Marine lithium battery BMS control method and system based on deep learning

Abstract

The invention relates to the technical field of lithium battery management systems, and discloses a marine lithium battery BMS control method and system based on deep learning, and the method comprises the steps: constructing a three-layer causal structure of an environment layer, a system layer and a fault layer; learning a stable causal structure by using an invariant risk minimization algorithm of multi-environment data and gradient penalty; evaluating causal edge strength based on a mutual information theory, and extracting a stable causal skeleton; constructing a cross-environment enhanced data set through an anti-fact data generation technology; constructing a meta-learning framework to realize rapid adaptation of the model to a new environment; according to the method, the problem of generalization of fault diagnosis of the marine lithium battery in different global sea area environments is solved, the fault root cause identification accuracy is improved, the misdiagnosis rate is reduced, and rapid adaptation to a new environment is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery management systems, and more specifically, to a marine lithium battery BMS control method and system based on deep learning. Background Art

[0002] The deployment of marine lithium battery management systems (BMS) in different sea areas around the world faces huge challenges of environmental differences. When the same battery system sails in different sea areas such as the Arctic, tropical, and temperate zones, it is affected by multiple factors such as sea conditions, climate, and navigation missions, and exhibits differentiated operating characteristics and failure modes. This difference is not only reflected in the basic characteristics of the battery, but also affects the accuracy and reliability of fault diagnosis.

[0003] Traditional marine lithium battery fault diagnosis methods mainly rely on correlation analysis or simple rule reasoning, which makes it difficult to effectively distinguish the true cause of the fault from the symptom. Faced with the combined influence of multiple factors such as environmental conditions, operating modes and maintenance history, the accuracy of diagnosis is significantly reduced. For example, in tropical high temperature environments, the battery derating protection mechanism may be misdiagnosed as a battery cell failure; while in polar low temperature environments, the normal decrease in charging and discharging efficiency may be mistakenly judged as accelerated aging. This environmental dependence causes the same diagnostic algorithm to perform inconsistently in different deployment environments, seriously affecting the reliability of the battery management system and the accuracy of maintenance decisions.

[0004] In the existing technology, there is still a lack of methods that can effectively solve the generalization problem of marine lithium battery fault diagnosis in different sea environments, especially in identifying the root cause of the fault and quickly adapting to the new environment. Summary of the invention

[0005] The present invention provides a deep learning-based marine lithium battery BMS control method and system to solve the technical problem of the generalization difficulty of fault diagnosis of marine lithium batteries in different environments in the related art.

[0006] The present invention provides a marine lithium battery BMS control method based on deep learning, comprising the following steps: Construct a three-layer causal structure including the environment layer, system layer and fault layer to form a hierarchical causal graph; Based on the hierarchical causal graph, we use the invariant risk minimization algorithm with multi-environment data and gradient penalty to learn stable causal structures. For the learned stable causal structure, based on the mutual information theory, the strength of the causal edge is evaluated and a stable causal skeleton is extracted; Based on the extracted stable causal skeleton, we construct an enhanced dataset across environments through counterfactual data generation technology. By utilizing the constructed cross-environment enhanced dataset, a meta-learning framework is constructed to enable the model to quickly adapt to the new environment.

[0007] In a preferred embodiment, the step of constructing a three-layer causal structure including an environment layer, a system layer, and a fault layer comprises: Construct an environment layer variable set to collect variables related to the ship's operating environment; Construct a system-level variable set to collect the operating parameters of the marine lithium battery system; Construct a set of fault layer variables to define the possible fault types of marine lithium batteries; Based on domain knowledge and historical data, causal relationships between variables are established to form a hierarchical causal graph.

[0008] In a preferred embodiment, the environmental layer variables include sea area type, sea state, climate conditions, navigation mission and ship load status.

[0009] In a preferred embodiment, the system-level variables include battery pack voltage, single cell voltage, charge and discharge current, surface temperature, internal temperature, state of charge and state of health.

[0010] In a preferred embodiment, the fault layer variables include overcharge, over discharge, internal short circuit, external short circuit, thermal runaway, insulation failure and balance circuit failure.

[0011] In a preferred embodiment, the step of learning a stable causal structure comprises: Construct multi-environment datasets; Constructing a deep learning model in which a subset of parameters related to system-level variables are represented as system-level parameters; Design a gradient-penalized invariant risk minimization learning objective; Use the optimization algorithm to optimize the objective function and obtain the model parameters that meet the invariance.

