Container battery system intelligent monitoring method and system based on internet of things

Abstract

The application discloses a container type battery system intelligent monitoring method and system based on the Internet of Things, relates to the technical field of energy storage battery safety monitoring, and comprises the following steps: acquiring transportation stage characteristic parameters, warehouse stage chemical escape characteristic parameters and use stage baseline characteristic parameters respectively; performing cross-stage alignment and normalization processing on each characteristic data based on a container identity and a timestamp, constructing a full-cycle state parameter set, and determining a mechanical damage component value, a chemical damage component value and an electrochemical aging component value; then, performing risk grade division according to damage indexes of each stage, performing transportation scheduling, warehouse isolation re-inspection and charge-discharge constraint control, and realizing data integrity checking and anti-tampering tracing through alliance chain hash recording; and the application significantly improves the cross-stage risk identification, collaborative control and auxiliary responsibility analysis capability of the container type battery system.

Description

Technical Field

[0001] This invention relates to the field of energy storage battery safety monitoring technology, specifically to an intelligent monitoring method and system for containerized battery systems based on the Internet of Things. Background Technology

[0002] With the continuous expansion of new energy power generation and industrial and commercial energy storage scenarios, containerized battery systems are widely used in energy storage peak shaving, valley filling and emergency power supply scenarios due to their characteristics of concentrated capacity, convenient transportation and high on-site deployment efficiency. The intelligent monitoring method and system of containerized battery systems based on the Internet of Things is developed around the data perception, status analysis, risk judgment and scheduling control needs of such battery systems in the continuous process of transportation, warehousing and commissioning after leaving the factory. Through sensors, communication gateways and monitoring platforms, information such as vibration, environment, gas, electrochemical baseline, etc. are continuously collected and correlated, so that the containerized battery system is no longer monitored only in the operation stage, but can form a status perception throughout the whole life cycle from the beginning of logistics transportation.

[0003] However, the monitoring of existing containerized battery systems is mostly focused on operational data such as voltage, current, temperature, SOC, and SOH after connection to the power station. There is a lack of continuous recording and unified analysis of road impacts, loading and unloading collisions, continuous vibrations during transportation, and the closed environment, excessive temperature and humidity, and the tendency of hydrogen, carbon monoxide, or volatile organic compounds to escape during storage. This results in fragmented transportation data, storage data, and usage data, making it difficult to determine whether there is a correlation between mechanical damage, chemical anomalies, and electrochemical aging that have occurred before commissioning. It is also difficult to implement timely measures such as adjusting transportation routes, warehouse isolation and re-inspection, commissioning rate restrictions, or traceability analysis based on the risk level at different stages. Therefore, there is a need for an intelligent monitoring solution that can align, normalize, and convert cross-stage data into comparable damage indices based on container identification and timestamps. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent monitoring method and system for containerized battery systems based on the Internet of Things, thus solving the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring method for containerized battery systems based on the Internet of Things, comprising the following steps: S1. Obtain characteristic parameters of the transportation stage, chemical escaping characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage, respectively. S2. Align the characteristic parameters of the transportation stage, the chemical emission characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage across stages using container identification and timestamp as indexes, and perform normalization processing to construct a full-cycle state parameter set. S3. Based on the full-cycle state parameter set, determine the mechanical damage component value in the transportation stage, the chemical damage component value in the storage stage, and the electrochemical aging component value in the use stage, and use them as the damage index for the corresponding stage. S4. Compare the damage index of each stage with the preset risk threshold set to determine the risk level classification result, and perform cross-stage collaborative scheduling operation based on the risk level classification result. S5. Record the status data of each stage to the consortium blockchain in a hash-anchored manner to obtain the data integrity verification log, and determine the stage risk contribution ranking results and auxiliary responsibility analysis results based on the data integrity verification log.

[0006] Preferably, the logic for obtaining characteristic parameters of the transportation stage, chemical efflux characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage is as follows: During the transportation phase, vibration data is continuously collected at a preset transportation vibration sampling frequency using a triaxial accelerometer installed at the bottom of the container. Environmental temperature and humidity data are continuously collected at preset transportation environment sampling intervals using a cargo hold temperature and humidity sensor installed outside the container. The vibration data is processed using a Fast Fourier Transform (FFT) to calculate the impact energy value for each time period. Continuous time periods where the impact energy value exceeds a preset single-frame impact threshold are marked as abnormal candidate time periods. The impact energy values ​​corresponding to consecutive abnormal candidate time periods are accumulated to obtain the cumulative impact energy value of the abnormal impact event. When the cumulative impact energy value of the abnormal impact event exceeds a preset event impact threshold, the consecutive abnormal candidate time periods are confirmed as abnormal impact events. Simultaneously, the duration, peak acceleration, and cumulative impact energy value of the abnormal impact event are recorded as characteristic parameters of the transportation phase. During the storage phase, a gas sensor array installed inside the container continuously collects hydrogen, carbon monoxide, and volatile organic compound (VOC) concentration data at preset gas collection intervals. Simultaneously, a storage temperature and humidity sensor continuously collects ambient temperature and humidity data at preset temperature and humidity collection intervals. Baseline drift correction is performed on the hydrogen, carbon monoxide, and VOC concentration data, and the linear fitting slope of each gas concentration during the storage phase is calculated. The linear fitting slope of each gas concentration, the target escaping gas type, and the linear fitting slope of the target gas concentration are used as chemical escaping characteristic parameters during the storage phase. During the first charge-discharge cycle of the containerized battery system in the usage phase, the initial DC internal resistance and initial capacity of each cell are obtained as baseline characteristic parameters for the usage phase.

[0007] Preferably, the logic for constructing the full-cycle state parameter set is as follows: Based on the container identification, the characteristic parameters of the transportation stage, the chemical escaping characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage are respectively appended with the corresponding stage start timestamp and stage end timestamp. Based on the cumulative impact energy, peak acceleration, and duration of each abnormal impact event during the transportation phase, the raw mechanical damage data is constructed. The raw mechanical damage data also includes the total cumulative impact energy during the transportation phase. Based on the chemical escape characteristic parameters and cumulative temperature and humidity values ​​during the storage stage, the original chemical state data is constructed. Among them, the cumulative temperature and humidity values ​​are obtained by normalizing and summing the cumulative temperature exceedance of the storage stage ambient temperature data relative to the preset standard temperature range and the cumulative humidity exceedance of the storage stage ambient humidity data relative to the preset standard humidity range. Based on the baseline characteristic parameters of the usage phase, construct the raw data of the electrochemical baseline; The mechanical damage raw data, chemical state raw data, and electrochemical baseline raw data belonging to the same container identity are combined in stages according to time sequence to generate a full-cycle raw state table for the corresponding container identity. The parameters in the original state table for the entire cycle are normalized according to their corresponding parameter types to generate a set of state parameters for damage analysis. Among them, mechanical damage parameters are normalized based on the impact energy baseline value of events under non-destructive transportation conditions, which is the statistical value of the impact energy of events under non-destructive transportation conditions. Chemical emission parameters are normalized based on the baseline of the ambient gas concentration at the factory standard. Electrochemical baseline parameters are normalized based on the rated internal resistance and rated capacity in the factory test report.

[0008] Preferably, the logic for determining the damage index at the corresponding stage is as follows: Feature identification is performed on the full-cycle state parameter set, and the cumulative impact energy value of each abnormal impact event is extracted and statistically analyzed. The cumulative impact energy value, peak acceleration and duration of the abnormal impact event corresponding to the i-th abnormal impact event are multiplied to obtain the single mechanical damage component value of the i-th abnormal impact event. The single mechanical damage component values ​​of all abnormal impact events are accumulated to determine the mechanical damage component value, which is used as the mechanical damage index for the transportation stage. The slope of the linear fitting of the target gas concentration during the storage stage is extracted from the set of state parameters throughout the entire cycle. When the slope of the linear fitting of the target gas concentration is positive and exceeds the preset escape slope threshold, the slope of the linear fitting of the target gas concentration is multiplied by the cumulative value of temperature and humidity to obtain the chemical damage component value, which is used as the chemical damage index during the storage stage. When the slope of the linear fitting of the target gas concentration is not positive, or the slope of the linear fitting of the target gas concentration does not exceed the preset escape slope threshold, the chemical damage component value is set to zero. The internal resistance increment and capacity decay of each cell during the service stage are extracted from the full-cycle state parameter set. The internal resistance increment and capacity decay of each cell are weighted and summed to obtain the electrochemical aging component value, which is used as the aging damage index during the service stage.

[0009] Preferably, the logic for performing cross-stage collaborative scheduling operations based on risk level classification results is as follows: Damage indices at different stages are compared with a set of risk thresholds to obtain risk level classification results; the set of risk thresholds includes a first-level warning threshold for the transportation stage, a second-level warning threshold for the storage stage, and multiple charge and discharge constraint thresholds for the usage stage. During the transportation phase, when the mechanical damage index exceeds the preset first-level warning threshold, a transportation suggestion signal is sent to the logistics scheduling system through the vehicle communication unit. The transportation suggestion signal includes reselecting a transportation route with road smoothness parameters that meet the preset road smoothness threshold and driving within the preset vibration reduction transportation speed range. When the mechanical damage index does not exceed the first-level warning threshold, no additional transportation suggestion signal is issued. During the warehousing phase, when the chemical damage index during warehousing is greater than or equal to the Level 2 warning threshold, the corresponding container identification is marked as a warehousing isolation re-inspection object, and the warehousing isolation re-inspection object is transferred from the regular outbound queue to the isolation re-inspection outbound queue, ventilation detection area, or fire monitoring area, triggering a manual review process. When there are multiple warehousing isolation re-inspection objects, an isolation re-inspection list is generated in descending order of the chemical damage index during warehousing. If the chemical damage index during warehousing is the same, they are arranged in ascending order of entry time. When the chemical damage index during warehousing is less than the Level 2 warning threshold, the corresponding container identification is marked as a regular warehousing object, and a regular outbound list is generated in ascending order of entry time. Based on the isolation re-inspection list and the regular outbound list, an updated warehousing re-inspection and outbound plan list is generated. During the usage phase, when the containerized battery system is connected to the power station and put into operation, the aging damage index of the usage phase is compared with multiple charge and discharge constraint thresholds. Among them, multiple charge and discharge constraint thresholds are respectively established with multiple rate upper limits. When the aging damage index of the usage phase exceeds the corresponding charge and discharge constraint threshold, the corresponding rate upper limit is determined in the charge and discharge constraint mapping relationship according to the risk level classification result, and the maximum allowable charge and discharge rate is limited to no more than the corresponding rate upper limit.

[0010] Preferably, the logic for determining the stage risk contribution ranking results and auxiliary responsibility analysis results is as follows: Establish a phase integrity relationship chain corresponding to the container identity in the consortium blockchain. The data integrity verification log consists of the hash value and corresponding state data index recorded by each node in the phase integrity relationship chain. At the end of the transportation phase, the transportation status data is recorded, and the first hash value of the transportation status data is determined. The first hash value is written into the first node of the phase integrity relation chain. The transportation status data includes container identification, transportation start and end timestamps, abnormal impact events, mechanical damage component values, and total impact energy accumulation value of the transportation phase. At the end of the storage phase, the storage status data is recorded, and the second hash value of the storage status data is determined. The second hash value is written into the second node of the phase integrity relation chain. The storage status data includes container identification, storage start and end timestamps, chemical escape characteristic parameters of the storage phase, cumulative temperature and humidity values, and chemical damage component values. During the usage phase, the usage status data is recorded, and the third hash value of the usage status data is determined. The third hash value is then written to the third node of the phase integrity relationship chain. The usage status data includes container identification, usage timestamp, initial DC internal resistance, initial capacity, and electrochemical aging component value. When a safety incident occurs in the containerized battery system at any stage, the last stage node that was written into the stage integrity relation chain at the time of the incident is determined, and the stage integrity relation chain is traced back from the last stage node. The state data corresponding to the hash values ​​of each stage written into the stage integrity relation chain is verified in turn. The trace hash value is obtained by re-hash calculation on the state data, and the trace hash value is compared with the corresponding hash value written into the consortium chain. When the traceability hash value of any stage is inconsistent with the corresponding hash value written to the consortium blockchain, that stage is identified as a data anomaly stage and a data anomaly prompt is output. When the traceability hash values ​​of all stages are consistent, the non-negative excess ratio of the mechanical damage component value relative to the first-level warning threshold is determined as the risk contribution value of the transportation stage, the non-negative excess ratio of the chemical damage component value relative to the second-level warning threshold is determined as the risk contribution value of the storage stage, and the non-negative excess ratio of the electrochemical aging component value relative to the lowest charge / discharge constraint threshold among multiple charge / discharge constraint thresholds is determined as the risk contribution value of the usage stage. Among these, when the corresponding damage component value does not exceed the corresponding risk threshold, the non-negative excess ratio of the corresponding stage is marked as zero; when the corresponding damage component value exceeds the corresponding risk threshold, the non-negative excess ratio of the corresponding stage is determined according to the ratio of the difference between the corresponding damage component value and the corresponding risk threshold to the corresponding risk threshold. The risk contribution ranking results are generated in descending order of risk contribution value in the transportation stage, storage stage, and usage stage. The risk contribution ranking results are then used together with the manual review rules as auxiliary liability analysis results.

