Marine battery fault grading processing method

Abstract

The embodiment of the application provides a kind of marine battery fault grading processing method, applied to the power distribution and fault management technical field of marine movable battery system, by obtaining interface connection state signal and position positioning signal analysis operation stage identification, collect battery monomer voltage, temperature and current parameter and construct stage correlation battery feature vector, deviation enhancement is generated in combination with stage historical data steady reference value, again according to mechanical vibration state, temperature gradient state and current impact state generate environmental disturbance weight, carry out stage adjustment to basic determination baseline, obtain fault grading result based on deviation degree, and match reachable execution end to generate grading disposal instruction set.The application can improve the accuracy of fault grading under multi-stage ship-shore transfer working condition, reduce the risk of disturbance false alarm and early leakage, and improve the timeliness and safety of collaborative disposal.

Description

Technical Field

[0001] This application relates to the field of power distribution and fault management technology for marine portable battery systems, and in particular to a method for graded fault handling of marine batteries. Background Technology

[0002] A marine mobile battery platform is a large energy storage device used to provide power, replenishment, or shore power replacement services for ships in port. It typically carries lithium iron phosphate battery modules in standardized container form and coordinates with shore-ship interface equipment such as a ship battery compartment push-pull mechanism, a docking platform, a shore-based electric tractor, and a shore-based charging station to complete the transfer and replenishment of power from the ship's compartment to the shore-based charging station. In a port environment, this type of platform undergoes multiple operational stages, including in-hull power supply, launching from the ship's compartment, docking and transfer, land towing, shore-based charging, and retraction and reset. Each stage corresponds to different electrical load characteristics, mechanical vibration environments, and thermal boundary conditions.

[0003] Existing methods for detecting marine battery faults typically follow a static threshold comparison approach. This involves setting fixed upper and lower limits for voltage, temperature, and current protection based on the factory-rated parameters of individual battery cells. During platform operation, the parameters of each cell are collected in real time and compared with the fixed thresholds. Once a parameter exceeds the preset boundary, an alarm or protection action is triggered. Some solutions further incorporate timing average filtering or temperature rise rate judgment logic to reduce the probability of false triggers caused by transient interference.

[0004] However, fixed fault detection thresholds face significant adaptability limitations under the multi-stage operational conditions of marine battery platforms. Road vibrations and shocks experienced by the battery platform during land transport cause high-frequency noise to superimpose on the voltage signal. During charging operations, the rapid changes in current according to the charging curve significantly alter the temperature rise rate compared to the stable discharge phase. Using a uniform threshold across different operational stages makes it difficult to simultaneously ensure fault identification sensitivity and false alarm suppression. Furthermore, individual battery cells exhibit varying parameter drifts during service due to differences in charge-discharge cycles, aging levels, and operating environments. Static thresholds struggle to distinguish between background deviations caused by aging accumulation and abnormal jumps caused by internal micro-short circuits, easily leading to missed early faults or repeated false alarms for aging cells. This further results in inaccurate fault level assessments and untimely response actions. Summary of the Invention

[0005] This application provides a method for graded fault handling of marine batteries. It enables dynamic acquisition of battery parameters, adaptive adjustment of fault thresholds, accurate identification of fault levels, and collaborative handling by multiple execution terminals during various operational phases of a mobile battery platform, including shipboard power supply, pushback transition, docking, land towing, charging, and pullback reset. This reduces the risk of false alarms caused by mechanical vibration, temperature gradients, and current surges under complex ship-to-shore transfer conditions, and improves the timeliness and safety of fault handling for marine mobile battery platforms. To achieve the above objectives, this application adopts the following technical solution: A method for graded fault handling of marine batteries, the method comprising: The system acquires interface connection status signals between the mobile battery platform and the ship's battery compartment push-pull mechanism, docking platform, shore-based electric tractor, and shore-side charging station, and acquires the location signal of the mobile battery platform. It then performs joint analysis on the interface connection status signals and the location signal to obtain the operation phase identifier. The system collects voltage, temperature, and current parameters of individual battery cells within the mobile battery platform and associates them with the operation phase identifier to generate a phase-associated battery feature vector. It categorizes and records preceding voltage, temperature, and current parameters according to the operation phase identifier to construct a phase history dataset. Based on this dataset, it generates a phase steady-state reference value and calculates the difference between the phase-associated battery feature vector and the phase steady-state reference value. The features are then merged to obtain the enhanced stage-related battery feature vector; a basic judgment baseline corresponding to the rated operating parameters of the battery cell is obtained; mechanical vibration state, temperature gradient state, and current impact state are identified based on the operation stage identifier, and environmental disturbance weights are generated; the basic judgment baseline is expanded or tightened based on the environmental disturbance weights to obtain a staged fault judgment threshold; a fault classification result is generated based on the deviation between the enhanced stage-related battery feature vector and the staged fault judgment threshold; based on the fault classification result and the operation stage identifier, disposal actions are selected from the disposal action list, and a graded disposal instruction set is generated by combining the disposal capabilities of the reachable execution end, and distributed to the corresponding execution end to perform collaborative disposal actions.

[0006] As can be seen from the above technical solution, this application has the following beneficial effects: 1. This application, by jointly analyzing the interface connection status signals between the mobile battery platform and the ship's battery compartment push-pull mechanism, docking platform, shore-based electric tugboat, and shore-side charging station, along with the mobile battery platform's location signal, can accurately identify the actual status of the mobile battery platform during different operational stages, such as shipboard power supply, push-out transition, docking and mooring, land towing, charging operations, and pull-back reset. Compared to methods that rely solely on a single electrical parameter or interface signal for fault diagnosis, this application can first determine the specific operational stage of the battery platform before proceeding with subsequent battery parameter acquisition, threshold adjustment, and fault classification. This avoids mismatches between sampling strategies, judgment thresholds, and handling actions due to misjudgments of the operational stage, thereby improving the fundamental accuracy of fault identification in multi-stage ship-shore transfer scenarios.

[0007] 2. This application collects the voltage, temperature, and current parameters of individual battery cells according to the operating stage identifier, generates a stage steady-state reference value based on historical data within the same operating stage, and then performs differential enhancement by comparing the current stage's associated battery feature vector with the stage steady-state reference value. This weakens the background drift caused by long-term service, differences in cycle count, and normal aging of individual battery cells, making abnormal changes in current parameters relative to historical steady-state values ​​more prominent. Therefore, it can more accurately distinguish between slow deviations caused by normal aging and abnormal deviations caused by faults, reducing repeated false alarms for aging cells, and improving the ability to identify early faults, latent faults, and stage-specific abnormal fluctuations.

[0008] 3. This application generates environmental disturbance weights by acquiring mechanical vibration, temperature gradient, and current impact states. These weights are then used to expand or tighten the boundaries of the basic judgment baseline, resulting in a phased fault judgment threshold. This allows the fault judgment threshold to be dynamically adjusted according to the actual disturbance intensity at different operational stages. During periods of strong vibration or impact, such as land towing, the judgment boundary can be appropriately expanded to reduce false alarms caused by mechanical disturbances and current impacts. During relatively stable stages, such as shipboard power supply and docking, the judgment boundary can be tightened to maintain sensitivity to subtle anomalies. Furthermore, this application combines fault classification results, operational stage identifiers, and the handling capabilities of reachable execution terminals to generate a tiered handling instruction set. This enables actions such as early warning prompts, reduced power operation, emergency towing, standby locking, and station entry blockade to be assigned to execution terminals with corresponding capabilities and currently reachable locations. This achieves closed-loop processing from fault identification and tiered judgment to collaborative handling, improving the timeliness, accuracy, and safety of fault handling for marine mobile battery platforms. Attached Figure Description

[0009] The present application will be further described below with reference to the accompanying drawings.

[0010] Figure 1A flowchart of a fault classification and handling method for marine batteries provided in this application embodiment; Figure 2 A flowchart for parsing the operation stage identifier and generating the battery feature vector associated with the enhancement stage is provided in the embodiments of this application; Figure 3 A flowchart for generating staged fault determination thresholds and classifying faults provided in the embodiments of this application; Figure 4 A flowchart illustrating the generation and collaborative processing of hierarchical processing instructions provided in this application embodiment. Detailed Implementation

[0011] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are used to distinguish different objects, not to limit a specific order.

[0012] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate illustration or explanation. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0013] Research has revealed that existing marine battery fault detection methods typically set fixed voltage, temperature, and current thresholds based on the rated operating parameters of individual battery cells. During the operation of the mobile battery platform, real-time battery parameters are compared with these fixed thresholds to trigger alarms or protection actions. However, the mechanical vibrations, thermal boundary conditions, and current variation characteristics experienced by the mobile battery platform differ across various operational phases, including shipboard power supply, pushback transition, docking, land towing, charging, and pullback / reset. If a uniform fault threshold is applied across all operational phases, transient parameter fluctuations caused by disturbances may be misinterpreted as faults during high-vibration phases such as land towing. During periods of rapid current change, such as charging, differences in temperature rise and current variation characteristics can lead to false alarms. Conversely, in relatively stable phases such as shipboard power supply, overly broad thresholds may reduce the sensitivity to early anomalies. During long-term service, individual battery cells may experience individual parameter drift due to differences in cycle number, aging degree, and operating environment. Existing fixed threshold methods cannot distinguish between background deviations caused by normal aging and sudden deviations caused by internal anomalies, thus affecting the accuracy of fault classification results and the rationality of subsequent handling actions.