[0012] In a preferred embodiment, the step of evaluating the strength of the causal edge comprises: Calculate mutual information between variables; Calculate conditional mutual information; Compute the strength of causal edges; Based on the causal strength threshold, causal edges with strength greater than the threshold are extracted to form a core causal skeleton.

[0013] In a preferred embodiment, the step of constructing a cross-environment enhanced dataset includes: Based on the core causal framework, construct an intervention model; Construct a counterfactual data generator to generate counterfactual samples; For each environment, select representative samples and generate a set of counterfactual samples; The original dataset is merged with the counterfactual dataset to form an augmented dataset.

[0014] In a preferred embodiment, the step of constructing a meta-learning framework includes: Treat fault diagnosis in each environment as a task and define a set of meta-learning tasks; Based on the trained model parameters, they are used as the initial parameters for meta-learning; For new environments, use a small amount of support set data to update parameters through gradient descent; The adapted model performance is evaluated based on the query set and the initial parameters are updated via meta-optimization.

[0015] In a preferred embodiment, a marine lithium battery BMS control system based on deep learning is used to execute a marine lithium battery BMS control method based on deep learning, including: A multi-level causal structure building module is used to build a three-level causal structure including the environment layer, system layer and fault layer; Invariant causal structure learning module, used to learn stable causal structures; The causal strength assessment and core skeleton extraction module is used to assess the strength of causal edges and extract stable causal skeletons; Multi-source environment data enhancement module, used to build enhanced datasets across environments; The meta-learning fast adaptation module is used to enable the model to quickly adapt to the new environment.

[0016] The beneficial effects of the present invention are: Through hierarchical causal structure and invariant causal learning, the accuracy of fault root cause identification is improved and the misdiagnosis rate is reduced; Rapid adaptation mechanism with multi-source environment data enhancement and meta-learning; Simplified reasoning based on a core causal framework shortens diagnostic time; Optimize maintenance strategies through causal reasoning to reduce maintenance costs; The stable performance of the same model in different sea areas, seasons and operating modes is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a marine lithium battery BMS control method based on deep learning of the present invention; Figure 2 is a detailed flow chart of forming a hierarchical cause-effect graph of the present invention; Figure 3 is a detailed flow chart of the present invention for learning a stable causal structure; Figure 4 is a detailed flow chart of extracting a stable causal skeleton of the present invention; Figure 5 is a detailed flow chart of the present invention for constructing an enhanced data set across environments; Figure 6 It is a detailed flow chart of constructing the meta-learning framework of the present invention. DETAILED DESCRIPTION

[0018] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the present specification. Various examples may omit, replace, or add various processes or components as needed. In addition, the features described in some examples may also be combined in other examples.

[0019] At least one embodiment of the present invention discloses a marine lithium battery BMS control method based on deep learning, such as Figures 1 to 6 As shown, the following steps are included: Step 1: construct a three-layer causal structure including the environment layer, system layer and fault layer to form a hierarchical causal graph; The specific steps include: Step 1.1, construction of environment layer variable set; Collect variables related to the ship's operating environment, including sea area type, sea conditions, climate conditions, navigation tasks, and ship load status, to form an environmental layer variable set ,in, , , Respectively represent , , environment variables, Indicates the total number of environment variables.

[0020] In the marine lithium battery management system, these environmental variables are acquired through the ship's navigation status monitoring system, meteorological system and mission management system, and transmitted to the data preprocessing module in real time.

[0021] Step 1.2, construction of system-level variable set; Collect the operating parameters of the marine lithium battery system, including battery pack voltage, single cell voltage, charge and discharge current, surface temperature, internal temperature, state of charge (SOC), state of health (SOH) and other variables to form a system-level variable set ,in, , , Respectively represent , , system variables, Indicates the total number of system variables.

[0022] These variables are captured by the sensor network of the lithium battery management system and transmitted to the data acquisition module via the CAN bus or other ship communication protocols.

[0023] Step 1.3, construction of the fault layer variable set; Define the possible fault types of marine lithium batteries, including overcharge, over discharge, internal short circuit, external short circuit, thermal runaway, insulation failure, balance circuit failure and other variables, forming a set of fault layer variables ,in, , , Respectively represent , , Fault type variables, Indicates the total number of fault types.

[0024] This application subdivides the fault layer variables according to severity and impact range. For example, thermal runaway faults are subdivided into three stages: early temperature anomaly, hot spot formation, and heat diffusion, so as to accurately locate the root cause of the fault.

