A graded early warning method and related devices for ship power batteries based on operating conditions

By acquiring ship operating condition parameters and calculating ship operating condition coefficients, and dynamically adjusting early warning judgment parameters, the problem of not being able to distinguish between reversible voltage fluctuations and irreversible degradation in existing technologies is solved. This achieves efficient and safe monitoring of ship power batteries, reduces false alarms and accidental power outages, and is applicable to ocean-going vessels.

CN123120932APending Publication Date: 2026-11-17BEIJING ZERO ONE EVERYTHING INFORMATION TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611217166.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-12
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing ship power battery early warning systems rely solely on BMS electrochemical parameters, which cannot distinguish between reversible voltage fluctuations caused by normal operation and irreversible internal degradation of the battery, leading to frequent false alarms and accidental power outages.

Method used

By acquiring battery operating parameters and ship operating condition parameters from the battery management system, the ship operating condition coefficient is calculated, the early warning judgment parameters are dynamically adjusted, and graded early warning results are output. Combined with the NMEA 2000 communication bus, engine load, speed, and thruster status are acquired, normalized and fused rule calculations are performed, and the early warning judgment benchmark is dynamically adjusted.

Benefits of technology

It effectively distinguishes between normal operating condition fluctuations and actual battery faults, reduces false alarms and accidental power outages, improves the safety monitoring level of the power battery system under all navigation conditions, and is suitable for ocean-going vessel environments where communication is frequently interrupted.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123120932A_ABST
    Figure CN123120932A_ABST
Patent Text Reader

Abstract

This invention provides a graded early warning method and related device for ship power batteries based on operating condition linkage. The method involves acquiring battery operating parameters and ship operating condition parameters from the battery management system; calculating a ship operating condition coefficient based on the ship operating condition parameters; dynamically adjusting early warning judgment parameters based on the ship operating condition coefficient; and outputting graded early warning results for the ship power battery based on the battery operating parameters and the dynamically adjusted early warning judgment parameters. This invention enables the early warning system to adaptively adjust the judgment benchmark according to the ship's current operating state, effectively distinguishing between normal operating condition fluctuations and actual battery faults. Furthermore, the operating condition linkage mechanism of this invention does not rely on historical data training or real-time cloud communication; it can dynamically correct the early warning judgment parameters based solely on locally acquired ship operating condition parameters. This results in fast response speed and high reliability, making it suitable for ocean-going vessel environments with frequent communication interruptions, and improving the safety monitoring level of the power battery system under all navigation conditions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of marine power battery management technology, and in particular to a graded early warning method and related device for marine power batteries based on operating condition linkage. Background Technology

[0002] Currently, the safety monitoring of marine power battery systems mainly relies on the Battery Management System (BMS) for data acquisition and protection control. The BMS monitors physical parameters such as cell voltage, cell temperature, and charging / discharging current in real time, and triggers protective actions such as audible and visual alarms, power limiting operation, or disconnection of the main circuit relay when any parameter exceeds a preset hard threshold. In existing technical solutions, the judgment criteria for early warning triggering are based solely on the electrochemical parameters of the BMS itself, such as voltage and temperature, without considering the real-time operating conditions of the marine power system, nor incorporating any information about the ship's navigation or operational status into the pre-adjustment stage of the early warning logic. This technical deficiency leads to serious problems in actual ship operation scenarios. Under typical operating conditions such as side thruster startup, crane operation, and propeller entry and exit from the water, the power distribution bus will experience current pulses of hundreds or even thousands of amperes within seconds, directly causing a short-term sharp drop in the minimum voltage of individual cells and a sudden increase in the voltage difference between cells. Its external manifestations are very similar to micro-short circuits or severe over-discharge inside the battery. However, since the existing early warning system can only see the voltage and temperature values ​​transmitted by the BMS, and does not know whether these value changes are due to irreversible degradation inside the battery or simply due to normal ship operations such as the side thruster just starting, the only thing it can do is mechanically trigger alarms or trip the brakes according to preset thresholds. Summary of the Invention

[0003] This invention provides a graded early warning method and related device for marine power batteries based on operating condition linkage, which solves the defects of existing marine power battery early warning systems that use BMS electrochemical parameters as input, cannot sense the real-time operating conditions of the ship, and therefore cannot distinguish between reversible voltage fluctuations caused by normal operation and irreversible internal degradation of the battery, and can only mechanically trigger alarms.

[0004] This invention provides a graded early warning method for ship power batteries based on operating condition linkage, comprising: Obtain battery operating parameters and ship operating condition parameters from the battery management system; Calculate the ship's operating condition coefficient based on the aforementioned ship operating condition parameters; The early warning judgment parameters are dynamically adjusted based on the aforementioned ship operating condition coefficients; Based on the battery operating parameters and the dynamically adjusted warning judgment parameters, the graded warning results for the ship's power battery are output.

[0005] According to the ship power battery graded early warning method based on operating condition linkage provided by the present invention, the step of calculating the ship operating condition coefficient based on the ship operating condition parameters includes: Ship operating parameters are acquired in real time from the ship's NMEA 2000 communication bus. These parameters include at least engine load, speed, and thruster operating status. The engine load, speed, and thruster operating status are normalized to map parameters of different dimensions to a unified numerical range. Based on the normalized engine load, the speed, and the thruster operating status, the ship's operating condition coefficient is calculated according to a preset fusion rule.

[0006] According to the ship power battery graded early warning method based on operating condition linkage provided by the present invention, the step of calculating the ship operating condition coefficient according to a preset fusion rule based on the normalized engine load, the speed, and the thruster operation state includes: The basic dynamic condition value of the ship is calculated by weighting and summing the normalized engine load and speed. The side thruster's operating status is detected. If the side thruster is in the activated state, the preset impact coefficient is used as the working condition impact factor; otherwise, the working condition impact factor is set to the baseline value. The ship's basic dynamic operating condition value is multiplied by the operating condition impact factor to obtain the ship's operating condition coefficient.

