Distributed fault prediction system of military ship control unit

Abstract

The invention provides a distributed fault prediction system for a military ship control unit, and relates to the technical field of fault prediction, and the system comprises an architecture building module which is used for building a distributed fault prediction architecture; the feature extraction module is used for acquiring a sub-control unit working signal set and performing fault feature extraction on the sub-control unit working signal set; the path analysis module is used for mining to obtain a fault association rule base and activating the regional aggregation node to perform fault path analysis on the edge node fault feature set based on the fault association rule base; and the fault operation and maintenance module is used for calling the military ship fault knowledge base and performing integrated prediction on the distributed regional fault path based on the military ship fault knowledge base to obtain a fault integrated prediction result, and the military ship control unit performs fault early warning operation and maintenance based on the fault integrated prediction result. According to the invention, the technical problem of insufficient fault identification precision in the prior art can be solved, and the technical effect of improving the fault early warning accuracy is achieved.

Description

Technical Field

[0001] The present application relates to the field of fault prediction technology, and in particular to a distributed fault prediction system for military ship control units. Background Art

[0002] With the increasing complexity of modern military ship systems, their control units have become highly complex in terms of structural hierarchy, functional distribution and information interaction. Ensuring their efficient and reliable operation has become the key to the intelligent development of military equipment.

[0003] Currently, in the fault prediction and diagnosis of shipborne systems, traditional methods mostly rely on centralized data collection and rule-matching analysis strategies, that is, all monitoring signals are uniformly processed and analyzed through a single control center. This approach has obvious limitations in coping with the massive growth of information, the increasing real-time requirements, and the fusion of multi-source heterogeneous data.

[0004] In summary, the existing technology has technical problems such as reliance on a centralized data processing architecture for shipborne system fault prediction and a lack of regional-level collaborative fault modeling and dynamic knowledge updating capabilities, which leads to response delays and insufficient fault identification accuracy in a multi-source heterogeneous data environment, further affecting the real-time fault warning efficiency and intelligent operation and maintenance capabilities of the entire ship control system. Summary of the Invention

[0005] The purpose of this application is to provide a distributed fault prediction system for military ship control units to solve the technical problems in the existing technology, such as the reliance of shipborne system fault prediction on a centralized data processing architecture and the lack of regional-level collaborative fault modeling and dynamic knowledge updating capabilities, which leads to response delays and insufficient fault identification accuracy in a multi-source heterogeneous data environment, further affecting the real-time fault warning efficiency and intelligent operation and maintenance capabilities of the entire ship control system.

[0006] In view of the above problems, the present application provides a distributed fault prediction system for a military ship control unit, including: an architecture building module, used to build a distributed fault prediction architecture based on the military ship control unit, the distributed fault prediction architecture including edge computing nodes, regional aggregation nodes and central decision nodes; a feature extraction module, used to collect and obtain the sub-control unit working signal set through the edge computing node, extract fault features of the sub-control unit working signal set, and obtain the edge node fault feature set; a path analysis module, used to mine and obtain a fault association rule base, activate the regional aggregation node to perform fault path analysis on the edge node fault feature set based on the fault association rule base, and determine the distributed regional fault path; a fault operation and maintenance module, used to call the military ship fault knowledge base through the central decision node, perform integrated prediction on the distributed regional fault path based on the military ship fault knowledge base, and obtain a fault integrated prediction result, and the military ship control unit performs fault warning operation and maintenance based on the fault integrated prediction result.

[0007] Preferably, the distributed fault prediction system of the military ship control unit also includes: a sub-unit disassembly unit, which is used to disassemble the military ship control unit into sub-units to obtain a sub-control unit set; an edge computing node acquisition unit, which is used to perform sensor deployment and computing resource configuration in sequence based on the monitoring demand information of the sub-control unit set to obtain edge computing nodes; a regional division unit, which is used to collect and obtain historical fault data sets of military ships, and use the historical fault data sets of military ships to divide the edge computing nodes into regions to determine regional aggregation nodes; a global federated learning unit, which is used to perform global federated learning based on the regional aggregation nodes to obtain a central decision node, and build the distributed fault prediction architecture based on the edge computing nodes, the regional aggregation nodes and the central decision nodes.

[0008] Preferably, the distributed fault prediction system of the military ship control unit also includes: a fault label data set acquisition unit, which is used to normalize and add fault labels to the military ship historical fault data set to obtain a military ship fault label data set; a fault correlation feature set acquisition unit, which is used to extract correlation features based on the military ship fault label data set to obtain a military ship fault correlation feature set; a clustering result acquisition unit, which is used to perform K-means spatial clustering on the military ship fault correlation feature set to obtain a sub-control unit clustering result; a computing power resource allocation unit, which is used to allocate computing power resources based on the amount of correlation data of the sub-control unit clustering result to determine the regional aggregation node.