[0011] Preferably, the vibration data is subjected to a fast Fourier transform, and the cumulative impact energy value within each time period is calculated, specifically including: Based on the vibration data continuously collected by the triaxial accelerometer during the transportation phase, and combined with the vector magnitude synthesis method, a vibration acceleration time-domain sequence is obtained. The vibration acceleration time-domain sequence is then processed into frames, and each frame is windowed to obtain windowed vibration frame data. The windowed vibration frame data is then processed by Fast Fourier Transform to obtain the spectrum of each frame. The spectral amplitude within the preset vibration analysis frequency range is extracted, and the total energy of a single frame is calculated. This total energy is then used as the single-frame impact energy of that frame. The single-frame impact energies of all frames in the current transportation phase are accumulated to obtain the total cumulative impact energy of the transportation phase. When the single-frame impact energy of any frame exceeds the preset single-frame impact threshold, the frame is marked as an abnormal frame, and the single-frame impact energy of the abnormal frame is added to the single-frame impact energy of the subsequent consecutive frames marked as abnormal, until a frame that does not exceed the preset single-frame impact threshold appears. The combined impact energy of consecutive abnormal frames is used as the cumulative energy value of a candidate impact event. When the cumulative energy value of the candidate impact event exceeds the preset event impact threshold, the consecutive abnormal frames are confirmed as an abnormal impact event. The cumulative energy value of the candidate impact event is used as the cumulative impact energy value of the abnormal impact event. The duration of the consecutive abnormal frames is used as the duration of the abnormal impact event. The maximum peak value of the vibration acceleration time-domain sequence within the consecutive abnormal frames is used as the peak acceleration of the abnormal impact event.

[0012] Preferably, baseline drift correction is performed on the concentration data of each gas, and the slope of the linear fitting of the concentration of each gas during the storage stage is calculated, specifically including: The first gas concentration sample value after the start of the storage phase is obtained as the initial concentration baseline value of the corresponding gas; Subtract the initial concentration baseline value of the corresponding gas from the concentration sampling value of each gas within each gas sampling interval in the storage area to obtain the drift correction gas concentration value of the corresponding gas. A sliding observation window containing a preset number of gas collection intervals is set. The drift correction gas concentration values ​​of each gas within the sliding observation window are subjected to least squares linear fitting to obtain the unit concentration linear fitting slope of each gas within the sliding observation window. After obtaining the unit concentration linear fitting slope of each gas within the sliding observation window, the sliding observation window is moved forward by one storage gas collection interval, and the least squares linear fitting process is performed again to obtain the unit concentration linear fitting slope of each gas within the next sliding observation window. The maximum value of the linear fitting slope for the unit concentration of the same gas within each sliding observation window is taken as the linear fitting slope for the concentration of the corresponding gas. The slopes of the linear fitting of the concentrations of each gas are compared, and the gas with the largest linear fitting slope is identified as the target escaping gas. The slope of the linear fitting of the concentration of the target escaping gas is then identified as the slope of the linear fitting of the concentration of the target gas.

[0013] Preferably, the weighted summation of the internal resistance increment and capacity decay of each cell is performed, specifically including: Feature identification is performed on the electrochemical baseline parameters during the usage phase to obtain the initial DC internal resistance and initial capacity of each cell. Based on the rated internal resistance and rated capacity of each cell extracted from the factory test report, the initial DC internal resistance is subtracted from the rated internal resistance in the factory test report to obtain the internal resistance difference. The ratio of the internal resistance difference to the rated internal resistance in the factory test report is used as the internal resistance increment of the corresponding cell. When the internal resistance increment is less than zero, the internal resistance increment of the corresponding cell is marked as zero. The rated capacity in the factory test report is subtracted from the initial capacity to obtain the capacity difference. The ratio of the capacity difference to the rated capacity in the factory test report is used as the capacity decay of the corresponding cell. When the capacity decay is less than zero, the capacity decay of the corresponding cell is marked as zero. Based on the internal resistance increment and capacity decay of each cell, the average value of the internal resistance increment of each cell is calculated to obtain the average value of the internal resistance increment; the average value of the capacity decay of each cell is calculated to obtain the average value of the capacity decay, and these are respectively used as the normalized internal resistance increment and the normalized capacity decay. The normalized internal resistance increment is multiplied by a preset first aging coefficient to obtain the normalized internal resistance aging term. The normalized capacity decay is multiplied by a preset second aging coefficient to obtain the normalized capacity aging term. The normalized internal resistance aging term and the normalized capacity aging term are added together to obtain the electrochemical aging component value.

[0014] The IoT-based intelligent monitoring system for containerized battery systems includes: The data acquisition module is used to acquire characteristic parameters of the transportation stage, chemical emission characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage, respectively. The full-cycle construction module is used to align the characteristic parameters of the transportation stage, the chemical emission characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage across stages using container identification and timestamp as indexes, and then perform normalization processing to construct a full-cycle state parameter set. The damage analysis module determines the mechanical damage component value during the transportation stage, the chemical damage component value during the storage stage, and the electrochemical aging component value during the use stage based on the full-cycle state parameter set, and uses them as the damage index for the corresponding stage. The cross-stage monitoring module is used to compare the damage index of each stage with a preset set of risk thresholds, determine the risk level classification results, and perform cross-stage collaborative scheduling operations based on the risk level classification results. The traceability analysis module is used to record the status data of each stage to the consortium blockchain in a hash-anchored manner, obtain the data integrity verification log, and determine the stage risk contribution ranking results and auxiliary responsibility analysis results based on the data integrity verification log.

[0015] This invention provides an intelligent monitoring method and system for containerized battery systems based on the Internet of Things, which has the following beneficial effects: (1) By collecting data through the Internet of Things, aligning data across stages, analyzing damage indexes, judging risk thresholds, and verifying the integrity of blockchain data, the critical states of containerized battery systems from factory transportation and temporary storage to power plant commissioning are incorporated into a unified monitoring process. This solution can convert abnormal shocks during transportation, gas escape and temperature and humidity exceeding limits during storage, and increased internal resistance and capacity decay during use into quantifiable and comparable stage risk results. Based on these results, collaborative control measures such as transportation route suggestions, storage isolation and re-inspection, and charge / discharge rate constraints can be implemented. This not only enables the early detection of hidden risks that are difficult to identify in a timely manner by traditional operation monitoring, reducing the possibility of problems such as commissioning with damage, abnormal heating, abnormal capacity decay, and thermal runaway warnings, but also enables the determination of the authenticity and completeness of data at each stage through hash verification after an accident, assisting in the formation of stage risk contribution ranking results, thereby improving the reliability of the full-cycle safety management, operation control, and evidence retention of containerized battery systems.

[0016] (2) By collecting vibration, environmental, gas concentration, temperature and humidity and first charge-discharge cycle data during the transportation, storage and use stages respectively, and aligning them across stages with container identification and timestamps, the data that were originally scattered in the logistics, storage and power plant commissioning stages can be unified into a full-cycle state parameter set, thereby accurately identifying the correlation between transportation shock, storage chemical escape and electrochemical aging in the early stage of use, and improving the ability to discover hidden risks of containerized battery systems before and after commissioning.

[0017] (3) By constructing damage indices for each stage through mechanical damage component values, chemical damage component values ​​and electrochemical aging component values, and combining them with risk thresholds to implement transportation route suggestions, warehouse isolation re-inspection and charge / discharge rate constraint control, risk handling can be transformed from a single alarm to cross-stage collaborative scheduling; at the same time, the status data of each stage is recorded to the consortium blockchain in a hash anchoring manner to obtain data integrity verification logs, and the stage risk contribution ranking results and auxiliary responsibility analysis results are determined based on the data integrity verification logs to improve the reliability of safety management and traceability analysis. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the intelligent monitoring method for a containerized battery system based on the Internet of Things according to the present invention. Figure 2 This is a block diagram of the intelligent monitoring system for the containerized battery system based on the Internet of Things according to the present invention; Figure 3 This is a schematic diagram of the abnormal impact event identification process during the transportation phase of the present invention; Figure 4 This is a schematic diagram of the chemical escaping characteristic analysis process during the storage stage of this invention; Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1, please refer to Figures 1 to 4This embodiment uses a containerized battery system for energy storage and peak shaving at a new energy power station as an application example. This containerized battery system adopts a 40-foot standard container structure. Inside the container are several battery clusters, a battery management unit, a fire suppression unit, a temperature control unit, a gas detection unit, and a communication gateway. The battery type is lithium iron phosphate, and the system's rated capacity is approximately 2MWh. After leaving the factory, the containerized battery system needs to be transported by logistics vehicles from the battery manufacturer to a regional warehousing center. After temporary storage and outbound scheduling at the warehousing center, it is then transported to the new energy power station site and connected to the power station's energy storage system for operation.

[0021] In the aforementioned application scenarios, containerized battery systems may be affected by mechanical shocks such as road bumps, sudden braking, turning, and loading / unloading collisions during transportation; during storage, they may be affected by environmental factors such as high temperature, high humidity, and long-term sealed storage, and there may be a tendency for abnormal gases such as hydrogen, carbon monoxide, or volatile organic compounds to escape; during the usage phase, after the battery system is connected to the power station, the initial DC internal resistance and initial capacity of each cell need to be obtained through the first charge-discharge cycle in order to determine its initial electrochemical state before it is put into operation.

[0022] Therefore, this embodiment collects vibration and environmental data during transportation, gas concentration and temperature / humidity data during storage, and initial charge / discharge cycle data during use. Using container identification and timestamps as the basis for association, it performs cross-stage alignment, normalization, and damage analysis on the data from the three stages of transportation, storage, and use. This yields mechanical damage component values, chemical damage component values, and electrochemical aging component values, which are further used for risk level classification, transportation scheduling recommendations, storage isolation and re-inspection, charge / discharge rate constraints during use, and data integrity verification and auxiliary liability analysis after an accident.

[0023] Specifically, the container identification is generated at the time of system delivery and is linked to the container number, battery cluster number, batch number, and battery management system number. The transportation phase begins when the logistics vehicle leaves the manufacturing plant's electronic fence, and ends when the vehicle enters the regional warehousing center's electronic fence and completes its registration. The warehousing phase begins when the warehousing system completes its confirmation of inbound delivery, and ends when the warehousing system generates its confirmation of outbound delivery. The usage phase begins when the containerized battery system completes its pre-grid connection testing at the new energy power station and initiates its first charge-discharge cycle. Through these timeline settings, implementers can clearly distinguish the data sources and responsibility boundaries for different stages.

