A control system for managing a secondary fusion circuit breaker

Abstract

The application belongs to the technical field of circuit breaker management, and discloses a control system for managing a primary and secondary fusion circuit breaker. The system collects electrical parameters, mechanical states and environmental conditions of the circuit breaker through multiple channels, and constructs a standardized operation data set. A circuit breaker health feature map is established based on deep feature learning. An environmental adaptability control parameter library is established through environmental factor correlation analysis and multi-scenario simulation. Fault mode identification and predictive diagnosis are performed in combination with the health feature map, and a fault risk early warning matrix is generated. A multi-circuit breaker collaborative control strategy is optimized based on the early warning matrix, and an optimal control instruction sequence is generated. Through safe encryption verification and hierarchical execution mechanism, reliable execution and effect feedback of the control instruction are realized. The application solves the problems of traditional circuit breaker control systems, such as difficult data integration, insufficient environmental adaptability, weak fault prediction ability and imperfect collaborative control, and significantly improves the safety and reliability of power grid operation.

Description

Technical Field

[0001] This invention relates to the field of circuit breaker management technology, and more specifically, to a control system for managing a primary and secondary integrated circuit breaker. Background Technology

[0002] As a critical protection device in power systems, the operating status of circuit breakers directly affects the safe and stable operation of the power grid. With the continuous advancement of smart grid construction, integrated primary and secondary circuit breaker technology has emerged. This technology deeply integrates traditional primary equipment (such as the circuit breaker body and current transformers) with secondary equipment (such as protection devices and measurement and control units), achieving integrated design and application of the equipment. In recent years, the research and application of integrated primary and secondary circuit breakers in the domestic and international power industries have deepened, and related control technologies have also made significant progress.

[0003] However, in complex power grid environments, information barriers between primary and secondary equipment lead to severe data silos, preventing the system from fully grasping the true operating status of circuit breakers. This is especially pronounced during power grid fluctuations, where data asynchrony issues are exacerbated, delaying fault handling. When circuit breakers are in harsh environments such as high altitudes, high humidity, or high pollution, the control system lacks environmental adaptability mechanisms, resulting in a significant increase in maloperation rates. For example, circuit breakers in rainy southern regions frequently experience insulation performance degradation. Traditional fault judgment mechanisms based on simple thresholds are too crude to identify complex fault development patterns, causing many potential faults to be overlooked in their early stages, ultimately leading to large-scale power outages. In large substations with multiple circuit breakers operating collaboratively, the lack of system-level collaborative decision-making capabilities and isolated operation of each circuit breaker result in frequent problems such as uncoordinated protection actions and unreasonable load shedding. The security mechanisms in the control command execution stage are weak, making them particularly vulnerable to the increasing cybersecurity threats in the context of smart grid construction. Furthermore, the lack of an effective execution feedback mechanism prevents the system from optimizing control strategies based on historical experience, causing similar faults to recur, seriously affecting the stable operation of the power grid and the reliability of power supply, and resulting in high equipment maintenance costs.

[0004] In view of this, the present invention proposes a control system for managing primary and secondary fusion circuit breakers to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a control system for managing a primary and secondary integrated circuit breaker, comprising:

[0006] The data acquisition and synchronization module is used to comprehensively collect and synchronize the operating parameters of the primary and secondary integrated circuit breakers to obtain a standardized operating dataset.

[0007] The state perception and feature extraction module is used to perform multi-dimensional state perception and key feature extraction based on the standardized operating dataset to obtain a circuit breaker health feature map.

[0008] The operating environment adaptability assessment module is used to perform operating environment adaptability assessment and parameter adaptive optimization based on the circuit breaker health feature map, and obtain an environmental adaptability control parameter library;

[0009] The fault prediction and diagnosis module is used to perform fault mode identification and predictive diagnosis of the circuit breaker based on the environmental adaptability control parameter library and the circuit breaker health feature map, and obtain a fault risk warning matrix.

[0010] The collaborative control decision module is used to optimize the collaborative control strategy of multiple circuit breakers and generate decisions based on the fault risk warning matrix, so as to obtain the optimal control command sequence.

[0011] The safety execution and feedback module is used to perform secure encryption verification and hierarchical execution of control commands based on the optimal control command sequence, and to provide real-time feedback on the execution effect to obtain a circuit breaker control execution report; the various modules are connected through a secure communication network to realize secure data transmission and sharing between modules.

[0012] The technical effects and advantages of the control system for managing primary and secondary integrated circuit breakers of the present invention are as follows:

[0013] This invention significantly improves the safe and reliable operation of power systems, substantially reducing circuit breaker failure rates and power outage risks through proactive fault warning and intelligent prevention mechanisms. The system shifts from passive response to proactive predictive maintenance, greatly extending equipment lifespan and reducing maintenance costs. Under complex and changing environmental conditions, the system demonstrates excellent adaptability, effectively coping with external interference factors such as temperature fluctuations and changes in pollution levels. Particularly under grid disturbances and extreme weather conditions, the system's stability and reliability are significantly enhanced. Through coordinated optimization control among multiple circuit breakers, the system achieves efficient allocation of grid resources and load balancing, improving power quality and supply reliability. Simultaneously, the system's safe execution mechanism provides multiple safeguards for grid operation, effectively preventing risks of misoperation and malicious attacks. Furthermore, the system's self-learning and continuous optimization capabilities ensure its operational efficiency continuously improves over time. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of a control system for managing a primary and secondary integrated circuit breaker according to the present invention;

[0015] Figure 2 This is a detailed flowchart illustrating the implementation of the present invention to obtain the circuit breaker operating state feature set. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0018] This application provides a control system for managing a primary and secondary integrated circuit breaker. The executing entity of the control system for managing the primary and secondary integrated circuit breaker includes, but is not limited to, devices equipped with the system such as: intelligent terminals, power automation equipment, edge computing nodes, cloud server clusters, and smart grid monitoring centers, which can be considered as general computing nodes in this application. The power automation equipment includes, but is not limited to, at least one of: a circuit breaker control unit, an intelligent monitoring device, and an intelligent distribution terminal.

[0019] Please see Figure 1 The present invention provides a control system for managing a primary and secondary integrated circuit breaker, comprising:

[0020] The data acquisition and synchronization module is used to comprehensively collect and synchronize the operating parameters of the primary and secondary integrated circuit breakers to obtain a standardized operating dataset.

[0021] The state perception and feature extraction module is used to perform multi-dimensional state perception and key feature extraction based on a standardized operating dataset to obtain a circuit breaker health feature map.

[0022] The operating environment adaptability assessment module is used to perform operating environment adaptability assessment and parameter adaptive optimization based on the circuit breaker health characteristic map, and obtain an environmental adaptability control parameter library;

[0023] The fault prediction and diagnosis module is used to perform fault mode identification and predictive diagnosis of circuit breakers based on the environmental adaptability control parameter library and the circuit breaker health feature map, and obtain a fault risk warning matrix.

[0024] The collaborative control decision module is used to optimize the collaborative control strategy of multiple circuit breakers and generate decisions based on the fault risk early warning matrix, so as to obtain the optimal control command sequence.

[0025] The safety execution and feedback module is used to perform secure encryption verification and hierarchical execution of control commands based on the optimal control command sequence, and to provide real-time feedback on the execution effect to obtain a circuit breaker control execution report; the various modules are connected through a secure communication network to realize secure data transmission and sharing between modules.

[0026] This invention comprehensively collects diverse operating parameters of primary and secondary integrated circuit breakers, efficiently and synchronously processes the data to form a standardized operating dataset, and uses multi-dimensional state perception technology to extract key features to construct a circuit breaker health feature map, comprehensively characterizing the circuit breaker's operating status. Based on the health feature map, environmental adaptability assessment and adaptive parameter optimization are performed, constructing a control parameter library that can adapt to various environmental conditions, improving the system's environmental adaptability. Combining the environmentally adaptable control parameter library and the health feature map, accurate fault prediction and diagnosis are achieved, generating a detailed risk warning matrix to provide a basis for preventive maintenance. Through multi-circuit breaker collaborative control strategy optimization, the optimal control command sequence is generated, achieving coordinated and efficient operation of the overall system. A secure encryption verification and hierarchical execution mechanism are adopted to ensure the safe and reliable execution of control commands, and a complete execution report is generated through real-time feedback, providing a basis for continuous system optimization.

[0027] In this embodiment of the invention, the control system for managing the primary and secondary integrated circuit breaker includes:

[0028] The data acquisition and synchronization module is used to comprehensively collect and synchronize the operating parameters of the primary and secondary integrated circuit breakers to obtain a standardized operating dataset.

[0029] In this embodiment, a high-precision sensor network is deployed to monitor the electrical parameters of the fusion circuit breaker in real time, collecting electrical characteristic data such as current, voltage, power factor, and harmonic content, including key parameters such as short-circuit current, overload current, dynamic stability current, and thermal stability current. Mechanical condition monitoring devices (such as acceleration sensors and displacement sensors) are installed to capture the motion characteristics of the circuit breaker's mechanical components and record mechanical condition parameters such as operating stroke, speed, acceleration, and vibration characteristics. Environmental monitoring equipment is deployed to collect environmental conditions of the circuit breaker's operation, including environmental parameters such as temperature, humidity, air pressure, and salt spray concentration, and environmental factors that may affect the circuit breaker's performance are recorded in real time. Operation commands and feedback signals are extracted from the circuit breaker control system, and control process data such as operation sequence, response time, and execution effect are recorded. The electromagnetic compatibility performance of the circuit breaker is monitored, and relevant parameters such as electromagnetic interference level and immunity performance are collected.