[0009] Preferably, the distributed fault prediction system of the military ship control unit also includes: an association and diversion unit, which is used to perform association and diversion on the military ship historical fault data set based on the regional aggregation node to obtain a regional node historical fault data set; a fault identification training unit, which is used to use a deep neural network to perform fault identification training on the regional node historical fault data set to generate a regional node fault prediction network set; a fault prediction twin acquisition unit, which is used to extract the model parameters of the regional node fault prediction network set for global federated learning to obtain a global node fault prediction twin; a verification and tuning storage unit, which is used to verify, tune and store the global node fault prediction twin to obtain the central decision node.

[0010] Preferably, the distributed fault prediction system of the military ship control unit also includes: a program acquisition unit, used to obtain a node preprocessing calculation program and a node feature extraction calculation program according to the edge computing node; a standardization preprocessing unit, used to use the node preprocessing calculation program to perform standardization preprocessing on the sub-control unit working signal set to obtain a standard sub-control unit working signal set; a sub-unit associated working feature set acquisition unit, used to perform associated feature extraction on the standard sub-control unit working signal set based on the node feature extraction calculation program to obtain a sub-unit associated working feature set; an edge node fault feature set acquisition unit, used to perform fault feature extraction on the sub-unit associated working feature set according to the sub-unit normal working threshold to obtain the edge node fault feature set.

[0011] Preferably, the distributed fault prediction system of the military ship control unit also includes: a fault acquisition unit, used to identify the fault type and extract the fault characteristics of the historical fault data set of the military ship, and obtain a fault type data set and a fault characteristic data set; a timestamp alignment unit, used to perform timestamp alignment and association analysis on the fault type data set and the fault characteristic data set, and construct a fault type-feature association table; an association rule mining unit, used to perform association rule mining based on the fault type-feature association table, and obtain a fault association rule base.

[0012] Preferably, the distributed fault prediction system of the military ship control unit also includes: a format conversion unit, which is used to discretize and convert the fault type-feature association table into a format to obtain a fault-related feature transaction data set; a frequent item set mining unit, which is used to set a confidence benchmark threshold, and use the confidence benchmark threshold to perform frequent item set mining and association rule generation on the fault-related feature transaction data set to obtain the fault association rule base.

[0013] Preferably, the distributed fault prediction system of the military ship control unit also includes: an associated regional node set acquisition unit, which is used to activate the regional aggregation node according to the edge node fault feature set to obtain the associated regional node set; a regional matching fault rule set acquisition unit, which is used to call the fault association rule library through the associated regional node set to perform fault diagnosis on the associated regional node set to obtain the regional matching fault rule set; a distributed regional fault path determination unit, which is used to perform propagation path deduction based on the regional matching fault rule set to determine the distributed regional fault path.

[0014] Preferably, the distributed fault prediction system of the military ship control unit also includes: a fusion processing unit, which is used to fuse the distributed regional fault paths to construct a global fault path diagram; an associated path description unit, which is used to extract fault modes and describe associated paths on the military ship fault knowledge base to generate a fault path pattern prediction network; and an integrated prediction unit, which is used to perform integrated prediction on the global fault path diagram based on the fault path pattern prediction network to obtain the fault integrated prediction result.

[0015] Preferably, the distributed fault prediction system of the military ship control unit also includes: a simulation verification unit, which is used to use the global node fault prediction twin to simulate and verify the fault integrated prediction result to obtain a fault prediction accuracy parameter; and an adaptive learning optimization unit, which is used to adaptively learn and optimize the fault path pattern prediction network based on the fault prediction accuracy parameter.

[0016] The technical solution provided in this application has at least the following technical effects or advantages: by realizing the technical goal of building a multi-level collaborative distributed fault prediction architecture and realizing edge-region-center node linkage perception and intelligent judgment, the technical effect of improving fault detection response speed, enhancing model generalization ability and real-time warning accuracy is achieved.

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0019] Figure 1 This is a structural diagram of the distributed fault prediction system for the military ship control unit of this application.

[0020] Figure 2 This is a structural diagram of the architecture building module in the distributed fault prediction system of the military ship control unit in this application.

[0021] Explanation of the accompanying symbols: architecture building module 1, feature extraction module 2, path analysis module 3, fault operation and maintenance module 4, sub-unit disassembly unit 11, edge computing node acquisition unit 12, area division unit 13, global federated learning unit 14. DETAILED DESCRIPTION

[0022] This application provides a distributed fault prediction system for military ship control units, resolving the existing technical issues of shipboard system fault prediction relying on a centralized data processing architecture and lacking the ability to model regional collaborative faults and dynamically update knowledge. This leads to response delays and insufficient fault identification accuracy in a multi-source heterogeneous data environment, further impacting the real-time fault warning efficiency and intelligent operation and maintenance capabilities of the entire ship control system. This system achieves the technical goals of building a multi-level collaborative distributed fault prediction architecture and realizing edge-region-center node linkage perception and intelligent judgment, thereby improving fault detection response speed, enhancing model generalization capabilities, and real-time warning accuracy.

[0023] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0024] Please see the attached Figure 1 This application provides a distributed fault prediction system for military ship control units, specifically including:

[0025] Architecture building module 1 is used to build a distributed fault prediction architecture based on the military ship control unit. The distributed fault prediction architecture includes edge computing nodes, regional aggregation nodes and central decision nodes.