[0025] Step 1.4, construction of hierarchical causal diagram; Based on domain knowledge and historical data, causal relationships between variables are established to form a hierarchical causal graph: ; in, represents a hierarchical causal graph, represents a set of directed edges between variables, representing causal relationships, Represents a collection of environment layer variables, system layer variables, and fault layer variables.

[0026] Each edge Representation variables To variable The causal influence includes the causal relationship from environment to system, system to fault, environment to fault, and between variables within a layer.

[0027] In the specific implementation, this application uses an adjacency matrix or an adjacency list to represent the causal graph structure, supporting efficient graph operations and traversal.

[0028] Step 2: Based on the hierarchical causal graph, we use the invariant risk minimization algorithm with multi-environment data and gradient penalty to learn a stable causal structure. The specific steps include: Step 2.1, multi-environment dataset construction; Collect battery operation data from ships in different sea areas, seasons, and mission types to form a multi-environment data set ,in , , Respectively represent , , Data sets in this environment, Indicates the total number of environments.

[0029] For the application scenarios of marine lithium batteries, this application collects three types of typical environmental data: Arctic routes, equatorial routes and temperate routes. Each type of environment contains at least 500 hours of continuous operation data, covering normal operating conditions and various fault conditions.

[0030] Step 2.2, system level parameter representation; Construct a deep learning model where the subset of parameters related to system-level variables is represented as , which is the parameter part of the model responsible for processing system-level variables.

[0031] This application adopts a hybrid architecture of Multi-Layer Perceptron (MLP) and Long Short-Term Memory (LSTM), where MLP is responsible for processing static features and LSTM is responsible for capturing temporal features. System-level parameters include network weights and biases related to battery status, which can be optimized through the back-propagation algorithm.

[0032] Step 2.3, construction of the objective function of constant risk minimization; Design a gradient-penalized invariant risk minimization learning objective: ; in, For all environments The loss function Sum and find the model parameters that minimize this total loss ; Represents the model in the environment The loss function under Represents the gradient of the loss function with respect to the system layer parameters; is the gradient penalty weight coefficient, which is used to balance the empirical risk and invariance constraints; Represents all parameters of the model; represents the set of all environments; represents the sum over all environments; Indicates the square of the L2 norm, which is used to measure the size of the gradient. In the application of marine lithium battery fault diagnosis, this application determines the optimal The value is 0.1, which effectively improves the generalization ability while ensuring the accuracy of the model.

[0033] Step 2.4, invariant causal structure optimization; Use optimization algorithms such as stochastic gradient descent to optimize the invariant risk minimization objective function and obtain model parameters that meet the invariance requirement. In addition, by penalizing the risk gradient variation between different environments, the model will automatically ignore environment-specific spurious associations and retain the causal structure that remains unchanged across environments. During the optimization process, this application uses the Adam optimizer, sets the initial learning rate to 0.001, and combines the learning rate decay strategy to complete model convergence within 40 iterations.

[0034] Step 3: Based on the mutual information theory, the strength of the causal edge is evaluated for the learned stable causal structure, and a stable causal skeleton is extracted; The specific steps include: Step 3.1, calculation of mutual information between variables; For hierarchical cause-effect diagrams Each pair of variables in , calculate their mutual information : ; in, Representation variables and The mutual information between and Respectively represent variables and The specific value of and Respectively represent the variables and The sum of all possible values; Representation variables The value is And variable The value is The joint probability distribution of and Respectively represent variables The value is and variables The value is The marginal probability distribution of Represents the natural logarithm function.

[0035] Mutual information measures the interdependence between two variables, with larger values ​​indicating stronger dependence.

[0036] In practical applications, the present application uses the k-nearest neighbor estimation method to calculate the mutual information of high-dimensional continuous variables. This method shows good stability when the number of samples is limited.

[0037] For the battery temperature and SOC variables, the present application observed that they exhibited high mutual information values ​​in different environments, indicating that the effect of temperature on the battery state is a key factor that is invariant to the environment.

[0038] Step 3.2, conditional mutual information calculation; For each pair of variables and its parent node set , calculate the conditional mutual information : ; in, Indicates that in a given variable Under the condition of and The conditional mutual information between , and Respectively represent variables , and The specific value of , and Respectively represent the variables , and The sum of all possible values; Representation variables The value is ,variable The value is And variable The value is The joint probability distribution of Indicates in the variable The value is Under the condition of The value is And variable The value is The conditional probability of and Respectively expressed in variables The value is Under the condition of The value is and variables The value is The conditional probability of represents the natural logarithm function; Representation variables The parent node set of Representation variables Exclude variables from the parent node collection After the collection.