[0024] In this embodiment, during the transportation phase, vibration data can be collected using a triaxial accelerometer installed at the bottom crossbeam of the container, and transportation environment data can be collected using temperature and humidity sensors installed near the ventilation points on the outside of the container. During the storage phase, hydrogen concentration, carbon monoxide concentration, and volatile organic compound concentration can be collected using a gas sensor array installed near the return air duct at the top inside the container, and storage environment temperature and humidity can be collected using storage temperature and humidity sensors. During the usage phase, the initial DC internal resistance, initial capacity, current, voltage, and temperature data during the first charge-discharge cycle can be obtained through the battery management system and the power station energy management system. The above data is aggregated by the edge gateway and uploaded to the monitoring platform, where the monitoring platform completes cross-stage data association, damage index calculation, and collaborative scheduling decisions.

[0025] Through the above application scenario settings, this embodiment can cover the entire process of containerized battery systems from factory transportation and temporary storage to power plant commissioning, so that vibration data processing, gas concentration trend analysis, electrochemical baseline calculation, risk threshold comparison, cross-stage scheduling, data integrity verification and auxiliary responsibility analysis in subsequent embodiments all have a unified data source, a unified object identifier and a unified implementation environment.

[0026] This invention provides an intelligent monitoring method for containerized battery systems based on the Internet of Things, comprising the following steps: S1. Obtain characteristic parameters of the transportation stage, chemical escaping characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage, respectively. Specifically, during the transportation phase, vibration data is continuously collected at a preset transportation vibration sampling frequency using a triaxial accelerometer installed at the bottom of the container; and ambient temperature and humidity data during the transportation phase are continuously collected at preset transportation environment sampling intervals using a cargo hold temperature and humidity sensor installed on the outside of the container.

[0027] In the scenario described in this embodiment, a triaxial accelerometer is installed in the middle of the bottom crossbeam of the container and fixed with bolts to a mounting plate rigidly connected to the battery cluster base, enabling it to simultaneously sense vehicle bumps, sudden braking, road impacts, and loading / unloading impacts. The triaxial accelerometer's range is set to ±16g, with a resolution of no less than 0.01g, and the transportation vibration sampling frequency is set to 200Hz. The reason for choosing 200Hz is that the vibration frequencies that significantly affect the container's battery system during road transportation are typically concentrated in the range of 1Hz to 80Hz. The 200Hz sampling frequency meets the Nyquist sampling requirements and provides sufficient frequency margin.

[0028] Specifically, the cargo hold temperature and humidity sensors are installed on the outer wall of the container near the ventilation openings, avoiding direct contact with the battery cluster's heat dissipation area. The data collection interval for transport environment temperature and humidity is set to 60 seconds. This collection interval can reflect environmental changes during transportation, such as day-night temperature differences, tunnel entry and exit, rainfall, and high-temperature exposure, without generating excessive communication data.

[0029] The vibration data during the transportation phase is processed by Fast Fourier Transform, and the impact energy value of each time period is calculated. Continuous time periods with impact energy values ​​exceeding a preset single-frame impact threshold are marked as abnormal candidate time periods. The impact energy values ​​of the corresponding time periods of consecutive abnormal candidate time periods are accumulated to obtain the cumulative impact energy value of the abnormal impact event. When the cumulative impact energy value of the abnormal impact event exceeds the preset event impact threshold, the consecutive abnormal candidate time periods are confirmed as abnormal impact events. At the same time, the duration of the abnormal impact event, the peak acceleration, and the cumulative impact energy value of the abnormal impact event are recorded as characteristic parameters of the transportation phase.

[0030] This includes performing a Fast Fourier Transform on the vibration data and calculating the cumulative impact energy over each time period, specifically: Based on the vibration data continuously collected by the triaxial accelerometer during the transportation phase, and combined with the vector magnitude synthesis method, a vibration acceleration time-domain sequence is obtained. The vibration acceleration time-domain sequence is then processed into frames, and each frame is windowed to obtain windowed vibration frame data. The windowed vibration frame data is then processed by Fast Fourier Transform to obtain the spectrum of each frame. The spectral amplitude within the preset vibration analysis frequency range is extracted, and the total energy of a single frame is calculated. This total energy is then used as the single-frame impact energy of that frame. The single-frame impact energies of all frames in the current transportation phase are accumulated to obtain the total cumulative impact energy of the transportation phase.

[0031] The three-axis accelerations are denoted as x-axis, y-axis, and z-axis accelerations, respectively. During vector magnitude synthesis, the square roots of the summations of the squares of the three directional accelerations at the same sampling moment are taken to obtain the composite vibration acceleration at that moment. This process avoids omissions in single-axis judgments due to differences in container installation direction, vehicle turning direction, or impact direction.

[0032] In practice, the frame length is preferably set to 1 second, and the frame shift is also set to 1 second, meaning adjacent frames do not overlap. A Hanning window can be used for windowing to reduce spectral leakage and make the frequency domain energy after the Fast Fourier Transform more stable. The preset vibration analysis frequency range is set to 1Hz to 80Hz. Changes below 1Hz are mostly related to slow changes in vehicle attitude, while signals above 80Hz are mostly local high-frequency noise or minor structural vibrations, contributing little to the overall mechanical impact assessment of the container battery system.

[0033] When the single-frame impact energy of any frame exceeds the preset single-frame impact threshold, the frame is marked as an abnormal frame, and the single-frame impact energy of the abnormal frame is accumulated with the single-frame impact energy of subsequent consecutive frames marked as abnormal, until a frame that does not exceed the preset single-frame impact threshold appears.

[0034] Specifically, the identification of consecutive abnormal frames is used to merge multiple consecutive high-energy frames generated by the same turbulence or loading / unloading collision into a single event, preventing the same impact from being counted repeatedly. If there is only one non-abnormal frame between two abnormal frames, a fault-tolerant merging rule can also be set in the system to merge the abnormal frames before and after the non-abnormal frame into the same candidate impact event. This fault-tolerant rule is applicable to situations where the impact energy briefly decreases but still actually belongs to the same impact process.

[0035] The combined impact energy of consecutive abnormal frames is used as the cumulative energy value of a candidate impact event. When the cumulative energy value of the candidate impact event exceeds the preset event impact threshold, the consecutive abnormal frames are confirmed as an abnormal impact event. The cumulative energy value of the candidate impact event is used as the cumulative impact energy value of the abnormal impact event. The duration of the consecutive abnormal frames is used as the duration of the abnormal impact event. The maximum peak value of the vibration acceleration time-domain sequence within the consecutive abnormal frames is used as the peak acceleration of the abnormal impact event.

[0036] Specifically, a preset event impact threshold is used to filter out transient noise and short-term minor impacts in a single frame. In Example 1, the event impact threshold is preferably set to four times the single-frame impact threshold. When the combined impact energy of consecutive abnormal frames exceeds this value, it indicates that the impact not only reaches an abnormality in a single frame but also forms an identifiable event in terms of duration and energy accumulation. Peak acceleration is used to record the maximum impact amplitude, duration is used to record the impact duration, and impact energy accumulation is used to record the overall impact intensity; all three are used together in subsequent mechanical damage calculations.

[0037] During data processing, the continuously acquired triaxial acceleration data is divided into frames, each lasting one second, with 200 sampling points per frame. After windowing, a Fast Fourier Transform (FFT) is performed on each frame to extract the spectral amplitude within the 1Hz to 80Hz frequency band. The sum of the squares of the amplitudes at each frequency point within this band is taken as the single-frame impact energy. The single-frame impact threshold can be determined based on historical samples under non-destructive transportation conditions. For example, at least 30 normal transportation samples are selected, and the average value μ and standard deviation σ of the impact energy for each frame are calculated. μ+3σ is then set as the preset single-frame impact threshold. The reason for choosing 3σ is that most disturbances in normal transportation are distributed around the mean, and frames exceeding μ+3σ typically indicate abnormal impacts such as emergency braking, pothole impacts, and loading / unloading collisions.

[0038] The preset event impact threshold is set to 3 to 5 times the single-frame impact threshold. In this embodiment, 4 times is preferred. The reason is that a single instantaneous high-energy frame may be generated by sensor spike noise or transient interference. Only when the energy accumulation of multiple consecutive frames exceeds the event threshold can the impact event that actually affects the battery cluster bracket, connecting bar, and module fixing structure be more stably characterized. The duration of the abnormal impact event is obtained by multiplying the number of consecutive abnormal frames by the duration of a single frame. The peak acceleration is taken as the maximum value of the triaxial composite acceleration during the abnormal impact event. The cumulative impact energy of the abnormal impact event is taken as the sum of the impact energy of a single frame of consecutive abnormal frames.

[0039] During the storage phase, a gas sensor array installed inside the container continuously collects hydrogen concentration data, carbon monoxide concentration data, and volatile organic compound concentration data at preset storage gas collection intervals; simultaneously, storage temperature and humidity sensors continuously collect storage temperature and humidity data at preset storage temperature and humidity collection intervals.

[0040] Specifically, the gas sensor array is installed near the return air duct at the top inside the container and includes at least a hydrogen sensor, a carbon monoxide sensor, and a volatile organic compound (VOC) sensor. The hydrogen sensor detects light gases released during abnormal battery side reactions or early pressure leaks; the carbon monoxide sensor detects byproducts in the early stages of thermal runaway; and the VOC sensor detects organic gases released from electrolyte evaporation or seal failure. The gas sampling interval is set to 300 seconds, as are the temperature and humidity sampling intervals. The reason for choosing 300 seconds is that gas concentration changes during storage are typically not abrupt changes on a millisecond scale, but rather accumulate over a minute-level trend. Sampling every 5 minutes meets the trend recognition requirements and reduces sensor power consumption and communication load.

[0041] Baseline drift correction was performed on hydrogen concentration data, carbon monoxide concentration data, and volatile organic compound concentration data. The linear fitting slope of each gas concentration during the storage stage was calculated. The linear fitting slope of each gas concentration, the target escaping gas type, and the linear fitting slope of the target gas concentration were used as chemical escaping characteristic parameters during the storage stage.

[0042] The calculation of the linear fitting slope for the concentration of each gas during the storage stage specifically includes: The first gas concentration sample value after the start of the storage phase is obtained as the initial concentration baseline value of the corresponding gas.

[0043] Specifically, after the storage phase begins, the system collects the first gas concentration sample value 30 minutes after the container has been placed in the warehouse, and uses this value as the initial concentration baseline. The reason for setting a 30-minute settling time is that opening doors, handling, and ventilation during the warehouse entry process can cause short-term gas disturbances, and sampling after settling can better reflect the stable background concentration inside the container.

[0044] The drift-corrected gas concentration value of the corresponding gas is obtained by subtracting the initial concentration baseline value of the gas from the concentration sampling value of each gas in each gas sampling interval.

[0045] Specifically, drift-corrected gas concentration values ​​can eliminate differences in background air quality, sensor zero-point variations, and initial storage conditions between different warehouses. For example, if the initial stable hydrogen sampling value is 2 ppm, and a subsequent sampling value is 5 ppm, then the drift-corrected hydrogen concentration value is 3 ppm. Subsequent trend analysis is based on 3 ppm, rather than directly using 5 ppm, thus highlighting the new changes that occur during storage.

[0046] A sliding observation window containing a preset number of gas collection intervals is set. The drift correction gas concentration values ​​of each gas within the sliding observation window are subjected to least squares linear fitting to obtain the unit concentration linear fitting slope of each gas within the sliding observation window.

[0047] Specifically, the preset quantity is preferably 12 gas sampling intervals, corresponding to 60 minutes. Least squares linear fitting is used with sampling time as the independent variable and drift-corrected gas concentration as the dependent variable to obtain the rate of change of gas concentration per unit time. If the slope is greater than 0, it indicates that the gas concentration is rising; if the slope is close to 0, it indicates that the concentration is basically stable; if the slope is less than 0, it indicates that the concentration is falling or being diluted by ventilation.

[0048] After obtaining the linear fitting slope of the unit concentration of each gas within the sliding observation window, the sliding observation window is moved forward by one storage gas collection interval, and the least squares linear fitting process is performed again to obtain the linear fitting slope of the unit concentration of each gas within the next sliding observation window.

[0049] In this embodiment, the sliding step size is set to a sampling interval of 5 minutes. This updates the gas trend assessment every 5 minutes, which not only allows for continuous tracking of storage risks but also avoids frequent false alarms caused by concentration fluctuations at individual points.