[0026] Specifically, a military ship control unit is a control module used to manage and coordinate the operation of various mission systems within a ship, such as the power system, navigation system, and weapon control system. Each of these modules is composed of different sub-control units. Due to the complex working environment and high-intensity tasks, sub-control units are prone to potential failures. To achieve real-time monitoring and early fault prediction of sub-control units, it is necessary to build a systematic and hierarchical prediction framework to obtain a distributed fault prediction architecture, thereby improving response speed and system robustness. The distributed fault prediction architecture is deployed in a physical location close to the specific sub-control unit and is responsible for collecting sensor data and performing pre-processing and preliminary feature extraction, such as determining whether current and voltage are abnormal and identifying local vibration patterns.

[0027] Next, multiple edge computing nodes upload their extracted fault signature data to the regional aggregation node. The regional aggregation node is responsible for integrating and analyzing fault signatures within the same area, further establishing regional-level fault correlations and propagation paths. The central decision node is responsible for integrating the fault path information uploaded by multiple regional aggregation nodes, ultimately generating a unified fault prediction result for the entire ship.

[0028] The feature extraction module 2 is used to collect and obtain the sub-control unit working signal set through the edge computing node, extract fault features from the sub-control unit working signal set, and obtain an edge node fault feature set.

[0029] Specifically, edge computing nodes are used to collect and acquire the sub-control unit's operating signal set, and computing equipment deployed near the control unit of a military ship is used to collect and locally process the operating signals generated by each sub-control unit in real time. A sub-control unit refers to the smallest independent functional module within a military ship's control unit, such as those used to control propulsion, power distribution, and communications. The operating signal set is the physical or logical output data of the sub-control unit in normal or abnormal conditions, such as voltage, current, frequency, and operating status code.

[0030] Fault feature extraction is performed on the sub-control unit working signal set, and feature analysis is performed on the collected original signals to discover characteristic values or indicators closely related to potential fault states, including signal amplitude mutations beyond the normal fluctuation range, continuous oscillations, zero drift, spectral distortion, etc. The extraction method can adopt methods such as time domain statistical analysis, frequency domain filtering, principal component analysis or machine learning models, so as to convert the non-directly identifiable content in the original signal into indicators that can be used to judge the equipment status, and obtain the edge node fault feature set for subsequent fault diagnosis, predictive analysis and federated model training.

[0031] The path analysis module 3 is used to mine and obtain a fault association rule base, activate the regional convergence node to perform fault path analysis on the edge node fault feature set based on the fault association rule base, and determine the distributed regional fault path.

[0032] Specifically, mining and obtaining a fault association rule base refers to extracting a set of rules that can represent the causal relationship and frequent associations between different fault modes through in-depth analysis of the statistical relationship between fault types and related features in a large amount of historical fault data of military ships. Activating a regional convergence node means that when the fault characteristics reported by the edge computing node meet the triggering conditions of certain rules in the fault association rule base, the regional computing center is triggered to conduct a joint analysis of the status of multiple edge nodes in the relevant area. A regional convergence node is a type of mid-level node that performs data integration and computational processing on multiple edge nodes. It is used to collaboratively process complex events and fault trends across nodes to improve the comprehensiveness and accuracy of judgments.

[0033] Based on the fault association rule library, the fault path analysis of the edge node fault feature set is performed, and the rules in the fault association rule library are used to match and deduce the fault features uploaded by the current edge node, so as to identify the possible propagation path or chain reaction structure of a certain fault. Determining the distributed regional fault path means regionally integrating the above path analysis results to identify the fault evolution line composed of multiple sub-control units and regional nodes in the entire military ship system.

[0034] The fault operation and maintenance module 4 is used to call the military ship fault knowledge base through the central decision node, perform integrated prediction on the distributed regional fault path based on the military ship fault knowledge base, and obtain a fault integrated prediction result. The military ship control unit performs fault warning operation and maintenance based on the fault integrated prediction result.

[0035] Specifically, the military ship fault knowledge base, which stores a large amount of historical fault information, rules, and patterns, is accessed and called through the central decision-making node. The military ship fault knowledge base contains information such as various fault types, occurrence conditions, corresponding characteristics, propagation paths, and repair measures. Based on the military ship fault knowledge base, distributed regional fault paths are integrated and predicted. The logical associations between historical fault patterns and characteristics in the fault association rule base are utilized to comprehensively evaluate and determine the trends of distributed regional fault paths currently uploaded from multiple regions. The integrated prediction process integrates the commonalities, differences, and possible coupling effects between multiple paths to output intelligent prediction results for future system states. Modeling is typically performed using methods such as Bayesian networks, graph neural networks, and logistic regression integration to obtain integrated fault prediction results. The output includes a set of prediction results including information such as the probability of fault occurrence, potential impact range, and propagation trend paths, reflecting the severity and urgency of the risks faced by different control units.