[0030] The collected heterogeneous data streams are initially processed to establish a unified data format standard, ensuring that the data contains accurate timestamps and device identification information, forming the original heterogeneous data stream. GPS or Network Time Protocol (NTP) is used to synchronize the time of the multi-source data, ensuring precise alignment of data collected by different sensors in the time dimension, constructing time-synchronized data. Gaussian filtering or wavelet transform algorithms are applied to the time-synchronized data for noise suppression, removing noise introduced by factors such as power grid fluctuations and equipment vibrations. Outliers are automatically detected, and algorithms such as isolated forest or local anomaly factor are used to identify outliers in the data. Data interpolation techniques (such as multiple interpolation and K-nearest neighbor interpolation) are used to repair missing data, ensuring the integrity and continuity of the dataset, forming a cleaned dataset. Wavelet denoising or Kalman filtering techniques are applied to further optimize data quality, improve the signal-to-noise ratio, enhance signal characteristics, and generate optimized... The dataset undergoes standardized processing, such as Z-score standardization or MinMax standardization, to transform parameters with different dimensions to the same scale. A standard feature matrix is ​​constructed, and the correlation between parameters is calculated using methods such as Pearson correlation coefficient or mutual information. A parameter correlation matrix is ​​established to quantify the mutual influence between parameters. Feature importance assessment algorithms (such as random forest importance assessment and recursive feature elimination) are used to calculate the contribution of each parameter to the circuit breaker state, generating a parameter weight distribution. Based on the parameter weight distribution, features are prioritized, and the set of parameters most critical to the circuit breaker state characterization is selected to form a key parameter set. The key parameter set is then classified and organized, and multi-dimensional parameter subsets are constructed according to dimensions such as electrical parameters, mechanical state, and environmental conditions. These multi-dimensional parameter subsets are integrated to form a structured, standardized operational dataset, providing a complete, accurate, and consistent data foundation for subsequent analysis.

[0031] The state perception and feature extraction module is used to perform multi-dimensional state perception and key feature extraction based on a standardized operating dataset to obtain a circuit breaker health feature map.

[0032] In this embodiment, a deep learning network (such as a convolutional neural network, recurrent neural network, or autoencoder) is applied to perform feature learning on the standardized operating dataset, extracting deep-level feature representations embedded in the data, capturing the inherent patterns of the circuit breaker's operating state, and forming latent feature representations. A multi-level abstract modeling method is applied to these latent feature representations to construct a hierarchical representation model containing low-level, mid-level, and high-level features, achieving a multi-level representation of the circuit breaker's state and forming a hierarchical feature structure. An ensemble classifier (such as a random forest, gradient boosting tree, or deep neural network) is applied to this hierarchical feature structure to perform state pattern recognition, identifying different operating states of the circuit breaker (such as normal closure). The circuit breaker operating state feature set is generated, including normal disconnection, overload protection, and short-circuit protection. Based on the circuit breaker operating state feature set, the transition characteristics between states are extracted, such as the transition characteristics from closing to disconnection and the evolution characteristics from normal to overload. A state transition sequence is constructed, and time series correlation analysis is performed on the state transition sequence. Time series analysis techniques (such as autoregressive models and long short-term memory networks) are applied to mine the time patterns of state changes and form state evolution laws. Based on the state evolution laws, a circuit breaker state transition probability matrix is ​​constructed, and a state transition model based on Markov processes is established. The transition probabilities and conditions between different states are quantified to form a Markov state model.

[0033] Based on the Markov state model, a health measurement index system is defined, including reliability, stability, responsiveness, and lifespan indicators. A health assessment index system is constructed, and fuzzy logic or evidence theory methods are applied to calculate the health status index of each key component of the circuit breaker (such as the contact system, arc extinguishing device, operating mechanism, and control circuit), forming a health status distribution map. The health status distribution map is then topologically modeled, constructing a network structure with components as nodes and relationships between components as edges. The edge weights represent the correlation strength between components, and the node attributes include health status indicators, forming a circuit breaker health feature map that comprehensively characterizes the health status and evolution trend of the circuit breaker.

[0034] The operating environment adaptability assessment module is used to perform operating environment adaptability assessment and parameter adaptive optimization based on the circuit breaker health characteristic map, and obtain an environmental adaptability control parameter library;

[0035] In this embodiment, an association rule mining algorithm (such as the Apriori algorithm or the FP-Growth algorithm) is applied to the circuit breaker health feature map to analyze the correlation between environmental factors (such as temperature, humidity, operating frequency, etc.) and circuit breaker health status indicators, identify the impact patterns of environmental factors on circuit breaker performance, construct an environmental sensitivity matrix, and use feature selection algorithms (such as recursive feature elimination or model-based feature selection) to screen out the environmental factors that have the most significant impact on circuit breaker performance, such as extreme temperature, high humidity, and frequent operation, to form an environmental impact factor set. Multi-factor, multi-level scenario simulation experiments are designed for the environmental impact factor set to evaluate the response characteristics of the circuit breaker under different combinations of environmental conditions. Numerical simulation or physical testing methods are applied to collect the circuit breaker's operating data under various environmental conditions to construct an environmental change response model.

[0036] Based on an environmental change response model, an adaptive parameter adjustment strategy is designed. Adjustment rules for control parameters (such as operating timing, mechanical parameters, and protection setpoints) under different environmental conditions are defined, forming a parameter adjustment rule base. This rule base is then used to construct a parameter space, defining the parameter value ranges and constraints to form an adaptive control parameter space. Within this space, multi-objective optimization algorithms (such as genetic algorithms, particle swarm optimization, or differential evolution algorithms) are applied to find the optimal parameter combination that satisfies multiple objective requirements (such as reliability, response speed, and energy minimization), forming an optimized parameter combination. This optimized parameter combination is then analyzed. Simulation verification and experimental testing are conducted to evaluate the performance of parameter combinations under different operating conditions, generating performance evaluation results. Based on these results, sensitivity analysis and response surface methodology are applied to fine-tune the parameters, improving their robustness and stability, thus forming a refined control parameter set. A mapping relationship model between environmental conditions and control parameters is established for this refined control parameter set. Methods such as support vector regression, neural networks, or Gaussian process regression can be used to construct an adaptive model that can automatically generate optimal control parameters based on environmental conditions. Ultimately, an environmentally adaptive control parameter library is formed, enabling optimal control of the circuit breaker under varying environments.

[0037] The fault prediction and diagnosis module is used to perform fault mode identification and predictive diagnosis of circuit breakers based on the environmental adaptability control parameter library and the circuit breaker health feature map, and obtain a fault risk warning matrix.

[0038] In this embodiment, fault-related feature patterns, such as abnormal component health and changes in state transition probability, are extracted from the circuit breaker health feature map to construct a fault feature set. A typical fault mode library for the circuit breaker is established by integrating historical fault case libraries and expert knowledge, including common fault types such as contact wear, operating mechanism jamming, arc extinguishing device failure, and control circuit abnormalities. A set of typical fault modes is formed, and similarity calculation algorithms (such as cosine similarity, Euclidean distance, or Mahalanobis distance) are applied to quantify the similarity between different fault modes. A fault mode similarity network is constructed to describe the association and similarity between fault types. Cluster analysis and evolutionary calculation methods are applied to the fault mode similarity network to identify the classification structure and development trend of fault modes, constructing a fault development path diagram to reveal the evolution process of faults from early symptoms to severe faults. Combined with environmental impact models in the environmental adaptability control parameter library, the impact of different environmental conditions on fault development is analyzed, such as high temperature accelerating contact wear and high humidity promoting insulation degradation. An environmentally sensitive fault development model is constructed to describe the environmental impact of the fault. This study investigates the modulating effect of environmental factors on the rate and path of fault evolution. Based on an environmentally sensitive fault development model, a fault prediction model is established using machine learning algorithms (such as support vector machines, random forests, or deep learning networks). Inputting current state characteristics and environmental conditions, the model outputs future fault probabilities, forming a fault probability prediction function. This function is then used to assess the fault risks of key components of the circuit breaker (such as main contacts, auxiliary contacts, operating mechanisms, arc extinguishing devices, and control circuits), predicting the probability and time window of various faults, and generating fault risk assessment results. A risk matrix analysis method is applied to these results to classify risks based on both fault severity and urgency, constructing a risk level matrix. Based on this matrix, a tiered early warning strategy is designed, defining early warning levels, methods, and response measures for different risk levels, forming multi-level early warning rules. Finally, based on these rules, a fault risk early warning matrix is ​​generated, containing fault type, probability of occurrence, impact level, early warning level, and recommended measures, providing a scientific basis for operation and maintenance decisions.

[0039] The collaborative control decision module is used to optimize the collaborative control strategy of multiple circuit breakers and generate decisions based on the fault risk early warning matrix, so as to obtain the optimal control command sequence.