[0050] The maximum value of the linear fitting slope for unit concentration of the same gas within each sliding observation window is taken as the linear fitting slope for the concentration of the corresponding gas.

[0051] Specifically, the purpose of taking the maximum slope is to capture the most obvious primary gas escape trend during the storage stage. For battery systems, even if the average concentration does not change much throughout the storage stage, as long as there is a continuous increase in a certain period of time, it may correspond to local cell abnormalities, sealing failures, or side reactions induced by temperature and humidity. Therefore, using the maximum slope is more conducive to early risk identification.

[0052] The slopes of the linear fitting of the concentrations of each gas are compared, and the gas with the largest linear fitting slope is identified as the target escaping gas. The slope of the linear fitting of the concentration of the target escaping gas is then identified as the slope of the linear fitting of the concentration of the target gas.

[0053] If hydrogen, carbon monoxide, and volatile organic compounds all exhibit positive slopes, the gas with the largest slope is selected as the target escaping gas. This gas represents the most significant abnormal escaping type during the current storage phase. If the slopes of all gases are not greater than 0, the slope of the linear fit of the target gas concentration is recorded as 0, and it is determined that no continuous escaping trend has occurred during the storage phase.

[0054] During the first charge-discharge cycle of the containerized battery system in the usage phase, the initial DC internal resistance and initial capacity of each cell are obtained as baseline characteristic parameters for the usage phase.

[0055] In this embodiment, after the containerized battery system is connected to the power station, a controlled first charge-discharge cycle is performed. This cycle can be conducted at a 0.2C rate, charging to a set upper limit SOC, resting for 30 minutes, discharging to a set lower limit SOC, and then resting for another 30 minutes. The 0.2C rate is chosen because it has a relatively small impact on battery thermal performance and can more stably reflect the initial capacity and DC internal resistance of the cell. The DC internal resistance can be obtained using the current step method, that is, applying a short-time current step within the stable SOC range and collecting the ratio of the voltage change to the current change as the initial DC internal resistance. The initial capacity is obtained by integrating the current over time during the first discharge process.

[0056] S2. Align the characteristic parameters of the transportation stage, the chemical emission characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage across stages using container identification and timestamp as indexes, and perform normalization processing to construct a full-cycle state parameter set. Based on the container identification, the characteristic parameters of the transportation stage, the chemical escaping characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage are respectively affixed with the corresponding stage start timestamp and stage end timestamp.

[0057] Specifically, each containerized battery system is assigned a unique container identification identifier upon leaving the factory. This identifier can be formed by combining the manufacturing batch number, container number, and battery cluster number. For example, the identifier can adopt a structure of "manufacturing batch - container number - battery cluster batch". The transportation phase start timestamp is the time when the vehicle leaves the factory's electronic fence, and the transportation phase end timestamp is the time when the vehicle enters the warehouse park's electronic fence and completes the warehousing registration. The warehousing phase start timestamp is the time when the warehousing system completes the warehousing confirmation, and the warehousing phase end timestamp is the time when the warehousing system generates the outbound confirmation. The usage phase start timestamp is the time when the containerized battery system completes pre-grid connection testing and begins its first charge-discharge cycle.

[0058] Based on the cumulative impact energy, peak acceleration, and duration of each abnormal impact event during the transportation phase, raw mechanical damage data is constructed. This raw mechanical damage data also includes the total cumulative impact energy during the transportation phase.

[0059] Raw data on mechanical damage is stored in an event table format. Each record includes at least the container identification, event number, event start time, event end time, duration of the abnormal impact event, peak acceleration, cumulative impact energy of the abnormal impact event, vehicle location information at the time of the event, and the total cumulative impact energy during the transportation phase. The event number automatically increments in chronological order. By incorporating vehicle location information, abnormal impact events can be subsequently associated with specific road sections or loading / unloading nodes, facilitating transportation scheduling optimization and traceability analysis.

[0060] Based on the chemical escape characteristic parameters and cumulative temperature and humidity values ​​during the storage stage, raw chemical state data are constructed. The cumulative temperature and humidity values ​​are obtained by normalizing and summing the cumulative temperature exceedance of the storage stage ambient temperature data relative to the preset standard temperature range and the cumulative humidity exceedance of the storage stage ambient humidity data relative to the preset standard humidity range.

[0061] In this embodiment, the preset standard temperature range is set to 15°C to 30°C, and the preset standard humidity range is set to 30%RH to 70%RH. This range is used for routine safe storage judgment during the warehousing stage. The cumulative amount of temperature exceeding the limit is calculated as follows: when the sampling temperature is below 15°C, the difference between 15°C and the sampling temperature is taken; when the sampling temperature is above 30°C, the difference between the sampling temperature and 30°C is taken; when the sampling temperature is between 15°C and 30°C, the excess amount is 0. The cumulative amount of humidity exceeding the limit is calculated similarly: humidity exceeds the limit when it is below 30%RH or above 70%RH, and is 0 when it is within the range.

[0062] Specifically, the cumulative temperature exceedance is calculated as "exceeding temperature difference × duration", and the cumulative humidity exceedance is calculated as "exceeding humidity difference × duration". For normalization, the cumulative temperature exceedance can be divided by a preset temperature exceedance baseline value, and the cumulative humidity exceedance can be divided by a preset humidity exceedance baseline value. The two values ​​are then added together to obtain the cumulative temperature and humidity values. The preset temperature exceedance baseline value can be set to "5℃ × 24 hours", and the preset humidity exceedance baseline value can be set to "10%RH × 24 hours". This setting is because being in a significantly deviated storage environment for a continuous day will have a identifiable impact on battery seals, electrolyte stability, and sensor drift.

[0063] Based on the baseline characteristic parameters of the usage phase, the raw data of the electrochemical baseline are constructed.

[0064] In this embodiment, the electrochemical baseline raw data is recorded at the cell level as the smallest unit. Each record includes container identification, battery cluster number, module number, cell number, initial DC internal resistance, initial capacity, test SOC range, test temperature, test rate, rated internal resistance, and rated capacity. The test temperature is preferably controlled at 25℃±2℃ because both DC internal resistance and capacity are significantly affected by temperature, and controlling the test temperature can reduce the interference of environmental factors on the initial state judgment.

[0065] The raw mechanical damage data, raw chemical state data, and raw electrochemical baseline data belonging to the same container identity are combined in chronological order to generate a full-cycle raw state table for the corresponding container identity.

[0066] Specifically, the full-cycle raw status table uses the container's identity identifier as the primary key and the stage type and timestamp as a composite index. Transportation stage data, warehousing stage data, and usage stage data are written to different stage fields, and are linked sequentially by the stage start and end times. If the same container is stored or transported multiple times, multiple stage segments are generated in chronological order, with the stage order indicated in the segment number, such as T1, S1, T2, S2, U1. This avoids data confusion in scenarios involving multiple transfers.

[0067] The parameters in the original state table for the entire cycle are normalized according to their corresponding parameter types to generate a set of state parameters for damage analysis. Among them, mechanical damage parameters are normalized based on the impact energy baseline value of events under non-destructive transportation conditions, which is the statistical value of the impact energy of events under non-destructive transportation conditions. Chemical emission parameters are normalized based on the baseline of the ambient gas concentration at the factory standard. Electrochemical baseline parameters are normalized based on the rated internal resistance and rated capacity in the factory test report.

[0068] It should be noted that when normalizing mechanical damage parameters, the cumulative impact energy of abnormal impact events is divided by the baseline impact energy of events under non-damaging transportation conditions; the peak acceleration is divided by the baseline peak acceleration under non-damaging conditions; and the duration of abnormal impact events is divided by the baseline duration under non-damaging conditions. The baseline impact energy of events under non-damaging transportation conditions can be obtained by statistically analyzing at least 30 transportation records of the same type of container under normal loading and unloading conditions on smooth highways, with vibration-damped vehicles. The preferred baseline value is the statistical mean plus one standard deviation. The reason for choosing the mean plus one standard deviation is that this value can cover common fluctuations in normal transportation but will not include significant abnormal impacts within the normal range.

[0069] Specifically, when normalizing chemical efflux parameters, the drift correction concentration value or concentration linear fitting slope of each gas is divided by the corresponding gas's factory standard ambient gas concentration baseline. This baseline can be collected before the battery system is sealed at the factory, at 25℃±2℃, relative humidity 50%RH±10%RH, and after standing for 2 hours. When normalizing electrochemical baseline parameters, the initial DC internal resistance is compared with the factory rated internal resistance, and the initial capacity is compared with the factory rated capacity, thereby eliminating differences in specifications between different cells.

[0070] S3. Based on the full-cycle state parameter set, determine the mechanical damage component value in the transportation stage, the chemical damage component value in the storage stage, and the electrochemical aging component value in the use stage, and use them as the damage index for the corresponding stage. Specifically, feature identification is performed on the full-cycle state parameter set, the cumulative impact energy value of each abnormal impact event is extracted and statistically analyzed, the cumulative impact energy value, peak acceleration and duration of the abnormal impact event corresponding to the i-th abnormal impact event are multiplied to obtain the single mechanical damage component value of the i-th abnormal impact event, and the single mechanical damage component values ​​of all abnormal impact events are accumulated to determine the mechanical damage component value, which is used as the mechanical damage index for the transportation stage.

[0071] During data processing, records of the transportation stage type are first filtered from the full-cycle state parameter set, and then the list of abnormal impact events is read. For the i-th abnormal impact event, the normalized cumulative impact energy, normalized peak acceleration, and normalized duration of the abnormal impact event are multiplied together. The purpose of this multiplication is to simultaneously reflect the impact intensity, impact amplitude, and impact duration: only when all three are simultaneously high will the single mechanical damage component value increase significantly; if only a brief, slight jolt exists, the calculated result will be small.

[0072] The mechanical damage component value can be obtained as follows: Single mechanical damage component value = Normalized impact energy × Normalized peak acceleration × Normalized duration; Mechanical damage component value = sum of individual mechanical damage component values.

[0073] When no abnormal impact event is identified during the transportation phase, the mechanical damage component value is recorded as 0. This result is directly used as the mechanical damage index during the transportation phase for subsequent transportation risk assessment.

[0074] Furthermore, to demonstrate an acceptable engineering correlation between the cumulative impact energy, peak acceleration, and duration of abnormal impact events and the damage to the battery structure, this embodiment establishes an impact event-structure verification sample library during the transportation calibration phase of the same model of containerized battery system. The structural verification items in the sample library include the loosening of battery cluster fixing bolts, the displacement of module limiting components, the change in contact resistance at busbar connection points, the change in insulation resistance, and the peak strain of the container base. After each transportation, the identified abnormal impact events are recorded in relation to the above structural verification items, and samples with sensor disconnections, missing loading / unloading records, or incomplete verification data are removed.

[0075] During calibration, the normalized impact energy, normalized peak acceleration, and normalized duration are first obtained to determine the single-instance mechanical damage component value, specifically: Dmc = Ec × Ac × Tc; where Dmc is the single-instance mechanical damage component value, Ec is the normalized impact energy, Ac is the normalized peak acceleration, and Tc is the normalized duration. Multiplying these three values ​​does not equate mechanical damage to a single physical destruction quantity, but rather forms an engineering equivalent impact exposure quantity, making "high-energy, high-peak-value, long-duration" impact events more distinguishable from short-term noise and minor bumps.

[0076] The correlation between the single mechanical damage component value and the bolt loosening amount, module displacement amount, contact resistance change amount, and insulation resistance change amount in the post-transport structural review items is tested. When the single mechanical damage component value increases, the proportion of samples that simultaneously increase in at least one of the above structural review items reaches the preset consistency proportion, and the single mechanical damage component value is below the 95th percentile of the normal transportation sample without triggering structural review anomalies, then it is confirmed that the mechanical damage component value and the battery structural damage risk meet the engineering acceptable correspondence.

