery pack but also provides rich data support for the long-term evaluation and prediction of battery performance, timely discovers potential fault hazards, and ensures the reliable operation of the battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the overall process of the present invention.

[0036] Figure 2 It is a schematic diagram of the process for detecting the appearance of the battery of the present invention.

[0037] Figure 3 It is a schematic diagram of the process for detecting the interior of the battery of the present invention.

[0038] Figure 4 It is a schematic diagram of the process for classifying faults of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be further described below in conjunction with the drawings and embodiments.

[0040] Embodiment 1:

[0041] As Figures 1-4 shown, a battery intelligent monitoring and fault diagnosis method based on multi-source data fusion includes the following steps:

[0042] Step S10: Conduct an appearance inspection of the battery. If the inspection is qualified, proceed to the next step. If the inspection is unqualified, perform a fault analysis on the battery appearance and then proceed to the next step;

[0043] Step S20: Conduct an interior inspection of the battery. If the inspection is qualified, proceed to the next step. If the inspection is unqualified, perform a fault analysis on the battery interior and then proceed to the next step;

[0044] Step S30: Classify the battery according to different faults detected in the battery;

[0045] Step S40: Upload to the cloud system.

[0046] It should be noted that the battery state evaluation method of the present invention mainly relies on the following points:

[0047] Environmental interference elimination technology: By establishing an environmental compensation algorithm, the influence of external variables such as temperature and humidity on battery monitoring parameters is eliminated, and the characteristic parameters of the battery body are extracted. This compensation model can ensure that the detected values during fault diagnosis truly reflect the state of the battery body and avoid misjudgment caused by environmental factors.

[0048] Dynamic calibration system: Implement a dual-channel data correction mechanism to synchronously update the compensated core parameters to the monitoring interface to achieve visual monitoring of the parameters; at the same time, establish an automatic diagnosis channel and use the purified data for fault analysis; the calibration system is set with a parameter anomaly warning function, and an alarm is triggered when the corrected value deviates from the reference range.

[0049] Construction of a multi-dimensional evaluation system:

[0050] (1) Standard fusion mechanism: Integrate the battery factory technical specifications, industry detection standards, and historical operation data to establish a dynamic evaluation index library;

[0051] (2) Hierarchical threshold management: Set different parameter tolerance intervals for different battery types and usage cycles;

[0052] (3) Intelligent diagnosis engine: Use the compensated parameters to perform matching analysis with multiple levels of thresholds, and determine the abnormal state when the detected values exceed the allowable deviation three times in a row.

[0053] The specific battery appearance inspection includes the following steps:

[0054] Step S11, Lifting lug system verification, use a three-dimensional coordinate measuring instrument to detect the installation position deviation of the lifting lugs (required to be ≤ ±0.5 mm), and verify the depth of the threaded holes;

[0055] Step S12, Fastener inspection, use a digital display torque wrench to randomly inspect the bolts outside the housing, and focus on checking the integrity of the sealing rubber ring of the waterproof screw on the maintenance hole cover;

[0056] Step S13, Housing damage assessment, use a 3D profile scanner to detect the flatness of the housing (allowable error ≤ 0.3 mm / m 2 ), use an infrared thermal imager to detect hidden cracks;

[0057] Step S14, Waterproof performance test, conduct IP68 level verification on a rain test bench (water pressure 100 kPa, lasting 30 minutes);

[0058] Step S15, Corrosion detection, use an XRF spectrometer to analyze the composition of the oxide layer on the housing surface and determine the corrosion grade (according to the ASTM B117 standard);

[0059] Step S16, Identification system verification, use a barcode scanner to check the battery pack identity code and check the integrity of the safety warning label.

[0060] It should be noted that the three-dimensional coordinate measuring instrument uses laser ranging technology. By measuring the spatial coordinates of multiple measurement points on the lifting lug, the installation position deviation of the lifting lug is calculated, and the measurement accuracy can reach ±0.01 mm. In the waterproof performance test, if any slight leakage is found, the sealing condition of the rain test bench should be checked, and the suspicious parts of the battery pack should be marked for further detection later.

[0061] The specific internal battery detection includes the following steps:

[0062] Step S21: Establish a data connection with the battery pack using a sensing device, and obtain the battery operation state parameters in real time and present them visually through a human-machine interface.