[0040] In this embodiment, network analysis is performed on the fault risk warning matrix to identify control dependencies and influence links in the multi-circuit breaker system. A cooperative control network describing the interaction between circuit breakers is constructed. Based on the cooperative control network, a system-level state-space model of the multi-circuit breaker system is established to describe the overall state transition characteristics and dynamic behavior of the system, forming a system state representation. Based on the system state representation, a multi-objective optimization function for cooperative control is defined, including objectives such as maximizing system reliability, minimizing the number of operations, load balancing, and minimizing maintenance costs, forming a multi-objective optimization problem. Constraint analysis is performed on the multi-objective optimization problem, considering various constraints on the safe and stable operation of the power system, such as current carrying capacity, voltage stability, and power balance, forming a set of control constraints. Based on the multi-objective optimization problem and the set of control constraints, advanced optimization algorithms (such as multi-agent reinforcement learning, distributed optimization, or model-based optimization) are applied. A predictive control (CDC) algorithm is constructed to optimize the collaborative control strategy, generating a set of candidate control strategies that meet the constraints. The candidate strategies are then simulated and evaluated to analyze their performance under different scenarios, including normal operating conditions, fault conditions, and extreme conditions. A strategy performance evaluation matrix is ​​generated. Based on this matrix, a multi-criteria decision-making method (such as the Analytic Hierarchy Process, TOPSIS, or PROMETHEE) is used to select the optimal control strategy, forming a decision control scheme. This scheme is then decomposed into a series of control steps with defined execution times and sequences, forming a control instruction timing plan. Based on this plan, detailed control instruction parameters are generated, including information such as the specific operating equipment, execution time, operation type, parameter settings, and expected results. Finally, the optimal control instruction sequence is formed, providing precise guidance for the collaborative control of circuit breakers.

[0041] The safety execution and feedback module is used to perform security encryption verification and hierarchical execution of control commands based on the optimal control command sequence, and to provide real-time feedback on the execution effect to obtain a circuit breaker control execution report.

[0042] In this embodiment, the optimal control command sequence is classified according to its security risk level, such as high-risk commands (forced circuit breaker disconnection), medium-risk commands (protection parameter adjustment), and low-risk commands (status monitoring), forming a hierarchical command set. An identity authentication and authorization verification mechanism is applied to this hierarchical command set to ensure that the command executor has the corresponding operating authority. Digital signatures and two-factor authentication are used to verify the legitimacy of the command source, generating an authorized command set. This authorized command set is then securely encrypted using a high-strength encryption algorithm (such as AES-256, RSA, or elliptic curve cryptography) to ensure that the commands are not stolen or tampered with during transmission, forming a secure command packet. Based on this secure command packet, a detailed execution plan is formulated, including execution order, time window, resource allocation, and contingency plans, forming a command execution scheduling scheme. According to the command execution scheduling scheme, the control commands are distributed through a secure communication network. The system sends commands to each execution terminal, monitors the receipt and execution status of commands, captures execution process data in real time, forms an execution process data stream, performs real-time analysis of the execution process data stream, evaluates the accuracy, timeliness, and effectiveness of command execution, calculates execution success rate and performance indicators, and generates execution effect evaluation results. Based on the execution effect evaluation results, a closed-loop control strategy is adopted to adjust the execution of subsequent commands in real time, such as execution sequence optimization, parameter fine-tuning, or emergency response, forming adaptive execution feedback. The system records detailed information on the entire command execution process, including execution time, operation parameters, response status, abnormal situations, and handling measures, generating detailed execution process logs. The execution effect evaluation results and execution process logs are integrated to form a comprehensive circuit breaker control execution report, including execution statistical analysis, abnormal situation summary, performance improvement suggestions, and subsequent operation and maintenance suggestions, providing data support for continuous system optimization.

[0043] In this embodiment of the invention, the data acquisition and synchronization module is used to comprehensively collect and synchronize the operating parameters of the primary and secondary integrated circuit breakers to obtain a standardized operating dataset, specifically used for:

[0044] The electrical parameters, mechanical status and environmental conditions of the circuit breaker are collected through multiple channels to obtain the raw heterogeneous data stream. The raw heterogeneous data stream is then timestamped and the data quality is evaluated to obtain time-synchronized data.

[0045] Anomaly detection and missing value repair are performed on the time-synchronized data to obtain a cleaned dataset. Then, data denoising and signal enhancement are performed on the cleaned dataset to obtain an optimized dataset.

[0046] The optimized dataset is subjected to feature normalization and standardization to obtain a standard feature matrix. Based on the standard feature matrix, the correlation between each parameter is calculated to obtain the parameter correlation matrix.

[0047] The feature importance of the parameter correlation matrix is ​​evaluated to obtain the parameter weight distribution. Based on the parameter weight distribution, the features are prioritized and filtered to obtain the key parameter set.

[0048] The key parameter set is classified and organized to obtain a multi-dimensional parameter subset. Based on the multi-dimensional parameter subset, a standardized operation dataset is constructed, which includes electrical parameter subset, mechanical state subset, and environmental condition subset.

[0049] In this embodiment, a high-precision electrical parameter monitoring system is used to collect electrical parameters of the circuit breaker in real time, such as current, voltage, power factor, and harmonic content. Normal operating parameters and extreme operating condition parameters are recorded to form an electrical parameter database. An intelligent mechanical condition monitoring device is deployed to detect mechanical condition parameters of the circuit breaker, such as operating stroke, speed, acceleration, and vibration characteristics, to obtain the motion characteristics and wear features of the operating mechanism and establish a mechanical condition database. Environmental monitoring equipment is installed to collect environmental conditions of the circuit breaker, including environmental parameters such as temperature, humidity, air pressure, and pollution level. Environmental factors are correlated with circuit breaker performance to construct an environmental condition database. A multi-channel data acquisition system enables parallel acquisition of heterogeneous data, ensuring the comprehensiveness and real-time nature of data acquisition, forming a raw heterogeneous data stream. High-precision time synchronization technology (such as GPS clock, PTP protocol, or IEEE) is used. The 1588 standard performs timestamp calibration on multi-source data to ensure that time deviations between different acquisition devices are controlled within microseconds. It then performs quality assessment on synchronized data, calculating data integrity, accuracy, and consistency indicators, marking data segments with quality anomalies, and forming time-synchronized data. Anomaly detection algorithms based on statistics and machine learning (such as isolated forest, single-class SVM, or autoencoder anomaly detection) are applied to identify outliers in the data. Multi-dimensional analysis distinguishes between system anomalies and data anomalies. Advanced data repair techniques (such as tensor completion, deep generative models, or Bayesian inference) are used to intelligently repair missing data, ensuring data continuity and reliability, generating a cleaned, high-quality dataset. Signal processing techniques are applied to the cleaned dataset for noise reduction, using wavelet transform, empirical mode decomposition, or adaptive filtering to suppress random noise and interference signals. Signal enhancement algorithms (such as spectrum reconstruction, eigenre resonance, or mode separation) are applied to highlight key features in the data, improving the signal-to-noise ratio, forming an optimized dataset. Finally, the optimized dataset is uniformly normalized. Standardization and normalization processes eliminate dimensional differences and scale effects, enabling direct comparison and analysis of different types of parameters. A standard feature matrix is ​​constructed, and advanced correlation analysis methods (such as Pearson correlation coefficient, Kendall rank correlation, or mutual information) are applied to evaluate the linear and nonlinear correlations between parameters, constructing a parameter correlation degree matrix to form a relationship network between parameters. Feature importance evaluation algorithms (such as random forest importance, stability selection, or information gain-based methods) are applied to the parameter correlation degree matrix to calculate the contribution of each parameter to the circuit breaker's state characterization, forming a parameter weight distribution. Based on the parameter weight distribution, importance thresholds are set for feature screening, retaining high-information parameters and filtering redundant features, forming a feature subset containing key monitoring indicators, i.e., a key parameter set. The key parameter set is then organized into multi-dimensional categories, constructing a parameter classification system from different perspectives such as electrical characteristics, mechanical condition, and environmental conditions, forming a structured multi-dimensional parameter subset. Integrating the multi-dimensional parameter subsets, a unified data model and access interface are established, ultimately forming a standardized operational dataset.This provides a high-quality data foundation for subsequent state perception and feature extraction.

[0050] In this embodiment of the invention, the state perception and feature extraction module is used to perform multi-dimensional state perception and key feature extraction based on a standardized operational dataset to obtain a circuit breaker health feature map, specifically for:

[0051] Deep feature learning is performed on the standardized running dataset to obtain latent feature representations, and a multi-level feature abstraction model is built based on the latent feature representations to obtain a hierarchical feature structure;

[0052] State pattern recognition is performed on the hierarchical feature structure to obtain the circuit breaker operating state feature set, and state transition features are extracted based on the circuit breaker operating state feature set to obtain the state transition sequence.

[0053] A time-series correlation analysis was performed on the state transition sequence to obtain the state evolution law. Based on the state evolution law, the state transition probability matrix of the circuit breaker was constructed to obtain the Markov state model.

[0054] A health quantification analysis was performed on the Markov state model to obtain a health assessment index system. Based on the health assessment index system, the health index of each component of the circuit breaker was calculated to obtain a health status distribution map.

[0055] A topological model of the health status distribution map is performed to obtain the circuit breaker health feature map, which includes information on node health status, edge association strength, and state evolution trend.