[0007] According to the ship power battery graded early warning method based on operating condition linkage provided by the present invention, the step of dynamically adjusting the early warning judgment parameters according to the ship operating condition coefficient includes: Obtain preset baseline warning judgment parameters, which include baseline drift parameters and baseline judgment thresholds; Using the ship operating condition coefficient as a multiplier, calculate the product of the baseline drift parameter and the ship operating condition coefficient, and use the product result as the corrected drift parameter; Using the logarithmic transformation value of the ship's operating condition coefficient as the adjustment amount, calculate the sum or product of the benchmark judgment threshold and the adjustment amount, and use the calculation result as the corrected judgment threshold; The corrected drift parameters and the decision threshold are used as dynamically adjusted early warning decision parameters; Wherein, when the ship operating condition coefficient is equal to 1, the corrected drift parameter is equal to the baseline drift parameter, and the corrected decision threshold is equal to the baseline decision threshold; when the ship operating condition coefficient is greater than 1, the corrected drift parameter is greater than the baseline drift parameter, and the corrected decision threshold is greater than the baseline decision threshold.

[0008] According to the ship power battery graded early warning method based on operating condition linkage provided by the present invention, the step of outputting the ship power battery graded early warning result based on the battery operating parameters and the dynamically adjusted early warning judgment parameters includes: Obtain the monitored indicators from the battery operating parameters; The dynamically adjusted early warning judgment parameters are used as the judgment criteria to perform anomaly detection on the monitored indicators. When the monitored indicator meets the first preset condition, the first-level early warning result is output, which includes alarm prompts or port inspection suggestions. When the monitored indicator meets the second preset condition, a second-level early warning result is output. The second-level early warning result includes at least one of power-limited operation, limiting charging and discharging, or triggering local protection action.

[0009] According to the marine power battery classification and early warning method based on working condition linkage provided by the present invention, the monitored indicators include at least one of the following: minimum voltage of a single cell, maximum voltage of a single cell, voltage difference between cells, maximum temperature, temperature rise rate, charging and discharging current, and insulation resistance.

[0010] According to the ship power battery graded early warning method based on operating condition linkage provided by the present invention, after dynamically adjusting the early warning judgment parameters according to the ship operating condition coefficient, it further includes: When the ship's operating condition coefficient represents a high-load condition, the warning judgment parameter is increased to suppress false alarms caused by transient load shocks. Once the ship's operating condition coefficient returns to normal operating conditions, the early warning judgment parameter is smoothly reverted to the baseline value.

[0011] This invention also provides a graded early warning system for ship power batteries based on operating condition linkage, comprising: The acquisition module is used to acquire battery operating parameters and ship operating condition parameters of the battery management system; The calculation module is used to calculate the ship's operating condition coefficient based on the ship's operating condition parameters; The adjustment module is used to dynamically adjust the early warning judgment parameters according to the ship's operating condition coefficient; The output module is used to output the graded early warning result of the ship's power battery based on the battery operating parameters and the dynamically adjusted early warning judgment parameters.

[0012] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the graded early warning method for ship power batteries based on operating conditions as described above.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the graded early warning method for ship power batteries based on operating condition linkage as described above.

[0014] This invention provides a graded early warning method and related device for ship power batteries based on operating condition linkage. The method acquires battery operating parameters and ship operating condition parameters from the battery management system; calculates a ship operating condition coefficient based on the ship operating condition parameters; dynamically adjusts the early warning judgment parameters based on the ship operating condition coefficient; and outputs graded early warning results for ship power batteries based on the battery operating parameters and the dynamically adjusted early warning judgment parameters. This invention introduces real-time ship operating condition parameters to dynamically adjust the early warning judgment parameters, enabling the early warning system to adaptively adjust the judgment benchmark according to the ship's current operating state. This effectively distinguishes between normal operating condition fluctuations and actual battery faults, significantly reducing false alarms and accidental brake failures under transient high load impacts on the ship, ensuring continuous ship operations. Furthermore, the operating condition linkage mechanism of this invention does not rely on historical data training or real-time cloud communication; it can dynamically correct the early warning judgment parameters based solely on locally acquired ship operating condition parameters. This results in fast response speed and high reliability, making it suitable for ocean-going vessel environments with frequent communication interruptions, and improving the safety monitoring level of the power battery system under all navigation conditions. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the graded early warning method for ship power batteries based on operating condition linkage provided by the present invention. Figure 2 This is the health state machine and reversible / irreversible discrimination flowchart provided by the present invention; Figure 3 This is a flowchart of the cloud-based deep diagnostic process for network disconnection and reconnection provided by the present invention; Figure 4 This invention provides a knowledge graph of degradation mechanisms. Figure 5 This is a flowchart of the shadow mode unlocking process provided by the present invention; Figure 6 This is a schematic diagram of the structure of the ship power battery graded early warning system based on working condition linkage provided by the present invention; Figure 7 This is an architecture diagram of the ship power battery graded early warning system based on working condition linkage provided by the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0018] Figure 1 A flowchart of a graded early warning method for ship power batteries based on operating condition linkage provided in an embodiment of the present invention is shown below. Figure 1 As shown, the graded early warning method for ship power batteries based on operating condition linkage provided in this embodiment of the invention includes: Step 101: Obtain the battery operating parameters and ship operating condition parameters from the battery management system; Step 102: Calculate the ship's operating condition coefficient based on the ship's operating condition parameters; Step 103: Dynamically adjust the early warning judgment parameters according to the ship operating condition coefficient; Step 104: Based on the battery operating parameters and the dynamically adjusted warning judgment parameters, output the graded warning result of the ship's power battery.