[0036] The control unit of a military ship conducts fault warning and operation and maintenance based on the results of integrated fault prediction, and takes response measures in advance, such as switching to the backup system, reducing the operating load, issuing operation and maintenance warnings, etc., to prevent or alleviate possible impending failures, achieve proactive prevention, reduce losses, and improve system stability.

[0037] Furthermore, the present application also includes: a sub-unit disassembly unit 11, which is used to disassemble the military ship control unit into sub-units to obtain a sub-control unit set; an edge computing node acquisition unit 12, which is used to perform sensor deployment and computing resource configuration in sequence based on the monitoring demand information of the sub-control unit set to obtain edge computing nodes; a regional division unit 13, which is used to collect and obtain a historical fault data set of military ships, and use the historical fault data set of military ships to divide the edge computing nodes into regions to determine regional aggregation nodes; a global federated learning unit 14, which is used to perform global federated learning based on the regional aggregation nodes to obtain a central decision node, and build the distributed fault prediction architecture based on the edge computing nodes, the regional aggregation nodes and the central decision nodes.

[0038] Specifically, the control unit of a military ship is disassembled into sub-units, and the automatic control system of the entire ship is subdivided and decomposed according to its functional structure to form several relatively independent modular units with specific control responsibilities, such as propulsion control sub-units, cabin environment adjustment sub-units, weapon launch sub-units, etc., which constitute a sub-control unit collection, so that subsequent monitoring, analysis and operation and maintenance work can be carried out on specific functional blocks, thereby improving system management accuracy and response efficiency.

[0039] Next, based on the monitoring requirements of each sub-control unit, the corresponding type and number of sensors are deployed one by one, and localized computing resources are allocated to build edge computing nodes. Monitoring requirements include sampling frequency, data dimensions, and real-time requirements.

[0040] Subsequently, a historical fault dataset from past military vessel operations was collected and organized, documenting the types, characteristics, and evolution of faults occurring in different sub-control units under specific conditions. This dataset was then used to regionalize edge computing nodes. Edge nodes with similar functions, high data correlation, or similar historical fault types were grouped into a region. Within this region, nodes with stronger computing and communication capabilities were selected as regional aggregation nodes to determine the regional aggregation nodes.

[0041] Then, through the use of regional aggregation nodes, their local model parameters are uploaded to participate in global federated learning, thus establishing a unified central decision node. Federated learning is a distributed machine learning method that enables collaborative model training without centralizing raw data, protecting data privacy while enabling knowledge fusion. The central decision node aggregates the knowledge of each regional model to form a global prediction model, ultimately coordinating the fault prediction task for the entire military ship control unit.

[0042] Furthermore, the present application also includes: a fault label data set obtaining unit, which is used to normalize and add fault labels to the historical fault data set of military ships to obtain a military ship fault label data set; a fault association feature set obtaining unit, which is used to extract association features based on the military ship fault label data set to obtain a military ship fault association feature set; a clustering result obtaining unit, which is used to perform K-means spatial clustering on the military ship fault association feature set to obtain a sub-control unit clustering result; a computing power resource allocation unit, which is used to allocate computing power resources based on the amount of associated data of the sub-control unit clustering result to determine the regional aggregation node.

[0043] Specifically, the historical military ship fault dataset was normalized and fault labels were added. The collected raw fault data of different dimensions and dimensions was mathematically standardized. For example, values such as temperature, voltage, and vibration amplitude were converted to dimensionless values between 0 and 1, making different features comparable and on a uniform scale. Furthermore, clear fault type labels, such as short circuit, overheating, and seizure, were added to each data record in the historical military ship fault dataset. This resulted in a dataset of military ship fault labels, enabling subsequent machine learning algorithms to conduct supervised learning based on these labels.

[0044] Next, using the military ship fault label dataset, key features closely related to fault occurrence can be extracted through statistical analysis, information entropy screening, or model training, forming a set of associated features for military ship faults. Associated features are sensor signals or system state values that have historically co-occurred with specific fault labels, such as the significant correspondence between a specific pressure increase and a valve sticking fault.

[0045] Subsequently, a K-means spatial clustering operation is performed on the extracted fault-correlation feature set. K-means is an unsupervised clustering algorithm that can partition high-dimensional data into several closely related subgroups in feature space. This K-means spatial clustering operation can automatically group sub-control units with similar military ship fault-correlation features together, forming a sub-control unit clustering result.

[0046] Finally, computing resources are allocated based on the amount of associated data contained in each sub-cluster of the sub-control unit clustering results. Data pressure and computing requirements are comprehensively assessed based on indicators such as the number of nodes in each sub-cluster, data transmission frequency, and historical failure frequency. Computing resources are then dynamically allocated to each cluster, determining the regional aggregation nodes that will undertake the convergence analysis tasks.

[0047] Furthermore, the present application also includes: an association and diversion unit, which is used to associate and divert the military ship historical fault data set based on the regional aggregation node to obtain a regional node historical fault data set; a fault identification training unit, which is used to use a deep neural network to perform fault identification training on the regional node historical fault data set to generate a regional node fault prediction network set; a fault prediction twin acquisition unit, which is used to extract the model parameters of the regional node fault prediction network set for global federated learning to obtain a global node fault prediction twin; a verification and tuning storage unit, which is used to verify, tune and store the global node fault prediction twin to obtain the central decision node.

