Equipment intelligent connector management system

Abstract

The invention provides an equipment intelligent connector management system which comprises an underwater data acquisition module, an anti-interference data preprocessing module, an underwater intelligent diagnosis module, a closed-loop decision execution module, an underwater distributed storage module and an underwater cloud interaction module which are sequentially connected to form a closed-loop data stream. The system adapts to underwater complex working conditions, collected data are accurate and distortionless, the problems that a traditional system is insufficient in collection dimension and prone to data transmission misalignment are solved by integrating multiple types of special sensors, synchronously collecting underwater exclusive signals such as corrosion current and sealing pressure and matching cyclic redundancy check, and the improved LSTM model is combined with an underwater exclusive fault feature library, so that the data transmission accuracy is improved. Through cosine similarity comparison and accurate identification of three types of underwater specific faults such as sealing failure, the problems that a traditional system depends on a land fault library and is insufficient in identification capability are solved, a three-level emergency instruction is linked with a feedback verification mechanism, re-diagnosis is triggered if the standard is not reached, and the problems that the traditional system is lagged in response and disjointed in processing are solved.

Description

Technical Field

[0001] This invention relates to a management system, and more particularly to an intelligent connector management system for equipment, belonging to the field of marine engineering technology. Background Technology

[0002] In applications such as marine engineering and underwater exploration, underwater smart connectors are key components that ensure signal transmission and energy supply. Their operating status is directly related to the reliability and safety of the entire underwater system. Underwater smart connectors must work stably in high-pressure and highly corrosive seawater environments, while also dealing with strong signal attenuation and low-frequency noise interference. These special operating conditions place specific demands on the management of underwater smart connectors, especially the need to build a closed-loop management system to fully ensure the normal operation of the underwater system. Most current underwater connector management systems on the market adopt general technical architectures used on land, failing to optimize for the advantages and challenges of the underwater environment. This results in an inability to effectively manage underwater smart connectors. Traditional management systems often lack dedicated data acquisition dimensions to monitor underwater-specific fault signals, such as corrosion current and sealing pressure. Furthermore, the algorithms used for signal processing are not optimized for the strong interference environment in water, making raw data prone to distortion and affecting subsequent analysis and judgment. More seriously, these systems often rely on land-based fault feature databases and lack the ability to effectively identify and predict underwater-specific faults, such as seal failure or seawater intrusion. Due to the lack of effective feedback and verification mechanisms between data processing, fault diagnosis, and execution control, traditional underwater connector management systems often fail to form a closed-loop management system from data acquisition to fault diagnosis and execution processing. This disconnect in management leads to delayed fault response and low processing efficiency, failing to meet the high requirements of connector management under complex underwater conditions. To address the aforementioned technical issues, an intelligent connector management system for equipment is proposed. Summary of the Invention

[0003] In view of this, the present invention provides an intelligent connector management system for equipment to solve or alleviate the technical problems existing in the prior art, and at least provides a beneficial alternative.

[0004] The technical solution of the present invention is implemented as follows: an intelligent connector management system for equipment includes an underwater data acquisition module, an anti-interference data preprocessing module, an underwater intelligent diagnostic module, a closed-loop decision execution module, an underwater distributed storage module, and an underwater cloud interaction module, with each module connected in sequence to form a closed-loop data flow; The underwater data acquisition module acquires multi-dimensional physical signals from the smart connector, and after anti-interference processing, outputs distortion-free acquired data to the anti-interference data preprocessing module. The anti-interference data preprocessing module denoises and standardizes the data, and then outputs the anti-interference preprocessed data to the underwater intelligent diagnostic module. The underwater intelligent diagnostic module extracts features and compares them with underwater-specific fault features, and outputs diagnostic results containing fault type and level to the closed-loop decision execution module and the underwater distributed storage module. The closed-loop decision execution module generates and executes emergency commands, and outputs feedback signals to the anti-interference data preprocessing module. The underwater distributed storage module encrypts and partitions various types of data to form a structured database for the underwater cloud interaction module to call; The underwater cloud interaction module synchronizes data to the cloud and generates reports, and updates the fault feature database to the underwater intelligent diagnosis module.