[0019] Traditional early warning systems mainly rely on fixed thresholds or cloud-based batch calculations with delays, which are difficult to cope with transient fluctuations and cannot fundamentally distinguish between reversible operating condition fluctuations and irreversible degradation trends, leading to frequent false alarms and unexpected emergency shutdowns.

[0020] The present invention provides a graded early warning method for ship power batteries based on operating condition linkage. This method acquires battery operating parameters and ship operating condition parameters from the battery management system; calculates a ship operating condition coefficient based on the ship operating condition parameters; dynamically adjusts the early warning judgment parameters based on the ship operating condition coefficient; and outputs a graded early warning result for the ship power battery based on the battery operating parameters and the dynamically adjusted early warning judgment parameters. By introducing real-time ship operating condition parameters to dynamically adjust the early warning judgment parameters, the present invention enables the early warning system to adaptively adjust the judgment benchmark according to the ship's current operating state, effectively distinguishing between normal operating condition fluctuations and actual battery faults. This significantly reduces false alarms and accidental brake failures under transient high load impacts on the ship, ensuring continuous ship operations. Furthermore, the operating condition linkage mechanism of this invention does not rely on historical data training or real-time cloud communication; it can dynamically correct the early warning judgment parameters based solely on locally acquired ship operating condition parameters. This results in fast response speed, high reliability, and suitability for ocean-going vessel environments with frequent communication interruptions, improving the safety monitoring level of the power battery system under all navigation conditions.

[0021] Based on any of the above embodiments, the calculation of the ship operating condition coefficient according to the ship operating condition parameters includes: Step 201: Obtain ship operating condition parameters in real time from the ship's NMEA 2000 communication bus. The ship operating condition parameters include at least engine load, speed and thruster operation status. Specifically, the edge gateway synchronously acquires ship operating status messages from the ship's NMEA 2000 communication bus using a zero-copy parsing method. These operating status messages include at least engine load, engine speed, roll / pitch attitude, and thruster operation status. Simultaneously, the edge gateway also receives high-frequency timing data from the battery management system and aligns the ship operating status messages with this data. After data alignment, multi-frame filtering and transient peak loss compensation are performed to eliminate signal noise and sampling anomalies. Based on the processed data, the current operating state of the ship is identified, including at least one of charging, discharging, stationary, berthing, propulsion, and high-load operation.

[0022] Step 202: Normalize the engine load, speed and thruster operation status, and map parameters of different dimensions to a unified numerical range. In this embodiment of the invention, the engine load is expressed as a percentage representing the current engine output power relative to the rated power, the speed is expressed in revolutions per minute representing the engine crankshaft rotation speed, and the side thruster operation status is expressed as a Boolean value representing whether the side thruster is in the start-up and operation state. The above three parameters with different dimensions and numerical ranges are normalized and mapped to convert each parameter to a unified continuous numerical range, so as to eliminate the influence of the difference in dimensions on the subsequent calculation of the operating condition coefficient.

[0023] Step 203: Calculate the ship's operating condition coefficient according to the normalized engine load, the speed, and the thruster's operating state, following a preset fusion rule.

[0024] The ship's basic dynamic condition value is calculated by weighted summation of the normalized engine load and speed. The operation status of the side thrusters is detected. If the side thrusters are in operation, a preset impact coefficient is used as the condition impact factor; otherwise, the condition impact factor is set as the baseline value. The ship's basic dynamic condition value is multiplied by the condition impact factor to obtain a dimensionless ship condition coefficient. The lower limit of the ship condition coefficient is 1, indicating that the current condition is stable. When the engine load and / or speed increases, and / or the side thrusters are in operation, the ship condition coefficient is greater than 1 and increases with the increase of load and speed, indicating that the ship is currently in a high-load transient impact condition.

[0025] The ship operating condition coefficient is used for subsequent dynamic adjustment of early warning judgment parameters. Specifically, it includes: using the ship operating condition coefficient as input to the CUSUM (cumulative sum) dynamic anomaly index grid, and dynamically adjusting the cumulative sum drift parameters and anomaly judgment threshold according to the ship operating condition coefficient, so as to smooth and suppress the voltage and current abnormal pulses generated by the ship's transient heavy load impact.

[0026] Based on any of the above embodiments, the step of calculating the ship's operating condition coefficient according to a preset fusion rule based on the normalized engine load, the engine speed, and the thruster's operating state includes: Step 301: Calculate the ship's basic dynamic operating condition value by weighting and summing the normalized engine load and speed. Step 302: Detect the side thruster's operating status. If the side thruster is in the activated state, use the preset impact coefficient as the working condition impact factor; otherwise, set the working condition impact factor to the baseline value. Step 303: Multiply the ship's basic dynamic operating condition value by the operating condition impact factor to obtain the ship's operating condition coefficient.

[0027] Based on any of the above embodiments, the step of dynamically adjusting the early warning judgment parameters according to the ship operating condition coefficient includes: Step 401: Obtain preset benchmark warning judgment parameters, which include benchmark drift parameters and benchmark judgment thresholds; Step 402: Using the ship operating condition coefficient as a multiplier, calculate the product of the baseline drift parameter and the ship operating condition coefficient, and use the product result as the corrected drift parameter; Step 403: Using the logarithmic transformation value of the ship's operating condition coefficient as the adjustment amount, calculate the sum or product of the benchmark judgment threshold and the adjustment amount, and use the calculation result as the corrected judgment threshold; Step 404: Use the corrected drift parameters and the judgment threshold as the dynamically adjusted early warning judgment parameters; Wherein, when the ship operating condition coefficient is equal to 1, the corrected drift parameter is equal to the baseline drift parameter, and the corrected decision threshold is equal to the baseline decision threshold; when the ship operating condition coefficient is greater than 1, the corrected drift parameter is greater than the baseline drift parameter, and the corrected decision threshold is greater than the baseline decision threshold.