[0048] Specifically, the historical fault dataset of military ships is associated and distributed based on regional aggregation nodes. This data is categorized according to the functional area to which it belongs and divided into various aggregation nodes to obtain regional node historical fault datasets. Association and distribution utilizes factors such as time synchronization, physical location, signal similarity, or functional coupling between fault data to establish mapping relationships between data. This data is then divided into logically consistent regional nodes, making the data managed by each node more centralized and domain-specific, thus facilitating subsequent independent training.

[0049] Next, a deep neural network, a multilayer perceptron architecture, possesses nonlinear representation and automatic feature extraction capabilities. A deep neural network is used to train fault identification on a dataset of regional node historical faults to identify potential fault signature patterns within the data. During training, the dataset is first divided into a training set, a validation set, and a test set based on time or ratio, for example, in a ratio of 7:2:1. Based on the temporal and spatial correlation of the dataset, a deep neural network architecture is designed, consisting of an input layer, multiple hidden layers, and an output layer. The input layer receives the training set, while the hidden layers use activation functions such as ReLU to construct nonlinear mappings. The output layer uses a Softmax activation function to output the probability distribution of each fault. The training set is used for forward propagation to calculate the output, using a cross-entropy loss or mean squared error loss. Each training iteration is called an epoch, with a range of 50 to 200 epochs, and a preset number of iterations to prevent overfitting. Performance metrics such as accuracy are evaluated on the validation set after each training round. After completing training of the basic deep neural network model, a regional node fault prediction network set is obtained.

[0050] Subsequently, model parameters are extracted from the regional node fault prediction network, and a global federated learning operation is performed, ultimately constructing a global node fault prediction twin. Federated learning is a distributed collaborative modeling approach in which each regional node completes model training locally, uploading only model parameters rather than raw data, effectively protecting sensitive information. During the global federated learning process, all model parameters are aggregated using weighted averaging, difference constraints, and global consistency optimization methods to construct a prediction model that represents the overall operating status and fault trends. This is the global node fault prediction twin, which truly reflects the response pattern of the control system under different states.

[0051] Finally, the global node fault prediction twin is verified, tuned, and stored to test its prediction accuracy and generalization capabilities, ensuring its applicability in real-world decision-making scenarios. The verification and tuning process includes test set evaluation, error backpropagation correction, and parameter refinement. After passing the dual standards of stability and accuracy, the global node fault prediction twin is officially deployed as a central decision-making node. The central decision-making node possesses a global perspective and coordinates the prediction results of nodes in each region, enabling multi-regional coordinated judgment and ship-wide fault warning decisions.

[0052] Furthermore, the present application also includes: a program obtaining unit, used to obtain a node preprocessing calculation program and a node feature extraction calculation program according to the edge computing node; a standardization preprocessing unit, used to use the node preprocessing calculation program to perform standardization preprocessing on the sub-control unit working signal set to obtain a standard sub-control unit working signal set; a sub-unit associated working feature set obtaining unit, used to perform associated feature extraction on the standard sub-control unit working signal set based on the node feature extraction calculation program to obtain a sub-unit associated working feature set; an edge node fault feature set obtaining unit, used to perform fault feature extraction on the sub-unit associated working feature set according to the sub-unit normal working threshold to obtain the edge node fault feature set.

[0053] Specifically, edge computing nodes are miniature intelligent computing modules deployed in the local control areas of military vessels. They possess certain data processing and model calculation capabilities and can independently perform some tasks without relying on central servers. Depending on the task type and number of sensors connected to each sub-control unit, each edge node can be assigned two types of computing programs: one is a node preprocessing program, which performs basic processing on raw data, such as cleaning, alignment, and interpolation; the other is a node feature extraction program, which extracts key features related to the equipment's health status from the preprocessed data, such as fluctuation frequency, rate of increase, and mean change.

[0054] Next, the node preprocessing algorithm performs standardization on the sub-control unit's operating signal set. This refers to the raw data collected by the edge computing node, such as temperature, voltage, and pressure. Standardization eliminates differences in scale, unit, and sampling frequency between signals, allowing subsequent models to fairly process all data features.

[0055] The node feature extraction algorithm then performs feature extraction on the standardized signal set, extracting correlation features that reflect the operational status of the control unit in a military vessel. Correlation features refer to the coupling behavior between multiple signals, such as whether a current change is always accompanied by a temperature increase, or whether a certain pressure anomaly always occurs after a voltage drop. Extracting correlation features can reveal potential patterns of coordinated changes in the control unit in a military vessel, helping to build more interpretable health models.

[0056] Subsequently, fault signatures are extracted from the associated features based on the preset normal operating threshold of the subunit. The normal operating threshold of the subunit is a stable operating range defined by experience or historical data. Once a feature exceeds this range, it is marked as an edge node fault signature set.