[0005] More preferably, the underwater data acquisition module includes an edge acquisition terminal, which integrates an electrochemical sensor, a fiber optic grating sealing sensor, an underwater low-frequency accelerometer, and an optical fiber transmission unit; the multi-dimensional physical signals include electrical signals, corrosion current signals, sealing pressure signals, and vibration signals, and the edge acquisition terminal synchronously acquires various signals.

[0006] More preferably, the anti-interference processing of the underwater data acquisition module includes synchronous acquisition and integrity verification. The integrity of the transmitted data is verified using a cyclic redundancy check (CRC) algorithm. The core model of the CRC algorithm is: ; in, The original data is a polynomial. It generates a standard CRC-32 generator polynomial; the sending end calculates the CRC value during data transmission, and the receiving end verifies it.

[0007] More preferably, the anti-interference data preprocessing module includes a preprocessing server, and the denoising is performed using a wavelet threshold denoising algorithm. The wavelet threshold denoising uses db4 wavelet basis 3-level decomposition, and the improved soft thresholding function is: ; in, These are wavelet coefficients. For the threshold, The standard deviation of noise; Outlier removal during the denoising process employs the 3σ criterion. The standardization process maps the data to the [0, 1] interval using a normalization algorithm. The mathematical model of this normalization algorithm is as follows: ; in, , These represent the minimum and maximum values ​​of the data, respectively.

[0008] More preferably, the underwater intelligent diagnostic module includes a diagnostic server, which has a built-in improved LSTM neural network model and an underwater-specific fault feature library. The improved LSTM neural network model adds layer normalization and dropout regularization. The underwater-specific fault feature library includes standard feature vectors for three types of faults: sealing failure, seawater intrusion, and corrosion aging. Feature comparison uses cosine similarity calculation, and the calculation formula is as follows: ; in, The extracted fault feature vector, These are standard vectors from the fault feature library. This corresponds to sealing failure, seawater intrusion, and corrosion aging.

[0009] More preferably, the closed-loop decision execution module includes an underwater sealed execution controller, which integrates an underwater acoustic communication module and a CAN bus interface, and the emergency command is generated based on the fault level of the diagnostic results.

[0010] More preferably, the underwater distributed storage module includes waterproof distributed storage nodes. These nodes encrypt data using an encryption algorithm and implement partitioned storage according to four areas: a data acquisition area, a preprocessed data area, a diagnostic result area, and a feedback signal area. The encryption algorithm is as follows: ; in, A 256-bit key. Plain text data This is encrypted data.

[0011] More preferably, the underwater cloud interaction module includes an underwater 5G relay cloud platform, which employs an improved LZ77 data compression algorithm with a compression ratio of: ; in, The length of the original data. The length is the compressed value; the visualization report includes fault occurrence time, fault type statistics, and key signal trend curves.

[0012] More preferably, the feedback signal output by the closed-loop decision execution module includes the real-time value of the sealing pressure and the position information of the underwater sealing type actuator. The emergency commands of the closed-loop decision execution module include a first-level sealing reinforcement command, a second-level parameter adjustment command, and a third-level emergency recovery command. After receiving the feedback signal, if the sealing pressure corresponding to the failure of the first-level sealing reinforcement command and the second-level parameter adjustment command is not restored to ≥0.8MPa, or the emergency recovery operation is not executed in the third-level emergency recovery command, the underwater intelligent diagnostic module is triggered to re-execute the fault diagnosis process.

[0013] A further preferred embodiment includes an underwater edge-cloud collaborative architecture, which comprises an underwater edge cloud, a cloud platform, and a 5G relay unit. The underwater edge-cloud collaborative architecture uses the edge acquisition terminal as the data processing core and the cloud platform as the data optimization hub, with the two achieving data interaction through the 5G relay unit.