[0028] Based on any of the above embodiments, the step of outputting a graded early warning result for the ship's power battery according to the battery operating parameters and the dynamically adjusted early warning judgment parameters includes: Step 501: Obtain the monitored indicators from the battery operating parameters; In this embodiment of the invention, the monitored parameters include at least one of the following: minimum voltage of a single cell, maximum voltage of a single cell, voltage difference between cells, maximum temperature, rate of temperature rise, charge / discharge current, and insulation resistance.

[0029] Step 502: Using the dynamically adjusted early warning judgment parameters as the judgment benchmark, perform anomaly detection on the indicator to be monitored; Step 503: When the monitored indicator meets the first preset condition, output the first-level early warning result. The first-level early warning result includes alarm prompts or port inspection suggestions. Step 504: When the monitored indicator meets the second preset condition, output the second-level early warning result. The second-level early warning result includes at least one of power-limited operation, limiting charging and discharging, or triggering local protection action.

[0030] In this embodiment of the invention, the first preset condition includes a state where the monitored indicators (such as minimum voltage, maximum voltage, inter-cell voltage difference, maximum temperature, temperature rise rate, charge / discharge current, and insulation resistance) reach or exceed the warning trigger threshold, but have not yet reached the emergency safety threshold. For example, the cell voltage drops below the preset warning line (e.g., 2.8V), or the temperature rise rate exceeds the preset rate threshold, or the voltage difference exceeds the preset voltage difference threshold. When this condition is met, the system determines that the battery is abnormal but has not yet endangered safety, and outputs a first-level warning result, including an alarm prompt or port inspection suggestion, prompting maintenance personnel to pay attention and arrange planned inspections.

[0031] The second preset condition includes a state where the monitored indicators continue to deteriorate, reaching or exceeding the emergency safety threshold. Examples include: a single cell voltage further dropping below a dangerous threshold (e.g., 2.5V) for a preset time; the highest temperature reaching the emergency cut-off threshold; or the insulation resistance falling below the safety boundary. When this condition is met, the system determines that the battery has an emergency safety risk and outputs a second-level warning result, including power-limited operation, limiting charge and discharge, or triggering local protection actions to achieve emergency safety handling.

[0032] Based on any of the above embodiments, after dynamically adjusting the early warning judgment parameters according to the ship operating condition coefficient, the method further includes: Step 601: When the ship's operating condition coefficient represents a high load condition, increase the warning judgment parameter to suppress false alarms caused by transient load impacts; Step 602: When the ship's operating condition coefficient returns to normal operating conditions, the early warning judgment parameter is smoothly withdrawn to the benchmark value.

[0033] In this embodiment of the invention, the process for obtaining the dynamic threshold of CUSUM based on operating conditions includes: S1. The edge gateway reads the individual minimum voltage, maximum voltage, voltage difference, maximum temperature, temperature rise rate, current, insulation resistance and fire signal of the BMS at a frequency of 1 second or higher.

[0034] S2. The edge gateway synchronously parses NMEA 2000 messages, extracts real-time dynamic data such as engine load, speed (RPM), transmission status, and thruster action, and calculates the dimensionless ship operating condition coefficient alpha_ship. alpha_ship = f(Load, RPM, Thruster_Active) S3. Dynamically generate the baseline mu_t based on op_state, and dynamically adjust the cumulative and drift parameters k_t of CUSUM and the anomaly detection threshold h_t in the L3 anomaly index grid based on alpha_ship: k_t = k_base alpha_ship h_t = h_base (1 + beta ln(1 + alpha_ship)) Where k_base and h_base are the baseline parameters, and beta is the sensitivity adjustment factor.

[0035] S4. Input x_t for parameters such as minimum voltage, maximum voltage difference, and temperature rise rate of a single unit, and recursively calculate the cumulative sum S_t: S_t = max(0, S_{t-1} + x_t - mu_t - k_t) An early warning is triggered when S_t continuously exceeds h_t and remains within the preset time frame during the sliding window. During high-load conditions such as pusher startup, the system automatically increases k_t and h_t to suppress false alarms caused by short-term fluctuations in operating conditions; if the operating conditions return to stability, the parameters smoothly revert within 5-10 seconds.

[0036] S5. For L4 level extreme safety faults such as severe undervoltage, overtemperature, and multiple fire signal confirmations, the system bypasses CUSUM judgment, and the edge gateway directly triggers the dry contact and PCS shutdown and current limiting locally.

[0037] In this embodiment of the invention, during actual ship operation, the timing data collected at high frequency by the battery management system and the operating condition data output in real time by the ship's NMEA 2000 bus are first converged to an edge gateway deployed at the ship's end. The edge gateway is configured with a RAM ring buffer for cyclically storing the raw timing data of the BMS at high frequency. Under normal operating conditions, the data is only cyclically overwritten in the RAM ring buffer and is not written to the non-volatile flash memory. Only when a local warning event is triggered or a state upgrade transition occurs in the health state machine, will the edge gateway package the window data before and after the trigger time in the ring buffer and persistently write it to the non-volatile flash memory in one go. This reduces the erase and write losses of the flash memory under strong earthquakes and high salt spray environments on ships and extends the service life of the storage medium.

[0038] After the ship docks and network connectivity is restored, the cloud retrieves the idle operating condition time-series data cached by the edge gateway during the network outage. The cloud calculates the self-discharge voltage drop of each individual battery cell at the beginning and end of the idle period and compares it with the self-discharge voltage drop of other cells in the same cluster using Z-score outlier comparison. If the Z-score of the self-discharge voltage drop of a certain cell exceeds the preset abnormal threshold, even if the absolute voltage of the cell is still within the normal range, the cloud will determine that there is an early micro-internal short circuit and generate an early warning work order to push to the maintenance personnel.