[0057] Furthermore, the present application also includes: a fault acquisition unit, which is used to identify the fault type and extract the fault features of the historical fault data set of the military ship to obtain a fault type data set and a fault feature data set; a timestamp alignment unit, which is used to perform timestamp alignment and association analysis on the fault type data set and the fault feature data set to construct a fault type-feature association table; an association rule mining unit, which is used to perform association rule mining based on the fault type-feature association table to obtain a fault association rule library.

[0058] Specifically, the historical fault dataset for military vessels refers to all relevant data recorded when a fault event occurs during the operation of a military vessel, including sensor signals, alarm records, maintenance logs, and other information before and after the fault occurs. The dataset is used to identify fault types, classifying different fault phenomena into several distinct fault categories, such as "main motor overload," "hydraulic system leakage," and "control module failure." Fault feature extraction is performed on the dataset, identifying data features from the raw signals that represent fault behavior, such as sudden temperature rise, voltage drop, or abnormal vibration frequency. This allows the construction of a feature set that describes the behavioral characteristics of each type of fault. Ultimately, both a fault type dataset and a fault feature dataset are obtained.

[0059] Next, the fault type dataset and the fault feature dataset are timestamp aligned and analyzed for correlation. Timestamp alignment unifies the time information recorded in the two datasets, allowing fault events occurring at a specific moment to be mapped to feature changes extracted at that time. For example, if a "pressure sensor anomaly" is recorded at a specific moment, the feature changes at that moment can be traced. Correlation analysis involves data mining the aligned records to identify features whose change patterns consistently precede specific fault types, thereby forming a data foundation for "strongly correlated faults and features." Ultimately, a fault type-feature correlation table is constructed. Table 1 shows a partial record of the most recently obtained fault type-feature correlation table.

[0060] Table 1: Partial records of the most recently obtained fault type-feature association table

[0061]

[0062] Furthermore, association rule mining is performed based on the fault type-feature association table to generate a fault association rule base. Association rule mining is a data mining technique that aims to discover potential causal or co-occurrence relationships between multiple variables. For example, it can be used to infer that "if feature A and feature B are detected simultaneously, fault type C is very likely to occur." It can transform complex data into a set of high-probability rules, namely the fault association rule base.

[0063] Furthermore, the present application also includes: a format conversion unit, which is used to discretize and convert the features of the fault type-feature association table to obtain a fault-related feature transaction data set; a frequent item set mining unit, which is used to set a confidence benchmark threshold, and use the confidence benchmark threshold to perform frequent item set mining and association rule generation on the fault-related feature transaction data set to obtain the fault association rule base.

[0064] Specifically, the fault type-feature association table is discretized, and the original continuous numerical features (such as temperature, voltage, vibration frequency, etc.) are converted into several discrete intervals or classification labels. For example, the temperature can be divided into three discrete intervals: "below 50 degrees Celsius", "50 to 80 degrees Celsius", and "above 80 degrees Celsius". The purpose of feature discretization is to enable the subsequent data mining process to handle discrete values, thereby improving computational efficiency and facilitating rule extraction. Then, format conversion is performed, and the discretized features are organized according to the input format of the association rule mining algorithm, and each piece of data is represented as a transaction (that is, a group of feature items that appear at the same time), thereby constructing a fault association feature transaction data set.

[0065] Next, setting a confidence threshold involves customizing a minimum confidence level based on actual conditions by skilled practitioners to filter out unreliable rules generated during data mining. Confidence is a measure of a rule's reliability, indicating the probability of the target fault occurring given the presence of the prerequisite features. For example, if the rule "large current fluctuations → motor overheating" holds true in 80% of samples, the rule's confidence level is 80%. Setting a confidence threshold of 70%, for example, means that only rules with a confidence level of 70% or higher will be retained. Frequent item set mining is performed on the fault-related feature transaction dataset to identify frequently occurring feature combinations, such as "high temperature + high pressure" and "high vibration + abnormal voltage." Specific association rules are then derived from the frequent item sets, such as "if high temperature and high vibration occur, the probability of a control module failure is 85%." This ultimately forms a fault association rule base, a collection of predictive rules that meet the required confidence levels.

[0066] Furthermore, the present application also includes: an associated regional node set obtaining unit, used to activate the regional aggregation node according to the edge node fault feature set to obtain the associated regional node set; a regional matching fault rule set obtaining unit, used to call the fault association rule library through the associated regional node set to perform fault diagnosis on the associated regional node set to obtain the regional matching fault rule set; a distributed regional fault path determination unit, used to perform propagation path deduction based on the regional matching fault rule set to determine the distributed regional fault path.

[0067] Specifically, regional aggregation nodes are activated based on the edge node fault signature set. When an edge node detects a representative fault signature, it triggers its associated regional aggregation nodes to participate in analysis and processing, thereby obtaining a set of associated regional nodes. Regional aggregation nodes are intermediate nodes that centrally manage, process, and judge data uploaded by multiple edge nodes.