[0014] To address the above problems, this application provides a method for graded fault handling of marine batteries: Example: To solve the above problems, such as Figures 1-4As shown, this embodiment relates to a method for classifying and handling faults in marine batteries. It is particularly applicable to multi-stage operation scenarios where marine mobile battery platforms are used for ship-to-shore transfer, shore-based charging, and pullback reset between ship battery compartments, docking platforms, shore-based electric tractors, and shore-side charging stations. It is especially suitable for port marine battery replacement, transfer, and charging management scenarios that require dynamic judgment, classification, and collaborative handling of individual battery cell faults based on mechanical vibration, temperature gradient, and current impact conditions at different operation stages. The specific implementation process includes the following steps.

[0015] S1. Obtain the interface connection status signal between the mobile battery platform and the ship's battery compartment push-pull mechanism, docking platform, shore-based electric tractor and shore-side charging station, and obtain the position positioning signal of the mobile battery platform. Perform joint analysis on the interface connection status signal and the position positioning signal to obtain the operation stage identifier. In this embodiment, step S1 includes the following process: The system acquires feedback signals from the push-pull stroke of the ship's battery compartment push-pull mechanism, the load-bearing positioning signal of the docking platform, the traction locking signal of the shore-based electric tractor, and the electrical docking signal of the shore-side charging station, and aggregates them into a multi-source interface status set. The system then sorts the trigger times of each signal in the multi-source interface status set and identifies the physical migration trajectory of the mobile battery platform based on the sorting results. Finally, the system cross-validates the physical migration trajectory with the location positioning signal and outputs the operation stage identifier corresponding to the physical migration trajectory when the cross-validation passes.

[0016] Specifically, the central processing unit of the mobile battery platform establishes communication connections with the control unit of the ship's battery compartment push-pull mechanism, the signal processing unit of the docking platform, the vehicle-mounted control unit of the shore-based electric tractor, and the charging controller of the shore-side charging station. The push-pull stroke feedback signal of the ship's battery compartment push-pull mechanism is acquired by displacement sensors installed on the push-pull screw or push-pull guide rail. The analog signal output by the displacement sensor is converted into a stroke value by an analog-to-digital converter. The control unit compares the stroke value with a preset push-out stroke threshold or a preset pull-back stroke threshold. When the stroke value meets the corresponding threshold condition, the push-pull stroke feedback signal is set to the trigger state, and the state transition time is recorded. The load-bearing positioning signal of the docking platform is acquired by weighing sensors or pressure sensors arranged at the load-bearing position of the docking platform. The signal processing unit reads the load readings at each load detection point. When all load readings reach a preset load threshold and the load fluctuation amplitude within a preset stabilization time meets the preset stabilization condition, the load-bearing positioning signal is set to the trigger state, and the state transition time is recorded. The traction locking signal of the shore-based electric tractor is acquired by a proximity switch, limit switch, or locking status sensor located at the traction seat or locking pin mechanism. When the locking pin is inserted into place and the locking mechanism feedback indicates a locked state, the on-board control unit sets the traction locking signal to the triggered state and records the state transition time. The electrical connection signal of the shore charging station is obtained by the charging controller detecting the connection confirmation signal between the charging gun and the mobile battery platform socket, the electronic lock status, and the contactor pre-connection status. Among them, IEC 62196 is the conductive charging connection standard for electric vehicles formulated by the International Electrotechnical Commission. The charging controller reads the connection confirmation signal according to this standard, and after the connection confirmation signal is valid and the electronic lock is locked, it sets the electrical connection signal to the triggered state and records the state transition time.

[0017] The multi-source interface state set refers to a set of state data consisting of push-pull stroke feedback signals, load positioning signals, traction locking signals, and electrical connection signals. This set is not a single sensor signal, but rather structured interface state data stored according to signal type, state value, and system timestamp. After receiving the connection status signals from each interface, the central processing unit writes the signal type, current state value, and state transition time of each signal into a buffer queue, forming the multi-source interface state set. The state transition time is uniformly written by the system clock of the central processing unit to avoid sorting errors caused by inconsistencies in the local clocks of different execution terminals.

[0018] After forming a multi-source interface state set, the central processing unit extracts the first trigger moment when each interface connection state signal changes from an invalid state to an effective state, and sorts them according to the order of trigger moments. For interface connection state signals that have not been triggered, they are not included in the trigger order sorting, but their untriggered state is retained as the basis for judging the operation stage. The central processing unit compares the sorting results with the pre-stored operation stage timing pattern. The operation stage timing pattern is determined by the sequence of physical actions of the mobile battery platform between the ship's cabin, docking platform, towing path, and charging station: when no interface connection state signal is triggered, it corresponds to the shipboard power supply stage; when the push-pull stroke feedback signal is triggered but the load-bearing positioning signal is not triggered, it corresponds to the push-out transition stage; when the load-bearing positioning signal is triggered but the traction lock signal is not triggered, it corresponds to the docking and mooring stage; when the traction lock signal is triggered but the electrical docking signal is not triggered, it corresponds to the land towing stage; when the electrical docking signal is triggered, it corresponds to the charging operation stage; when the electrical docking signal, traction lock signal, load-bearing positioning signal, and push-pull stroke feedback signal are deactivated in sequence, it corresponds to the pull-back reset stage. The central processing unit matches and sorts the results one by one according to the above timing pattern to identify the physical migration trajectory of the mobile battery platform. The physical migration trajectory refers to the platform flow path inferred from the triggering or deactivation sequence of interface connection status signals, and is used to characterize the phase transition trend of the mobile battery platform from shipboard power supply, launch, docking, towing, charging to pull-back and reset.

[0019] The location signal is acquired using ultra-wideband (UWB) positioning. Specifically, an UWB positioning tag is fixedly installed on the outer shell of the mobile battery platform. Multiple positioning base stations are deployed in the port operation area. These base stations receive pulse signals from the positioning tag and calculate the two-dimensional coordinates of the mobile battery platform in the port's planar coordinate system using the time difference of arrival principle. The central processing unit receives the two-dimensional coordinates according to a preset positioning refresh cycle and combines the two-dimensional coordinates with a positioning timestamp to form the location signal. The spatial reference area refers to the port area space corresponding to each operational stage. It is predetermined by the port layout plan, equipment installation location, and actual survey coordinates, and stored in the central processing unit as polygon vertex coordinates. The shipboard power supply stage corresponds to the ship's battery compartment area; the push-out transition stage corresponds to the push-pull mechanism's travel coverage area; the docking and berthing stage corresponds to the docking platform's carrying area; the land towing stage corresponds to the towing road or safety passage area; the charging operation stage corresponds to the docking area of ​​the shore charging station; and the pull-back reset stage corresponds to the pull-back path area from the docking platform back to the ship's battery compartment.

[0020] The central processing unit cross-validates the physical migration trajectory with the location signals. The specific method of cross-validation is as follows: Based on the physical migration trajectory, determine the spatial reference region corresponding to the current inferred operation stage; extract multiple location signal sample points within a preset time range from the location signals; use the ray-mapping method (point within a polygon) to determine whether each location signal sample point falls within the spatial reference region; count the number of location signal sample points falling within the spatial reference region and calculate their proportion of the total number of location signal sample points to obtain the spatial fit. The spatial fit is used to characterize the degree of agreement between the operation stage inferred from the interface connection status signal and the actual spatial location.

[0021] ; In the formula, Indicates spatial fit; This indicates the total number of location signal sample points within a preset time range; Indicates the position positioning signal sample point number; Indicates the first The lateral coordinates of each location signal sample point in the port area's plane coordinate system; Indicates the first The longitudinal coordinates of each location signal sample point in the port area's plane coordinate system; The spatial reference area representing the operational stage corresponding to the physical migration trajectory; This represents the indicator function, which takes a value of 1 when the location signal sample point falls within the spatial reference area, and a value of 0 when the location signal sample point does not fall within the spatial reference area. The spatial fit is a dimensionless proportional value.

[0022] In the above formula, the more location signal sample points that fall within the spatial reference area, the greater the spatial fit, indicating a closer match between the operational stage inferred from the interface connection status signal and the actual spatial location. Conversely, with a constant number of location signal sample points falling within the spatial reference area, a larger total number of location signal sample points results in a smaller spatial fit, indicating a reduced degree of support from the spatial location for the operational stage. Each location signal sample point falling within the spatial reference area contributes positively to the spatial fit, while each location signal sample point not falling within the spatial reference area does not contribute positively.

[0023] When the spatial matching degree meets the preset spatial matching degree threshold, the central processing unit determines that the physical migration trajectory is consistent with the spatial movement path represented by the location positioning signal, the cross-validation passes, and the stage corresponding to the physical migration trajectory is output as the operation stage identifier. The operation stage identifier is a data marker used to represent the current operation status of the mobile battery platform, and its value includes at least one of the following: shipboard power supply stage, push-out transition stage, docking and mooring stage, land towing stage, charging operation stage, and pull-back and reset stage. When the spatial matching degree does not meet the preset spatial matching degree threshold, the central processing unit does not use the stage result corresponding to the current physical migration trajectory, but maintains the previous valid operation stage identifier and records the cross-validation anomaly event. The cross-validation anomaly event includes the anomaly occurrence time, the current sorting result, the current spatial matching degree, and the previous valid operation stage identifier, which are used for subsequent operation and maintenance traceability.