[0039] The cloud also performs wavelet feature analysis on the high-frequency time-series data cached by the edge gateway during the network outage: the Db4 wavelet basis is used to extract the high-frequency spike components of the voltage during the charging stage to characterize the sudden change characteristics of the battery internal resistance; the Sym5 wavelet basis is used to extract the second derivative of the temperature rise rate to capture the sudden change characteristics of the early temperature evolution in thermal runaway, so as to achieve in-depth identification of early hidden dangers before the voltage and temperature exceed the limits.

[0040] After triggering the local protection decision, the edge gateway executes local protection actions within seconds, including controlling the physical dry contact to disconnect or controlling the PCS to shut down and limit the flow, to ensure that emergency safety measures can still be taken in the event of communication interruption.

[0041] In the initial stage of system deployment, the edge gateway operates in shadow mode: even if the local protection judgment conditions are met, the edge gateway only illuminates the simulated indicator light and records the action log, without closing the actual physical dry contact, to avoid accidental power outages due to initial configuration deviations. The cloud aggregates and analyzes the simulated action logs in shadow mode, calculates the false trip rate, and after confirming the system reliability, sends an unlocking command to the edge gateway via OTA. Upon receiving the unlocking command, the edge gateway enables the power outage permission of the actual physical dry contact.

[0042] The system also receives real-world fault tags and cell replacement records entered by maintenance personnel and feeds them back to the cloud. The cloud adaptively calibrates the fusion weights of the Health Index (HI) and the baseline parameters of the CUSUM dynamic threshold based on the fed-back data, enabling continuous optimization of the early warning model and judgment criteria.

[0043] This invention embodiment also includes: performing cloud-edge collaboration and network outage compensation diagnosis, specifically including: S1. The edge gateway utilizes the edge memory bus to temporarily store high-frequency, high-precision BMS raw data. Only when a local warning event is triggered or the health state machine state is upgraded and transferred will the data of the preceding and following windows in the ring cache be packaged and persistently written to the non-volatile flash memory, which greatly reduces the physical loss of non-volatile flash memory in harsh environments such as strong earthquakes and high salt spray on ships.

[0044] S2, the cloud-based system is responsible for in-depth checks on low-online-rate port connection windows. It retrieves the terminal voltage of individual cells under static conditions during the network outage, eliminates polarization, and calculates the self-discharge drop DeltaV_leak. The cloud also calculates the outlier degree of the cluster's Z-score. If a cell's Z-score exceeds an outlier value, even if the absolute voltage is currently within the normal range, it is determined to have an early micro-internal short circuit, and a warning work order is generated.

[0045] S3. High-frequency spikes during the charging stage are extracted using the Db4 wavelet basis to characterize the abrupt change in internal resistance. The second derivative of the temperature rise rate is extracted using the Sym5 wavelet basis (without phase delay) to capture the abrupt change in temperature evolution in the early stage of thermal runaway.

[0046] S4, cloud-based maintenance of degradation mechanism library, including lithium plating, micro-internal short circuit, SEI thickening, connection loosening / loosening, etc., maps multi-source abnormal signals to failure mechanisms, and updates baseline parameters and recommended thresholds and sends them to the edge.

[0047] In some embodiments of the present invention, equipment health assessment and degradation trend extraction are performed according to the following steps: S1. The system automatically segments the time sequence according to op_state and builds baseline models for charging, discharging and resting conditions respectively, avoiding direct mixing and extrapolation of data from different conditions.

[0048] S2. Employing a multi-index fusion algorithm, this algorithm comprehensively considers differential pressure increase, internal resistance increase, SOH decrease rate, self-discharge, and temperature dispersion to calculate and output a continuous value of 0-100 for the entity health index HI.

[0049] S3, the health state machine, drives the battery physical state to migrate between four states: normal, slight degradation, obvious abnormality, and critical, based on HI and state transition conditions. Upgrades require multi-frame confirmation, and hysteresis debouncing is performed when the state falls back.

[0050] S4. Before performing lifetime predictions such as the time-to-threshold (TTD) or remaining useful life (RUL), the system identifies and separates the reversible voltage and temperature rise components from the original signal Y_t, including the charge-discharge polarization voltage V_pol, the temperature rise Delta T_env caused by ambient temperature, and the instantaneous voltage drop V_trans caused by high-power transient load.

[0051] S5. Calculate the net degradation component X_net, which reflects the irreversible physical degradation of the battery material structure: X_net = Y_t - (V_pol + Delta T_env + V_trans) S6. Extrapolation prediction is performed only on the net degradation component X_net after deduction. Lifetime warnings are only generated when X_net shows continuous monotonous deterioration and matches early failure signs in the degradation mechanism library, thereby greatly reducing false alarms. If the trend only shows reversible polarization fluctuations, the system outputs health observation or low-level prompts and does not trigger emergency shutdown or lifetime warnings.

[0052] like Figure 2 As shown, the health state machine and the reversible / irreversible determination process include: The system first performs a continuous health assessment of the battery, integrating multiple indicators such as battery differential pressure, internal resistance estimation, SOH decay rate, and static self-discharge rate, and calculates a weighted continuous health index HI. This health index is a continuous value between 0 and 100, with a higher value indicating a better battery health status.

[0053] Based on the continuous health index, the system drives the health state machine to gradually transition through the following four states: State 1: Normal. When HI is greater than or equal to 85, the battery is in normal condition and only routine checks are required. If HI continues to drop below 85, the state machine will transition from the normal state to the slightly degraded state after multiple consecutive scans and confirmations. If HI rises back above 90, the state machine will allow the state to revert to normal after multiple scans and confirmations to achieve hysteresis debouncing.

[0054] State 2: Slight Degradation. When HI is between 70 and 85, the battery is in a state of slight degradation, and the system outputs a recommendation to schedule a follow-up check within 7 days. If HI continues to drop below 70, the state machine will transition to a clearly abnormal state after multiple consecutive scans and confirmations; if HI rises back above 75, the state machine will allow the state to regress to slight degradation after multiple scans and confirmations.