[0077] The preset consistency ratio is denoted as Pc, and the value of Pc ranges from 70% to 90%, with 80% being preferred in this embodiment. The specific method for determining Pc is as follows: During the transportation calibration stage, at least 30 valid transportation samples of the same model of containerized battery system are selected. A correspondence is established between the single mechanical damage component value Dmc in each valid transportation sample and the structural verification item after transportation. The samples are sorted from low to high according to Dmc and divided into at least three damage grades according to the quartile interval or equal-width interval of Dmc. For any higher damage grade, if the mean value of at least one type of structural verification item in the grade is higher than the mean value of the corresponding structural verification item in the adjacent lower damage grade, and the increase exceeds the upper limit of the measurement repeatability error of the structural verification item, then the samples in the grade are included in the synchronous increase samples. The ratio of the number of synchronous increase samples Nsy to the number of valid samples Nva is taken as the measured consistency ratio Pm. When Pm ≥ Pc, it is determined that the single mechanical damage component value and the structural verification item meet the consistency requirements.

[0078] Furthermore, the 95th percentile of the normal transport sample is determined by the Dmc distribution of no less than 30 normal transport samples. If Dmc is less than or equal to the 95th percentile and no structural verification anomaly occurs after transport, the sample is used as a non-triggered structural anomaly sample to verify that the first-level warning threshold does not cause excessive alarm for normal transport. If Pm is lower than Pc, the benchmarks for normalized impact energy, normalized peak acceleration, or normalized duration are readjusted, or the first-level warning threshold is recalibrated and the consistency test is performed again.

[0079] The slope of the linear fitting of the target gas concentration during the storage stage is extracted from the set of state parameters throughout the entire cycle. When the slope of the linear fitting of the target gas concentration is positive and exceeds the preset escape slope threshold, the slope of the linear fitting of the target gas concentration is multiplied by the cumulative value of temperature and humidity to obtain the chemical damage component value, which is used as the chemical damage index during the storage stage. When the slope of the linear fitting of the target gas concentration is not positive, or the slope of the linear fitting of the target gas concentration does not exceed the preset escape slope threshold, the chemical damage component value is set to zero.

[0080] Specifically, a positive slope in the linear fitting of the target gas concentration indicates that the gas concentration is trending upward within the sliding window; exceeding the preset escape slope threshold indicates that the upward trend has exceeded the normal drift range of the sensor and the range of air disturbance in the storage area.

[0081] For any target gas g, n gas concentration data points are collected within a sliding window, and the i-th sampling time is denoted as . The corresponding gas concentration is The slope of the linear fitting of the target gas concentration Determined using the least squares method, specifically: ,in, This represents the average value of the sampling times within the sliding window. This represents the average gas concentration within the sliding window; The preset escape slope threshold can be determined jointly through empty container storage testing and healthy battery system storage testing; specifically, storage testing for no less than 7 days is conducted in an empty container of the same model to obtain the average slope value of the target gas g. and standard deviation ; Conduct a storage test for no less than 7 days with a healthy battery system installed, and obtain the average slope value of the target gas g. and standard deviation Then the preset escape slope threshold of the target gas g. Determined according to this formula: .

[0082] Among them, the empty container storage test is used to characterize the background slope caused by sensor zero-point drift, container material release, and storage air disturbance; the healthy battery system storage test is used to characterize the background slope of a normal battery system under conditions without abnormal emissions. The larger of the two is then selected. The upper limit, as a preset escape slope threshold, can reduce false alarms caused by sensor noise, short-term ventilation, personnel opening doors, or fluctuations in the storage environment.

[0083] Once the slope of the linear fit of the target gas concentration exceeds a threshold, this slope is multiplied by the cumulative temperature and humidity value to obtain the chemical damage component value. This product is used to express the combined effect of "continuous gas escape" and "adverse storage environment". If the gas concentration increases but the temperature and humidity environment remains stable, the chemical damage component value will not be too large; if the gas concentration increases while the storage environment is hot, humid, or excessively dry, the chemical damage component value will increase significantly, indicating a higher risk of battery sealing, electrolyte stability, or side reactions.

[0084] The internal resistance increment and capacity decay of each cell during the service stage are extracted from the full-cycle state parameter set. The internal resistance increment and capacity decay of each cell are weighted and summed to obtain the electrochemical aging component value, which is used as the aging damage index during the service stage.

[0085] The process involves a weighted summation of the internal resistance increment and capacity decay of each cell, specifically including: Feature identification is performed on the electrochemical baseline parameters during the usage phase to obtain the initial DC internal resistance and initial capacity of each cell. Based on the rated internal resistance and rated capacity of each cell extracted from the factory test report, the initial DC internal resistance is subtracted from the rated internal resistance in the factory test report to obtain the internal resistance difference. The ratio of the internal resistance difference to the rated internal resistance in the factory test report is used as the internal resistance increment of the corresponding cell. When the internal resistance increment is less than zero, the internal resistance increment of the corresponding cell is marked as zero. The rated capacity in the factory test report is subtracted from the initial capacity to obtain the capacity difference. The ratio of the capacity difference to the rated capacity in the factory test report is used as the capacity decay of the corresponding cell. When the capacity decay is less than zero, the capacity decay of the corresponding cell is marked as zero.

[0086] Specifically, the internal resistance increment represents the percentage increase in the cell's initial DC internal resistance relative to its factory-rated internal resistance. If the initial DC internal resistance is lower than the factory-rated internal resistance, it is usually caused by differences in test temperature, measurement errors, or differences in the cell's activation state, and does not indicate a reduction in damage; therefore, this internal resistance increment is set to 0. The capacity decay represents the percentage decrease in initial capacity relative to the factory-rated capacity. If the initial capacity is higher than the factory-rated capacity, it is not treated as negative damage; instead, the capacity decay is set to 0. This truncation process prevents abnormally low internal resistance or high capacity data from offsetting actual damage.

[0087] Based on the internal resistance increment and capacity decay of each cell, the average value of the internal resistance increment of each cell is calculated to obtain the average value of the internal resistance increment; the average value of the capacity decay of each cell is calculated to obtain the average value of the capacity decay, and these are respectively used as the normalized internal resistance increment and the normalized capacity decay.

[0088] It should be noted that the average internal resistance increment reflects the overall impedance deviation of the containerized battery system during its initial commissioning, while the average capacity decay reflects the overall decrease in usable capacity. If it is necessary to enhance the identification of individual cell anomalies, invalid sampling points can be removed before the average calculation, and the maximum single-cell internal resistance increment and the maximum single-cell capacity decay can be retained as auxiliary alarm parameters. However, the electrochemical aging component value is still based on the average value as the main calculation basis to ensure the stability of the results.

[0089] The normalized internal resistance increment is multiplied by a preset first aging coefficient to obtain the normalized internal resistance aging term. The normalized capacity decay is multiplied by a preset second aging coefficient to obtain the normalized capacity aging term. The normalized internal resistance aging term and the normalized capacity aging term are added together to obtain the electrochemical aging component value. The sum of the preset first aging coefficient and the second aging coefficient is 1.

[0090] In this embodiment, the first aging factor is set to 0.6, and the second aging factor is set to 0.4. This setting is suitable for scenarios where energy storage power stations are sensitive to power response and safe operation, because increased internal resistance directly affects heat generation, voltage drop, and rate capability; while capacity degradation mainly affects available power and revenue. If subsequent application scenarios focus more on capacity revenue, the first aging factor can also be set to 0.4, and the second aging factor to 0.6. Regardless of the setting, the sum of the two remains 1, ensuring that the electrochemical aging component value is maintained within a dimension range that facilitates threshold comparison.

[0091] The setting process for each parameter is as follows: The first-level warning threshold is determined by the mechanical damage index distribution of normal transportation samples, preferably taking the 95th percentile value and calibrated by the post-transport structural verification results; the second-level warning threshold is determined by the chemical damage index distribution of healthy storage samples, preferably taking the 95th percentile value and calibrated in conjunction with the upper limit of zero-point drift of the gas sensor; the road smoothness threshold is determined by the correspondence between the vehicle's historical vibration data and the road smoothness score, so that the single-frame impact energy generated by the candidate route that meets the threshold at the same speed is lower than the safety margin corresponding to the first-level warning threshold; the escape slope threshold is determined by the mean of the target gas slope of the empty container storage sample and the healthy battery storage sample plus 3 times the standard deviation; the first aging coefficient and the second aging coefficient are determined by the sensitivity of the operation limit to the increase in heat generation, abnormal pressure drop and capacity deviation in the commissioning verification sample.

[0092] The above parameters are used to control the intensity of false alarms, missed alarms, and response measures, respectively: the mechanical damage component value is used to identify whether a change in route or speed is needed due to transportation shocks; the chemical damage component value is used to identify whether isolation, ventilation, and fire monitoring are needed during the storage phase; the electrochemical aging component value is used to identify whether charging and discharging rates need to be limited during the initial operation phase; the risk threshold set is used to convert continuous values ​​into actionable actions such as transportation recommendations, storage re-inspections, and rate limits; the road smoothness threshold is used to reduce subsequent shock inputs during transportation; the escape slope threshold is used to filter sensor drift and short-term ventilation disturbances; and the aging coefficient is used to reflect the differences in the operational impact of internal resistance deviation and capacity deviation in the current application scenario.

[0093] S4. Compare the damage index of each stage with the preset risk threshold set to determine the risk level classification result, and perform cross-stage collaborative scheduling operation based on the risk level classification result. The damage index at different stages is compared with the risk threshold set to obtain the risk level classification result; the risk threshold set includes a first-level warning threshold for the transportation stage, a second-level warning threshold for the storage stage, and multiple charge and discharge constraint thresholds for the usage stage.

[0094] Specifically, the risk threshold set can be determined using historical transportation, storage, and operation data of the same model of containerized battery system. The first-level warning threshold during transportation can be taken as the 95th percentile of the mechanical damage index of a normal transportation sample; the second-level warning threshold during storage can be taken as the 95th percentile of the chemical damage index of a healthy storage sample; multiple charge / discharge constraint thresholds during use can be divided into three ranges—low risk, medium risk, and high risk—according to the power plant's operation and maintenance rules. For example, when the aging damage index is less than 0.05, normal rate operation is allowed; between 0.05 and 0.10, it is limited to no more than 0.5C; between 0.10 and 0.20, it is limited to no more than 0.25C; and when it is greater than 0.20, it enters the manual review or shutdown inspection process. These values ​​can be calibrated before commissioning based on the cell model, operating rate, and power plant safety strategy.

[0095] During the transportation phase, when the mechanical damage index exceeds the preset first-level warning threshold, a transportation suggestion signal is sent to the logistics scheduling system through the vehicle communication unit. The transportation suggestion signal includes reselecting a transportation route with road smoothness parameters that meet the preset road smoothness threshold and driving within the preset vibration reduction transportation speed range. When the mechanical damage index does not exceed the first-level warning threshold, no additional transportation suggestion signal is issued.

[0096] Specifically, the vehicle-mounted communication unit can use 4G, 5G, or BeiDou short message communication to send container identification, current mechanical damage index, location of abnormal impact events, peak acceleration, and suggested handling measures to the logistics dispatch system. Road smoothness parameters can be obtained from the road maintenance system, historical vehicle vibration data, or road condition scores from the logistics platform. The preset road smoothness threshold can be set to prioritize roads with a smoothness level no lower than 80 points on the platform score, or roads with an international smoothness index lower than the preset limit. The preset vibration-damping transport speed range can be set from 40 km / h to 70 km / h. Speeds below 40 km / h may affect logistics efficiency, and frequent low-speed starts and stops will increase impact; speeds above 70 km / h tend to amplify vibrations on uneven road surfaces. Therefore, this speed range is suitable for vibration control in road transport.

[0097] During the warehousing phase, when the chemical damage index during warehousing is greater than or equal to the Level 2 warning threshold, the corresponding container identification is marked as a warehousing isolation re-inspection object, and the warehousing isolation re-inspection object is transferred from the regular outbound queue to the isolation re-inspection outbound queue, ventilation detection area, or fire monitoring area, triggering a manual review process. When there are multiple warehousing isolation re-inspection objects, an isolation re-inspection list is generated in descending order of the chemical damage index during warehousing. If the chemical damage index during warehousing is the same, they are arranged in ascending order of entry time. When the chemical damage index during warehousing is less than the Level 2 warning threshold, the corresponding container identification is marked as a regular warehousing object, and a regular outbound list is generated in ascending order of entry time. Based on the isolation re-inspection list and the regular outbound list, an updated warehousing re-inspection and outbound plan list is generated.