[0056] In this embodiment, a deep learning architecture (such as convolutional neural networks, long short-term memory networks, or variational autoencoders) is applied to perform nonlinear feature learning on the standardized running dataset to extract deep-level patterns and implicit relationships contained in the data, generating high-dimensional latent feature representations. Feature dimensionality reduction and visualization techniques (such as t-SNE, UMAP, or principal component analysis) are applied to these latent feature representations for low-dimensional mapping, visually displaying feature distribution and clustering structures. A feature representation model with multiple levels of abstraction is constructed, forming a hierarchical structure from low-level signal features to high-level semantic features. Different levels capture different granular features of the circuit breaker state, forming a hierarchical feature structure. A multi-model ensemble classifier (such as a combination of random forest, gradient boosting tree, and support vector machine) is applied to this hierarchical feature structure for state pattern recognition, identifying the circuit breaker's operating state category and state parameters, generating a circuit breaker operating state feature set. Feature change patterns during state transitions are extracted from this feature set, analyzing the triggering conditions, transition process, and completion markers of state transitions to form a state transition feature description. The state transition features are organized chronologically to construct a serialized representation reflecting the historical evolution of the circuit breaker's state. This process involves representing the state transition sequence. Time-series pattern mining algorithms (such as recurrent neural networks and sequence association rule mining) are applied to this sequence to identify periodicity, regularity, and anomalies in state changes, uncovering the inherent laws governing state evolution and forming state evolution laws. Based on these laws, a probability matrix describing the circuit breaker's state transition process is constructed, quantifying the transition probabilities and conditional probabilities between different states. A Markov process-based state transition model is then established, forming a Markov state model. Based on this model, a circuit breaker health measurement index system is defined, including multi-dimensional evaluation indicators such as reliability indicators (failure rate, mean time between failures), stability indicators (parameter fluctuation, state maintenance capability), responsiveness indicators (operation time, action characteristics), and life indicators (cumulative number of operations, wear degree). A health assessment index system is then constructed. Based on this system, fuzzy logic reasoning or evidence theory methods are used to calculate the health status index of each key component of the circuit breaker (such as the contact system, arc extinguishing device, operating mechanism, control circuit, etc.), generating a distribution map reflecting the health status of each component, i.e., a health status distribution map. The specific process is as follows:

[0057] The key components of the circuit breaker are systematically decomposed to identify the core components for health status assessment, including the main contact system (moving and stationary contacts, arc-extinguishing contacts), arc-extinguishing device (arc-extinguishing chamber, arc-extinguishing grid), operating mechanism (energy storage mechanism, transmission mechanism, auxiliary switch), control circuit (secondary circuit, protection unit, monitoring unit), and auxiliary system (insulation support, sealing system), forming a hierarchical component assessment structure system.

[0058] The implementation process of the fuzzy logic reasoning method involves: constructing a fuzzy linguistic variable set for the circuit breaker's health status, defining health status levels such as "excellent," "good," "average," "slightly deteriorated," "moderately deteriorated," "severely deteriorated," and "dangerous," and designing corresponding membership functions. For the characteristic parameters of each component, such as contact resistance, opening distance, overtravel, contact pressure, and contact temperature rise, a dedicated fuzzy set and membership function are established to map precise measurement values ​​onto the fuzzy set. A fuzzy rule base is designed, integrating expert knowledge and historical operating experience, to formulate reasoning rules such as "if the contact resistance is high and the temperature rise is large, then the contact status is moderately deteriorated," with 30-50 dedicated rules established for each component type.

[0059] The input component monitoring data is processed using a fuzzy inference engine, employing Mamdani or Sugeno inference models to perform rule matching, condition evaluation, and conclusion synthesis. Taking a contact system as an example, with an input contact resistance of 135 μΩ (close to the high threshold) and a temperature rise of 42℃ (moderately high), the membership degree for the "normal" state is calculated to be 0.35, and the membership degree for the "slightly deteriorated" state is 0.65, based on rule activation. The inference results are then defuzzified using methods such as the centroid method, the maximum membership method, or the weighted average method to convert the fuzzy output into a precise health status index value. For example, the fuzzy result for the contact system described above is converted into a health index of 68 (out of 100).

[0060] After the component-level health assessment is completed, a hierarchical weighted fusion method is used to consider the contribution weight of each component to the overall circuit breaker health status, such as 0.35 for the contact system, 0.25 for the arc extinguishing device, 0.25 for the operating mechanism, and 0.15 for the control circuit, to generate an overall health status index. To improve the interpretability of the assessment results, various forms such as health dashboards, radar charts, and heat maps are used to visually display the health status of each component; subsequently, a distribution map reflecting the health status of each component is generated.

[0061] By applying complex network modeling methods to the health status distribution map, a network structure is constructed with components as nodes and relationships between components as edges. The node attributes include health status indicators, and the edge weights represent the correlation strength between components. Time dimension information representing the evolution trend of the status is added to form a complete health feature map of the circuit breaker, which comprehensively represents the health status, component relationships and evolution trend of the circuit breaker.

[0062] In the embodiments of this invention, please refer to Figure 2 To obtain a detailed implementation flowchart for the circuit breaker operating state feature set, state pattern recognition is performed on the hierarchical feature structure to obtain the circuit breaker operating state feature set, which specifically includes:

[0063] Multiple classification algorithms are integrated for learning on the hierarchical feature structure to obtain initial state classification results. The initial state classification results are then fused by voting to obtain the state recognition fusion results.

[0064] By combining the pre-set historical state database and expert knowledge base, the state recognition fusion results are verified and corrected to obtain accurate working state identifiers. Based on the working state identifiers, state feature vectors are extracted to obtain state feature representations.

[0065] Based on the state feature representation, a state feature space is constructed, and density clustering analysis is performed in the state feature space to obtain the working state cluster centers;

[0066] Boundary learning is performed on the cluster centers of working states to obtain state boundary descriptors. Based on the state boundary descriptors, state transition critical points are identified to obtain state transition triggering conditions.

[0067] By integrating the working state cluster center, state boundary descriptor, and state transition triggering conditions, a complete set of circuit breaker working state features is formed.

[0068] In this embodiment, multiple advanced classification algorithms (such as support vector machines, random forests, gradient boosting trees, and deep neural networks) are applied in parallel to process the hierarchical feature structure. Each algorithm captures state features from different angles based on its characteristics, generating multiple independent state classification results to form an initial state classification result set. A voting fusion strategy (such as hard voting, soft voting, or weighted voting) is applied to the initial state classification results to integrate the judgment results of multiple classifiers, improving the accuracy and robustness of classification and forming a state recognition fusion result. A preset state library constructed from historical operating data and a knowledge base formed from industry expert experience are introduced as reference benchmarks to verify and correct the state recognition fusion result. Knowledge graph reasoning or rule matching methods are applied to identify potential misclassifications, adjust the classification results, and obtain a more accurate result. The circuit breaker uses a state identifier to extract feature vectors that comprehensively characterize the state's properties, including multi-dimensional information such as electrical, mechanical, and temporal characteristics. This forms a structured state feature representation. Based on this representation, a multi-dimensional state feature space is constructed, where each dimension represents a type of feature attribute. Different operating states of the circuit breaker form different distribution patterns within this space. Density clustering algorithms (such as DBSCAN, OPTICS, or average moving clustering) are applied to the state feature space to identify high-density regions. These regions represent typical operating states of the circuit breaker, such as normal closing, normal breaking, and overload protection states. The center point of each cluster is calculated to form the operating state cluster center. Boundary learning algorithms (such as One-Class SVM, isolated forest, or boundary recognition neural network) are applied to the operating state cluster centers to learn the boundary features of each state. The normal range of state changes and abnormal judgment criteria are defined, generating state boundary descriptors. Based on these descriptors, the transition regions between different states are analyzed to identify critical features and triggering conditions for state transitions, such as the operating current threshold from closing to breaking and the current-time characteristics from normal to overload, thus forming state transition triggering conditions.

[0069] The detailed process of identifying the critical features and triggering conditions of state transitions in a specific embodiment is as follows:

[0070] For each operating state cluster center, a boundary representation model in the state feature space is constructed. An adaptive statistical process control method is employed to calculate the normal fluctuation range of each state characteristic parameter based on historical operating data. For key features such as opening and closing time, peak current, operating energy, and contact speed, a control limit calculation method is applied, and the upper and lower threshold values ​​of the characteristic parameters are determined using the μ±kσ principle (μ is the parameter mean, σ is the standard deviation, and k is the tolerance coefficient). Different tolerance coefficients are adopted for different operating states of the circuit breaker; for example, k=3 is used for the normal closing state, and k=2.5 is used for the overload protection state, ensuring the rationality of the boundary definition.

[0071] For complex states characterized by multiple feature combinations, a multivariate statistical process control method is adopted to construct a T²-Hotelling control chart to monitor the multivariate covariance structure. The Mahalanobis distance between the state feature vector x and the cluster center μ is calculated.

[0072] For nonlinear boundary features, anomaly detection algorithms such as One-Class Support Vector Machine (SVM) or Isolation Forest are employed to construct closed boundaries of states in a high-dimensional feature space. By adjusting the kernel function parameters and anomaly scaling parameters, precise control of boundary tightness is achieved. Specifically, for continuous features such as the characteristic curves of the operating mechanism, a function envelope method is used to construct upper and lower envelope curves as dynamic boundaries, enabling boundary monitoring of the entire process characteristics.

[0073] A multi-level anomaly judgment standard is defined, establishing a hierarchical boundary from slight deviation, obvious anomaly to severe fault. A progressive anomaly judgment mechanism based on fuzzy membership is designed to avoid decision jitter caused by traditional binary judgment. For each type of state feature, a warning threshold and a fault threshold are defined respectively. For example, for contact resistance of a contactor, 120μΩ is the warning threshold and 150μΩ is the fault threshold. An anomaly severity evaluation function ES(x) is constructed, mapping the feature deviation degree to the [0,1] interval to achieve quantitative expression of anomalies.

[0074] By integrating boundary definitions and anomaly detection criteria, a complete state boundary descriptor data structure is constructed, including: state center vector, state covariance matrix, boundary model parameters, anomaly detection threshold, and state stability index. A multi-scale boundary description system is established, setting differentiated boundary constraints for characteristic fluctuations at different time granularities (such as instantaneous state, short-term steady state, and long-term trend).