[0055] State 3: Obvious Abnormality. When HI is between 50 and 70, the battery is in an obviously abnormal state, and the system recommends scheduling an inspection within 24 hours. If HI continues to drop below 50, the state machine will transition to a critical state after multiple consecutive scans and confirmations; if HI rises back above 55, the state machine will allow the state to revert to the obviously abnormal state after multiple scans and confirmations.

[0056] State 4: Critical. When HI drops below 50, the battery is in a critical state. The system output recommends immediate shutdown for inspection and activation of safety controls (such as tripping the circuit breaker or PCS shutdown current limiting). In this state, no fallback conditions to lower states are set to ensure the seriousness of safety procedures.

[0057] All of the above state transitions adopt a multi-frame confirmation mechanism, that is, the state transition can only be executed after multiple scans to meet the transition conditions; at the same time, the state rollback adopts a hysteresis design, and the rollback threshold is higher than the entry threshold, in order to prevent the health index from frequently fluctuating near the boundary, which would cause the state machine to oscillate.

[0058] Subsequently, the system identified and separated three reversible components from the original signal: first, the charge / discharge polarization voltage, which is the voltage shift caused by the battery impedance during high-current charging and discharging, which can be recovered after the current is removed; second, the daily variation of ambient temperature, which is the voltage temperature drift caused by the daily variation of the ship's cabin ambient temperature, which can be recovered after the temperature returns to normal; and third, the transient high load drop, which is the instantaneous voltage drop caused by the impact of transient high-power loads on the ship, such as the start of the side thrusters and the operation of the crane, which can be recovered after the load is removed.

[0059] The system subtracts all the reversible components mentioned above from the original time-series signal and extracts the net degradation component that truly reflects the irreversible deterioration of the internal material structure of the battery.

[0060] Ultimately, the system performs linear or nonlinear extrapolation prediction only on the net degradation component to calculate the remaining lifetime margin (TTD). If the net degradation component shows a continuous monotonically worsening trend, the system determines that there is an irreversible degradation risk and upgrades the lifetime warning; if it is only a reversible polarization fluctuation, the system only outputs a low-level warning or observation suggestion and does not trigger a high-level warning.

[0061] Taking the suppression of undervoltage malfunction during thruster startup as an example, when the ship berths and the thrusters start, the current rises rapidly within seconds, causing a brief drop in the minimum voltage of the individual unit. The edge gateway analyzes the engine load and speed in the NMEA 2000 and, combined with the bus voltage drop rate, raises alpha_ship to the high load range. The system simultaneously increases the CUSUM drift parameter k_t and the threshold h_t, ensuring that a short-term explainable voltage drop does not immediately trigger a level 4 trip. After the thrusters stop, alpha_ship smoothly returns to normal within 5-10 seconds.

[0062] Taking the health state machine's identification of progressive degradation as an example, a battery box did not trigger the Level 3 hard threshold, but for several consecutive days, it exhibited increased differential pressure, decreased SOH slope, increased self-discharge rate, and increased temperature dispersion. The system calculates the health index HI and transitions the state from normal to slight degradation, and then to obvious abnormality. At this point, the system generates an L3 health degradation warning and docking troubleshooting suggestions, instead of waiting for the hard threshold to be crossed.

[0063] Taking TTD reversible / irreversible discrimination as an example, at the end of charging, polarization causes the single-cell voltage to briefly approach the overvoltage threshold. Traditional TTD linear extrapolation would predict that it is about to cross the threshold. The system first identifies that this change occurs in the high SOC range of charging and is consistent with the synchronous fluctuations of current, temperature and cells in the same cluster, and determines it as a reversible polarization component. After deducting this component, the net degradation trend is insufficient to trigger a lifetime warning, so only an observation prompt is output, and it is not upgraded to an L4 candidate.

[0064] Taking cloud-based short-circuit tracing during network outages as an example, such as... Figure 3 As shown, the ship experienced a continuous 12-hour network outage during navigation, during which the battery remained in a low-current or quiescent state for an extended period. The edge device retained critical windows in memory and uploaded the first and last sampled data upon reconnection. The cloud calculated the quiescent voltage drop DeltaV_leak for each cell and compared it with the median of the same cluster. If a cell's Z-score exceeded the threshold and the degradation mechanism library matched early signs of a micro-internal short circuit, a Level 3 micro-internal short circuit warning work order was generated, even if the current absolute voltage remained within a safe range.

[0065] Taking the degradation mechanism library to assist in root cause diagnosis as an example, such as Figure 4 As shown, when multi-cluster cross-validation detects an abnormal voltage difference in a cluster but the temperature does not rise synchronously, the system compares the observed signal with the mechanism library. If the anomaly manifests as a jump in individual cell voltage, abnormal local internal resistance estimation, and asynchronous current changes, it is more likely to indicate a loose connection or sampling anomaly. If it manifests as a persistently large static voltage drop, a decrease in SOH, and abnormal temperature rise acceleration, it is more likely to indicate a micro-internal short circuit or a precursor to thermal runaway. Based on this, the system outputs the root cause, the chain of evidence, and handling recommendations.

[0066] Taking the secure deployment and closed-loop backfeedback of shadow mode as an example, such as Figure 5 As shown, during the first 30 days of deployment on a new ship or project, the edge gateway only illuminates the simulated indicator light and records logs even if it identifies a candidate, without closing the actual dry contact. The cloud aggregates simulated action logs under conditions such as high winds and waves, berthing, and heavy-load operations, calculating the true positive rate, false action rate, and lead time. After auditing is passed, the cloud OTA issues enable_trip=true, allowing the edge terminal to execute the actual power outage. Subsequent maintenance results and false alarm tags are fed back to the cloud to update the health baseline and mechanism matching weights.