[0014] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions: This invention is adapted to complex underwater operating conditions, collecting accurate and distortion-free data. By integrating multiple types of specialized sensors, it simultaneously collects underwater-specific signals such as corrosion current and sealing pressure. Coupled with cyclic redundancy check, it solves the problems of insufficient data acquisition dimensions and inaccurate data transmission in traditional systems, laying a reliable foundation for subsequent analysis. It employs db4 wavelet denoising, 3σ outlier removal, and normalization algorithms to effectively filter underwater low-frequency noise and abnormal data. The standardized data is more suitable for diagnostic models, avoiding the influence of original data distortion on judgment. The improved LSTM model, combined with an underwater-specific fault feature library, accurately identifies three types of underwater-specific faults, including sealing failure, through cosine similarity comparison. This solves the problems of traditional systems relying on land-based fault libraries and insufficient identification capabilities. The three-level emergency command and feedback verification mechanism are linked, triggering re-diagnosis if the standard is not met, solving the problems of delayed response and disconnected processing in traditional systems, improving the timeliness of fault handling. The underwater cloud collaborative architecture enables data synchronization and dynamic updates of the feature library, continuously optimizing system performance and comprehensively improving the operational reliability and management intelligence level of underwater intelligent connectors.

[0015] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is the core closed-loop main flowchart of the present invention; Figure 2 This is a connection diagram of the various modules in the intelligent connector management system of the present invention. Detailed Implementation

[0018] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0019] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0020] like Figure 1 and Figure 2 As shown, this embodiment of the invention provides an intelligent connector management system for equipment, based on an underwater edge-cloud collaborative architecture, including an underwater data acquisition module, an anti-interference data preprocessing module, an underwater intelligent diagnostic module, a closed-loop decision execution module, an underwater distributed storage module, and an underwater cloud interaction module, with each module connected in sequence to form a closed-loop data flow; The underwater data acquisition module collects multi-dimensional physical signals from the smart connector, and after anti-interference processing, outputs distortion-free acquired data to the anti-interference data preprocessing module. After the anti-interference data preprocessing module denoises and standardizes the data, it outputs the anti-interference preprocessed data to the underwater intelligent diagnostic module. The underwater intelligent diagnostic module extracts features and compares them with underwater-specific fault features, and outputs diagnostic results containing fault type and level to the closed-loop decision execution module and the underwater distributed storage module. The closed-loop decision execution module generates and executes emergency commands, and outputs feedback signals to the anti-interference data preprocessing module. The underwater distributed storage module encrypts and partitions various types of data to form a structured database for the underwater cloud interaction module to access; The underwater cloud interaction module synchronizes data to the cloud and generates reports, and updates the fault feature database to the underwater intelligent diagnostic module.

[0021] In one embodiment, the underwater data acquisition module serves as the core of the system data input, undertaking the functions of synchronous acquisition of multi-dimensional physical signals and anti-interference transmission. The core component is the edge acquisition terminal, which adopts a titanium alloy sealed package and integrates an electrochemical sensor, a fiber optic sealed sensor, an underwater low-frequency acceleration sensor, and an optical fiber transmission unit. The four components work together to achieve comprehensive capture of underwater exclusive signals. The electrochemical sensor adopts a structure design of working electrode, reference electrode and auxiliary electrode, with a measurement range of 0-100μA and a resolution of 0.01μA. Its core function is to collect corrosion current signals of the metal contact parts of underwater smart connectors in real time. This signal directly reflects the corrosion rate of the metal material. When the corrosion current exceeds 50μA, the risk of corrosion aging needs to be focused on, providing a core data source for subsequent fault diagnosis. The fiber optic grating sealing sensor is based on the fiber optic strain-wavelength coupling effect. It has a measurement range of 0-5MPa and an accuracy of ±0.01MPa. It is specifically designed to collect the sealing pressure signal of the connector sealing structure. It can accurately capture the pressure change caused by wear and deformation of the sealing surface, and is a key basis for identifying sealing failure faults. The underwater low-frequency accelerometer is optimized for underwater low-frequency noise, with a measurement range of 0-10g and a sensitivity of 100mV / g. It mainly collects vibration signals during the operation of the connector and analyzes the vibration frequency, amplitude and other characteristics to determine whether there are abnormalities such as looseness or wear in the mechanical structure. The fiber optic transmission unit uses single-mode fiber as the transmission medium, with a transmission rate of 1Gbps and a transmission distance of up to 10km. It is responsible for synchronously transmitting the above three types of sensor signals and connector working electrical signals to the subsequent modules. Fiber optic transmission can effectively resist underwater electromagnetic interference and ensure signal transmission stability. The edge acquisition terminal uses a built-in high-precision synchronous clock to achieve synchronous acquisition of four types of signals, ensuring that multi-dimensional data at the same timestamp correspond one-to-one, avoiding misjudgment of faults due to acquisition timing deviations. The acquisition frequency is set to 100Hz, which can be adjusted within the range of 50-200Hz according to actual working conditions.