[0075] Based on the generated state boundary descriptors, cross-state transition region analysis is conducted. Data on the entire process of the circuit breaker transitioning from one operating state to another is collected to construct a state transition dataset. The state transition sequence is segmented temporally to identify the pre-steady-state region, transition region, and post-steady-state region. The sliding window method is applied to calculate the characteristic change rate, locating the key time points where characteristic changes are most dramatic, as potential trigger points for state transitions.

[0076] A sequence pattern mining algorithm is employed to extract temporal feature change patterns from state transition data. Key transition event chains are identified, such as temporal correlation patterns like "current surge → contact action → arc extinguishing process → mechanical locking". For each type of state transition, a decision tree or rule set is constructed to formally express the combination of conditions triggering the state transition.

[0077] For different types of state transition processes, such as "normal closure → disconnection" and "normal closure → overcurrent protection action", dedicated transition characteristic models are established. Through statistical analysis of a large number of transition samples, the most discriminative critical features are identified. For example, for overcurrent protection triggering, "the duration t of the phase current exceeding the setting value > threshold T" is identified as the key triggering feature.

[0078] Based on dynamic systems theory, stability analysis is performed on critical points in the state space to distinguish between stable transitions, unstable transitions, and critically unstable states. A state transition hyperplane equation is established to describe the boundaries between different state domains, and the contribution weights of each feature to the hyperplane are calculated. Through confusion matrix analysis of boundary samples, the feature combinations most prone to misclassification are identified, and the boundary description is optimized accordingly.

[0079] For state transitions that are easily affected by environmental factors, an adaptive threshold adjustment mechanism is designed. A threshold correction model is established based on environmental parameters (such as temperature and humidity) to automatically adjust state transition conditions under different environments. For example, in low-temperature environments, the friction of the circuit breaker operating mechanism increases, and the response time boundary should be appropriately relaxed. The state boundary descriptor and state transition triggering conditions are integrated into a unified state transition model. A function for calculating state transition probabilities is established to achieve an end-to-end mapping from characteristic changes to state transition judgments.

[0080] By systematically integrating the working state cluster centers (characterizing typical features of the state), state boundary descriptors (defining the range of state changes), and state transition triggering conditions (determining the conditions for state transitions), a complete set of circuit breaker working state features is constructed, comprehensively characterizing the working state features of the circuit breaker under various operating conditions.

[0081] In this embodiment of the invention, the operating environment adaptability assessment module is used to perform operating environment adaptability assessment and parameter adaptive optimization based on the circuit breaker health feature map, and to obtain an environmental adaptability control parameter library, specifically used for:

[0082] An environmental factor correlation analysis was performed on the health characteristic map of the circuit breaker to obtain an environmental sensitivity matrix. Based on the environmental sensitivity matrix, key environmental impact factors were identified to obtain a set of environmental impact factors.

[0083] Multi-scenario simulations were performed on the set of environmental impact factors to obtain an environmental change response model. Based on the environmental change response model, an adaptive parameter adjustment strategy was defined to obtain a parameter adjustment rule base.

[0084] Based on the parameter adjustment rule base, an adaptive control parameter space is constructed, and multi-objective optimization is performed in the adaptive control parameter space to obtain the optimal parameter combination.

[0085] Verification tests were conducted on the optimized parameter combination to obtain performance evaluation results. Based on the performance evaluation results, the parameters were fine-tuned to obtain a refined set of control parameters.

[0086] Based on a refined set of control parameters, an environment-parameter mapping model is established to obtain an environmentally adaptive control parameter library, supporting optimal control of circuit breakers under different environmental conditions.

[0087] In this embodiment, graph mining algorithms and association rule analysis are applied to perform environmental factor correlation analysis on the circuit breaker health feature map. This identifies the correlation and influence patterns between environmental parameters (such as temperature, humidity, altitude, and pollution level) and circuit breaker health status indicators. An environmental sensitivity matrix is ​​constructed to quantify the degree of influence of environmental factors. Feature importance assessment methods (such as information gain-based assessment, stability selection, or model-based feature importance) are applied to analyze the environmental sensitivity matrix, identifying the environmental factors that have the most significant impact on circuit breaker performance and reliability, such as extreme temperatures, high humidity, and severely polluted environments. This forms a set of key environmental impact factors. Multi-factor, multi-level experimental schemes are designed to evaluate the circuit breaker's performance response under different combinations of environmental conditions. Orthogonal experimental design or response surface methodology is used to optimize the experimental scheme and improve experimental efficiency. Based on historical operating data and experimental results, regression analysis, neural networks, or Gaussian processes are applied to construct an environment-response relationship model, quantifying the influence function of different environmental conditions on circuit breaker performance parameters, thus forming an environmental change response model.

[0088] Based on an environmental change response model, an adaptive parameter adjustment strategy is designed. The optimal values ​​or adjustment methods for control parameters (such as operating speed, operating timing, and protection setpoints) under different environmental conditions are defined, forming a systematic parameter adjustment rule base. This rule base is then used to define the adaptive space of control parameters, clarifying the value range, adjustment step size, and constraints for each parameter. A multi-dimensional parameter space reflecting the relationships between parameters, i.e., the adaptive control parameter space, is constructed. In this space, advanced optimization algorithms (such as genetic algorithms, particle swarm optimization, simulated annealing, or differential evolution algorithms) are applied for multi-objective optimization, balancing various performance objectives such as reliability, response speed, and operational lifespan, to find the comprehensive optimal parameter combination. The optimized parameter set is then... Simulation verification and experimental testing are conducted to evaluate the performance of parameter combinations under different operating conditions, including reliability, stability, and economic indicators, generating comprehensive performance evaluation results. Based on these results, sensitivity analysis and response surface methodology are applied to fine-tune the parameters, optimizing their robustness and adaptability, resulting in a finely tuned set of control parameters. Based on this set, machine learning methods (such as support vector regression, random forest regression, or deep neural networks) are used to establish a mapping model between environmental conditions and optimal control parameters, achieving an end-to-end mapping from environmental monitoring data to the automatic generation of control parameters. Ultimately, this forms an environmentally adaptable control parameter library, supporting optimal control of circuit breakers under varying environmental conditions.

[0089] In this embodiment of the invention, multi-scenario simulations are performed on a set of environmental impact factors to obtain an environmental change response model, specifically including:

[0090] Based on pre-collected historical operational data and environmental records, a typical environmental scenario dataset is constructed, and data augmentation and mutation processing are performed on the typical environmental scenario dataset to obtain an extended environmental scenario library.

[0091] The simulation experiment scheme is defined using the extended environment scenario library, the scenario test matrix is ​​obtained, and the circuit breaker response is simulated based on the scenario test matrix to obtain multi-scenario response data.

[0092] Regression analysis was performed on response data from multiple scenarios to obtain a quantitative relationship model between environmental factors and circuit breaker performance. Nonlinear characteristic analysis was then performed on the quantitative relationship model to obtain a set of environmental critical points.

[0093] Based on the set of environmental critical points and quantitative relationship model, an environmental change sensitivity model is constructed, and the environmental change sensitivity model is verified and calibrated to obtain the calibrated sensitivity model.

[0094] By integrating the calibrated sensitivity model and quantitative relationship model, a complete environmental change response model is constructed, which can predict the impact of different environmental conditions on circuit breaker performance.

[0095] In this embodiment, historical operational data is collected and organized, including circuit breaker operation records, fault records, and maintenance records under different environmental conditions. Environmental parameters and performance indicators are correlated to form an initial environment-performance relationship dataset. Statistical analysis methods are applied to identify typical environmental scenarios, such as high-temperature and high-humidity environments, low-temperature and low-pressure environments, high-altitude environments, and highly corrosive environments. A typical environmental scenario dataset covering the circuit breaker's operating environment range is constructed. Data augmentation techniques (such as noise injection, parameter perturbation, or Monte Carlo simulation) are applied to this dataset to generate richer environmental scenario variations, expanding the scenario coverage and improving the model's generalization ability, thus forming an extended environmental scenario library. Based on this library, multi-factor experimental schemes are designed, test scenarios and evaluation indicators are defined, and orthogonal experimental design or optimal experimental design methods are used to optimize the experimental workload, generating a scenario test matrix. Based on this matrix, multi-physics coupling simulation tools (such as COMSOL) are used. Using software such as Multiphysics and Ansys, circuit breakers are simulated to evaluate their response characteristics under different environmental conditions, including mechanical, electrical, and thermal characteristics. Multi-scenario response data is generated, and regression analysis methods (such as multiple linear regression, support vector regression, or neural network regression) are applied to this data to establish a mapping relationship between environmental factors and circuit breaker performance indicators. This quantifies the degree and direction of the impact of different environmental factors on performance, forming a quantitative relationship model. Nonlinear characteristic analysis is then performed on this model to identify the nonlinear regions and abrupt change points in the impact of environmental parameter changes on performance. Environmental thresholds that may lead to significant changes in circuit breaker performance, such as contact icing temperature and insulation material softening, are identified. Temperature and other factors form a set of environmental critical points. Combining this set with a quantitative relationship model, a model describing the circuit breaker's sensitivity to environmental changes is constructed. This quantifies the performance change caused by a unit environmental change, generating an environmental change sensitivity model. Through experimental verification or historical data comparison, the environmental change sensitivity model is validated and calibrated. Model parameters are corrected to improve prediction accuracy, resulting in a calibrated sensitivity model. Integrating the calibrated sensitivity model with the quantitative relationship model, a unified environmental change response model framework is constructed, achieving end-to-end mapping from environmental parameter input to performance impact output. Ultimately, a complete environmental change response model is formed, providing a theoretical basis for the environmental adaptive control of circuit breakers.