[0068] Next, the fault association rule base is called up through the associated regional node set for fault diagnosis. This system's pre-established fault association rule base is then queried to match the observed characteristic patterns. An associated regional node set refers to a collection of geographically or functionally related nodes that may jointly influence or mutually transmit a particular type of fault. The fault diagnosis process involves determining whether the current observation matches a specific fault pattern through rule matching, thereby outputting a regional matching fault rule set.

[0069] Subsequently, propagation path deduction is performed based on the regional matching fault rule set, analyzing the possible spread of the fault and forming a path diagram from a certain starting point to multiple affected points. Propagation path deduction relies on the connection relationship between control units, the characteristic impact relationship, and the time sequence. For example, a cooling system fault may first affect the adjacent temperature control module and then spread to the energy supply module. The distributed regional fault path is the finalized fault propagation path composed of multiple nodes and a causal chain structure, which is used to guide emergency response or precision maintenance.

[0070] Furthermore, the present application also includes: a fusion processing unit, which is used to fuse the distributed regional fault paths and construct a global fault path map; an associated path description unit, which is used to extract fault modes and describe associated paths for the military ship fault knowledge base to generate a fault path pattern prediction network; and an integrated prediction unit, which is used to perform integrated prediction on the global fault path map based on the fault path pattern prediction network to obtain the fault integrated prediction result.

[0071] Specifically, distributed regional fault paths are fused to construct a global fault path map, integrating the fault propagation paths derived from convergence nodes in each region. Fusion analysis analyzes the overlapping nodes, associated logic, and propagation order of the causal chain between different fault nodes in multiple local regions, piecing together fragmented information into a complete global map to obtain a global fault path map.

[0072] Next, the military ship fault knowledge base is used to extract fault patterns and describe associated paths. Leveraging the extensive historical fault data, typical and representative fault evolution patterns are identified and their propagation paths within the system structure are annotated. The military ship fault knowledge base is a structured database containing fault records, feature descriptions, and processing results collected from past operations. Fault pattern extraction uses statistical analysis or machine learning methods to identify regular combinations. Associated path descriptions complement the spatial, functional, or temporal propagation relationships between nodes, ultimately constructing a fault path pattern prediction network.

[0073] Based on the fault path pattern prediction network, an integrated prediction is performed on the global fault path map. The extracted typical patterns are compared with the current global path map to assess possible future fault evolution trends. Integrated prediction combines historical empirical knowledge with current observational data to produce more accurate and robust prediction results. The integrated fault prediction result is a comprehensive assessment that reflects the severity of the current fault state, the scope of impact, and the potential subsequent risks.

[0074] Furthermore, the present application also includes: a simulation verification unit, which is used to use the global node fault prediction twin to simulate and verify the fault integrated prediction result to obtain a fault prediction accuracy parameter; and an adaptive learning optimization unit, which is used to adaptively learn and optimize the fault path pattern prediction network based on the fault prediction accuracy parameter.

[0075] Specifically, the global node fault prediction twin is used to simulate and verify the integrated fault prediction results to assess their accuracy. The simulation verification process includes setting fault trigger conditions, restoring sensor behavior, and simulating control responses. By comparing with real historical data or simulation target results, the deviation of the fault prediction is calculated, and the final output is fault prediction accuracy parameters such as mean absolute error and prediction hit rate.

[0076] Then, based on the fault prediction accuracy parameters, the fault path pattern prediction network is adaptively learned and optimized, and targeted structural adjustments or parameter updates are performed, so that the fault path pattern prediction network can be continuously fine-tuned according to its own prediction results. For example, the backpropagation algorithm is used to reduce the loss value, update the association weight, and enhance the weak feature signal processing capability, thereby improving its performance in future prediction tasks.

[0077] To sum up, the distributed fault prediction system for military ship control units provided in this application has the following technical effects: by realizing the technical goals of building a multi-level collaborative distributed fault prediction architecture and realizing edge-region-center node linkage perception and intelligent judgment, the technical effects of improving fault detection response speed, enhancing model generalization capabilities and real-time warning accuracy are achieved.

[0078] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0079] Obviously, those skilled in the art may make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A distributed fault prediction system for military ship control units, characterized by: The system comprises: An architecture building module for building a distributed fault prediction architecture based on a military ship control unit, the distributed fault prediction architecture comprising edge computing nodes, regional aggregation nodes, and a central decision-making node; A feature extraction module is used to collect and obtain a sub-control unit working signal set through the edge computing node, extract fault features from the sub-control unit working signal set, and obtain an edge node fault feature set; A path analysis module is used to mine and obtain a fault association rule base, activate the regional convergence node to perform fault path analysis on the edge node fault feature set based on the fault association rule base, and determine the distributed regional fault path; A fault operation and maintenance module is used to call the military ship fault knowledge base through the central decision node, perform integrated prediction of the distributed regional fault path based on the military ship fault knowledge base, and obtain a fault integrated prediction result. The military ship control unit performs fault early warning operation and maintenance based on the fault integrated prediction result.