[0039] The conditional mutual information measures the In this case, the variable and interdependence between them.

[0040] This application applies conditional mutual information analysis in thermal runaway fault diagnosis and finds that battery thermal runaway fault is directly correlated with internal temperature, while the correlation with ambient temperature is significantly weakened after controlling the internal temperature, indicating that internal temperature is the direct cause of thermal runaway, while ambient temperature is an indirect influencing factor.

[0041] Step 3.3, causal strength calculation; Using mutual information and conditional mutual information, causal edges are calculated Strength: ; in, Indicates that from the variable To variable The strength of the causal edge; Representation variables and The mutual information between them is used to quantify the degree of mutual dependence between two variables; Indicates in the variable All parent nodes (except Take the maximum value among the two; Representation variables The set of all parent nodes of ; Representation variables Exclude variables from the parent node collection The collection after Indicates that in a given variable under conditions and The conditional mutual information between .

[0042] This formula quantifies the variable For variables The direct causal influence strength is calculated by subtracting the mutual information conditional on other parent nodes to eliminate the influence of indirect associations.

[0043] The causal strength calculation module implemented in the present application can automatically identify key causal relationships. For example, in the low-temperature environment of the Arctic, the causal strength between the battery internal resistance and SOC is significantly higher than that in the normal temperature environment, indicating that the battery internal resistance under low-temperature conditions is a key factor affecting the battery status and should be monitored closely.

[0044] Step 3.4, core causal skeleton extraction; Based on causal strength threshold , extract causal edges with strength greater than the threshold to form a core causal skeleton: ; ; in, Represents the causal strength threshold, which is used to screen important causal relationships. Only causal edges with strength exceeding this threshold will be retained; Represents the core causal skeleton, which is a simplified version of the original causal diagram, containing only the most important causal relationships; Represents a variable set, including all variables in the environment layer, system layer, and fault layer; Represents the core causal edge set, which is the original edge set A subset that only contains causal strength greater than the threshold The edge of Indicates that from the variable To variable The directed edges of ,represent causal relationships; Indicates that from the variable To variable causal strength.

[0045] Therefore, the core causal skeleton contains causal relationships that are stable across environments, providing a structural basis for environmentally robust fault diagnosis.

[0046] In the marine lithium battery management system, this application determines the optimal The threshold is dynamically adjusted according to the complexity of the environment. A higher threshold is used in a complex environment to retain the most stable causal relationship, and a lower threshold is used in a simple environment to retain more information.

[0047] Step 4: Based on the extracted stable causal skeleton, a cross-environment enhanced dataset is constructed through counterfactual data generation technology; The specific steps include: Step 4.1, construction of intervention model; Based on the core causal skeleton extracted in step 3 , build the intervention model: ; in, represents the target variable; represents the intervention variable; represents the intervention value; Represents the intervention operator, which means the variable Perform human intervention and set to value The operation is different from simple conditional probability; Indicates that an intervention operation is being performed After that, the variable The probability distribution of is used to evaluate the effect of the intervention.

[0048] In a marine lithium battery management system, the intervention model implemented in this application can simulate the impact of different environmental conditions on battery performance. For example, by simulating the intervention of "disconnecting the causal edge from ambient temperature to SOC", the direct impact of temperature changes on the battery state is evaluated, providing a reliable counterfactual basis for fault diagnosis.

[0049] Step 4.2, counterfactual data generator implementation; Building a counterfactual data generator , generate counterfactual samples ,in is the original data, For environment variables, For intervention indicators, specify the intervention variable and target value, represents a counterfactual data generator function, which is used to generate counterfactual samples that meet specific intervention conditions. represents the generated counterfactual samples, representing simulated data under specific intervention conditions.

[0050] The counterfactual data generator of this application adopts a variational autoencoder (VAE) architecture, which includes three parts: encoder, latent space operation and decoder.

[0051] The encoder maps raw data to a latent space that adjusts the variable distribution according to the intervention indicator, and the decoder generates counterfactual samples that comply with physical constraints.

[0052] For marine lithium batteries, this application specifically adds electrochemical consistency constraints to ensure that the generated counterfactual samples conform to the basic physical and chemical laws of batteries.