[0024] The beneficial effects of this step are as follows: by identifying the physical flow process of the platform through the triggering timing of the interface connection status signal, and by performing spatial verification of the physical migration trajectory through the location positioning signal, it is possible to avoid stage misjudgment caused by relying solely on a single interface status signal; when the positioning signal occasionally changes or a certain interface signal briefly jitters, the spatial consistency is used for proportional judgment, which can maintain the previous valid operation stage identifier and reduce the impact of frequent operation stage changes on subsequent sampling frequency adjustment and fault threshold setting; at the same time, accurate operation stage identifiers provide a unified time reference for the staged acquisition in step S2 and the comparison with historical same stages, enabling battery parameters to be classified and processed according to actual operating conditions.

[0025] S2. Collect the voltage, temperature, and current parameters of individual battery cells within the mobile battery platform, and associate them with the operation stage identifier to generate a stage-associated battery feature vector; classify and record the preceding voltage, temperature, and current parameters according to the operation stage identifier to construct a stage historical dataset; generate a stage steady-state reference value based on the stage historical dataset; calculate and merge the difference between the stage-associated battery feature vector and the stage steady-state reference value to obtain an enhanced stage-associated battery feature vector; In this embodiment, step S2 includes the following process: The sampling frequency of voltage, temperature, and current parameters is adjusted according to the operation stage identifier to obtain a staged sampling strategy. This staged sampling strategy is then executed to synchronously sample each battery cell in the mobile battery platform, obtaining a time-series parameter sequence for each battery cell. The time-series parameter sequence is concatenated with the operation stage identifier and aligned according to the time windows corresponding to the start and end times of the operation stage identifier, generating a stage-associated battery feature vector. Voltage, temperature, and current parameters from the preceding operation of the mobile battery platform are recorded according to the operation stage identifier, resulting in a stage-historical dataset. Historical segments identical to the current operation stage identifier are extracted from the stage-historical dataset, and the historical segments are averaged to obtain a stage-steady-state reference value. The difference between the stage-associated battery feature vector and the stage-steady-state reference value is calculated to obtain a staged deviation feature. Finally, the staged deviation feature is merged with the stage-associated battery feature vector to obtain the enhanced stage-associated battery feature vector. Specifically, the operation phase identifier is written to the message queue by the central processing unit, and the acquisition strategy management module reads the operation phase identifier and its corresponding start time from the message queue. The phased acquisition strategy refers to the mapping relationship between the operation phase identifier and the acquisition frequencies of voltage, temperature, and current parameters. The acquisition strategy management module queries the mapping table based on the current operation phase identifier to obtain the target acquisition frequency corresponding to the current operation phase. The adjustment amount of the acquisition frequency is determined by the difference between the target acquisition frequency corresponding to the current operation phase and the actual acquisition frequency of the previous time window; when the target acquisition frequency is higher than the actual acquisition frequency of the previous time window, the acquisition strategy management module shortens the sampling trigger interval of the analog-to-digital converter; when the target acquisition frequency is lower than the actual acquisition frequency of the previous time window, the acquisition strategy management module extends the sampling trigger interval of the analog-to-digital converter; when the two are the same, the original sampling trigger interval is maintained. The sampling trigger interval is determined by the reciprocal of the target acquisition frequency and written to the sampling chip configuration register of the battery management system.

[0026] The Battery Management System (BMS) within the mobile battery platform connects to the positive and negative sampling terminals of each battery cell via individual cell voltage acquisition lines to read the voltage parameters of each cell. It also reads temperature parameters using thermocouple temperature probes, thermistor temperature probes, or digital temperature sensors located on the surface of the battery cells or inside the battery modules. Furthermore, it reads current parameters using Hall effect current sensors installed in the main circuit or individual cell branches. Upon receiving the phased acquisition strategy, the BMS synchronously sends sampling trigger pulses to each sampling channel, ensuring that the voltage, temperature, and current parameters at the same sampling moment have the same time reference. After synchronous sampling is completed, the BMS sends the sampled data, battery cell number, and sampling timestamp to the data aggregation layer.

[0027] A timing parameter sequence refers to the sequence of voltage, temperature, and current parameters of the same battery cell within the same operating phase and time window, arranged according to the order of sampling time. Assume the mobile battery platform includes... The first battery cell, the first The first battery cell in the... The voltage parameters at each sampling time are Temperature parameters are The current parameters are ,in, This indicates the serial number of the individual battery cell. This indicates the sampling sequence number, counting from the start time of the current operation phase. This indicates the total number of sampling points accumulated within the current operation phase time window.

[0028] No. The time-series parameter sequence of each battery cell is represented as follows: ; In the formula, Indicates the first The time-series parameter sequence of each battery cell; Indicates the first The first battery cell in the... Voltage parameters at each sampling time; Indicates the first The first battery cell in the... Temperature parameters at each sampling time; Indicates the first The first battery cell in the... Current parameters at each sampling time; This represents the total number of sampling points accumulated within the current operation phase's time window. Voltage, temperature, and current parameters in the time-series parameter sequence are stored side-by-side only at the same sampling time, without cross-dimensional addition or subtraction operations.

[0029] The data aggregation layer maintains a time window queue, using the start and end times of the job stage identifier as the time window boundaries. When step S1 outputs a new job stage identifier, the data aggregation layer closes the time window corresponding to the previous job stage and opens the time window corresponding to the current job stage. The time series parameter sequences falling within the same time window are bound to the job stage identifier corresponding to that time window to generate a stage-associated battery feature vector.

[0030] For the The stage-related battery feature vector of a single battery cell is represented as follows: ; In the formula, Indicates the first Stage-related battery feature vectors of individual battery cells; The coded value indicating the work phase identifier; , and The meaning is consistent with the above; the dimension of the stage-related battery feature vector is... The encoded value of the operation stage identifier is written as the first element of the vector as the stage attribute. Subsequent components are arranged in order of sampling time: voltage parameters, temperature parameters, and current parameters. Since the time window boundary coincides with the start and end times of the operation stage identifier, the battery feature vector associated with each stage only contains battery parameters within that single operation stage, without mixing in sampling data from other operation stages.

[0031] The stage history dataset refers to the collection of preceding operational data categorized and stored according to battery cell number and operation stage identifier. After each complete cycle of the mobile battery platform, the central processing unit writes the stage-related battery feature vectors of each battery cell within that cycle into local solid-state storage, using the battery cell number and operation stage identifier as a combined index. The storage depth of the stage history dataset uses a preset storage depth. When the number of historical segments for the same battery cell and the same operation stage exceeds the preset storage depth, the earliest written historical segment is deleted, and the most recent historical segment is retained. After the current operation stage begins, the central processing unit reads historical segments for the same battery cell and the same operation stage from the stage history dataset based on the current battery cell number and the current operation stage identifier, using these as the data source for generating the stage steady-state reference value.

[0032] The stage steady-state reference value refers to the average historical voltage, temperature, and current parameters of the same battery cell under the same operating stage. It is used to characterize the steady-state operating background of the battery cell under similar operating stages. After the historical segments are extracted, the central processing unit aligns the historical segments according to the sampling time and parameter dimensions. When the number of sampling points in different historical segments is inconsistent, the sampling data at the corresponding time position in the historical segment is truncated or matched by the nearest timestamp based on the number of sampling points in the current operating stage, so that the feature vector of the historical segment and the battery associated with the current stage are comparable at the same sampling position. Subsequently, the historical voltage, temperature, and current parameters at the same sampling position are averaged to obtain the stage steady-state reference vector.

[0033] ; ; In the formula, Indicates the first Individual battery cells are marked during the operation phase. The next stage steady-state reference vector; This indicates the identifier of the battery cell in the historical dataset during the operation phase. The total number of historical fragments below; Indicates the sequence number of a historical segment; Indicates the first The first historical segment The first battery cell in the... Voltage parameters at each sampling location; Indicates the first The first historical segment The first battery cell in the... Temperature parameters at each sampling location; Indicates the first The first historical segment The first battery cell in the... Current parameters at each sampling location; Indicates the first Individual battery cells are marked during the operation phase. Next Historical average voltage at each sampling location; Indicates the first Individual battery cells are marked during the operation phase. Next Historical average temperature at each sampling location; Indicates the first Individual battery cells are marked during the operation phase. Next The historical average current at each sampling location. Each historical average is in the same unit as its corresponding original parameter and is not added to parameters of other dimensions.

[0034] In calculating the steady-state reference values ​​for each stage, the larger the historical voltage parameter at the same sampling location, the larger the historical voltage mean; the larger the historical temperature parameter at the same sampling location, the larger the historical temperature mean; and the larger the historical current parameter at the same sampling location, the larger the historical current mean. The larger the total number of historical segments, the smaller the impact of a single historical segment on the corresponding historical mean; conversely, the smaller the total number of historical segments, the greater the impact of a single historical segment on the corresponding historical mean. The above mean statistics are only performed between the same battery cell, the same operating stage, the same parameter dimension, and the same sampling location to avoid mixing between different operating stages or different parameter dimensions.