[0096] In this embodiment of the invention, the fault prediction and diagnosis module is used to perform fault mode identification and predictive diagnosis of the circuit breaker based on the environmental adaptability control parameter library and the circuit breaker health feature map, and to obtain a fault risk warning matrix, specifically used for:

[0097] Based on the circuit breaker health feature map, fault-related features are extracted to obtain a fault feature set, and a fault mode library is established based on the fault feature set to obtain a set of typical fault modes.

[0098] Similarity calculations are performed on a set of typical failure modes to obtain a failure mode similarity network. Based on the failure mode similarity network, failure mode clustering and evolution analysis are performed to obtain a failure development path map.

[0099] By combining the environmental adaptability control parameter library, an environmental impact analysis is performed on the fault development path diagram to obtain an environmentally sensitive fault development model. Based on the environmentally sensitive fault development model, a fault prediction algorithm is constructed to obtain a fault probability prediction function.

[0100] The failure risk of different components of the circuit breaker is predicted by using the failure probability prediction function, and the failure risk assessment results are obtained. The severity and urgency of the failure risk assessment results are then classified to obtain a risk level matrix.

[0101] Risk warning strategies are defined based on risk level matrices, resulting in multi-level warning rules. A fault risk warning matrix is ​​then generated based on these rules, including fault type, probability of occurrence, degree of impact, and recommended measures.

[0102] In this embodiment, feature patterns related to fault states are extracted from the circuit breaker health feature map, such as abnormal node health index, changes in edge association strength, and abnormal state transition probability. A fault characterization feature set is constructed, integrating historical fault case libraries and industry expert knowledge to establish a fault mode library covering common circuit breaker fault types. This library includes various fault modes such as contact erosion, operating mechanism jamming, arc extinguishing device failure, and control circuit failure, forming a typical fault mode set. Multiple similarity measurement methods (such as cosine similarity, Euclidean distance, Mahalanobis distance, or DTW distance) are applied to calculate the similarity between different fault modes, constructing a quantitative... A fault mode similarity network is used to correlate fault types. Community detection algorithms (such as the Louvain method, spectral clustering, or InfoMap) are applied to this network to identify closely related fault type groups. Fault mode clustering is then performed to analyze the evolutionary relationships and development patterns among fault types. A fault development path map reflecting the evolution of faults from early symptoms to severe failures is constructed. Combined with an environment-performance mapping model from an environmental adaptability control parameter library, the influence mechanism of different environmental conditions on the fault development process is analyzed, such as high-temperature environments accelerating contact material degradation and humid environments promoting insulation aging. An environmentally sensitive fault development model is proposed. Based on this model, advanced machine learning algorithms (such as random forests, gradient boosting trees, deep learning networks, or survival analysis models) are applied to construct a fault prediction algorithm. The model takes current state characteristics and environmental conditions as input and outputs the future fault probability and expected occurrence time, forming a fault probability prediction function. This function is then used to predict the fault risk of key components of the circuit breaker (such as the main contact system, operating mechanism, arc extinguishing device, and control circuit). The probability of occurrence, confidence interval, and expected occurrence time window for various faults are calculated, resulting in a fault risk assessment. The risk assessment matrix method is applied to evaluate the risk from two dimensions: the severity of the consequences of the fault (such as safety risks, economic losses, system impact, etc.) and the urgency (such as the probability of occurrence, expected time window, etc.). A risk level matrix is ​​constructed, and based on the risk level matrix, a graded early warning strategy is designed. The early warning level, early warning method, response time limit and disposal measures corresponding to different risk levels are defined to form a multi-level early warning rule system. According to the multi-level early warning rules, a fault risk early warning matrix containing fault type, probability of occurrence, degree of impact, early warning level and recommended measures is generated, providing a scientific basis for the decision-making of circuit breaker preventive maintenance.

[0103] In this embodiment of the invention, a fault prediction algorithm is constructed based on an environmentally sensitive fault development model to obtain a fault probability prediction function, specifically including:

[0104] A parameter sensitivity analysis was performed on the environmentally sensitive fault development model to identify key predictive factors, obtain a subset of predictive features, and construct multiple predictive models based on the subset of predictive features to obtain a set of predictive models.

[0105] The prediction model set is validated using historical data, the prediction accuracy index is calculated, the model evaluation results are obtained, and the optimal base model is selected based on the model evaluation results to obtain the base prediction model.

[0106] Based on the basic prediction model, a time window sliding mechanism is introduced to enhance the ability to capture time series characteristics, resulting in a time series enhanced prediction model. An uncertainty quantification method is then introduced into the time series enhanced prediction model to obtain a probability distribution prediction function.

[0107] Online learning optimization of the probability distribution prediction function improves the model's adaptability, resulting in an adaptive prediction algorithm. Furthermore, multiple adaptive prediction algorithms are integrated through ensemble learning to obtain a robust prediction system.

[0108] Based on a robust prediction system, an end-to-end fault probability prediction function is constructed, which can output the occurrence probability and confidence interval of different fault types in real time.

[0109] In this embodiment, a sensitivity analysis is performed on the environmentally sensitive fault development model to assess the influence of different parameters on the model's prediction results. Methods such as analysis of variance, partial derivatives, or the Sobol index are used to quantify parameter sensitivity, identifying the key factors most influential on fault prediction, such as the number of electrical operations, contact erosion degree, and mechanism friction coefficient, forming a subset of prediction features. Based on this subset, various fault prediction models are constructed, including statistical models (such as time series models and lifetime analysis models), machine learning models (such as support vector machines, random forests, and gradient boosting trees), and deep learning models (such as long short-term memory networks, convolutional neural networks, and variational autoencoders), forming a set of prediction models. Historical data is used to cross-validate the prediction model set, evaluating the prediction accuracy, stability, and generalization ability of each model. Performance indicators such as ROC curves, F1 scores, and mean absolute error are calculated to generate model evaluation results. Based on these results, the model with the best overall performance is selected as the base prediction model, balancing prediction accuracy and computational complexity. A time window sliding mechanism is introduced into the base prediction model to dynamically adjust feature extraction. The window captures the long-term and short-term dependencies of time-series data, enhancing the model's ability to learn time-series characteristics and generating a time-series augmented prediction model. Uncertainty quantification techniques such as Bayesian methods, ensemble learning, or Monte Carlo simulation are introduced into this model, enabling it to not only output prediction results but also provide confidence intervals and probability distributions, forming a probability distribution prediction function. This effectively quantifies the uncertainty of prediction results. The probability distribution prediction function is deployed in an online learning framework, allowing the model to learn from continuously generated new data, constantly updating and adjusting model parameters, improving the model's adaptability and timeliness, and generating adaptive prediction algorithms. Ensemble learning methods (such as Bagging, Boosting, or Stacking) are applied to integrate multiple adaptive prediction algorithms with different characteristics, leveraging the complementarity between algorithms to improve overall prediction performance and robustness, constructing a robust prediction system. Based on this robust prediction system, an end-to-end fault probability prediction function is built, achieving a complete mapping from multi-source monitoring data input to fault probability and its confidence interval output, ultimately forming a prediction function capable of accurately predicting the probability and uncertainty of different fault types.

[0110] In a specific embodiment, the process of constructing an end-to-end fault probability prediction function is as follows:

[0111] Design a multi-level fusion end-to-end fault probability prediction function architecture. Define the input space to include multi-source heterogeneous monitoring data streams, such as electrical parameters, mechanical characteristics, and environmental conditions; define the output space to include the occurrence probability of various faults and their confidence intervals, realizing a full-process mapping from data to prediction results.

[0112] Unified processing of multi-source heterogeneous data is implemented. A sliding window method is applied to extract time-domain features (mean, standard deviation, skewness, kurtosis, etc.) and frequency-domain features (spectral power, band energy, etc.) from time-series data. A dynamic feature selection mechanism is implemented, comprehensively evaluating feature importance based on Fisher scores, mutual information, and recursive feature elimination methods to select the optimal feature subset.

[0113] Dedicated prediction base models are designed for different fault types, including gradient boosting trees, random forests, and deep neural networks. Uncertainty quantification is achieved using Bayesian methods, incorporating techniques such as Bayesian neural networks or Monte Carlo dropout, and the prediction distribution is calculated through multiple sampling. A deep ensemble learning framework is constructed to integrate the outputs of multiple heterogeneous prediction models, and the integrated prediction probability is calculated as follows:

[0114] ;in, Indicates the first The base model for the first Predicted probability of class of faults For model weights; The multidimensional feature vector representing the circuit breaker is the input parameter of the fault prediction model. The model weights are dynamically optimized and adjusted through performance evaluation on the validation set.

[0115] An adaptive weight adjustment mechanism is introduced to dynamically adjust the weight allocation based on the performance of each model under different operating conditions, thereby improving the system's robustness. A timescale transformation for fault prediction is implemented, extending from short-term state prediction to long-term trend prediction. A remaining useful life (RUL) estimation model is constructed, and survival analysis methods are combined to predict the fault occurrence time window.

[0116] Based on the ensemble output of the prediction model, a complete confidence interval for the failure probability is constructed. Multiple sets of predicted values ​​are generated through Bootstrap sampling and model parameter perturbation, and empirical quantiles are calculated to form the confidence interval.

[0117] ;in, Indicates the sorted order of the first... One predicted value, For the sorted number One predicted value, The significance level is usually set to 0.05. For the sample size, Indicates the first The confidence interval for the predicted probability of a fault type is as follows: For example, if the system predicts that the probability of a circuit breaker contact burn-out fault is 30%, its confidence interval CI may be [25%, 35%]. This means that at a 95% confidence level (assuming α=0.05), the actual fault probability falls within this interval.