[0067] The ship power battery graded early warning method based on operating condition linkage provided in this invention significantly reduces transient false alarms and accidental power outages. By using the operating condition linkage CUSUM dynamic index grid and reversible / irreversible discrimination, reversible voltage fluctuations such as propeller water discharge and thruster start-up are separated from irreversible degradation, reducing the false alarm rate by more than 90%. Through the RAM ring cache mechanism of the intelligent agent edge memory bus, high-frequency data is avoided from being continuously written to Flash, extending the Flash erase and write life by more than 5 times, perfectly adapting to the harsh ship strong vibration environment. Through cloud-based network outage compensation and micro-short circuit tracing analysis, using static self-discharge Z-score outlier degree and wavelet glitch extraction, lithium plating and micro-internal short circuit hazards can be captured in the very early stage when the absolute voltage has not exceeded the limit and there is no significant temperature rise. Combined with local second-level life-saving control, shadow mode safety anti-accidental power outage audit, and closed-loop feedback adaptive calibration, a complete upgrade from passive emergency alarm to active state machine intelligent agent evaluation, prediction, action, and self-evolution is completed.

[0068] The following describes the ship power battery graded early warning system based on operating condition linkage provided by the present invention. The ship power battery graded early warning system based on operating condition linkage described below can be referred to in correspondence with the ship power battery graded early warning method based on operating condition linkage described above.

[0069] Figure 6 This is a schematic diagram of the structure of a graded early warning system for ship power batteries based on operating condition linkage provided in an embodiment of the present invention, as shown below. Figure 6 As shown, the ship power battery graded early warning system based on operating condition linkage provided in this embodiment of the invention includes: The acquisition module 601 is used to acquire the battery operating parameters and ship operating condition parameters of the battery management system; Calculation module 602 is used to calculate the ship operating condition coefficient based on the ship operating condition parameters; Adjustment module 603 is used to dynamically adjust the early warning judgment parameters according to the ship operating condition coefficient; The output module 604 is used to output the graded early warning result of the ship's power battery based on the battery operating parameters and the dynamically adjusted early warning judgment parameters.

[0070] like Figure 7 As shown, the ship power battery graded fault early warning system proposed in this embodiment of the invention includes: The L1 multi-source physical / logical integration layer parses high-frequency timing data from the BMS and NMEA 2000 ship bus operating condition messages (engine load, speed, roll / pitch, thruster action, etc.) with zero replication; the operating condition segmentation module identifies op_states such as charging, discharging, stationary, berthing, propulsion, and high-load operation, and performs signal preprocessing using multi-frame filtering and transient peak loss compensation.

[0071] The L2 unified reality entity layer integrates differential pressure expansion rate, internal resistance growth rate, SOH decrease rate, temperature dispersion, self-discharge rate, and alarm frequency, using multi-indicator fusion calculation to output a continuous health index (HI). The entity state machine driver module controls the battery entity state to transition between normal, mild degradation, significant abnormality, and critical states.

[0072] The reversible / irreversible discrimination module identifies and subtracts the charge / discharge polarization voltage V_pol, temperature rise Delta T_env, and ship transient load drop V_trans from the original signal, and extracts the net degradation component X_net, which reflects the irreversible degradation trend of the battery, for TTD / RUL prediction.

[0073] The L3 unified semantic index layer calculates the dimensionless ship condition coefficient alpha_ship in real time. In the CUSUM dynamic anomaly index grid, the cumulative and drift parameters k_t and the anomaly judgment threshold h_t are dynamically adjusted according to alpha_ship to smooth and suppress voltage / current anomaly pulses generated by transient heavy load impacts.

[0074] In the L4 agent's long-term memory layer, high-frequency timing data in the edge gateway RAM circular buffer module is temporarily stored in the edge gateway memory bus. It is only written to non-volatile flash memory once when a local L3 / L4 event is triggered or a health state transition occurs, protecting the medium from damage caused by strong shocks.

[0075] The cloud-based network outage compensation micro-short circuit analysis module retrieves the static timing sequence during the network outage when the network is connected, calculates the self-discharge voltage drop DeltaV_leak, and diagnoses early micro-internal short circuits by comparing the outlier of the same cluster Z-score.

[0076] The L5 intelligent collaboration and OSDK action control layer enable the second-level local control module to determine the L4 local protection through the edge gateway and control the physical dry contact to trip or the PCS to stop and limit current.

[0077] The grayscale security control module (shadow mode) runs in shadow mode during the initial on-site deployment to perform security audits for malfunctions. After passing the audit, it unlocks the control of the real physical dry contact via OTA.

[0078] The closed-loop feedback module receives real-time on-site fault tags and replacement records from maintenance personnel, and feeds them back to the cloud for adaptive calibration of HI weights and CUSUM threshold baselines.

[0079] The present invention provides a graded early warning system for ship power batteries based on operating condition linkage. This system acquires battery operating parameters and ship operating condition parameters from the battery management system; calculates a ship operating condition coefficient based on the ship operating condition parameters; dynamically adjusts the early warning judgment parameters based on the ship operating condition coefficient; and outputs graded early warning results for ship power batteries based on the battery operating parameters and the dynamically adjusted early warning judgment parameters. By introducing real-time ship operating condition parameters to dynamically adjust the early warning judgment parameters, the present invention enables the early warning system to adaptively adjust the judgment benchmark according to the ship's current operating state, effectively distinguishing between normal operating condition fluctuations and actual battery faults. This significantly reduces false alarms and accidental brake failures under transient high load impacts on the ship, ensuring continuous ship operations. Furthermore, the operating condition linkage mechanism of this invention does not rely on historical data training or real-time cloud communication; it can dynamically correct the early warning judgment parameters based solely on locally acquired ship operating condition parameters. This results in fast response speed, high reliability, and suitability for ocean-going vessel environments with frequent communication interruptions, improving the safety monitoring level of the power battery system under all navigation conditions.