[0022] In one embodiment, the anti-interference processing of the underwater data acquisition module includes synchronous acquisition and integrity verification. The integrity of the transmitted data is verified by a cyclic redundancy check (CRC) algorithm. The core model of the CRC algorithm is as follows: ; in, The original data is a polynomial. Generate a standard CRC-32 generator polynomial; the sending end calculates the CRC value during data transmission, and the receiving end verifies it. The verification process is as follows: the sending end is an edge acquisition terminal, and the receiving end is a preprocessing server. Before transmitting data, the sending end calculates the CRC-32 checksum according to the formula mentioned above and appends the checksum to the end of the original data frame to form a complete data frame containing the original data and the checksum. After receiving the data frame, the receiving end extracts the original data and the checksum and uses the same generator polynomial... The checksum is recalculated. If the two checksums match, the data transmission is considered complete and undistorted, and the data is received normally. If they do not match, the data is considered lost or distorted, and a resampling command is immediately sent to the terminal to trigger data resampling until the checksum passes, ensuring that the output data is accurate and undistorted. Generate a polynomial for standard CRC-32: ; The original data polynomial is shifted left by 32 bits to reserve 32 bits of space to store the CRC check value; This means that a modulo-2 division operation is performed between the original data polynomial after left shift and the generator polynomial, and the resulting 32-bit remainder is the CRC check value.

[0023] In one embodiment, the anti-interference data preprocessing module includes a preprocessing server. The preprocessing server is an industrial-grade rack-mount server with an underwater environment adaptability design. Denoising is performed using a wavelet threshold denoising algorithm. The wavelet threshold denoising uses db4 wavelet basis 3-level decomposition, and the improved soft threshold function is: ; in, These are wavelet coefficients. For the threshold, The standard deviation of noise; Outlier removal during the denoising process employs the 3σ criterion, and the procedure is as follows: Calculate the mean of the denoised data with standard deviation ; Determine the normal data range as follows This range covers 99.73% of normal data; Data points outside this range are identified as outliers, removed, and filled with linear interpolation of adjacent data points to avoid data loss. Standardization maps data to the interval [0, 1] using a normalization algorithm. The mathematical model of the normalization algorithm is as follows: ; in, , These are the minimum and maximum values ​​of the data, respectively. The specific noise reduction process is as follows: After the preprocessing server receives the collected data, it first performs a 3-level decomposition of the db4 wavelet basis to obtain the approximation coefficients and detail coefficients; then, it calculates the threshold. Thresholding is applied to the detail coefficients of each layer, with the absolute value greater than 1. The coefficients are corrected according to the formula, and the absolute value is less than or equal to The coefficients are set to 0; finally, the processed coefficients are subjected to inverse wavelet transform to reconstruct the denoised clean data. After the above three-level processing, the preprocessing server outputs anti-interference preprocessing data in CSV format, which includes timestamp, standardized electrical signal, corrosion current signal, sealing pressure signal, and vibration signal fields, and is transmitted to the underwater intelligent diagnostic module.