[0118] A prediction reliability scoring function is designed, comprehensively considering the confidence interval width and the entropy of the prediction probability distribution to quantify the credibility of the current prediction. For different operating conditions, an environmentally sensitive prediction correction mechanism is constructed, adaptively adjusting the prediction probability based on the environmental condition vector to make the prediction results more adaptable to the actual operating environment.

[0119] An online learning and adaptive update mechanism for the predictive model is implemented, employing an incremental learning method to continuously update model parameters. An elastic weight consolidation (EWC) mechanism is introduced to prevent catastrophic forgetting, maintaining the model's ability to remember historical patterns while adapting to new data distribution characteristics.

[0120] The design employs a three-tiered fusion strategy: feature layer, model layer, and decision layer. The feature layer integrates features extracted from different data sources; the model layer integrates prediction results from multiple algorithms; and the decision layer combines expert rules and a historical failure case library for final judgment. An interpretability module is added, using feature importance analysis, SHAP value calculation, and decision path visualization to make the prediction results traceable and understandable.

[0121] The final end-to-end fault probability prediction function achieves an accurate mapping from monitoring data to predicted probabilities and confidence intervals for multiple fault types.

[0122] In this embodiment of the invention, the collaborative control decision module is used to optimize and generate decisions for multi-circuit breaker collaborative control strategies based on a fault risk warning matrix, thereby obtaining an optimal control command sequence. Specifically, it is used for:

[0123] Multi-circuit breaker correlation analysis is performed on the fault risk warning matrix to identify control dependencies, obtain the circuit breaker collaborative control network, and construct a system-level state space model based on the circuit breaker collaborative control network to obtain the system state representation;

[0124] Based on the system state representation, a cooperative control objective function is defined to obtain a multi-objective optimization problem. Constraint analysis is then performed on the multi-objective optimization problem to obtain a set of control constraints.

[0125] Based on a multi-objective optimization problem and a set of control constraints, a collaborative control strategy optimization algorithm is constructed to obtain a set of candidate control strategies. The candidate control strategy set is then evaluated through simulation to obtain a strategy performance evaluation matrix.

[0126] Based on the strategy performance evaluation matrix, the optimal control strategy is selected to obtain the decision control scheme, and the decision control scheme is decomposed into a timing plan of control commands.

[0127] Detailed control command parameters are generated based on the control command timing plan to obtain the optimal control command sequence, which includes execution time, execution device, operation parameters and expected results.

[0128] In this embodiment, a complex network analysis is performed on the fault risk warning matrix to identify control dependencies, influence propagation paths, and key control nodes in the multi-circuit breaker system. A collaborative control network reflecting the interaction and cooperation among circuit breakers is constructed. Based on the collaborative control network, a state-space modeling method is applied to establish a system-level state-space model describing the overall dynamic behavior of the multi-circuit breaker system, including state variables, control variables, disturbance variables, and observation variables, forming a system state representation. Based on the system state representation, a collaborative control objective function reflecting the overall system performance goal is defined, including multiple objective dimensions such as maximizing system reliability, minimizing operating costs, optimizing maintenance efficiency, and load balancing. A multi-objective optimization problem description is constructed, and constraint analysis is performed on the multi-objective optimization problem, considering various constraints on the safe and stable operation of the power system, such as current carrying capacity constraints, voltage stability constraints, power balance constraints, and equipment operation constraints, forming a complete set of control constraints. Based on the multi-objective optimization problem and the set of control constraints, advanced optimization algorithms (such as genetic algorithms, particle swarm optimization, differential evolution algorithms, or...) are applied. A multi-agent reinforcement learning (MAL) algorithm is used to construct a collaborative control strategy optimization algorithm, generating multiple candidate control strategies that meet the constraints, forming a candidate control strategy set. The candidate control strategy set is then evaluated through multi-scenario simulations, testing the performance of each strategy under normal, fault, and extreme conditions. Evaluation indicators include reliability, economy, real-time performance, and robustness, generating a strategy performance evaluation matrix. Based on this matrix, a multi-criteria decision-making method (such as analytic hierarchy process, TOPSIS, or fuzzy comprehensive evaluation) is applied to select the control strategy with the best overall performance, forming the final decision control scheme. The decision control scheme is then decomposed into a time sequence, breaking it down into a sequence of control steps with a clear execution order and time nodes. Considering dependencies and the possibility of parallel execution, a control instruction timing plan is formed. Based on this plan, control instructions containing detailed execution parameters are generated, specifying the execution time, execution device, operation type, parameter settings, and expected results. These are then integrated to form the optimal control instruction sequence, providing precise guidance for the actual execution of circuit breaker collaborative control.

[0129] In this embodiment of the invention, the secure execution and feedback module is used to perform secure encryption verification and hierarchical execution of control commands based on the optimal control command sequence, and to provide real-time feedback on the execution effect to obtain a circuit breaker control execution report. Specifically, it is used for:

[0130] The optimal control instruction sequence is classified by security level to obtain a hierarchical instruction set, and the hierarchical instruction set is then authenticated and authorized to obtain an authorized instruction set.

[0131] The authorized instruction set is encrypted to ensure secure transmission, resulting in a secure instruction packet. An execution plan is then generated based on the secure instruction packet to obtain the instruction execution scheduling scheme.

[0132] The control commands are executed step by step based on the instruction execution scheduling scheme. The execution status is captured in real time, the execution process data stream is obtained, and the execution process data stream is analyzed in real time to obtain the execution effect evaluation result.

[0133] Based on the performance evaluation results, the execution of instructions is adjusted and subsequent instructions are optimized to obtain adaptive execution feedback. Based on the adaptive execution feedback, an execution process log is generated to record the detailed execution situation.

[0134] Integrate the performance evaluation results and execution process logs to generate a circuit breaker control execution report, which includes execution success rate, abnormal situations, performance improvements and follow-up suggestions.

[0135] In this embodiment, the optimal control command sequence is classified according to the operational risk level and the degree of security impact. Commands are divided into different security levels, such as high-risk commands (e.g., circuit breaker opening and closing operations), medium-risk commands (e.g., protection parameter modifications), and low-risk commands (e.g., status queries), forming a hierarchical command set. Identity authentication and authorization verification mechanisms are applied to the hierarchical command set. Multi-factor authentication technologies (e.g., digital certificates, dynamic passwords, biometrics) are used to verify the operator's identity. Role-based access control (RBAC) ensures matching of operation permissions, generating a verified authorized command set. High-strength encryption algorithms (e.g., AES-256, RSA, or elliptic curve cryptography) are applied to the authorized command set for data encryption, generating a digital signature to ensure command integrity. A secure communication tunnel is constructed to prevent man-in-the-middle attacks, forming a secure command package. Based on the secure command package, a detailed execution plan is formulated, including execution order, time window, resource allocation, and contingency plans. Considering the dependencies between commands and the possibility of parallel execution, a command execution scheduling scheme is formed. According to the command execution scheduling scheme, the command is executed through a secure communication tunnel. The communication network distributes control commands to each execution terminal, monitors the receipt and confirmation of commands and their execution status, and collects status data, response signals and feedback information in real time during the execution process, forming an execution process data stream. Real-time analysis algorithms are applied to this data stream to evaluate the accuracy, timeliness, and effectiveness of command execution, calculate key performance indicators such as execution success rate and response time, and generate execution effect evaluation results. Based on these evaluation results, adaptive control strategies are used to adjust subsequent command execution, such as adjusting execution timing, optimizing operating parameters, or activating emergency plans, forming a closed-loop adaptive execution feedback mechanism. Detailed information about the entire command execution process is recorded, including command content, execution time, operation object, execution result, abnormal situations and handling measures, etc., constructing a structured execution process log. The execution effect evaluation results and execution process logs are integrated to form a comprehensive circuit breaker control execution report, including execution statistical analysis (such as success rate, average response time), anomaly summary, performance improvement suggestions, and subsequent operation and maintenance suggestions, providing data support for continuous system optimization and management decisions.

[0136] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0137] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0138] In the description of this invention, it should be understood that the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0139] In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0140] In the description of this invention, "several" means one or more, and "a large number" means two or more.