[0035] When calculating the difference between the battery feature vector associated with the stage and the steady-state reference value of the stage, the central processing unit only performs item-by-item difference on the same parameter dimension, that is, the current voltage parameter is subtracted from the corresponding historical voltage average, the current temperature parameter is subtracted from the corresponding historical temperature average, and the current current parameter is subtracted from the corresponding historical current average. No cross-dimensional addition or subtraction is performed on the voltage parameter, temperature parameter, and current parameter.

[0036] This yields the staged deviation feature vector: ; In the formula, Indicates the first This is a staged deviation feature vector for each battery cell. The first digit being 0 indicates that the operating stage identifier code is not included in the battery parameter deviation calculation. The other components represent the differences between the current voltage parameter, current temperature parameter, and current current parameter and the historical average value for the same stage. A larger current voltage parameter and a smaller historical voltage average value correspond to a larger voltage deviation; a larger current temperature parameter and a smaller historical temperature average value correspond to a larger temperature deviation; and a larger current parameter and a smaller historical current average value correspond to a larger current deviation. When the current parameter is equal to the corresponding historical average value, the corresponding deviation is zero; when the current parameter is lower than the corresponding historical average value, the corresponding deviation is negative, and subsequent fault classification is based on the direction and degree of deviation. Since each difference component is obtained by subtracting the historical average value from the current parameter in the same unit, there is no issue of inconsistent unit dimensions.

[0037] After obtaining the staged deviation feature vector, the central processing unit concatenates the stage-related battery feature vector with the staged deviation feature vector to obtain the enhanced stage-related battery feature vector. ; In the formula, Indicates the first The enhanced stage associated battery feature vector of each battery cell; Indicates the first Stage-related battery feature vectors of individual battery cells; Indicates the first The staged deviation feature vector of each battery cell. The enhanced stage-related battery feature vector retains both the original battery parameters within the current operating stage and the deviation data relative to the historical steady state of the same stage, which is used for subsequent staged fault determination threshold comparison and fault classification processing.

[0038] The beneficial effects of this step are as follows: By adjusting the acquisition frequency through the operation stage identifier, the acquisition density of voltage, temperature, and current parameters is adapted to the electrical change rate of different operation stages, reducing redundant sampling in the stable stage and improving the anomaly detection capability in the disturbance stage; through the time window alignment mechanism, battery parameters are limited to a single operation stage for vectorization processing, avoiding feature contamination caused by the mixture of load characteristics, mechanical vibration, and thermal boundary conditions in different stages; by generating stage steady-state reference values ​​for the same battery cell and the same operation stage through the stage historical dataset, and using the difference between the current parameters and the historical steady state to form stage-based deviation features, the background differences caused by battery cell aging, cycle number differences, and long-term parameter drift can be weakened, making the abnormal changes that truly deviate from the steady state of the same stage more prominent; furthermore, the original stage-related battery feature vector is merged with the stage-based deviation feature vector, which not only retains the current real operating parameters, but also provides dynamic deviations relative to the historical state, providing a more stable data foundation for subsequent stage-based fault judgment thresholds and fault classification results.

[0039] S3. Obtain the basic judgment baseline corresponding to the rated operating parameters of the battery cell, identify the mechanical vibration state, temperature gradient state and current impact state based on the operation stage identifier, generate environmental disturbance weights, and expand or tighten the boundary of the basic judgment baseline based on the environmental disturbance weights to obtain the staged fault judgment threshold; generate fault classification results based on the degree of deviation between the enhanced stage associated battery feature vector and the staged fault judgment threshold. In this embodiment, step S3 includes the following process: The voltage, temperature, and current baselines in the basic judgment baselines are analyzed to obtain a multi-dimensional basic baseline set. The mechanical vibration state, temperature gradient state, and current impact state of the mobile battery platform under the operation stage identifier are obtained. The mechanical vibration state, temperature gradient state, and current impact state are normalized according to historical reference values ​​under the current operation stage, and the normalized results are combined to obtain the environmental disturbance weight. Based on the environmental disturbance weight, the boundaries of each dimension of the baselines in the multi-dimensional basic baseline set are expanded or tightened to obtain the staged fault judgment threshold. Tightening refers to reducing the boundary adjustment value relative to the expanded staged fault judgment threshold under the previous high disturbance state when the environmental disturbance weight decreases, thus reducing the staged fault judgment threshold. The baseline for determining the value is recovered; each component in the enhanced stage-related battery feature vector corresponding to the stage-related battery feature vector is compared with the corresponding component in the stage-based fault determination threshold to obtain a set of component deviations; based on the stage-based deviation features, the component deviations in the set of component deviations corresponding to the sampling time and the corresponding parameter dimension are weighted and adjusted, and the weighted set of component deviations is weighted and aggregated according to the environmental disturbance weight to obtain a comprehensive deviation index; the comprehensive deviation index is compared with a preset classification threshold to determine the corresponding fault level, and the fault level is bound to the cell number of the corresponding battery cell to obtain the fault classification result; the severity of the third-level fault is higher than that of the second-level fault, and the severity of the second-level fault is higher than that of the first-level fault.

[0040] Specifically, the baseline determination is obtained from the battery management system's built-in factory database, battery cell specifications, or type test reports. The battery management system reads the corresponding rated operating parameters based on the battery cell serial number and parses the voltage, temperature, and current baselines from these parameters. The voltage baseline includes the lower and upper limits of the allowable voltage; the temperature baseline includes the lower and upper limits of the allowable temperature; and the current baseline includes the lower and upper limits of the allowable current. During parsing, the battery management system reads the boundary values ​​according to the parameter name and unit field and writes each boundary value into a multi-dimensional baseline set. The multi-dimensional baseline set refers to the benchmark boundary data composed of the upper and lower limits of the three parameter dimensions: voltage, temperature, and current, used as the basis for subsequent phased adjustments.

[0041] The multidimensional baseline set is represented as follows: ; In the formula, Represents a multidimensional set of basic baselines; Indicates the lower limit of the allowable voltage in the voltage baseline; Indicates the upper limit of the allowable voltage in the voltage baseline; Indicates the allowable lower temperature limit within the temperature baseline; Indicates the upper limit of the allowable temperature in the temperature baseline; Indicates the lower limit of the allowable current in the current baseline; This represents the upper limit of the allowable current in the current baseline. Voltage boundaries, temperature boundaries, and current boundaries each have the same unit, and no cross-dimensional addition or subtraction is performed.

[0042] Mechanical vibration status is acquired using a triaxial microelectromechanical system (MEMS) accelerometer mounted at the bottom of the battery pack. The triaxial accelerometer outputs acceleration samples in the lateral, longitudinal, and vertical directions. The central processing unit reads these triaxial acceleration samples within a preset statistical time window and calculates the root mean square (RMS) of the composite triaxial acceleration values, which is then used as the mechanical vibration status. The mechanical vibration status represents the vibration intensity experienced by the mobile battery platform during the current operational phase.

[0043] The state of mechanical vibration is represented as: ; In the formula, Indicates the state of mechanical vibration; This indicates the total number of acceleration sampling points within the preset state statistical time window; Indicates the sequence number of the acceleration sampling point; Indicates the first The lateral acceleration value at each acceleration sampling point; Indicates the first The longitudinal acceleration value at each acceleration sampling point; Indicates the first The vertical acceleration values ​​at each acceleration sampling point. The units for lateral, longitudinal, and vertical acceleration values ​​are consistent, therefore the synthesized mechanical vibration state has acceleration units. The greater the absolute amplitude of the lateral, longitudinal, or vertical acceleration values, the stronger the mechanical vibration state; the more strong vibration sampling points within the preset state statistical time window, the stronger the mechanical vibration state.

[0044] The temperature gradient state is obtained by measuring the temperature values ​​at multiple points inside the battery pack. The battery management system reads temperature sensor data located at different spatial positions inside the battery pack and acquires the temperature values ​​of each measuring point at the same statistical moment. The difference between the highest and lowest temperature values ​​is used as the temperature gradient state. The temperature gradient state is used to characterize the degree of uneven heat distribution inside the battery pack.

[0045] The temperature gradient state is represented as: ; In the formula, Indicates the temperature gradient state; This represents the highest temperature value among multiple temperature measurement points at the same statistical time. This represents the lowest temperature value among multiple temperature measurement points at the same statistical moment. The larger the highest temperature value and the smaller the lowest temperature value, the greater the temperature gradient, indicating a more significant temperature difference inside the battery pack.

[0046] The current surge state is calculated from the current sampling data of the battery management system. The central processing unit reads the total current sampling sequence within a preset state statistical time window, calculates the absolute value of the current change rate between adjacent sampling points, and counts the number of times the absolute value of the current change rate reaches a preset current change rate threshold. This count is used as the current surge state. The current surge state characterizes the frequency of current surges caused by load switching, charging curve climbs, or mechanical disturbances.

[0047] The current surge state is represented as: ; In the formula, Indicates a current surge condition; Indicates the total number of current sampling points within the preset state statistical time window; Indicates the current sampling point number; Indicates the first The total current value at each current sampling point; Indicates the first The total current value at each current sampling point; This indicates the sampling time interval between two adjacent current sampling points; Indicates the preset current change rate threshold; This represents a decision function; it takes a value of 1 when the condition within the parentheses is true, and a value of 0 when the condition within the parentheses is false. The more drastic the current change between adjacent sampling points, the more times the preset current change rate threshold is reached, and the greater the current surge.