[0024] In one embodiment, the underwater intelligent diagnostic module includes a diagnostic server. The diagnostic server has a built-in improved LSTM neural network model and an underwater-specific fault feature library. The improved LSTM neural network model adds layer normalization and dropout regularization. The underwater-specific fault feature library includes standard feature vectors for three types of faults: sealing failure, seawater intrusion, and corrosion aging. Feature comparison uses cosine similarity calculation, with the following formula: ; in, The extracted fault feature vector, These are standard vectors from the fault feature library. Corresponding to sealing failure, seawater intrusion, corrosion and aging; The improved LSTM neural network model performs two key optimizations on the traditional LSTM model to adapt to the complex characteristics of underwater data: Layer normalization: Normalize the input data of each layer of LSTM so that the mean of the input data is 0 and the variance is 1. This avoids gradient vanishing or gradient explosion caused by data distribution fluctuations during model training, and improves training stability and convergence speed. Layer normalization is deployed in the input gate, forget gate, output gate and cell state update of LSTM. Dropout regularization: Randomly discard 20% of neuron connections during training to prevent the model from overfitting the training data, improve the model's generalization ability to data of different depths and corrosion environments, and ensure stable diagnosis in unknown working conditions. Model training process: 100,000 sets of underwater working condition data were used as the training set, 50,000 sets of data were used as the test set, the training batch size was 64, the learning rate was 0.001, and the model converged after 100 iterations. The fault identification accuracy of the test set was ≥98.5%.

[0025] In one embodiment, the closed-loop decision execution module is the system's emergency response unit. Its core function is to generate and execute emergency commands based on diagnostic results, while simultaneously forming a closed loop through feedback verification. The core component is an underwater sealed execution controller, integrating an underwater acoustic communication module and a CAN bus interface to adapt to underwater high-pressure environments. Emergency commands are generated based on the fault level of the diagnostic results, with a total of three levels. The specific correspondence and execution logic are as follows: Level 1 Sealing Reinforcement Command: Corresponds to minor sealing failure. The command activates the electric sealing gasket built into the control connector, increasing the sealing pressure from the current value to 1.0MPa to enhance sealing performance. The controller sends a PWM control signal to the sealing reinforcement mechanism via the CAN bus interface, and the drive mechanism completes the reinforcement action within 5 seconds. Level 2 parameter adjustment command: corresponding to moderate faults. The command content is to adjust the connector operating parameters: when the seal fails, maintain the sealing pressure at 1.2MPa; when corrosion and aging occur, reduce the working voltage from 24V to 22V and limit the working current to within 3A to mitigate the development of the fault. The parameter adjustment is achieved by communicating with the connector main control unit through the CAN bus interface, and the adjustment response time is ≤3s. Level 3 Emergency Recovery Command: This command is for severe malfunctions. It controls the underwater recovery mechanism to start and recover the connector to the surface support vessel or underwater safety compartment within 30 seconds. The controller sends a start signal to the recovery mechanism via the underwater acoustic communication module and simultaneously receives feedback on the mechanism's position.

[0026] In one embodiment, the underwater distributed storage module includes waterproof distributed storage nodes, employing a distributed cluster architecture consisting of 5 storage nodes, supporting RAID5 redundancy backup. Each node features a waterproof and corrosion-resistant design, with a storage capacity of 1TB per node and a total storage capacity of 5TB. The waterproof distributed storage nodes encrypt data using an encryption algorithm. The encryption algorithm is as follows: ; in, A 256-bit key. Plain text data Encrypted data; The storage node divides the storage area into four independent data areas, storing data according to data type to form a structured database: Data Acquisition Area: Stores the distortion-free data acquired by the underwater data acquisition module. It is indexed by year-month-day-hour, retains the timestamps and integrity verification information of the original data, and has a storage period of 1 year. Preprocessed data area: Stores anti-interference preprocessed data, which is associated with the collected data through a unique ID, facilitating comparative analysis of noise reduction and standardization effects, with a storage period of 6 months; Diagnostic Results Area: Stores fault diagnosis results output by the diagnostic server, categorized by fault type, and records the fault occurrence time, duration, and severity level changes, with a storage period of 2 years; Feedback signal area: Stores feedback signals from the closed-loop decision execution module, which correspond one-to-one with emergency commands, record command execution time and changes in feedback parameters, and have a storage period of 1 year; The structured database supports fast queries based on timestamps, data types, fault types, and other criteria, with a data read rate of ≥100MB / s, meeting the calling requirements of the underwater cloud interaction module.