2. The distributed fault prediction system for a military ship control unit according to claim 1, characterized in that: The architecture building module includes: A sub-unit disassembling unit, configured to disassemble the military ship control unit into sub-units to obtain a sub-control unit set; An edge computing node acquisition unit, configured to sequentially perform sensor deployment and computing resource configuration based on the monitoring requirement information of the sub-control unit set to obtain an edge computing node; A regional division unit is used to collect and obtain a historical fault data set of military ships, divide the edge computing nodes into regions using the historical fault data set of military ships, and determine regional aggregation nodes; A global federated learning unit is used to perform global federated learning based on the regional aggregation node to obtain a central decision node, and to build the distributed fault prediction architecture based on the edge computing node, the regional aggregation node and the central decision node.

3. The distributed fault prediction system for a military ship control unit according to claim 2, characterized in that: The area division unit includes: a fault label data set obtaining unit, configured to perform normalization processing and fault label addition on the military ship historical fault data set to obtain a military ship fault label data set; a fault correlation feature set obtaining unit, configured to extract correlation features based on the military ship fault label dataset to obtain a military ship fault correlation feature set; A clustering result obtaining unit, configured to perform K-means spatial clustering on the military ship fault correlation feature set to obtain a sub-control unit clustering result; A computing power resource allocation unit is used to allocate computing power resources based on the amount of associated data of the clustering results of the sub-control units and determine the regional aggregation node.

4. The distributed fault prediction system for a military ship control unit according to claim 2, characterized in that: The global federated learning unit includes: an association and shunting unit, configured to associate and shunt the military ship historical fault data set based on the regional aggregation node to obtain a regional node historical fault data set; A fault identification training unit is used to perform fault identification training on the regional node historical fault data set using a deep neural network to generate a regional node fault prediction network set; A fault prediction twin acquisition unit is used to extract model parameters of the regional node fault prediction network set for global federated learning to obtain a global node fault prediction twin; The verification and tuning storage unit is used to verify, tune and store the global node fault prediction twin to obtain the central decision node.

5. The distributed fault prediction system for a military ship control unit according to claim 1, characterized in that: The feature extraction module includes: A program obtaining unit, configured to obtain a node preprocessing calculation program and a node feature extraction calculation program according to the edge computing node; a standardization preprocessing unit, configured to perform standardization preprocessing on the sub-control unit working signal set using the node preprocessing calculation program to obtain a standard sub-control unit working signal set; a subunit associated working feature set obtaining unit, configured to extract associated features from the standard sub-control unit working signal set based on the node feature extraction calculation program to obtain a subunit associated working feature set; The edge node fault feature set obtaining unit is used to extract fault features from the subunit associated working feature set according to the subunit normal working threshold to obtain the edge node fault feature set.

6. The distributed fault prediction system for a military ship control unit according to claim 2, characterized in that: The path analysis module includes: A fault acquisition unit is used to identify the fault type and extract the fault features of the historical fault data set of the military ship to obtain a fault type data set and a fault feature data set; A timestamp alignment unit, configured to perform timestamp alignment and correlation analysis on the fault type dataset and the fault feature dataset, and construct a fault type-feature correlation table; The association rule mining unit is used to perform association rule mining based on the fault type-feature association table to obtain a fault association rule library.

7. The distributed fault prediction system for a military ship control unit according to claim 6, characterized in that: The association rule mining unit includes: a format conversion unit, configured to discretize and convert the fault type-feature association table into a format to obtain a fault-related feature transaction data set; The frequent item set mining unit is used to set a confidence benchmark threshold, and use the confidence benchmark threshold to perform frequent item set mining and association rule generation on the fault association feature transaction data set to obtain the fault association rule base.

8. The distributed fault prediction system for a military ship control unit according to claim 1, characterized in that: The path analysis module further includes: an associated regional node set obtaining unit, configured to activate the regional aggregation node according to the edge node fault feature set to obtain an associated regional node set; A regional matching fault rule set obtaining unit is configured to call the fault association rule library through the associated regional node set to perform fault diagnosis on the associated regional node set to obtain a regional matching fault rule set; A distributed regional fault path determination unit is configured to perform propagation path deduction based on the regional matching fault rule set to determine the distributed regional fault path.

9. The distributed fault prediction system for a military ship control unit according to claim 4, characterized in that: The fault operation and maintenance module includes: A fusion processing unit, configured to fuse the distributed regional fault paths to construct a global fault path map; An associated path description unit is used to extract fault modes and describe associated paths from the military ship fault knowledge base to generate a fault path pattern prediction network; An integrated prediction unit is used to perform integrated prediction on the global fault path graph based on the fault path pattern prediction network to obtain the fault integrated prediction result.

10. The distributed fault prediction system for a military ship control unit according to claim 9, characterized in that: The integrated prediction unit comprises: A simulation verification unit, configured to simulate and verify the integrated fault prediction result using the global node fault prediction twin to obtain a fault prediction accuracy parameter; An adaptive learning optimization unit is used to perform adaptive learning optimization on the fault path pattern prediction network based on the fault prediction accuracy parameter.

Images