[0063] Step S22: After the basic parameter collection is completed, the system synchronously starts the environmental monitoring unit to dynamically capture environmental variables such as temperature, humidity, and air pressure at the work site, and generate a multi-dimensional environmental data map on the monitoring interface.

[0064] Step S23: By constructing a correlation model between environmental factors and battery performance parameters, the system can accurately quantify the influence weight of the external environment on the battery working characteristics, and then perform dynamic compensation and parameter calibration on the original detection data.

[0065] The battery core parameters corrected by environmental factors will enter a dual diagnosis process. First, primary fault screening is implemented based on the threshold comparison of environment-independent parameters, and then in-depth fault diagnosis is carried out through the analysis of the temporal variation of the compensated parameters. Finally, a hierarchical battery health status assessment report is formed.

[0066] It should be noted that this solution uses a multi-channel sensing network to establish two-way data interaction with the battery pack, real-time collects the battery body parameters through the CAN bus, and presents them visually through a human-machine interface. The sensing device includes:

[0067] Electrochemical monitoring module: Equipped with a TIBQ40Z80 chipset, it obtains the single-cell voltage in real time with a resolution of 0.1 mV (0 - 5V range), and uses a four-wire Kelvin connection to eliminate the influence of contact resistance.

[0068] Thermal management unit: Configured with a 16-channel PT1000 temperature sensor array, using the T-type thermocouple compensation algorithm to achieve three-dimensional temperature field monitoring with an accuracy of ±0.2℃.

[0069] Mechanical state detection: Integrated with an ADXL357 three-axis vibration sensor (±40g range) and a Kistler601C pressure sensor, and captures the structural deformation characteristics of the battery pack through FFT spectrum analysis.

[0070] The environmental monitoring unit uses an SHT31 temperature and humidity sensor (±1.5% RH accuracy) and a BMP280 barometric pressure sensor (±0.12 hPa accuracy), combined with the Kalman filtering algorithm to remove high-frequency noise. With a data acquisition frequency of 10 Hz, conditioning circuits and data acquisition chips are used to achieve fast and accurate acquisition of environmental variables. In terms of software algorithms, the Kalman filtering algorithm is used to denoise the acquired data, improving the stability and accuracy of the data.

[0071] After the acquired data is preprocessed by the STM32H743 main control chip, multi-dimensional visualization is achieved through an industrial touch screen, including: dynamic polarization curves (supporting zooming and trend prediction); three-dimensional heat maps (based on the OpenGLES3.0 rendering engine); health state (SOH) circular indicators (0 - 100% dynamic filling);

[0072] An association model between environmental factors and battery performance parameters is constructed. The system can accurately quantify the influence weight of the external environment on the battery's operating characteristics, and then perform dynamic compensation and parameter calibration on the data, as shown below:

[0073] Temperature compensation voltage:

[0074] V 校准 = V 原始 + K T ·(T - T 基准 )(K T = -0.005 V / °C, T 基准 = 25°C)

[0075] Humidity compensation resistance:

[0076] R 校准 = R 原始 + K H ·(H - H 基准 )(K H = 0.0001 Ω / %RH, H 基准 = 50%)

[0077] Barometric pressure compensation internal resistance:

[0078] R 内阻校准 = R 内阻原始 + K P ·(P - P 基准 )(K P = 0.0001 Ω / kPa, P 基准 = 101.325 kPa).

[0079] The interference law of environmental parameters on battery performance indicators is as follows:

[0080] Temperature effect:

[0081] Positive temperature change: When the temperature rises, the voltage shows a downward trend, and the capacity value increases synchronously (the capacity improvement will accelerate the battery aging process);

[0082] Negative temperature change: When the temperature drops, the voltage level rises accordingly, and the capacity value decreases correspondingly;

[0083] Humidity effect:

[0084] Increasing humidity directly leads to an increase in the measured resistance value;

[0085] Decreasing humidity causes a linear decrease in the measured resistance value;

[0086] Air pressure effect:

[0087] When the air pressure increases, the voltage reading increases positively with the air pressure value, and at the same time, the internal resistance parameter decreases;

[0088] When the air pressure decreases, the voltage indication decreases in the same direction as the air pressure change. At this time, the measured internal resistance value will increase synchronously;

[0089] The cross - effects of each environmental variable on battery parameters show a clear one - way characteristic: temperature is negatively correlated with voltage and positively correlated with capacity; humidity is positively correlated with resistance; air pressure is positively correlated with voltage and negatively correlated with internal resistance. This coupling relationship requires the establishment of a corresponding environmental compensation algorithm when evaluating battery performance. Through the analysis of the time - series changes of the compensated parameters, in - depth fault diagnosis is carried out.