[0053] Step 4.3, multi-environment counterfactual sample generation; For each environment , select representative samples and generate counterfactual sample sets: ; in, Indicates a specific environment; represents the set of all environments; Indicates the environment The counterfactual sample set under ; Indicates in the environment The following generated Counterfactual input samples, including various characteristic parameters of the battery system; Indicates in the environment The following generated The labels corresponding to the counterfactual samples; Indicates the environment The total number of counterfactual samples generated under

[0054] The multi-environment counterfactual sample generation scheme designed in this application pays special attention to rare cross-environment fault scenarios. For example, for the rare high-temperature runaway faults in polar environments, intervention transformations are performed based on normal temperature environmental data to generate high-temperature runaway samples that meet polar characteristics, thereby making up for the lack of actual data.

[0055] Step 4.4, enhanced dataset construction; The original data set With the counterfactual dataset Merge to form an enhanced dataset , used for model training to improve robustness to environmental changes.

[0056] in, represents the original dataset; Indicates a specific environment The counterfactual dataset generated below; represents the merging of counterfactual datasets of all environments, i.e. , contains counterfactual samples generated in all environments; The final constructed enhanced dataset is the union of the original dataset and the counterfactual dataset; represents the set of all environments; Represents a set union operation.

[0057] The example of enhancing the data set construction implemented in this application on ships on the Arctic route shows that by introducing counterfactual samples, the fault detection accuracy under extreme low temperature conditions is improved from the original 78% to 92%, especially for short-circuit fault types with sparse data, the recognition accuracy is improved by more than 30%.

[0058] Step 5: Use the constructed cross-environment enhanced dataset to build a meta-learning framework to enable the model to quickly adapt to the new environment; The specific steps include: Step 5.1, meta-learning task definition; Each environment The following fault diagnosis is considered as a task , define the set of meta-learning tasks , where each task contains a support set and queryset , , , Represents the first , , Specific tasks, Indicates the total number of tasks.

[0059] In the marine lithium battery system, this application defines the fault diagnosis problems in different environments such as polar routes, equatorial routes and temperate routes as independent tasks, and learns the common knowledge between tasks through the meta-learning framework while retaining the characteristics of each task. For each task, this application selects 100 samples as the support set for intra-task adaptation and 1000 samples as the query set for evaluating the adaptation effect.

[0060] Step 5.2, model initialization; Model parameters trained based on steps 2 and 4 , as the initial parameters for meta-learning, the parameters already contain causal structure information that is invariant across environments.

[0061] The model initialization strategy designed in this application combines the idea of ​​transfer learning and uses the knowledge extracted from the invariant causal structure as the prior information for model initialization to accelerate the meta-learning process. Experiments show that compared with random initialization, the initialization strategy based on the invariant causal structure can shorten the meta-learning convergence time by 50% and achieve better final performance.

[0062] Step 5.3, inner loop adaptation; For new environment , using a small amount of support set data , update the parameters by gradient descent: ; in, Indicates a new environment; Indicates new environment The support set data under is a small set of samples used for rapid model adaptation; represents the initial parameters of the model; Indicates the updated parameters after the inner loop adaptation, which have been adapted to the new environment Adjustments were made; Represents the inner loop learning rate, controls the step size of parameter updates, and affects the speed and stability of model adaptation; Indicates about the parameters The gradient operator is used to calculate the partial derivative of the loss function with respect to the parameters; represents the meta-learning loss function.

[0063] In the actual application of marine lithium battery systems, the internal cycle adaptation mechanism implemented in this application can quickly adjust the model parameters to adapt to new environmental characteristics using only 5 to 10 new environmental samples. For example, when a ship enters tropical waters from temperate waters, the system only needs to collect 10 minutes of battery operation data to complete model adaptation, which significantly improves the system's environmental adaptability.

[0064] Step 5.4, outer loop optimization; Based on query set Evaluate the performance of the adapted model and update the initial parameters via meta-optimization : ; in, Indicates new environment The query set data below is used to evaluate the performance of the model after adaptation; represents the initial parameters of the meta-learning model; Represents the model parameters after inner loop adaptation; Represents the outer loop learning rate, which controls the step size of the outer loop optimization; Indicates about the parameters The gradient operator is used to calculate the gradient of the loss function with respect to the parameters; represents the meta-learning loss function, which is used to evaluate the performance of the model on the query set; Indicates a parameter update operation, which assigns the calculation result on the right side to the variable on the left side.