[0048] Environmental disturbance weights are dimensionless weights used to characterize the impact of mechanical vibration, temperature gradient, and current surge on battery parameter determination during the current operating phase. The central processing unit (CPU) retrieves the historical reference values ​​for the mechanical vibration state, temperature gradient state, and current surge state for the same operating phase from the phase's historical dataset, based on the current operating phase identifier. These historical reference values ​​are obtained by statistically analyzing preceding operational data for the same operating phase, selecting the historical peak value or historical stability upper limit for each state parameter within the same operating phase, and writing them into the phase's environmental reference table. Subsequently, the CPU normalizes the current mechanical vibration state, current temperature gradient state, and current current surge state by dividing them by their corresponding historical reference values; if any normalization result exceeds one, it is truncated to one. The average of the three normalized results yields the environmental disturbance weight.

[0049] The environmental disturbance weights are expressed as follows: ; In the formula, Indicates the weight of environmental disturbances; Indicates the state of mechanical vibration; This represents the historical reference value of the mechanical vibration state during the current operation phase; Indicates the temperature gradient state; This represents a historical reference value indicating the temperature gradient state during the current operation phase. Indicates a current surge condition; This represents the historical reference value for the current impact state under the current operating phase. The environmental disturbance weight is dimensionless. The greater the mechanical vibration, temperature gradient, or current impact, the greater the environmental disturbance weight; correspondingly, the larger the historical reference value, the smaller the environmental disturbance weight when the current state remains unchanged. A larger environmental disturbance weight indicates a stronger transient disturbance to the battery parameters under the current operating conditions.

[0050] When expanding or tightening the boundaries of a multidimensional baseline set based on environmental disturbance weights, the central processing unit first calculates the adjustment values ​​of each dimension's boundaries according to the environmental disturbance weights and expansion coefficients. The expansion coefficients are stored in a parameter configuration table and are calibrated using historical operating data containing known fault labels and different disturbance states. During calibration, candidate expansion coefficients are iterated, and the fault detection rate and false alarm rate are calculated separately. The candidate expansion coefficient that achieves a comprehensive balance between the detection rate and the false alarm rate is selected as the final expansion coefficient. The voltage boundary adjustment value, temperature boundary adjustment value, and current boundary adjustment value are obtained by multiplying the absolute value of the baseline boundary in the same dimension by the environmental disturbance weight and the expansion coefficient, respectively.

[0051] The boundary adjustment values ​​for each dimension are expressed as follows: ; ; ; In the formula, This indicates the lower boundary adjustment value of the voltage; This indicates the upper limit adjustment value of the voltage; Indicates the boundary adjustment value at temperature; This indicates the upper limit adjustment value for temperature; This indicates the lower boundary adjustment value of the current; This indicates the upper boundary adjustment value of the current; Indicates the expansion coefficient; This represents the environmental disturbance weight; the meanings of the other symbols are the same as described above. The larger the environmental disturbance weight, the larger the adjustment value of each boundary; the larger the expansion coefficient, the larger the adjustment value of each boundary; the larger the absolute value of the corresponding basic boundary, the larger the adjustment value of that boundary.

[0052] The staged fault determination threshold is expressed as follows: ; In the formula, This represents the staged fault determination threshold; the meanings of each boundary adjustment value are consistent with those described above. When the environmental disturbance weight is large, the boundary adjustment values ​​increase, and the staged fault determination threshold expands outward relative to the basic determination baseline, which can reduce transient false alarms caused by mechanical vibration, temperature gradient, and current surge. When the environmental disturbance weight is small, the boundary adjustment values ​​decrease, and the staged fault determination threshold tightens towards the basic determination baseline, which can maintain sensitivity to minor anomalies under stable operating conditions.

[0053] After obtaining the staged fault determination threshold, the central processing unit extracts each component corresponding to the stage-related battery feature vector from the enhanced stage-related battery feature vector, and compares it with the corresponding boundary in the staged fault determination threshold according to the sampling time and parameter dimension. If a voltage parameter is located between the lower boundary and the upper boundary of the staged voltage, the voltage component deviation at that sampling time is zero; if it is below the lower boundary of the staged voltage, the voltage component deviation is calculated as the ratio of the magnitude below the lower boundary to the voltage boundary span; if it is above the upper boundary of the staged voltage, the voltage component deviation is calculated as the ratio of the magnitude above the upper boundary to the voltage boundary span. The calculation methods for temperature component deviation and current component deviation are consistent with those for voltage component deviation, both using the boundary span within the same dimension for dimensionless processing.

[0054] The deviations of the voltage component, temperature component, and current component are expressed as follows: ; ; ; In the formula, Indicates the first The first battery cell in the... The deviation of the voltage component at each sampling time; Indicates the first The first battery cell in the... Temperature component deviation at each sampling time; Indicates the first The first battery cell in the... Deviation of current component at each sampling time; Indicates the lower boundary of the staged voltage; Indicates the upper boundary of the staged voltage; Indicates the boundary at the staged temperature; Indicates the upper boundary of the staged temperature; Indicates the lower boundary of the staged current; Indicates the upper boundary of the staged current; , and They represent the first The first battery cell in the... The voltage, temperature, and current parameters at each sampling time are recorded. The greater the magnitude of the parameter exceeding the corresponding stage boundary, the greater the deviation of the corresponding component; when the parameter is within the corresponding stage boundary, the deviation of the corresponding component is zero.

[0055] After obtaining the set of component deviations, the central processing unit (CPU) adjusts the weights of the set of component deviations based on the staged deviation characteristics. The staged deviation characteristics were obtained in step S2 from the difference between the current battery parameters and the historical steady-state reference values ​​of the same battery cell during the same operating stage. For the same sampling time and the same parameter dimension, the CPU normalizes the staged deviation characteristics according to the staged boundary span of the corresponding dimension and generates a deviation amplification coefficient. The larger the deviation amplification coefficient, the more significant the deviation of the component not only exceeds the staged boundary but also deviates more significantly from its own historical steady-state value. The deviation amplification coefficient is controlled by a preset deviation adjustment coefficient, which is calibrated using historical operating data with known fault labels.

[0056] The voltage deviation amplification factor, temperature deviation amplification factor, and current deviation amplification factor are expressed as follows: ; ; ; In the formula, Indicates the first The first battery cell in the... Voltage deviation amplification factor at each sampling time; Indicates the first The first battery cell in the... Temperature deviation amplification factor at each sampling time; Indicates the first The first battery cell in the... Current deviation amplification factor at each sampling time; This indicates the preset deviation adjustment coefficient; , and They represent the first The first battery cell in the current operating phase The historical average voltage, historical average temperature, and historical average current at each sampling location. The larger the absolute value of the difference between the current parameter and the corresponding historical average, the larger the corresponding deviation amplification factor; the larger the preset deviation adjustment factor, the stronger the amplification effect of the staged deviation characteristics on the component deviation.

[0057] The comprehensive deviation index is obtained by weighted aggregation of the component deviations of each sampling time and parameter dimension for the same battery cell within the current operating time window. The central processing unit expands the environmental disturbance weights corresponding to each sampling time into a weight sequence aligned with the battery parameter sampling times; when the environmental disturbance weights are updated only at a low frequency, the environmental disturbance weight closest to that sampling time is used as the environmental disturbance weight for that sampling time. The environmental disturbance weights participate in the comprehensive deviation index calculation as a suppression factor, reducing the contribution of high disturbance times to the component deviations and retaining the contribution of low disturbance times to the component deviations. The voltage dimension weights, temperature dimension weights, and current dimension weights are read from a weight configuration table, which is determined through training or grid traversal using historical operating data containing known fault labels, ensuring that the classification accuracy of different fault levels meets preset requirements.

[0058] The comprehensive deviation index is expressed as follows: ; In the formula, Indicates the first The overall deviation index of each battery cell; This indicates the total number of sampling points within the current operation phase time window; Indicates the sampling time sequence number; Indicates the relationship with the first Environmental disturbance weights aligned to each sampling time; Indicates the voltage dimension weight; Indicates the weight of the temperature dimension; The symbol represents the weight of the current dimension; the meanings of the other symbols are the same as described above. The larger the deviation of the voltage component, temperature component, or current component, the larger the overall deviation index; the larger the corresponding deviation amplification factor, the larger the overall deviation index; the larger the corresponding dimension weight, the greater the contribution of that dimension deviation to the overall deviation index; the larger the environmental disturbance weight, the smaller the contribution of the same component deviation to the overall deviation index. The overall deviation index is a dimensionless index.

[0059] The fault classification result is obtained by comparing the comprehensive deviation index with a preset classification threshold. The preset classification threshold includes a first classification threshold and a second classification threshold, with the first classification threshold being lower than the second classification threshold. The first and second classification thresholds are calibrated using historical operational data containing known fault tags. During calibration, candidate threshold combinations are traversed, and the identification results for Level 1, Level 2, and Level 3 faults under different threshold combinations are calculated. The threshold combination that achieves a balance between detection rate and false alarm rate is selected as the preset classification threshold. The central processing unit determines the fault level according to the magnitude of the comprehensive deviation index and binds the fault level to the corresponding battery cell number to form the fault classification result. The fault classification result includes at least the cell number, fault level, comprehensive deviation index, and current operation stage identifier.