[0098] Specifically, the warehouse management system recalculates the chemical damage index during the storage phase daily or after each gas slope update. If the chemical damage index of a container reaches the second-level warning threshold, the container is removed from the regular first-in-first-out queue and transferred to the isolation re-inspection outbound queue, ventilation testing area, or fire monitoring area, and a task awaiting manual review is generated. This process does not mean prioritizing high-risk containers for transportation or operation, but rather prioritizing containers with a tendency for gas escape or a higher risk of environmental accumulation for ventilation, fire monitoring, gas retesting, insulation testing, and battery management system alarm checks, to prevent them from remaining in the warehouse for extended periods and accumulating further risks.

[0099] Specifically, when multiple containers are simultaneously marked as objects for warehouse isolation and re-inspection, they are first sorted from highest to lowest chemical damage index during the storage phase, as a higher chemical damage index indicates a more significant combined effect of gas escape and adverse environmental factors. When the chemical damage indices are the same, they are then sorted from earliest to latest entry time to ensure that the first-in, first-out principle is still followed under the same risk conditions. The updated warehouse re-inspection and outbound plan list should include at least the container identification, risk level, chemical damage index, entry time, recommended re-inspection sequence, isolation storage location, ventilation testing requirements, fire monitoring requirements, and handling recommendations.

[0100] During the usage phase, when the containerized battery system is connected to the power station and put into operation, the aging damage index of the usage phase is compared with multiple charge and discharge constraint thresholds. Among them, multiple charge and discharge constraint thresholds are respectively established with multiple rate upper limits. When the aging damage index of the usage phase exceeds the corresponding charge and discharge constraint threshold, the corresponding rate upper limit is determined in the charge and discharge constraint mapping relationship according to the risk level classification result, and the maximum allowable charge and discharge rate is limited to no more than the corresponding rate upper limit.

[0101] In this embodiment, the charge / discharge constraint mapping relationship can be set as follows: when the aging damage index is less than 0.05, the maximum charge / discharge rate does not exceed 1C; when the aging damage index is greater than or equal to 0.05 and less than 0.10, the maximum charge / discharge rate does not exceed 0.5C; when the aging damage index is greater than or equal to 0.10 and less than 0.20, the maximum charge / discharge rate does not exceed 0.25C; when the aging damage index is greater than or equal to 0.20, automatic high-power operation is prohibited, and manual review is triggered. 0.05 serves as the primary constraint threshold, which can identify cells that slightly deviate from their factory condition; 0.10 serves as the intermediate constraint threshold, used to limit systems that may have significant aging or transportation / warehousing impacts; 0.20 serves as the high-risk threshold, used to isolate obviously abnormal systems from regular scheduling.

[0102] S5. Record the status data of each stage to the consortium blockchain in a hash-anchored manner to obtain the data integrity verification log, and determine the stage risk contribution ranking results and auxiliary responsibility analysis results based on the data integrity verification log.

[0103] In the application scenario described in this embodiment, the containerized battery system, after being shipped from the manufacturer, sequentially undergoes the transportation, storage, and usage stages. To avoid data breakpoints, data tampering, or unclear evidence boundaries between stages, this embodiment establishes a corresponding stage integrity relationship chain for each container's identity in the consortium blockchain. The stage integrity relationship chain records the hash values ​​corresponding to key status data in the transportation, storage, and usage stages, as well as the storage index of this status data in the off-chain database. This stage integrity relationship chain is used for data integrity verification and tamper-proof recording, but does not independently prove the cause of an accident.

[0104] Specifically, the consortium blockchain consists of manufacturing nodes, logistics carrier nodes, warehouse management nodes, power plant operator nodes, and regulatory nodes. Manufacturing nodes have the authority to create a complete relationship chain and write factory binding information; logistics carrier nodes have the authority to write transportation status data; warehouse management nodes have the authority to write warehouse status data; power plant operator nodes have the authority to write usage status data; and regulatory nodes have the authority to read, verify, and trigger review. Each node confirms the written transaction through a Byzantine fault-tolerant or voting confirmation consensus mechanism, and the written transaction becomes effective after confirmation by at least a preset number of nodes.

[0105] It should be noted that the consortium blockchain adopts a permissioned consortium blockchain structure. Each participating node authenticates its identity through digital certificates. On-chain deployment of stage-anchored smart contracts and off-chain database storage of complete state data are used. The stage integrity relationship chain uses the container identifier containerId as the primary key. The data fields of each stage node on the chain include at least containerId, stageType, stageOrder, previousStageHash, dataHash, offChainIndex, dataSchemaVersion, parameterVersion, submitterNodeId, submitTimestamp, and nodeSignature. Among them, previousStageHash points to the previous stage node, dataHash is the hash value of the stage state data, and offChainIndex is the off-chain state data index.

[0106] In this embodiment, a write transaction becomes effective after being confirmed by at least three types of participating nodes and at least two-thirds of the total number of authorized nodes in the consortium blockchain; if there are fewer than five authorized nodes, it must be confirmed jointly by at least the manufacturing node, the current stage responsible party node, and the regulatory node. The current stage responsible party node is the logistics carrier node in the transportation stage, the warehouse management node in the warehousing stage, or the power plant operator node in the usage stage.

[0107] The status data update rules are as follows: Stage hash values ​​already written to the consortium blockchain cannot be overwritten or modified. If stage status data needs to be corrected due to sensor retransmission, manual verification, or updates to calculation parameters, a supplementary status data packet is generated and written to a supplementary node. The supplementary node records the original node hash value, the supplementary data hash value, the reason for supplementation, the operating node, and the operation time. The stage integrity relationship chain retains the reference relationship between the original node and the supplementary node. This allows for the differentiation between the original write facts and subsequent verification facts, preventing the traceability chain from being disrupted by overwriting.

[0108] When different nodes submit inconsistent hash values ​​to be written for the same stage, the system marks the write request as a hash collision event and suspends the automatic confirmation process for that stage. The supervisory node then reads the off-chain state data index, the original sensor records, and the stage handover records for comparison. When off-chain state data is lost or the state data index becomes invalid, that stage is marked as an evidence-deficient stage. The monitoring platform only retains the hash values ​​that have been uploaded to the chain and the existing sensor summary information, and does not output a separate judgment on the cause of the accident based on these.

[0109] It's important to note that the hash value is a fixed-length data digest calculated from the stage state data using a hash algorithm. The hash value is not the original state data itself, but rather a unique verification result of the original state data. If any field in the stage state data changes—for example, if the number of abnormal impact events changes, the mechanical damage component value is modified, the storage gas slope is replaced, or the initial capacity data is deleted—the recalculated hash value will be inconsistent with the original hash value written to the consortium blockchain. Therefore, the hash value is used to determine whether the stage state data has been tampered with or replaced after being written.

[0110] In this embodiment, the hash algorithm used is SHA-256. The SHA-256 algorithm has a fixed output length of 256 bits, capable of converting stage status data of arbitrary length into a fixed-length data digest. SHA-256 is chosen because it has high collision resistance and mature engineering applications, making it suitable for integrity verification of status data during transportation, warehousing, and usage stages.

[0111] It should be noted that the specific process of hash anchoring is as follows: Upon completion of the transportation, warehousing, and usage phases, or upon reaching preset write conditions, the system standardizes the status data for the corresponding phase and concatenates them according to a preset field order to form a data string to be hashed. Let the standardized status data string for the j-th phase be... The corresponding stage hash value Determine as follows: Where j represents the transportation stage, storage stage, or usage stage. This indicates the SHA-256 hash algorithm.

[0112] The standardization process includes field sorting, unit unification, numerical value retention digit unification, timestamp format unification, target gas type encoding unification, and null value filling rules unification. For example, peak acceleration during transportation is uniformly expressed in g, and the duration of abnormal impact events is uniformly expressed in seconds; gas concentration during storage is uniformly expressed in ppm, and the slope of gas concentration linear fitting is uniformly expressed in ppm / h; initial DC internal resistance during use is uniformly expressed in milliohms, and initial capacity is uniformly expressed in ampere-hours.

[0113] After generating the stage hash value, the system will... The container identification, stage type, stage timestamp, state data index, and previous stage hash value are written to the corresponding node in the consortium blockchain. The state data itself is stored in an off-chain database, monitoring platform database, or trusted object storage system, while the state data index points to the specific storage location of the off-chain data. In this way, the consortium blockchain does not directly store large volumes of raw sensor data; instead, it stores hash digests and index information that can verify the integrity of the off-chain state data.

[0114] The subsequent integrity verification process is as follows: When a safety incident, abnormal shutdown, thermal runaway warning, or pre-commissioning re-inspection anomaly occurs, the monitoring platform queries the stage integrity relationship chain based on the container's identity identifier and determines the last stage node that was written at the time of the incident. Subsequently, the system traces backward from the last stage node, sequentially reading the on-chain hash values ​​and off-chain state data indexes written in the transportation, storage, and usage stages.

[0115] For any given stage, the system reads the off-chain state data based on the state data index recorded on the chain, and regenerates the state data string to be verified according to the same field order, unit rules, numerical precision, and null padding method used when generating the original hash value for that stage. Let the regenerated state data string to be verified be... Then trace the hash value Determine as follows: .

[0116] Then, the hash value will be traced. The original hash value stored in the corresponding stage of the consortium blockchain Perform a character-by-character comparison. If satisfied... This indicates that the off-chain state data in this stage has not been tampered with, replaced, or missing since it was written to the consortium blockchain; if it meets the following conditions... If the data is found to be abnormal, it indicates that the off-chain state data at this stage may have been tampered with, missing, replaced, incorrectly uploaded, or the index may be invalid. The system will then identify this stage as a data abnormality stage and output a data abnormality message.

[0117] At the end of the transportation phase, the transportation status data is recorded, and the first hash value of the transportation status data is determined. The first hash value is written into the first node of the phase integrity relation chain. The transportation status data includes container identification, transportation start and end timestamps, abnormal impact events, mechanical damage component values, and total impact energy accumulation value of the transportation phase. Specifically, the system determines the end of the transportation phase when a logistics vehicle enters the electronic fence of the regional warehousing center and completes the handover. The logistics carrier's onboard communication unit or logistics dispatch system aggregates all key data from the transportation phase into transportation status data. Transportation status data includes at least: container identification, transport vehicle number, driving task number, transportation start timestamp, transportation end timestamp, transportation route number, abnormal impact event number, occurrence time of each abnormal impact event, peak acceleration, duration of the abnormal impact event, cumulative impact energy of the abnormal impact event, total cumulative impact energy of the transportation phase, and mechanical damage component value.

[0118] In the data processing, to ensure consistent results when recalculating hash values ​​for the same transportation status data on different devices, the transportation status data needs to be standardized first. Standardization includes field sorting, unit unification, uniformity of the number of decimal places retained, and uniform null value filling rules. For example, peak acceleration is uniformly set to g as the unit, retaining three decimal places; the duration of abnormal impact events is uniformly set to seconds as the unit, retaining two decimal places; timestamps are uniformly set to year-month-day-hour-minute-second format; and when no abnormal impact event has occurred, the abnormal impact event list is recorded as an empty array instead of directly deleting the field.

[0119] Specifically, after standardization, the transportation status data is concatenated into a data string to be hashed according to a preset field order. The preset field order can be as follows: container identification, transport vehicle number, transportation start timestamp, transportation end timestamp, transportation route number, list of abnormal impact events, total cumulative impact energy during transportation, and mechanical damage component value. Then, a SHA-256 hash calculation is performed on this data string to obtain the first hash value.