[0065] By repeatedly iterating inner-loop adaptation and outer-loop optimization, the model’s ability to quickly adapt to new environments is improved.

[0066] This application uses a second-order optimization algorithm in the outer loop optimization stage, taking into account the gradient effect of parameter adaptation on the initial parameters, and improving the stability of meta-learning. In an actual deployment case of a large container ship, after the system completed data collection and meta-learning training on three different routes, it was able to use only 10% of the data when entering a new route, and quickly increased the fault diagnosis accuracy to more than 95%, which is equivalent to the expert level.

[0067] Application examples of this implementation: The implementation methods of the present application have been tested in actual applications on a number of large commercial ships. A typical application example is introduced below.

[0068] A large container transport ship, mainly engaged in ocean shipping business from Asia to Europe, faces the following challenges in its lithium battery management system: The route spans tropical, temperate and subarctic waters, with ambient temperatures ranging from -10°C to 40°C and sea conditions ranging from calm waters to force 9 winds and waves; The battery system needs to operate stably under different ship load conditions (slow-speed sailing, full-speed sailing, port standby); The fault diagnosis accuracy of traditional BMS systems is less than 70%, especially during cross-sea navigation, with frequent false alarms; Faults manifest themselves differently in different sea areas, making it difficult for maintenance personnel to determine the root cause of the fault.

[0069] Implementation process example: Multi-level causal structure construction: Based on the existing sensor network of the ship, 38 environmental layer variables, 52 system layer variables and 24 fault layer variables were collected.

[0070] Environmental layer variables include location information (latitude and longitude), sea conditions (wind force, wave height), ambient temperature and humidity, ship load and other data; System-level variables include total battery pack voltage, 96 groups of single cell voltages, charge and discharge currents, temperatures at each point, SOC, SOH, etc.; Fault layer variables include various fault types and their subtypes such as overcharge, over discharge, internal short circuit, external short circuit, thermal runaway, etc.

[0071] By analyzing the ship's operation logs and fault records over the past three years and combining expert knowledge, a hierarchical causal graph with 114 nodes and 376 edges was initially constructed, establishing a complex correlation network between environmental conditions, battery status and fault types.

[0072] Invariant causal structure learning: Battery operation data were collected on three typical routes (Asia to Europe North, Asia to Europe South, and Asia to the Middle East). Each route contains about 750 hours of continuous records, covering various environmental conditions.

[0073] Through the invariant risk minimization algorithm, a causal structure that is stable across environments is extracted, and the gradient penalty weight coefficient λ is set to 0.1.

[0074] After 40 iterations of training, the model successfully identified the causal relationships of battery failures that remained stable under environmental changes.

[0075] Causal strength assessment and core skeleton extraction: The mutual information theory was used to assess the strength of causal edges, and 46 key causal edges with significant strength and stability in different environments were identified. These causal edges constitute the core causal skeleton, which only accounts for 12.2% of the original causal edges, but contains more than 85% of the key diagnostic information. For example, it was found that the causal relationship strength between abnormal internal battery temperature and thermal runaway failure remained above 0.72 in all environments, which is one of the most stable diagnostic bases.

[0076] Multi-source environmental data enhancement: For the rare high-temperature failures on the Nordic route and the rare low-temperature failures on the Middle East route, about 2,000 high-quality virtual samples were synthesized through counterfactual data generation technology. These samples maintain the rationality of the battery's physical properties, while simulating possible failure modes in extreme environments, effectively expanding the model's training data.

[0077] Meta-learning rapid adaptation mechanism: In order to cope with the situation of ships entering new routes, a meta-learning framework was built to achieve rapid adaptation to the new environment. When entering the Baltic Sea route for the first time, the system only collected 12 hours of operating data (about 240 samples), and quickly adapted to the characteristics of the new environment through the meta-learning mechanism, and the fault diagnosis accuracy was improved from the initial 76% to 94%.

[0078] Technical effect verification: Improved diagnostic accuracy: The accuracy of fault root cause identification has been significantly improved in the test of each route. The following is a comparison with the traditional diagnostic method:

[0079] Improved environmental adaptability: The method of this application has significantly improved adaptability to different environmental changes, especially when encountering extreme environmental changes. The following are the test results of environmental adaptability:

[0080] This application example fully verifies the effectiveness and superiority of the deep learning-based marine lithium battery BMS control method proposed in this application in the actual ship operation environment. Especially in the face of complex and changeable marine environments, this method can accurately identify the root cause of faults, quickly adapt to environmental changes, and improve the safety and reliability of ship battery systems.