[0060] The fault level determination rule is expressed as follows: ; In the formula, Indicates the first Fault level of each individual battery cell; This indicates a Level 1 fault; Indicates a level 2 fault; This indicates a level three fault; Indicates the first-level threshold; This represents the second-level threshold. The larger the overall deviation index, the higher the corresponding fault level; the smaller the first-level threshold, the easier it is for minor deviations to be identified as level one faults; the smaller the second-level threshold, the easier it is for severe deviations to be identified as level three faults.

[0061] The beneficial effects of this step are as follows: By generating environmental disturbance weights based on mechanical vibration, temperature gradient, and current surge states, and dynamically adjusting the fault judgment boundaries for voltage, temperature, and current using these weights, different judgment scales can be applied to the same battery cell at different operational stages. During periods of high disturbance, such as land towing and pushback transitions, the staged fault judgment threshold is appropriately expanded to reduce false alarms caused by vibration noise and transient current surges. During more stable periods, such as shipboard power supply and docking, the boundaries are tightened towards the basic judgment baseline, maintaining sensitivity to early anomalies. Simultaneously, by adjusting the weights of component deviations using staged deviation characteristics, abnormal jumps in current parameters relative to their historical steady state can be highlighted, reducing repeated misjudgments caused by aging accumulation offsets. This ensures that the fault classification results simultaneously reflect both "whether it has exceeded the boundary" and "whether it has deviated from the historical state of the same stage."

[0062] S4. Based on the fault classification result and the operation stage identifier, filter the disposal actions from the disposal action list, and generate a graded disposal instruction set by combining the disposal capabilities of the reachable execution end; In this embodiment, step S4 includes the following process: Identify the execution terminals that establish communication links with the mobile battery platform during the current operation phase, and aggregate the identified execution terminals to obtain a set of reachable execution terminals; based on the highest fault level in the fault classification result and the operation phase identifier, match the corresponding disposal actions from the disposal action list to obtain a set of disposal actions; obtain the set of capabilities required for each disposal action in the disposal action set, and obtain the set of disposal capability tags for each execution terminal in the set of reachable execution terminals; when the set of disposal capability tags covers the set of capabilities required for the corresponding disposal action, assign the disposal action to the corresponding execution terminal, and generate a graded disposal instruction containing the target execution terminal identifier, action code, and action parameters; aggregate all graded disposal instructions to obtain the graded disposal instruction set.

[0063] Specifically, the central processing unit of the mobile battery platform maintains an execution-end communication status table. This table records the network address, device identifier, real-time online flag, and most recent heartbeat response time for the ship's battery compartment push-pull mechanism, docking platform, shore-based electric tractor, and shore-side charging station. The central processing unit sends heartbeat request data packets to registered execution ends via a wireless LAN (IEEE 802.11n standard). The heartbeat request data packet includes the mobile battery platform identifier, execution end identifier, and an incrementing sequence number. Upon receiving the heartbeat request data packet, the execution end returns a heartbeat response data packet. If the central processing unit receives a heartbeat response data packet within a preset heartbeat timeout window, it sets the online flag of the corresponding execution end to valid; otherwise, it sets the online flag to invalid.

[0064] The reachable execution end set refers to the set of execution ends that have established a valid communication link with the mobile battery platform and are capable of receiving tiered handling instructions during the current operation phase. The central processing unit reads the online flags of valid execution ends based on the execution end communication status table and excludes execution ends that lack physical reachability or safe execution conditions during the current operation phase, based on the current operation phase identifier. For example, during the land towing phase, the mobile battery platform, which has detached from the ship's battery compartment push-pull mechanism, will no longer be assigned to traction actions by the push-pull mechanism; during the charging operation phase, shore-based charging stations and shore-based electric tractors in standby status are prioritized as reachable execution ends. The central processing unit aggregates the identifiers of execution ends that meet the conditions of valid communication and phase reachability to obtain the reachable execution end set.

[0065] The action list refers to a structured action mapping table pre-stored in the central processing unit of the mobile battery platform. The action list includes at least the following fields: fault level, applicable work stage, action name, action code, required capability set, and default action parameters. The central processing unit reads the fault level of each battery cell from the fault classification results and determines the highest fault level in order of priority: Level 3 faults are higher than Level 2 faults, and Level 2 faults are higher than Level 1 faults. The highest fault level represents the highest risk state that the mobile battery platform needs to address first. Subsequently, the central processing unit uses the highest fault level and the current work stage identifier as joint search criteria to find records in the action list that match both the fault level and the applicable work stage, and reads the corresponding action name and action code to form a set of action actions.

[0066] The required capability set for an action refers to the set of capability types that the target execution end must possess to perform a certain action. The action capability tag set refers to the set of action function types that an execution end can independently complete. The action capability tag set for a ship's battery compartment push-pull mechanism is reported by its control unit during equipment registration and must include at least push-pull and mechanical locking functions; the action capability tag set for a docking platform is reported by its signal processing unit and must include at least load-bearing locking and area blocking functions; the action capability tag set for a shore-based electric tractor unit is reported by its onboard control unit and must include at least traction, transfer, and local early warning functions; the action capability tag set for a shore-side charging station is reported by its charging controller and must include at least power regulation, electrical isolation, and local early warning functions. The central processing unit writes the action capability tag sets reported by each execution end into the execution end capability table and calls them up in real time during action allocation.

[0067] During action allocation, the central processing unit (CPU) reads each action in the action set and the required capability set for that action from the action list. Simultaneously, it reads the capability tag set for each execution terminal in the reachable execution terminal set. If the capability tag set of an execution terminal covers all the required capabilities for that action, the action is determined to be available for allocation to that terminal. If multiple execution terminals meet the capability coverage criteria, the CPU prioritizes the execution terminal that is spatially closer to the current work phase, has lower communication latency, or better matches the urgency of the action. When multi-terminal collaboration is allowed for an action, the CPU simultaneously allocates the action to multiple eligible execution terminals. If no reachable execution terminal meets the capability coverage criteria, the CPU records an action allocation failure event and reports it to the monitoring center, while continuing to generate other action instructions that meet the allocation criteria.

[0068] Tiered handling instructions are structured control instructions generated by the central processing unit and sent to the target execution end. Each tiered handling instruction includes a target execution end identifier, an action code, and action parameters. Action parameters are derived from the fault classification results, the current operation stage identifier, and the default action parameters from the handling action list. These parameters include at least the fault level, the faulty battery cell number, the comprehensive deviation index, the operation stage identifier, and the power rollback parameters, traction target area parameters, or early warning parameters corresponding to the handling action. The central processing unit generates a tiered handling instruction for each set of handling actions and execution ends that meet the capability coverage conditions, and sorts these instructions according to their urgency. The urgency is determined by the action code; emergency towing actions are prioritized over power reduction actions, and power reduction actions are prioritized over early warning actions. All sorted tiered handling instructions together constitute the tiered handling instruction set.

[0069] After the tiered handling instruction set is generated, the central processing unit writes it into the transmission buffer. The transmission buffer records the instruction generation timestamp, target execution terminal identifier, action code, action parameters, and transmission retry count for each tiered handling instruction. Subsequent step S5 distributes the instructions according to the target execution terminal identifier and action urgency level in the tiered handling instruction set.

[0070] The beneficial effects of this step are as follows: by filtering out execution terminals that cannot communicate or do not have the physical execution conditions under the current operation stage through the set of reachable execution terminals, the disposal actions are avoided from being assigned to non-executable equipment; by matching the coverage relationship between the set of required capabilities for the action and the set of disposal capability tags, it can be ensured that actions such as emergency towing, power reduction operation, area blockade, and early warning are assigned to execution terminals with corresponding capabilities; by jointly filtering disposal actions by the highest fault level and the operation stage identifier, the same fault level can correspond to different disposal strategies under different operation stages, thereby transforming the fault classification result obtained in step S3 into executable, distributable, and stage-adaptive collaborative disposal instructions.

[0071] S5. Distribute the graded handling instruction set to the corresponding execution terminal, trigger the corresponding execution terminal to perform collaborative handling actions, and complete the graded fault handling of the mobile battery platform; In this embodiment, step S5 includes the following process: When the operation phase identifier indicates that the mobile battery platform is in the land towing phase, a tiered handling instruction with the target execution end being the shore-based electric tractor and the action code corresponding to emergency towing is extracted from the tiered handling instruction set as a priority transfer instruction, and the priority transfer instruction is issued to the shore-based electric tractor; real-time position data and battery parameter change data of the mobile battery platform are collected during the towing process of the shore-based electric tractor, and the real-time position data is transmitted back to the ship's battery compartment push-pull mechanism and the docking platform, so that the ship's battery compartment push-pull mechanism and the docking platform enter a standby locked state; it is determined whether the real-time position data indicates that the mobile battery platform has arrived at the safety isolation zone, and when it arrives, an entry blockade instruction is sent to the shore charging station based on the evolution trend of the battery parameter change data; when The operation phase identifier indicates that the mobile battery platform is in the charging operation phase. A tiered handling instruction with the target execution end being the shore charging station and the action code corresponding to reduced power operation is extracted from the tiered handling instruction set as a power rollback instruction, and this power rollback instruction is sent to the shore charging station. Battery parameter change data of the mobile battery platform during the power rollback process at the shore charging station is collected, and the battery parameter change data is compared with the execution feedback of the power rollback instruction to obtain a rollback closed-loop response result. When the rollback closed-loop response result indicates that the battery parameter change data has not returned to within the tiered fault judgment threshold, an emergency towing instruction is sent to the shore-based electric tractor, triggering the shore-based electric tractor to tow the mobile battery platform from the shore charging station to a safe isolation zone.