[0080] This invention also provides an electronic device, which may include a processor, a communication interface, a memory, and a communication bus, wherein the processor, communication interface, and memory communicate with each other via the communication bus. The processor can call logical instructions in the memory to execute a ship power battery graded early warning method based on operating condition linkage. This method includes: acquiring battery operating parameters and ship operating condition parameters from the battery management system; calculating a ship operating condition coefficient based on the ship operating condition parameters; dynamically adjusting early warning judgment parameters based on the ship operating condition coefficient; and outputting a ship power battery graded early warning result based on the battery operating parameters and the dynamically adjusted early warning judgment parameters.

[0081] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the above-described method for graded early warning of ship power batteries based on operating condition linkage. The method includes: acquiring battery operating parameters and ship operating condition parameters from a battery management system; calculating a ship operating condition coefficient based on the ship operating condition parameters; dynamically adjusting early warning judgment parameters based on the ship operating condition coefficient; and outputting a graded early warning result of the ship power battery based on the battery operating parameters and the dynamically adjusted early warning judgment parameters.

[0082] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0083] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A graded early warning method for ship power batteries based on operating condition linkage, characterized in that, include: Obtain battery operating parameters and ship operating condition parameters from the battery management system; Calculate the ship's operating condition coefficient based on the aforementioned ship operating condition parameters; The early warning judgment parameters are dynamically adjusted based on the aforementioned ship operating condition coefficients; Based on the battery operating parameters and the dynamically adjusted warning judgment parameters, the graded warning results for the ship's power battery are output.

2. The graded early warning method for ship power batteries based on operating condition linkage according to claim 1, characterized in that, The calculation of the ship operating condition coefficient based on the ship operating condition parameters includes: Ship operating parameters are acquired in real time from the ship's NMEA 2000 communication bus. These parameters include at least engine load, speed, and thruster operating status. The engine load, speed, and thruster operating status are normalized to map parameters of different dimensions to a unified numerical range. Based on the normalized engine load, the speed, and the thruster operating status, the ship's operating condition coefficient is calculated according to a preset fusion rule.

3. The graded early warning method for ship power batteries based on operating condition linkage according to claim 2, characterized in that, The step of calculating the ship's operating condition coefficient according to a preset fusion rule based on the normalized engine load, the engine speed, and the thruster's operating state includes: The basic dynamic condition value of the ship is calculated by weighting and summing the normalized engine load and speed. The side thruster's operating status is detected. If the side thruster is in the activated state, the preset impact coefficient is used as the working condition impact factor; otherwise, the working condition impact factor is set to the baseline value. The ship's basic dynamic operating condition value is multiplied by the operating condition impact factor to obtain the ship's operating condition coefficient.

4. The graded early warning method for ship power batteries based on operating condition linkage according to claim 1, characterized in that, The dynamic adjustment of the early warning judgment parameters based on the ship's operating condition coefficient includes: Obtain preset baseline warning judgment parameters, which include baseline drift parameters and baseline judgment thresholds; Using the ship operating condition coefficient as a multiplier, calculate the product of the baseline drift parameter and the ship operating condition coefficient, and use the product result as the corrected drift parameter; Using the logarithmic transformation value of the ship's operating condition coefficient as the adjustment amount, calculate the sum or product of the benchmark judgment threshold and the adjustment amount, and use the calculation result as the corrected judgment threshold; The corrected drift parameters and the decision threshold are used as dynamically adjusted early warning decision parameters; Wherein, when the ship operating condition coefficient is equal to 1, the corrected drift parameter is equal to the baseline drift parameter, and the corrected decision threshold is equal to the baseline decision threshold; when the ship operating condition coefficient is greater than 1, the corrected drift parameter is greater than the baseline drift parameter, and the corrected decision threshold is greater than the baseline decision threshold.

5. The graded early warning method for ship power batteries based on operating condition linkage according to claim 1, characterized in that, The step of outputting a graded early warning result for the ship's power battery based on the battery operating parameters and the dynamically adjusted early warning judgment parameters includes: Obtain the monitored indicators from the battery operating parameters; The dynamically adjusted early warning judgment parameters are used as the judgment criteria to perform anomaly detection on the monitored indicators. When the monitored indicator meets the first preset condition, the first-level early warning result is output, which includes alarm prompts or port inspection suggestions. When the monitored indicator meets the second preset condition, a second-level early warning result is output. The second-level early warning result includes at least one of power-limited operation, limiting charging and discharging, or triggering local protection action.

6. The graded early warning method for ship power batteries based on operating condition linkage according to claim 5, characterized in that, The monitored parameters include at least one of the following: minimum voltage of a single cell, maximum voltage of a single cell, voltage difference between cells, maximum temperature, rate of temperature rise, charging and discharging current, and insulation resistance.

7. The graded early warning method for ship power batteries based on operating condition linkage according to claim 1, characterized in that, After dynamically adjusting the early warning judgment parameters based on the ship's operating condition coefficient, the method further includes: When the ship's operating condition coefficient represents a high-load condition, the warning judgment parameter is increased to suppress false alarms caused by transient load shocks. Once the ship's operating condition coefficient returns to normal operating conditions, the early warning judgment parameter is smoothly reverted to the baseline value.

8. A graded early warning system for ship power batteries based on operating condition linkage, characterized in that, include: The acquisition module is used to acquire battery operating parameters and ship operating condition parameters of the battery management system; The calculation module is used to calculate the ship's operating condition coefficient based on the ship's operating condition parameters; The adjustment module is used to dynamically adjust the early warning judgment parameters according to the ship's operating condition coefficient; The output module is used to output the graded early warning result of the ship's power battery based on the battery operating parameters and the dynamically adjusted early warning judgment parameters.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the graded early warning method for ship power batteries based on operating condition linkage as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the graded early warning method for ship power batteries based on operating condition linkage as described in any one of claims 1 to 7.