[0090] The specific steps of step S30 are as follows:

[0091] Step S31, insulation failure. If the insulation resistance < 500 kΩ (the lowest national standard threshold), immediately trigger the highest - priority sorting and transfer it to the "insulation failure area";

[0092] Step S32, relay adhesion. If the relay cannot disconnect after being energized (the current continues to be > 1 A for more than 10 s), it is determined as a relay fault and sorted to the "BMS fault area";

[0093] Step S33, abnormal internal resistance. If the internal resistance value exceeds 20% of the nominal value (for example, if the nominal internal resistance is 1 mΩ and the measured value is ≥ 1.2 mΩ), it is determined that the battery cell is aging or the connector is corroded, and sorted to the "battery cell aging area" or the "connector failure area";

[0094] Step S34, voltage imbalance. If ΔV > 50 mV and it does not recover after 3 consecutive charge - discharge cycles, it is determined that the consistency of the battery cells deteriorates and sorted to the "battery cell balance repair area";

[0095] Step S35, multiple faults occurring simultaneously. If insulation failure and abnormal internal resistance are detected at the same time, the insulation fault is processed first;

[0096] Step S36, unknown fault. If the data does not meet the preset threshold, start the manual re-inspection mode, sort it to the "to be confirmed area" and trigger an alarm signal.

[0097] It should be noted that step S31 and step S32 belong to the first-level determination (safety first), step S33 and step S34 belong to the second-level determination (performance-related faults), and step S35 and step S36 belong to the third-level determination (comprehensive diagnosis);

[0098] The operation details of the sorting actuator are as follows:

[0099] Path switching mechanism: The central controller outputs a sorting instruction according to the fault type, and the PLC controls the electromagnetic guide to adjust the branch path of the conveyor track:

[0100] Cell fault area: Guide A is energized, and the track deflects 15° to the left;

[0101] BMS fault area: Guide B is energized, and the track deflects 15° to the right;

[0102] Connector fault area: Guide C is energized, and the track remains straight.

[0103] Sorting action timing:

[0104] Positioning synchronization: When the battery pack reaches the sorting point, the photoelectric sensor triggers a position signal, and the controller reads the current tag information;

[0105] Instruction issuance: According to the bound fault type, issue an electromagnetic guide action instruction within 50 ms;

[0106] Mechanical execution: The action response time of the guide is ≤20 ms to ensure that the battery pack accurately enters the target area (position error ±2 mm).

[0107] Exception handling:

[0108] Sorting failure: If the battery pack does not move along the expected path (detected by the photoelectric sensor in the passing area), the system automatically pauses the conveyor track and starts the correction robotic arm to push it into the correct track;

[0109] Equipment conflict: If the distance between two battery packs is too small (<300 mm), the controller dynamically reduces the track speed to avoid collision.

[0110] Embodiment 2:

[0111] A new energy vehicle battery intelligent monitoring and fault diagnosis system, characterized in that it includes a multi-source data acquisition module, an environmental dynamic calibration module, a multi-level fault diagnosis module, a cloud collaborative analysis platform, and an intelligent sorting actuator.

[0112] The environmental dynamic calibration module adopts temperature, humidity, and air pressure compensation algorithms. The specific formulas are as follows:

[0113] Temperature compensation voltage: V 校准 = V 原始 + K T ·(T - T 基准 )(K T = -0.005V / °C, T 基准 = 25°C)

[0114] Humidity compensation resistance: R 校准 = R 原始 + K H ·(H - H 基准 )(K H = 0.0001Ω / %RH, H 基准 = 50%)

[0115] Air pressure compensation internal resistance: R 内阻校准 = R 内阻原始 + K P ·(P - P 基准 )(K P = 0.0001Ω / kPa, P 基准 = 101.325kPa).

[0116] The multi-level fault diagnosis module adopts a three-level decision tree model, with the priorities being safety faults, performance faults, and comprehensive diagnosis in sequence.