[0072] Specifically, the communication scheduler of the central processing unit of the mobile battery platform polls the transmission buffer according to the sorting result of the hierarchical handling instruction set. The communication scheduler reads the target execution terminal identifier from each hierarchical handling instruction and retrieves the network address of the target execution terminal from the execution terminal communication status table. The communication scheduler sends instruction messages to the target execution terminal via User Datagram Protocol (UDP). The instruction messages are in JSON string format. The instruction message includes at least an action code, action parameters, target execution terminal identifier, instruction generation timestamp, and verification field. After receiving the instruction message, the target execution terminal parses the action code and action parameters and returns a response message containing a receipt code. When the central processing unit receives a receipt code indicating acceptance confirmation or commencement of execution within a preset response waiting time, it records that the instruction distribution was successful; if no valid receipt code is received, it performs retransmission according to the retry count. If the communication fails after reaching the preset number of retries, it records a communication failure event and reports it to the monitoring center.

[0073] When the operation phase indicator signifies that the mobile battery platform is in the land towing phase, the central processing unit extracts the tiered handling instruction from the tiered handling instruction set, with the target execution end being the shore-based electric tractor and the action code corresponding to emergency towing. This is designated as the priority transfer instruction and issued to the shore-based electric tractor. The priority transfer instruction is a high-priority control instruction used during the land towing phase to instruct the shore-based electric tractor to tow the mobile battery platform to a safe isolation zone. The action parameters of the priority transfer instruction include the fault level, the faulty battery cell number, the current location of the mobile battery platform, and the safe isolation zone indicator. After receiving the priority transfer instruction, the onboard control unit of the shore-based electric tractor verifies the validity of the traction lock signal. If the traction lock signal is valid, the onboard traction control program is initiated, and the mobile battery platform is towed along a preset safe passage. If the traction lock signal is invalid, a mechanical lock confirmation is performed first, and then the traction control program is initiated.

[0074] During shore-based electric tugboat towing, the mobile battery platform continuously collects real-time location data and battery parameter change data. Real-time location data is acquired via ultra-wideband positioning tags, Differential Global Positioning System (DGPS), or port base station positioning devices. Battery parameter change data is continuously collected by the battery management system according to a phased acquisition strategy for the current operational stage, including voltage, temperature, and current parameters of the faulty battery cell and its associated battery cells. The central processing unit transmits the real-time location data back to the ship's battery compartment push-pull mechanism and the docking platform, putting these mechanisms into a standby locked state. The standby locked state means that the relevant equipment ceases normal scheduling operations and locks its current mechanical position, retaining only the higher-priority emergency unlocking interface. When the ship's battery compartment push-pull mechanism enters the standby locked state, the control unit switches the servo drive to position holding mode; when the docking platform enters the standby locked state, the signal processing unit locks the load-bearing locking mechanism or hydraulic holding mechanism.

[0075] The central processing unit (CPU) determines whether the mobile battery platform has arrived at the safety isolation zone. The safety isolation zone is a pre-defined, dedicated isolation area within the port area, stored in the CPU as a polygonal geofence. The CPU reads the two-dimensional coordinates from the real-time location data and uses a ray-mapping method (where the point lies within the polygon) to determine if these coordinates fall within the polygonal geofence of the safety isolation zone. The ray-mapping method is an existing geometric determination method. Specifically, it involves drawing a ray from the point to be determined along a fixed direction and counting the number of intersections between the ray and the polygon boundary. If the number of intersections is odd, the point is considered to be inside the polygon; if the number is even, the point is considered to be outside the polygon. When a predetermined number of consecutive real-time location data points fall within the safety isolation zone, the CPU confirms that the mobile battery platform has arrived.

[0076] Upon confirmation of arrival at the safety isolation zone, the central processing unit determines whether to send an entry blockade command to the shore charging station based on the evolution trend of battery parameter change data. The evolution trend is obtained by linearly fitting the highest temperature parameter within a preset trend analysis time window before and after arrival at the safety isolation zone. The linear fitting uses the existing least squares straight line slope formula to calculate the slope of the highest temperature parameter changing with time, which is taken as the temperature evolution slope.

[0077] The slope of temperature evolution is expressed as: ; In the formula, Indicates the slope of temperature evolution; This indicates the total number of highest temperature sampling points within the preset trend analysis time window; Indicates the serial number of the highest temperature sampling point; Indicates the first The sampling time corresponding to the highest temperature sampling point; Indicates the first The highest temperature parameter corresponding to each highest temperature sampling point; This represents the mean value at each sampling time within the preset trend analysis time window; This represents the average of the highest temperature parameters within the preset trend analysis time window. The faster the highest temperature parameter rises over time, the greater the temperature evolution slope; when the highest temperature parameter remains stable, the temperature evolution slope is close to zero; when the highest temperature parameter decreases over time, the temperature evolution slope is negative. When the temperature evolution slope exceeds the preset temperature rise slope threshold, the central processing unit determines that the fault thermal risk has not been eliminated and sends an entry blockade command to the shore charging station. The action parameters of the entry blockade command include the safety isolation zone identifier, the blockade range identifier, and the blockade reason code. After receiving the entry blockade command, the shore charging station locks the electronic lock of the corresponding charging connection channel, closes the channel entrance gate, and broadcasts a channel occupancy signal to the port area dispatch system to prevent other mobile battery platforms from entering the charging connection area surrounding the safety isolation zone.

[0078] When the operation phase identifier indicates that the mobile battery platform is in the charging operation phase, the central processing unit extracts the graded handling instruction from the graded handling instruction set, which has the target execution end as a shore charging station and the action code corresponding to power reduction operation. This is used as a power backoff instruction, which is then sent to the shore charging station. The power backoff instruction is a control instruction used to instruct the shore charging station to reduce the charging output power or cut off the charging output. The action parameters of the power backoff instruction include the fault level, target holding power, backoff rate, faulty battery cell number, and execution timeout condition. The target holding power and backoff rate are read from the default action parameters in the handling action list and determined according to the fault level; the higher the fault level, the lower the target holding power and the greater the backoff rate. After receiving the power backoff instruction, the charging controller of the shore charging station adjusts the power instruction value according to the action parameters and sends the actual output power, current setpoint, and contactor status back to the mobile battery platform as execution feedback.

[0079] The mobile battery platform continuously collects battery parameter change data during the power back-off process at the shore charging station and compares this data with the feedback from the power back-off command to obtain the back-off closed-loop response result. The comparison process includes two aspects: First, the central processing unit compares the actual output power fed back by the shore charging station with the target power in the power back-off command. If the difference is within the instrument's allowable error range, the power execution is considered consistent. Second, the central processing unit compares the total charging current in the battery parameter change data with the measured current fed back by the shore charging station. If the difference is within the instrument's allowable error range, the current response is considered consistent. The instrument's allowable error range is read from the calibration certificates of the charging station's metering module and the battery management system's current sampling module. If both the power execution and current response are consistent, the shore charging station is determined to have truly executed the power back-off command; otherwise, the back-off closed-loop response result is marked as an execution anomaly.

[0080] After confirming that the power rollback command has been executed, the central processing unit continues to compare the voltage, temperature, and current parameters of the faulty battery cell with the staged fault judgment threshold obtained in step S3, and recalculates the comprehensive deviation index. If the recalculated comprehensive deviation index is lower than the first-level threshold, and the voltage, temperature, and current parameters of the faulty battery cell have all returned to within the staged fault judgment threshold, the rollback closed-loop response result is marked as normal, and the shore charging station continues to operate and monitor according to the target power. If the recalculated comprehensive deviation index still reaches the first-level threshold, or any battery parameter of the faulty battery cell has not returned to within the staged fault judgment threshold, the rollback closed-loop response result is marked as not returned. When the rollback closed-loop response result is marked as not returned, the central processing unit generates an emergency towing command and sends it to the shore-based electric tractor, triggering the shore-based electric tractor to tow the mobile battery platform from the shore charging station to the safety isolation zone. The action parameters of the emergency towing command include the current location of the mobile battery platform, the target safety isolation zone identifier, the fault level, and the faulty battery cell number. After receiving the emergency towing command, the shore-based electric tractor completes mechanical docking according to the current location of the mobile battery platform and tows away along the preset safety channel.

[0081] The beneficial effects of this step are as follows: By distributing the tiered handling instruction set according to the target execution end, and combining it with the receipt, retransmission, and failure reporting mechanisms, it can ensure that the handling actions have clear execution targets and communication loops; during the land towing phase, the priority transfer instruction enables the shore-based electric tractor to directly undertake the physical isolation task, and triggers the push-pull mechanism and the docking platform to enter the standby lock state through real-time location data, which can avoid the faulty platform from generating new mechanical conflicts with other equipment during the transfer process; after arriving at the safe isolation zone, it can determine whether to block the entrance channel based on the temperature evolution slope, and decide whether to expand the isolation range based on whether the thermal risk continues to rise; during the charging operation phase, the power backoff instruction first reduces or cuts off the charging energy input, and then forms a backoff closed-loop response result based on the execution feedback and battery parameter change data. If the parameters do not return, an emergency towing is triggered, thus forming a collaborative handling closed loop of "electrical load reduction - parameter verification - physical isolation", reducing the risk that simple alarms or single power outage measures cannot cover the continuously deteriorating fault.