[0120] It should be noted that after the first hash value is written to the first node of the consortium blockchain, the transportation status data itself can be stored on the logistics carrier's server, the monitoring platform's database, or a trusted data storage system. In addition to recording the first hash value, the first node of the consortium blockchain also records a transportation status data index. This index points to the specific storage location of the off-chain transportation status data, such as a database table name, record number, file number, or object storage address. In this way, the consortium blockchain does not need to store large amounts of raw sensor data, but the raw data can still be read and the hash value recalculated subsequently through the index.

[0121] At the end of the storage phase, the storage status data is recorded, and the second hash value of the storage status data is determined. The second hash value is written into the second node of the phase integrity relation chain. The storage status data includes container identification, storage start and end timestamps, chemical escape characteristic parameters of the storage phase, cumulative temperature and humidity values, and chemical damage component values. Specifically, when the warehouse management system generates an outbound confirmation record, the system determines that the warehousing phase has ended. The warehouse management node extracts the warehouse status data corresponding to the container's identification from the warehouse monitoring platform. The warehouse status data includes at least: container identification, inbound timestamp, outbound timestamp, warehouse location number, linear fitting slope of hydrogen concentration, linear fitting slope of carbon monoxide concentration, linear fitting slope of volatile organic compound concentration, target escaping gas type, linear fitting slope of target gas concentration, cumulative temperature exceedance, cumulative humidity exceedance, cumulative temperature and humidity values, and chemical damage component value.

[0122] Specifically, before generating the second hash value, the warehouse status data also needs to be standardized. For example, gas concentration is uniformly represented by ppm, the slope of the linear fitting of gas concentration is uniformly represented by ppm / h, the cumulative amount of temperature exceeding the limit is uniformly represented by ℃·h, and the cumulative amount of humidity exceeding the limit is uniformly represented by %RH·h. The type of target escaping gas is represented by a unified code, such as hydrogen as H2, carbon monoxide as CO, and volatile organic compounds as VOC. This avoids the hash value being unreproducible due to inconsistencies in field names or units across different systems.

[0123] The standardized warehouse status data is concatenated into a hash string according to a preset field order, and then subjected to SHA-256 hash calculation to obtain a second hash value. This second hash value is written to the second node of the stage integrity relationship chain and simultaneously to the warehouse status data index. This index points to sensor records, gas slope calculation results, cumulative temperature and humidity results, and warehouse outbound sorting records for each stage of the warehouse process.

[0124] It should be noted that the role of the second hash value is to lock in the critical factual state during the storage phase. If a safety incident subsequently occurs, and the storage management attempts to modify the gas concentration slope, delete high-temperature and high-humidity records, or adjust the chemical damage component values, the recalculated traceability hash value will not match the on-chain second hash value, thus enabling the detection of data anomalies during the storage phase. However, this anomaly only indicates a problem with data integrity; the cause of the incident still needs to be analyzed in conjunction with the incident type, time window, failure mode, and on-site verification conclusions.

[0125] During the usage phase, the usage status data is recorded, and the third hash value of the usage status data is determined. The third hash value is then written to the third node of the phase integrity relationship chain. The usage status data includes container identification, usage timestamp, initial DC internal resistance, initial capacity, and electrochemical aging component value. Specifically, after the containerized battery system is connected to a new energy power station and completes its first charge-discharge cycle, the power station operator generates usage status data. This usage status data includes at least: container identification, power station number, grid connection testing timestamp, first charge-discharge cycle start timestamp, first charge-discharge cycle end timestamp, initial DC internal resistance of each cell, initial capacity of each cell, rated internal resistance, rated capacity, average internal resistance increment, average capacity decay, and electrochemical aging component value.

[0126] Before calculating the third hash value, the power plant operator's system standardizes the usage status data. For example, the initial DC internal resistance is uniformly expressed in milliohms, with three decimal places; the initial capacity is uniformly expressed in ampere-hours, with two decimal places; the cell numbers are arranged in ascending order of battery cluster number, module number, and cell serial number; if a cell's data is invalid, the cell number is retained in the status data and marked as invalid, rather than the cell record being deleted. This ensures that the same status data can be completely reproduced during traceability analysis.

[0127] Specifically, the standardized usage status data is concatenated into a data string to be hashed according to a preset field order, and then SHA-256 hash calculation is performed to obtain a third hash value. The third hash value is written to the third node of the stage integrity relationship chain, and simultaneously written to the usage status data index. The usage status data index is used to point to the first charge-discharge cycle data, internal resistance calculation results, capacity calculation results, and aging damage calculation results in the power plant energy management system or battery management system.

[0128] When a safety incident occurs in any stage of the containerized battery system, the last stage node that was written into the stage integrity relation chain at the time of the incident is determined. Then, the process is traced back along the stage integrity relation chain from the last stage node, and the state data corresponding to the hash values ​​of each stage written into the stage integrity relation chain is verified in turn. The trace hash value is obtained by re-hash calculation on the state data, and the trace hash value is compared with the corresponding hash value written into the consortium chain. If a stage has not yet completed the stage writing at the time of the incident, the stage that has not been written into the chain is only considered as a stage to be supplemented with evidence and does not participate in the determination of the integrity consistency of the data already on the chain.

[0129] Specifically, safety incidents can include structural abnormalities during transportation, gas anomaly alarms during storage, fires in storage, battery cluster overheating during use, thermal runaway warnings, or abnormal shutdowns. After an incident occurs, the monitoring platform queries the stage integrity chain based on the container's identification and determines the last stage node that was written at the time of the incident.

[0130] In this embodiment, the final stage node refers to the last stage node that has been written at the time of the incident. If the incident occurs during the warehousing process, but the warehousing stage has not yet ended and the second hash value has not been written, then the final stage node is the transportation stage node; if the incident occurs after the warehousing process has ended and the second hash value has been written, then the final stage node is the warehousing stage node; if the incident occurs during the usage stage and the third hash value has been written, then the final stage node is the usage stage node.

[0131] Specifically, after determining the final stage node, the system reads the on-chain nodes one by one from that node backwards. For example, if the final stage node is the usage stage node, the backtracking order is the usage stage node, the warehousing stage node, and the transportation stage node. The system reads the corresponding off-chain state data based on the state data index in each node, and re-executes the SHA-256 hash calculation according to the same standardization rules and field order used when generating the hash value for that stage to obtain the trace hash value.

[0132] The trace hash value is compared character by character with the hash value stored by the corresponding stage node in the consortium blockchain. If the two are completely consistent, it means that the state data of that stage has not been modified since it was written to the consortium blockchain; if the two are inconsistent, it means that the state data of that stage may have been tampered with, missing, replaced, uploaded incorrectly, or stored corrupted.

[0133] When the trace hash value of any stage is inconsistent with the corresponding hash value written to the consortium blockchain, the stage is identified as a data anomaly stage and a data anomaly prompt is output. When the traceability hash values ​​of each stage are consistent, the non-negative excess ratio of the mechanical damage component value relative to the first-level warning threshold is determined as the risk contribution value of the transportation stage; the non-negative excess ratio of the chemical damage component value relative to the second-level warning threshold is determined as the risk contribution value of the storage stage; and the non-negative excess ratio of the electrochemical aging component value relative to the lowest charge / discharge constraint threshold among multiple charge / discharge constraint thresholds is determined as the risk contribution value of the use stage. Specifically, when the corresponding damage component value does not exceed the corresponding risk threshold, the non-negative excess ratio of the corresponding stage is set to zero; when the corresponding damage component value exceeds the corresponding risk threshold, the non-negative excess ratio of the corresponding stage is determined based on the ratio of the difference between the corresponding damage component value and the corresponding risk threshold to the corresponding risk threshold. The risk contribution ranking results are generated in descending order of risk contribution value in the transportation stage, storage stage, and usage stage. The risk contribution ranking results are then used together with the manual review rules as auxiliary liability analysis results.

[0134] In one optional implementation, when a containerized battery system experiences a safety accident, abnormal shutdown, thermal runaway warning, or pre-commissioning inspection anomaly, the monitoring platform first determines the last stage node already written into the stage integrity relationship chain based on the accident occurrence time. Then, it traces back from this last stage node to the transportation stage node, storage stage node, and usage stage node. For each stage node, the monitoring platform retrieves off-chain status data based on the status data index recorded in that node, and regenerates the status data string to be verified according to a preset field order, unified data unit, and unified encoding format. A hash calculation is then performed on this status data string to obtain the traceability hash value for the corresponding stage. Subsequently, the traceability hash value is compared with the corresponding hash value already written in the consortium blockchain. If the traceability hash value of any stage is inconsistent with the corresponding hash value written to the consortium blockchain, it indicates that there is a risk of missing, tampered, written, or indexed abnormalities in the off-chain state data corresponding to that stage. The monitoring platform will identify that stage as a data abnormal stage and output a data abnormality prompt. The data abnormality prompt will include at least the container identification, abnormal stage type, abnormal node timestamp, state data index, and hash comparison result, so that operation and maintenance personnel can supplement, review, or manually confirm the data of that stage.

[0135] In another optional implementation, when the traceability hash values ​​for the transportation, storage, and usage stages are all consistent with the corresponding hash values ​​in the consortium blockchain, it indicates that the status data for each stage has passed integrity verification, and the monitoring platform further calculates the risk contribution value for each stage. Specifically, the mechanical damage component value is compared with the first-level warning threshold. When the mechanical damage component value does not exceed the first-level warning threshold, the non-negative over-limit ratio for the transportation stage is set to zero; when the mechanical damage component value exceeds the first-level warning threshold, the difference between the mechanical damage component value and the first-level warning threshold is divided by the first-level warning threshold to obtain the risk contribution value for the transportation stage. The chemical damage component value is compared with the second-level warning threshold. When the chemical damage component value does not exceed the second-level warning threshold, the non-negative over-limit ratio for the storage stage is set to zero; when the chemical damage component value exceeds the second-level warning threshold, the difference between the chemical damage component value and the second-level warning threshold is divided by the second-level warning threshold to obtain the risk contribution value for the storage stage. The electrochemical aging component value is compared with the lowest charge / discharge constraint threshold among multiple charge / discharge constraint thresholds. When the electrochemical aging component value does not exceed the lowest charge / discharge constraint threshold, the non-negative over-limit ratio of the usage stage is set to zero. When the electrochemical aging component value exceeds the lowest charge / discharge constraint threshold, the difference between the electrochemical aging component value and the lowest charge / discharge constraint threshold is divided by the lowest charge / discharge constraint threshold to obtain the risk contribution value of the usage stage. The monitoring platform generates a stage risk contribution ranking result in descending order of risk contribution value for the transportation stage, storage stage, and usage stage. This stage risk contribution ranking result, together with the manual review rules, serves as an auxiliary responsibility analysis result. The manual review rules include verifying the accident type, accident occurrence time window, failure mode, sensor evidence chain, on-site detection records, and operation and maintenance records to avoid directly determining the responsibility stage based solely on a single numerical result.

[0136] Example 2, please refer to Figure 1 and Figure 2 Specifically: an IoT-based intelligent monitoring system for containerized battery systems, including: The data acquisition module is used to acquire characteristic parameters of the transportation stage, chemical emission characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage, respectively. The full-cycle construction module is used to align the characteristic parameters of the transportation stage, the chemical emission characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage across stages using container identification and timestamp as indexes, and then perform normalization processing to construct a full-cycle state parameter set. The damage analysis module determines the mechanical damage component value during the transportation stage, the chemical damage component value during the storage stage, and the electrochemical aging component value during the use stage based on the full-cycle state parameter set, and uses them as the damage index for the corresponding stage. The cross-stage monitoring module is used to compare the damage index of each stage with a preset set of risk thresholds, determine the risk level classification results, and perform cross-stage collaborative scheduling operations based on the risk level classification results. The traceability analysis module is used to record the status data of each stage to the consortium blockchain in a hash-anchored manner, obtain the data integrity verification log, and determine the stage risk contribution ranking results and auxiliary responsibility analysis results based on the data integrity verification log.