[0117] The intelligent sorting actuator adjusts the conveying path through a PLC-controlled electromagnetic guide, with a response time ≤ 20ms and a positioning error of ±2mm.

[0118] The cloud collaborative analysis platform supports full-life cycle tracking, cross-platform data sharing, and dynamic optimization of AI models.

[0119] Cloud monitoring system is as follows:

[0120] 1. System architecture and functional modules

[0121] (1) Cloud monitoring system architecture

[0122] Data acquisition layer: Deployed on the edge computing gateway of the local classification device, it receives classification results, detection data (insulation resistance, internal resistance, voltage balance), and maintenance records in real time; the edge computing gateway uses an industrial-grade wireless router with a high-performance ARM processor, equipped with 4 Ethernet ports, 2 USB interfaces, and a wireless communication module, supporting multiple communication protocols. In terms of software, it integrates functions such as data acquisition, preprocessing, and encrypted transmission, and can achieve efficient processing and secure transmission of local data.

[0123] Transport Layer: Encrypted transmission to ensure data timeliness and security.

[0124] Storage Layer:

[0125] ■ Time-Series Database: Store high-frequency detection data (sampling interval 1s);

[0126] ■ Relational Database: Store structured data such as battery pack ID, classification results, and maintenance history

[0127] ■ Object Storage: Archive unstructured data such as maintenance process images and test report PDFs.

[0128] Application Layer:

[0129] ■ Visualization Dashboard: Dynamically display the heat map of battery pack distribution, statistics of fault types, and analysis of maintenance efficiency;

[0130] ■ Early Warning Module: Predict the risk of secondary battery pack failures based on historical data (such as the trend of sudden increase in internal resistance);

[0131] ■ API Interface: Open data query service, supporting vehicle manufacturers and recycling enterprises to retrieve data according to permissions.

[0132] The visualization dashboard displays the distribution of the number of battery packs of different fault types through bar charts, shows the changing trend of maintenance efficiency over time through line charts, and simultaneously displays the heat map of the location distribution of battery packs in real time, facilitating managers to intuitively understand the overall situation of battery packs;

[0133] The data of the entire life cycle of the battery pack is stored in JSON format, facilitating data transmission and parsing. The query interface adopts the RESTful API design, supporting data query through the HTTP protocol. The interface provides rich query parameters, such as battery pack ID, time range, etc., to meet the query needs of different users.

[0134] (2) Core Functions

[0135] Full Life Cycle Tracking: Record the status of the battery pack throughout the process from factory return inspection, repair, re-testing to secondary installation; support quick query of historical data by scanning the code (maintenance report QR code).

[0136] Cross-Platform Collaboration: The maintenance center shares data with vehicle manufacturers to optimize battery design defects (such as abnormal aging rate of a certain batch of battery cells); recycling enterprises obtain the health state (SOH) of retired batteries to accurately evaluate the value of cascade utilization; the maintenance center obtains the internal resistance abnormality rate of a certain batch of batteries (such as reaching 15% in the Q2 batch of 2023) through the API. After feedback to the vehicle manufacturer, the vehicle manufacturer optimizes the battery sealing process, reducing the internal resistance abnormality rate of subsequent batches to 5%.

[0137] AI optimization feedback: Utilizes cloud-based big data to train fault prediction models and regularly updates the decision tree algorithm of the local classification system; automatically generates maintenance recommendations (such as "Replace the connecting bolts with anti-corrosion coated ones").

[0138] 2. Data transmission and processing flow

[0139] (1) Data upload rules

[0140] Trigger conditions: After the battery pack is classified and sorted, the basic information (ID, fault type, sorting time) is uploaded immediately; after each maintenance operation is completed, the maintenance record is updated synchronously; the full inspection data compressed package is automatically uploaded at dawn every day.

[0141] 2) Exception handling mechanism

[0142] Resume data after network disconnection: The edge gateway caches data that has not been successfully uploaded and gives priority to retransmission after the network is restored;

[0143] Data Verification: The cloud receiving end verifies data integrity, discards abnormal data packets, and triggers local retransmission. Cloud data transmission uses the AES-256 encryption protocol. The edge gateway caches unuploaded data (capacity 16GB) and retransmits it in timestamp order after network recovery. Data integrity verification is implemented using the CRC32 algorithm.