[0082] The foregoing has shown and described the basic principles, main features, and advantages of this application. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this application. Various changes and modifications can be made to this application without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of this application as claimed. The scope of protection of this application is defined by the appended claims and their equivalents.

Claims

1. A method for graded fault handling of marine batteries, characterized in that, The method includes: The interface connection status signal between the mobile battery platform and the ship's battery compartment push-pull mechanism, docking platform, shore-based electric tractor and shore charging station is obtained, and the position positioning signal of the mobile battery platform is obtained. The interface connection status signal and the position positioning signal are jointly analyzed to obtain the operation stage identifier. The voltage, temperature, and current parameters of individual battery cells within the mobile battery platform are collected and associated with the operation stage identifier to generate a stage-associated battery feature vector. According to the operation stage identifier, the preceding voltage parameters, temperature parameters and current parameters are recorded to construct a stage history dataset. Based on the stage history dataset, a stage steady-state reference value is generated. The difference between the stage-related battery feature vector and the stage steady-state reference value is calculated and merged to obtain an enhanced stage-related battery feature vector. Obtain a basic judgment baseline corresponding to the rated operating parameters of the battery cell, identify the mechanical vibration state, temperature gradient state and current impact state based on the operation stage identifier, generate environmental disturbance weights, and expand or tighten the boundary of the basic judgment baseline based on the environmental disturbance weights to obtain the staged fault judgment threshold. Based on the degree of deviation between the enhanced stage-related battery feature vector and the staged fault determination threshold, a fault classification result is generated. Based on the fault classification results and the operation stage identifier, the disposal actions are selected from the disposal action list, and a graded disposal instruction set is generated by combining the disposal capabilities of the reachable execution end, and distributed to the corresponding execution end to perform collaborative disposal actions.

2. The marine battery fault classification and handling method according to claim 1, characterized in that, The process of obtaining the work stage identifier specifically includes: The push-pull stroke feedback signal of the ship's battery compartment push-pull mechanism, the load-bearing positioning signal of the docking platform, the traction locking signal of the shore-based electric tractor, and the electrical docking signal of the shore charging station are acquired and aggregated into a multi-source interface status set. The trigger times of each signal in the multi-source interface state set are sorted, and the physical migration trajectory of the mobile battery platform is identified based on the sorting result. The physical migration trajectory is cross-validated with the location positioning signal, and when the cross-validation passes, the operation stage identifier corresponding to the physical migration trajectory is output. The operation stage identifier includes at least one of the following: shipboard power supply stage, push-out transition stage, docking and mooring stage, land towing stage, charging operation stage, and pull-back and reset stage.

3. The marine battery fault classification and handling method according to claim 2, characterized in that, The step of cross-validating the physical migration trajectory with the location positioning signal specifically includes: The spatial reference area for the corresponding operation stage is determined based on the physical migration trajectory. Extract multiple location signal sample points within a preset time range from the location positioning signal, and determine whether each location signal sample point falls within the spatial reference area. The spatial fit is obtained by the ratio of the number of location signal sample points falling into the spatial reference area to the total number of location signal sample points; When the spatial matching degree meets the preset threshold, the operation stage corresponding to the physical migration trajectory is used as the operation stage identifier; When the spatial fit does not meet the preset threshold, the previous valid operation stage identifier is maintained, and the cross-validation anomaly event is recorded.

4. The method for graded fault handling of marine batteries according to claim 1, characterized in that, The battery feature vector associated with the generation stage specifically includes: The sampling frequency of the voltage, temperature, and current parameters is adjusted according to the operation stage identifier to obtain a staged sampling strategy; The phased acquisition strategy is executed to synchronously sample each battery cell in the mobile battery platform to obtain the time-series parameter sequence of each battery cell. The time sequence parameter sequence is concatenated with the operation stage identifier and aligned with the time window corresponding to the start and end times of the operation stage identifier to generate the stage-associated battery feature vector.

5. The marine battery fault classification and handling method according to claim 1, characterized in that, The obtained enhanced stage associated battery feature vector specifically includes: According to the operation stage identifier, the voltage parameters, temperature parameters, and current parameters of the mobile battery platform during the preceding operation process are recorded to obtain the stage historical dataset; Extract historical segments with the same identifier as the current operation stage from the historical dataset of the stage, and perform mean statistical processing on the historical segments to obtain the steady-state reference value of the stage; The difference between the stage-related battery feature vector and the stage steady-state reference value is calculated to obtain the stage-specific deviation feature; The staged deviation feature is merged with the stage-related battery feature vector to obtain the enhanced stage-related battery feature vector.

6. The marine battery fault classification and handling method according to claim 1, characterized in that, The obtained staged fault determination threshold specifically includes: By analyzing the voltage baseline, temperature baseline, and current baseline in the aforementioned basic determination baseline, a multidimensional set of basic baselines is obtained; The mechanical vibration state, temperature gradient state, and current impact state of the mobile battery platform under the specified operation stage identifier are obtained. The mechanical vibration state, the temperature gradient state, and the current impact state are normalized according to the historical reference values ​​under the current operation stage, and the results after normalization are combined to obtain the environmental disturbance weight. Based on the environmental disturbance weights, the boundaries of each dimension of the multidimensional basic baseline set are expanded or tightened to obtain the staged fault determination threshold. The tightening refers to reducing the boundary adjustment value relative to the expanded staged fault judgment threshold under the previous high disturbance state when the environmental disturbance weight is reduced, so that the staged fault judgment threshold is recovered to the basic judgment baseline.

7. The marine battery fault classification and handling method according to claim 6, characterized in that, The generation of fault classification results specifically includes: Each component in the enhanced stage-related battery feature vector corresponding to the stage-related battery feature vector is compared with the corresponding component in the stage-based fault determination threshold to obtain a set of component deviations. Based on the staged deviation characteristics, the component deviations at the corresponding sampling time and corresponding parameter dimensions in the component deviation set are weighted and adjusted, and the weighted component deviation set is weighted and aggregated according to the environmental disturbance weight to obtain a comprehensive deviation index. The comprehensive deviation index is compared with a preset grading threshold to determine the corresponding fault level, and the fault level is bound to the cell number of the corresponding battery cell to obtain the fault grading result. The fault levels include level 1 fault, level 2 fault and level 3 fault. The severity of a Level 3 fault is higher than that of a Level 2 fault, and the severity of a Level 2 fault is higher than that of a Level 1 fault.

8. The method for graded fault handling of marine batteries according to claim 1, characterized in that, The generation of the hierarchical processing instruction set specifically includes: Identify the execution terminals that establish communication links with the mobile battery platform during the current operation phase, and aggregate the identified execution terminals to obtain a set of reachable execution terminals; Based on the highest fault level in the fault classification result and the operation stage identifier, the corresponding handling action is matched from the handling action list to obtain the handling action set; Obtain the set of capabilities required for each action in the set of actions to be handled, and obtain the set of handling capability tags for each execution terminal in the set of reachable execution terminals; When the set of handling capability tags covers the set of capabilities required for the corresponding handling action, the handling action is assigned to the corresponding execution end, and a hierarchical handling instruction containing the target execution end identifier, action code and action parameters is generated; The various graded handling instructions are collected to obtain the graded handling instruction set.

9. The method for graded fault handling of marine batteries according to claim 1, characterized in that, The process of distributing data to the corresponding execution terminal to perform collaborative processing actions specifically includes: When the operation phase identifier indicates that the mobile battery platform is in the land towing phase, the graded handling instruction with the target execution end being the shore-based electric tractor and the action code corresponding to emergency towing is extracted from the graded handling instruction set as the priority transfer instruction, and the priority transfer instruction is issued to the shore-based electric tractor. The system collects real-time location data and battery parameter change data of the mobile battery platform during the traction process of the shore-based electric tractor, and transmits the real-time location data back to the ship battery compartment push-pull mechanism and the docking platform, so that the ship battery compartment push-pull mechanism and the docking platform enter a standby locked state. Determine whether the real-time location data indicates that the mobile battery platform has arrived at the safe isolation zone. If it has arrived, send an entry blockade command to the shore charging station based on the evolution trend of the battery parameter change data.

10. The method for graded fault handling of marine batteries according to claim 1, characterized in that, The step of distributing the data to the corresponding execution terminal to perform collaborative processing actions also includes: When the operation phase identifier indicates that the mobile battery platform is in the charging operation phase, the tiered handling instruction with the target execution terminal being the shore charging station and the action code corresponding to power reduction operation is extracted from the tiered handling instruction set as a power rollback instruction, and the power rollback instruction is sent to the shore charging station. The battery parameter change data of the mobile battery platform during the power rollback process of the shore charging station is collected, and the battery parameter change data is compared with the execution feedback of the power rollback command to obtain the rollback closed-loop response result; When the back-off closed-loop response result indicates that the battery parameter change data has not returned to within the staged fault determination threshold, an emergency towing command is sent to the shore-based electric tractor, triggering the shore-based electric tractor to tow the mobile battery platform from the shore charging station to a safe isolation zone.

Images