[0137] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0138] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0139] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0140] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0141] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent monitoring method for containerized battery systems based on the Internet of Things, characterized in that: Includes the following steps: S1. Obtain characteristic parameters of the transportation stage, chemical escaping characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage, respectively. S2. Align the characteristic parameters of the transportation stage, the chemical emission characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage across stages using container identification and timestamp as indexes, and perform normalization processing to construct a full-cycle state parameter set. S3. Based on the full-cycle state parameter set, determine the mechanical damage component value in the transportation stage, the chemical damage component value in the storage stage, and the electrochemical aging component value in the use stage, and use them as the damage index for the corresponding stage. S4. Compare the damage index of each stage with the preset risk threshold set to determine the risk level classification result, and perform cross-stage collaborative scheduling operation based on the risk level classification result. S5. Record the status data of each stage to the consortium blockchain in a hash-anchored manner to obtain the data integrity verification log, and determine the stage risk contribution ranking results and auxiliary responsibility analysis results based on the data integrity verification log.

2. The intelligent monitoring method for a containerized battery system based on the Internet of Things according to claim 1, characterized in that: The logic for obtaining characteristic parameters of the transportation stage, chemical escaping characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage is as follows: During the transportation phase, vibration data is continuously collected at a preset transportation vibration sampling frequency using a triaxial accelerometer installed at the bottom of the container. Environmental temperature and humidity data are continuously collected at preset transportation environment sampling intervals using a cargo hold temperature and humidity sensor installed outside the container. The vibration data is processed using a Fast Fourier Transform (FFT) to calculate the impact energy value for each time period. Continuous time periods where the impact energy value exceeds a preset single-frame impact threshold are marked as abnormal candidate time periods. The impact energy values ​​corresponding to consecutive abnormal candidate time periods are accumulated to obtain the cumulative impact energy value of the abnormal impact event. When the cumulative impact energy value of the abnormal impact event exceeds a preset event impact threshold, the consecutive abnormal candidate time periods are confirmed as abnormal impact events. Simultaneously, the duration, peak acceleration, and cumulative impact energy value of the abnormal impact event are recorded as characteristic parameters of the transportation phase. During the storage phase, a gas sensor array installed inside the container continuously collects hydrogen, carbon monoxide, and volatile organic compound (VOC) concentration data at preset gas collection intervals. Simultaneously, a storage temperature and humidity sensor continuously collects ambient temperature and humidity data at preset temperature and humidity collection intervals. Baseline drift correction is performed on the hydrogen, carbon monoxide, and VOC concentration data, and the linear fitting slope of each gas concentration during the storage phase is calculated. The linear fitting slope of each gas concentration, the target escaping gas type, and the linear fitting slope of the target gas concentration are used as chemical escaping characteristic parameters during the storage phase. During the first charge-discharge cycle of the containerized battery system in the usage phase, the initial DC internal resistance and initial capacity of each cell are obtained as baseline characteristic parameters for the usage phase.

3. The intelligent monitoring method for containerized battery systems based on the Internet of Things according to claim 2, characterized in that: The logic for constructing the full-cycle state parameter set is as follows: Based on the container identification, the characteristic parameters of the transportation stage, the chemical escaping characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage are respectively appended with the corresponding stage start timestamp and stage end timestamp. Based on the cumulative impact energy, peak acceleration, and duration of each abnormal impact event during the transportation phase, the raw mechanical damage data is constructed. The raw mechanical damage data also includes the total cumulative impact energy during the transportation phase. Based on the chemical escape characteristic parameters and cumulative temperature and humidity values ​​during the storage stage, the original chemical state data is constructed. Among them, the cumulative temperature and humidity values ​​are obtained by normalizing and summing the cumulative temperature exceedance of the storage stage ambient temperature data relative to the preset standard temperature range and the cumulative humidity exceedance of the storage stage ambient humidity data relative to the preset standard humidity range. Based on the baseline characteristic parameters of the usage phase, construct the raw data of the electrochemical baseline; The mechanical damage raw data, chemical state raw data, and electrochemical baseline raw data belonging to the same container identity are combined in stages according to time sequence to generate a full-cycle raw state table for the corresponding container identity. The parameters in the original state table for the entire cycle are normalized according to their corresponding parameter types to generate a set of state parameters for damage analysis. Among them, mechanical damage parameters are normalized based on the impact energy baseline value of events under non-destructive transportation conditions, which is the statistical value of the impact energy of events under non-destructive transportation conditions. Chemical emission parameters are normalized based on the baseline of the ambient gas concentration at the factory standard. Electrochemical baseline parameters are normalized based on the rated internal resistance and rated capacity in the factory test report.

4. The intelligent monitoring method for containerized battery systems based on the Internet of Things according to claim 3, characterized in that: The logic for determining the damage index at the corresponding stage is as follows: Feature identification is performed on the full-cycle state parameter set, and the cumulative impact energy value of each abnormal impact event is extracted and statistically analyzed. The cumulative impact energy value, peak acceleration and duration of the abnormal impact event corresponding to the i-th abnormal impact event are multiplied to obtain the single mechanical damage component value of the i-th abnormal impact event. The single mechanical damage component values ​​of all abnormal impact events are accumulated to determine the mechanical damage component value, which is used as the mechanical damage index for the transportation stage. The slope of the linear fitting of the target gas concentration during the storage stage is extracted from the set of state parameters throughout the entire cycle. When the slope of the linear fitting of the target gas concentration is positive and exceeds the preset escape slope threshold, the slope of the linear fitting of the target gas concentration is multiplied by the cumulative value of temperature and humidity to obtain the chemical damage component value, which is used as the chemical damage index during the storage stage. When the slope of the linear fitting of the target gas concentration is not positive, or the slope of the linear fitting of the target gas concentration does not exceed the preset escape slope threshold, the chemical damage component value is set to zero. The internal resistance increment and capacity decay of each cell during the service stage are extracted from the full-cycle state parameter set. The internal resistance increment and capacity decay of each cell are weighted and summed to obtain the electrochemical aging component value, which is used as the aging damage index during the service stage.

5. The intelligent monitoring method for a containerized battery system based on the Internet of Things according to claim 4, characterized in that: The logic for performing cross-stage collaborative scheduling operations based on risk level classification results is as follows: Damage indices at different stages are compared with a set of risk thresholds to obtain risk level classification results; the set of risk thresholds includes a first-level warning threshold for the transportation stage, a second-level warning threshold for the storage stage, and multiple charge and discharge constraint thresholds for the usage stage. During the transportation phase, when the mechanical damage index exceeds the preset first-level warning threshold, a transportation suggestion signal is sent to the logistics scheduling system through the vehicle communication unit. The transportation suggestion signal includes reselecting a transportation route with road smoothness parameters that meet the preset road smoothness threshold and driving within the preset vibration reduction transportation speed range. When the mechanical damage index does not exceed the first-level warning threshold, no additional transportation suggestion signal is issued. During the warehousing phase, when the chemical damage index during warehousing is greater than or equal to the Level 2 warning threshold, the corresponding container identification is marked as a warehousing isolation re-inspection object, and the warehousing isolation re-inspection object is transferred from the regular outbound queue to the isolation re-inspection outbound queue, ventilation detection area, or fire monitoring area, triggering a manual review process. When there are multiple warehousing isolation re-inspection objects, an isolation re-inspection list is generated in descending order of the chemical damage index during warehousing. If the chemical damage index during warehousing is the same, they are arranged in ascending order of entry time. When the chemical damage index during warehousing is less than the Level 2 warning threshold, the corresponding container identification is marked as a regular warehousing object, and a regular outbound list is generated in ascending order of entry time. Based on the isolation re-inspection list and the regular outbound list, an updated warehousing re-inspection and outbound plan list is generated. During the usage phase, when the containerized battery system is connected to the power station and put into operation, the aging damage index of the usage phase is compared with multiple charge and discharge constraint thresholds. Among them, multiple charge and discharge constraint thresholds are respectively established with multiple rate upper limits. When the aging damage index of the usage phase exceeds the corresponding charge and discharge constraint threshold, the corresponding rate upper limit is determined in the charge and discharge constraint mapping relationship according to the risk level classification result, and the maximum allowable charge and discharge rate is limited to no more than the corresponding rate upper limit.

6. The intelligent monitoring method for a containerized battery system based on the Internet of Things according to claim 5, characterized in that: The logic for determining the ranking of risk contributions at each stage and the results of the auxiliary responsibility analysis is as follows: Establish a phase integrity relationship chain corresponding to the container identity in the consortium blockchain. The data integrity verification log consists of the hash value and corresponding state data index recorded by each node in the phase integrity relationship chain. At the end of the transportation phase, the transportation status data is recorded, and the first hash value of the transportation status data is determined. The first hash value is written into the first node of the phase integrity relation chain. The transportation status data includes container identification, transportation start and end timestamps, abnormal impact events, mechanical damage component values, and total impact energy accumulation value of the transportation phase. At the end of the storage phase, the storage status data is recorded, and the second hash value of the storage status data is determined. The second hash value is written into the second node of the phase integrity relation chain. The storage status data includes container identification, storage start and end timestamps, chemical escape characteristic parameters of the storage phase, cumulative temperature and humidity values, and chemical damage component values. During the usage phase, the usage status data is recorded, and the third hash value of the usage status data is determined. The third hash value is then written to the third node of the phase integrity relationship chain. The usage status data includes container identification, usage timestamp, initial DC internal resistance, initial capacity, and electrochemical aging component value. When a safety incident occurs in the containerized battery system at any stage, the last stage node that was written into the stage integrity relation chain at the time of the incident is determined, and the stage integrity relation chain is traced back from the last stage node. The state data corresponding to the hash values ​​of each stage written into the stage integrity relation chain is verified in turn. The trace hash value is obtained by re-hash calculation on the state data, and the trace hash value is compared with the corresponding hash value written into the consortium chain. When the traceability hash value of any stage is inconsistent with the corresponding hash value written to the consortium blockchain, that stage is identified as a data anomaly stage and a data anomaly prompt is output. When the traceability hash values ​​of all stages are consistent, the non-negative excess ratio of the mechanical damage component value relative to the first-level warning threshold is determined as the risk contribution value of the transportation stage, the non-negative excess ratio of the chemical damage component value relative to the second-level warning threshold is determined as the risk contribution value of the storage stage, and the non-negative excess ratio of the electrochemical aging component value relative to the lowest charge / discharge constraint threshold among multiple charge / discharge constraint thresholds is determined as the risk contribution value of the usage stage. Among these, when the corresponding damage component value does not exceed the corresponding risk threshold, the non-negative excess ratio of the corresponding stage is marked as zero; when the corresponding damage component value exceeds the corresponding risk threshold, the non-negative excess ratio of the corresponding stage is determined according to the ratio of the difference between the corresponding damage component value and the corresponding risk threshold to the corresponding risk threshold. The risk contribution ranking results are generated in descending order of risk contribution value in the transportation stage, storage stage, and usage stage. The risk contribution ranking results are then used together with the manual review rules as auxiliary liability analysis results.

7. An IoT-based intelligent monitoring system for containerized battery systems, used to implement the IoT-based intelligent monitoring method for containerized battery systems as described in any one of claims 1 to 6, characterized in that: include: The data acquisition module is used to acquire characteristic parameters of the transportation stage, chemical emission characteristic parameters of the storage stage, and baseline characteristic parameters of the usage stage, respectively. The full-cycle construction module is used to align the characteristic parameters of the transportation stage, the chemical emission characteristic parameters of the storage stage, and the baseline characteristic parameters of the usage stage across stages using container identification and timestamp as indexes, and then perform normalization processing to construct a full-cycle state parameter set. The damage analysis module determines the mechanical damage component value during the transportation stage, the chemical damage component value during the storage stage, and the electrochemical aging component value during the use stage based on the full-cycle state parameter set, and uses them as the damage index for the corresponding stage. The cross-stage monitoring module is used to compare the damage index of each stage with a preset set of risk thresholds, determine the risk level classification results, and perform cross-stage collaborative scheduling operations based on the risk level classification results. The traceability analysis module is used to record the status data of each stage to the consortium blockchain in a hash-anchored manner, obtain the data integrity verification log, and determine the stage risk contribution ranking results and auxiliary responsibility analysis results based on the data integrity verification log.

Images