[0081] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation mode. The above-mentioned specific implementation mode is merely illustrative and not restrictive. Under the guidance of this embodiment, ordinary technicians in this field can also make more forms of equivalent embodiments, all of which are within the protection of this embodiment.

Claims

1. A marine lithium battery BMS control method based on deep learning, characterized in that: The following steps are involved: Construct a three-layer causal structure including the environment layer, system layer and fault layer to form a hierarchical causal graph; Based on the hierarchical causal graph, we use the invariant risk minimization algorithm with multi-environment data and gradient penalty to learn stable causal structures. For the learned stable causal structure, based on the mutual information theory, the strength of the causal edge is evaluated and a stable causal skeleton is extracted; Based on the extracted stable causal skeleton, we construct an enhanced dataset across environments through counterfactual data generation technology. By utilizing the constructed cross-environment enhanced dataset, a meta-learning framework is constructed to enable the model to quickly adapt to the new environment.

2. According to a deep learning-based marine lithium battery BMS control method according to claim 1, it is characterized in that: The steps of constructing a three-layer causal structure including an environment layer, a system layer, and a fault layer include: Construct an environment layer variable set to collect variables related to the ship's operating environment; Construct a system-level variable set to collect the operating parameters of the marine lithium battery system; Construct a set of fault layer variables to define the possible fault types of marine lithium batteries; Based on domain knowledge and historical data, causal relationships between variables are established to form a hierarchical causal graph.

3. A marine lithium battery BMS control method based on deep learning according to claim 2, characterized in that: The environmental layer variables include sea area type, sea state, climate conditions, navigation mission and ship load status.

4. A marine lithium battery BMS control method based on deep learning according to claim 2, characterized in that: The system-level variables include battery pack voltage, single cell voltage, charge and discharge current, surface temperature, internal temperature, state of charge, and state of health.

5. A marine lithium battery BMS control method based on deep learning according to claim 2, characterized in that: The fault layer variables include overcharge, over discharge, internal short circuit, external short circuit, thermal runaway, insulation failure and balance circuit failure.

6. A marine lithium battery BMS control method based on deep learning according to claim 1, characterized in that: The steps of learning a stable causal structure include: Construct multi-environment datasets; Constructing a deep learning model in which a subset of parameters related to system-level variables are represented as system-level parameters; Design a gradient-penalized invariant risk minimization learning objective; Use the optimization algorithm to optimize the objective function and obtain the model parameters that meet the invariance.

7. A marine lithium battery BMS control method based on deep learning according to claim 1, characterized in that: The step of evaluating the strength of the causal edge comprises: Calculate mutual information between variables; Calculate conditional mutual information; Compute the strength of causal edges; Based on the causal strength threshold, causal edges with strength greater than the threshold are extracted to form a core causal skeleton.

8. A marine lithium battery BMS control method based on deep learning according to claim 1, characterized in that: The steps of constructing a cross-environment enhanced dataset include: Based on the core causal framework, construct an intervention model; Construct a counterfactual data generator to generate counterfactual samples; For each environment, select representative samples and generate a set of counterfactual samples; The original dataset is merged with the counterfactual dataset to form an augmented dataset.

9. A marine lithium battery BMS control method based on deep learning according to claim 1, characterized in that: The steps of constructing the meta-learning framework include: Treat fault diagnosis in each environment as a task and define a set of meta-learning tasks; Based on the trained model parameters, they are used as the initial parameters for meta-learning; For new environments, use a small amount of support set data to update parameters through gradient descent; The adapted model performance is evaluated based on the query set and the initial parameters are updated via meta-optimization.

10. A deep learning-based marine lithium battery BMS control system, used to execute a deep learning-based marine lithium battery BMS control method according to any one of claims 1 to 9, characterized in that: include: A multi-level causal structure building module is used to build a three-level causal structure including the environment layer, system layer and fault layer; Invariant causal structure learning module, used to learn stable causal structures; The causal strength assessment and core skeleton extraction module is used to assess the strength of causal edges and extract stable causal skeletons; Multi-source environment data enhancement module, used to build enhanced datasets across environments; The meta-learning fast adaptation module is used to enable the model to quickly adapt to the new environment.

Images