[0141] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0142] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0143] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A control system for managing a primary and secondary integrated circuit breaker, characterized in that, include: The data acquisition and synchronization module is used to comprehensively collect and synchronize the operating parameters of the primary and secondary integrated circuit breakers to obtain a standardized operating dataset. The state awareness and feature extraction module is used to perform deep feature learning on the standardized running dataset to obtain latent feature representations, and to establish a multi-level feature abstraction model based on the latent feature representations to obtain a hierarchical feature structure. State pattern recognition is performed on the hierarchical feature structure to obtain the circuit breaker operating state feature set, and state transition features are extracted based on the circuit breaker operating state feature set to obtain the state transition sequence. A time-series correlation analysis is performed on the state transition sequence to obtain the state evolution law, and based on the state evolution law, a circuit breaker state transition probability matrix is ​​constructed to obtain a Markov state model. A health quantification analysis is performed on the Markov state model to obtain a health assessment index system. Based on the health assessment index system, the health index of each component of the circuit breaker is calculated to obtain a health status distribution map. For the characteristic parameters of each component, a special fuzzy set and membership function are established to map the precise measurement values ​​onto the fuzzy set; Design a fuzzy rule base, apply a fuzzy inference engine to process the input component monitoring data, defuzzify the inference results, and convert the fuzzy output into a health status index value. A hierarchical weighted fusion method is adopted to consider the contribution weight of components to the overall health status of the circuit breaker and generate a health status index. Topological modeling is performed on the health status distribution map to obtain the circuit breaker health feature map, which includes information on node health status, edge association strength and state evolution trend. The environmental adaptability assessment module is used to perform environmental factor correlation analysis on the health characteristic map of the circuit breaker to obtain an environmental sensitivity matrix, and based on the environmental sensitivity matrix, to identify key environmental impact factors and obtain a set of environmental impact factors. Multi-scenario simulations are performed on the set of environmental impact factors to obtain an environmental change response model. Based on the environmental change response model, an adaptive parameter adjustment strategy is defined to obtain a parameter adjustment rule base. Based on the parameter adjustment rule base, an adaptive control parameter space is constructed, and multi-objective optimization is performed in the adaptive control parameter space to obtain the optimal parameter combination. The optimized combination of parameters is verified and tested to obtain performance evaluation results. Based on the performance evaluation results, the parameters are fine-tuned to obtain a refined set of control parameters. Based on the refined control parameter set, an environment-parameter mapping relationship model is established to obtain an environmental adaptive control parameter library; The fault prediction and diagnosis module is used to perform fault mode identification and predictive diagnosis of the circuit breaker based on the environmental adaptability control parameter library and the circuit breaker health feature map, and obtain a fault risk warning matrix. The collaborative control decision module is used to optimize the collaborative control strategy of multiple circuit breakers and generate decisions based on the fault risk warning matrix, so as to obtain the optimal control command sequence. The safety execution and feedback module is used to perform security encryption verification and hierarchical execution of control commands based on the optimal control command sequence, and to provide real-time feedback on the execution effect to obtain a circuit breaker control execution report.

2. The control system for managing a primary and secondary integrated circuit breaker according to claim 1, characterized in that, The operation parameters of the primary and secondary integrated circuit breakers are comprehensively collected and time-synchronized to obtain a standardized operation dataset, including: The electrical parameters, mechanical status, and environmental conditions of the circuit breaker are acquired through multiple channels to obtain a raw heterogeneous data stream. The raw heterogeneous data stream is then timestamped and its data quality is evaluated to obtain time-synchronized data. Anomaly detection and missing value repair are performed on the time-synchronized data to obtain a cleaned dataset. Then, data denoising and signal enhancement processing are performed on the cleaned dataset to obtain an optimized dataset. The optimized dataset is subjected to feature normalization and standardization to obtain a standard feature matrix, and the correlation between each parameter is calculated based on the standard feature matrix to obtain a parameter correlation matrix. The feature importance of the parameter correlation matrix is ​​evaluated to obtain the parameter weight distribution. The features are then prioritized and filtered according to the parameter weight distribution to obtain the key parameter set. The key parameter set is classified and organized to obtain a multi-dimensional parameter subset. Based on the multi-dimensional parameter subset, a standardized operation dataset is constructed, which includes a subset of electrical parameters, a subset of mechanical states, and a subset of environmental conditions.

3. The control system for managing a primary and secondary integrated circuit breaker according to claim 1, characterized in that, The state pattern recognition of the hierarchical feature structure yields a circuit breaker operating state feature set, including: Multiple classification algorithms are integrated for learning on the hierarchical feature structure to obtain initial state classification results, and the initial state classification results are then fused by voting to obtain state recognition fusion results. By combining a pre-set historical state database and an expert knowledge base, the state recognition fusion result is verified and corrected to obtain an accurate working state identifier. Based on the working state identifier, a state feature vector is extracted to obtain a state feature representation. Based on the state feature representation, a state feature space is constructed, and density clustering analysis is performed in the state feature space to obtain the working state cluster centers. Boundary learning is performed on the cluster centers of the working states to obtain state boundary descriptors, and state transition critical points are identified based on the state boundary descriptors to obtain state transition triggering conditions. The working state cluster center, the state boundary descriptor, and the state transition triggering condition are integrated to form a complete circuit breaker working state feature set.

4. The control system for managing a primary and secondary integrated circuit breaker according to claim 1, characterized in that, The process of performing multi-scenario simulations on the set of environmental impact factors to obtain an environmental change response model includes: Based on pre-collected historical operational data and environmental records, a typical environmental scenario dataset is constructed, and data augmentation and mutation processing are performed on the typical environmental scenario dataset to obtain an extended environmental scenario library. Using the extended environment scenario library, a simulation experiment scheme is defined to obtain a scenario test matrix. Based on the scenario test matrix, the circuit breaker response is simulated to obtain multi-scenario response data. Regression analysis was performed on the multi-scenario response data to obtain a quantitative relationship model between environmental factors and circuit breaker performance. Nonlinear characteristic analysis was then performed on the quantitative relationship model to obtain a set of environmental critical points. Based on the set of environmental critical points and the quantitative relationship model, an environmental change sensitivity model is constructed, and the environmental change sensitivity model is verified and calibrated to obtain a calibrated sensitivity model. By integrating the calibrated sensitivity model and the quantitative relationship model, a complete environmental change response model is constructed.

5. The control system for managing a primary and secondary integrated circuit breaker according to claim 1, characterized in that, Based on the environmental adaptability control parameter library and the circuit breaker health feature map, fault mode identification and predictive diagnosis are performed on the circuit breaker to obtain a fault risk warning matrix, including: Based on the circuit breaker health feature map, fault-related features are extracted to obtain a fault feature set, and a fault mode library is established based on the fault feature set to obtain a set of typical fault modes. Similarity calculation is performed on the set of typical failure modes to obtain a failure mode similarity network. Based on the failure mode similarity network, failure mode clustering and evolution analysis are performed to obtain a failure development path map. By combining the environmental adaptability control parameter library, an environmental impact analysis is performed on the fault development path diagram to obtain an environmentally sensitive fault development model. Based on the environmentally sensitive fault development model, a fault prediction algorithm is constructed to obtain a fault probability prediction function. The fault probability prediction function is used to predict the fault risk of different components of the circuit breaker to obtain the fault risk assessment result. The severity and urgency of the fault risk assessment result are then classified to obtain a risk level matrix. Based on the risk level matrix, a risk warning strategy is defined to obtain multi-level warning rules. A fault risk warning matrix is ​​then generated according to the multi-level warning rules, which includes fault type, probability of occurrence, degree of impact, and recommended measures.

6. The control system for managing a primary and secondary integrated circuit breaker according to claim 5, characterized in that, The fault prediction algorithm, constructed based on the environmentally sensitive fault development model, yields a fault probability prediction function, including: A parameter sensitivity analysis is performed on the environmentally sensitive fault development model to identify key predictive factors, obtain a subset of predictive features, and construct multiple predictive models based on the subset of predictive features to obtain a set of predictive models. The prediction model set is validated using historical data, the prediction accuracy index is calculated, the model evaluation results are obtained, and the optimal basic model is selected based on the model evaluation results to obtain the basic prediction model. A time window sliding mechanism is introduced on the basis of the basic prediction model to enhance the ability to capture time series characteristics, resulting in a time series enhanced prediction model. An uncertainty quantification method is then introduced into the time series enhanced prediction model to obtain a probability distribution prediction function. The probability distribution prediction function is optimized through online learning to obtain an adaptive prediction algorithm. Multiple adaptive prediction algorithms are then integrated through an ensemble learning method to obtain a robust prediction system. Based on the robust prediction system, an end-to-end fault probability prediction function is constructed.

7. The control system for managing a primary and secondary integrated circuit breaker according to claim 1, characterized in that, Based on the fault risk warning matrix, the multi-circuit breaker collaborative control strategy optimization and decision generation are performed to obtain the optimal control command sequence, including: Multi-circuit breaker correlation analysis is performed on the fault risk warning matrix to identify control dependencies, obtain the circuit breaker collaborative control network, and construct a system-level state space model based on the circuit breaker collaborative control network to obtain the system state representation; Based on the system state representation, a cooperative control objective function is defined to obtain a multi-objective optimization problem. Constraint analysis is then performed on the multi-objective optimization problem to obtain a set of control constraints. Based on the multi-objective optimization problem and the set of control constraints, a collaborative control strategy optimization algorithm is constructed to obtain a set of candidate control strategies. The candidate control strategy set is then simulated and evaluated to obtain a strategy performance evaluation matrix. Based on the strategy performance evaluation matrix, the optimal control strategy is selected to obtain the decision control scheme, and the decision control scheme is decomposed into a timing plan to obtain the control command timing plan. Based on the control command timing plan, detailed control command parameters are generated to obtain the optimal control command sequence, which includes execution time, execution device, operation parameters and expected results.

8. The control system for managing a primary and secondary integrated circuit breaker according to claim 1, characterized in that, Based on the optimal control command sequence, the control commands are securely encrypted and verified, and executed in a hierarchical manner. The execution effect is then fed back in real time, resulting in a circuit breaker control execution report, including: The optimal control instruction sequence is classified by security level to obtain a hierarchical instruction set, and the hierarchical instruction set is then authenticated and authorized to obtain an authorized instruction set. The authorized instruction set is encrypted to obtain a secure instruction package, and an execution plan is generated based on the secure instruction package to obtain an instruction execution scheduling scheme; Based on the instruction execution scheduling scheme, the control instructions are executed step by step to obtain the execution process data stream, and the execution process data stream is analyzed in real time to obtain the execution effect evaluation result; Based on the execution effect evaluation results, the instruction execution is adjusted and subsequent instruction execution is optimized to obtain adaptive execution feedback. Based on the adaptive execution feedback, an execution process log is generated to record detailed execution information. The execution effect evaluation results and the execution process log are integrated to generate a circuit breaker control execution report.

Images