[0144] Security protection: Data transmission is encrypted; database access rights are assigned by role (e.g. maintenance personnel can only read data in this area).

[0145] The present invention is described above by way of example in conjunction with the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as various non-substantial improvements are made using the method concept and technical solution of the present invention; or the above-mentioned concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the scope of protection of the present invention.

Claims

1. A battery intelligent monitoring and fault diagnosis method based on multi-source data fusion, characterized in that, It includes the following steps: Step S10: Conduct battery appearance inspection. If the inspection is qualified, proceed to the next step. If the inspection is unqualified, conduct fault analysis on the battery appearance and then proceed to the next step; Step S20: Conduct battery internal inspection. If the inspection is qualified, proceed to the next step. If the inspection is unqualified, conduct fault analysis on the battery internal and then proceed to the next step; Step S30: Classify the batteries according to different faults detected in the battery inspection; Step S40: Upload to the cloud system.

2. The battery intelligent monitoring and fault diagnosis method based on multi-source data fusion according to claim 1, wherein, The specific battery appearance inspection includes the following steps: Step S11: Lug system verification. Use a three-dimensional coordinate measuring instrument to detect the installation position deviation of the lugs (required to be ≤ ±0.5 mm), and verify the depth of the threaded holes; Step S12: Fastener inspection. Use a digital display torque wrench to randomly inspect the bolts outside the housing, and focus on checking the integrity of the sealing rubber ring of the waterproof screw of the maintenance hole cover plate; Step S13: Shell damage assessment. Use a 3D profile scanner to detect the flatness of the outer shell, and use an infrared thermal imager to detect hidden cracks; Step S14: Waterproof performance test. Conduct IP68 level verification on the rain test bench; Step S15: Corrosion detection. Use an XRF spectrometer to analyze the composition of the oxide layer on the surface of the housing and determine the corrosion grade; Step S16: Identification system verification. Check the battery pack identity code through a barcode scanner and check the integrity of the safety warning label.

3. A battery intelligent monitoring and fault diagnosis method based on multi-source data fusion according to claim 1, characterized in that, The specific battery internal inspection includes the following steps: Step S21: Establish a data connection with the battery pack using a sensing device, obtain the battery operation status parameters in real time and present them visually through a human-machine interface; Step S22: After completing the basic parameter collection, the system synchronously starts the environmental monitoring unit to dynamically capture environmental variables such as temperature, humidity, and air pressure at the work site, and generate a multi-dimensional environmental data map on the monitoring interface; Step S23: By constructing an association model between environmental factors and battery performance parameters, the system can accurately quantify the influence weight of the external environment on the battery working characteristics, and then perform dynamic compensation and parameter calibration on the original detection data; The battery core parameters corrected by environmental factors will enter a dual diagnosis process. First, primary fault screening is implemented based on the threshold comparison of environment-independent parameters, and then in-depth fault diagnosis is performed through the time-series change analysis of the compensated parameters. Finally, a hierarchical battery health status assessment report is formed.

4. A battery intelligent monitoring and fault diagnosis method based on multi-source data fusion according to claim 1, characterized in that The specific Step S30 includes the following steps: Step S31: Insulation failure. If the insulation resistance < 500 kΩ (the lowest national standard threshold), immediately trigger the highest-priority sorting and transfer it to the "insulation fault area"; Step S32: Relay adhesion. If the relay cannot be disconnected after being attracted (the current continues to be > 1 A for more than 10 s), it is determined as a relay fault and sorted to the "BMS fault area"; Step S33: Abnormal internal resistance. If the internal resistance value exceeds 20% of the nominal value (for example, the nominal internal resistance is 1 mΩ, and the actual measurement is ≥ 1.2 mΩ), it is determined as cell aging or connector corrosion and sorted to the "cell aging area" or "connector fault area"; Step S34, voltage imbalance. If ΔV > 50 mV and it does not recover after 3 consecutive charge-discharge cycles, it is determined that the cell consistency deteriorates, and it is sorted to the "cell balance repair area"; Step S35, multiple faults occurring simultaneously. If insulation failure and abnormal internal resistance are detected at the same time, the insulation fault is preferentially processed; Step S36, unknown fault. If the data does not meet the preset threshold, start the manual re-inspection mode, sort it to the "to be confirmed area" and trigger an alarm signal.

Images