[0027] In one embodiment, the underwater cloud interaction module includes an underwater 5G relay cloud platform. The underwater 5G relay cloud platform employs an improved LZ77 data compression algorithm with a compression ratio of: ; in, The length of the original data. The data is the compressed length; the visualization report includes the fault occurrence time, fault type statistics, and key signal trend curves; the compressed data is transmitted to the cloud data processing center through an underwater 5G relay unit, with a synchronization frequency of 1 minute / time. The synchronization process uses the TCP / IP protocol to ensure data transmission reliability; for fault data, a priority transmission mechanism is adopted, with a synchronization delay of ≤10s, to ensure real-time monitoring of fault status in the cloud. The cloud-based data processing center performs statistical analysis on the synchronized data, generating three types of visual reports that can be viewed on both web and mobile devices: Fault Occurrence Time Report: Presents the fault trigger time, duration, and handling results in a timeline format, and supports filtering and querying by date; Fault Type Statistics Report: Statistics on the number of occurrences and frequency of three types of faults by day / week / month, displayed in bar charts and pie charts. For example, this week there were 3 sealing failure faults, accounting for 60%. Key signal trend curves: Plot the curves of corrosion current, sealing pressure and vibration signal with time as the horizontal axis, and mark the signal abrupt change points when the fault occurs, so as to help staff trace the cause of the fault. Fault Feature Library Update: The cloud data processing center performs in-depth mining on massive synchronous data, combines new underwater operating condition data and fault cases, optimizes standard feature vector parameters, and updates the underwater-specific fault feature library once a quarter. The updated feature library is transmitted to the diagnostic server through an underwater 5G relay unit to achieve continuous optimization of model performance.

[0028] In one embodiment, the feedback signal output by the closed-loop decision execution module includes the real-time value of the sealing pressure and the position information of the underwater sealing type execution controller; After the controller executes the command, it collects two types of feedback signals in real time and transmits them to the anti-interference data preprocessing module: Real-time sealing pressure: acquired by fiber optic grating sealing sensor, sampling frequency 50Hz, accuracy ±0.01MPa; Controller position information: collected by the Hall displacement sensor built into the controller, reflecting whether the actuator has reached the correct position; The preprocessing module verifies the feedback signal according to the following rules: When executing Level 1 / Level 2 instructions, verify whether the real-time value of the sealing pressure is ≥0.8MPa and monitor continuously for 10 seconds. If the condition is met, the treatment is deemed to have met the standard; if the pressure is lower than 0.8MPa at any time, the treatment is deemed to have failed to meet the standard. When executing a Level 3 command, check whether the location information shows that the recycling mechanism has reached the recycling completion position. If it arrives within 30 seconds, it is determined to meet the standard; otherwise, it is determined to fail to meet the standard. If the verification fails, the preprocessing module immediately sends a re-diagnosis command to the underwater intelligent diagnostic module to restart the fault identification and command generation process until the problem is resolved, thus completely solving the problems of delayed response and disconnected processing in traditional systems.

[0029] In one embodiment, the underwater edge-cloud collaborative architecture includes an underwater edge cloud, a cloud platform, and a 5G relay unit. The underwater edge cloud is centered on an edge acquisition terminal and has local data preprocessing and rapid diagnostic capabilities. For minor faults, there is no need to wait for cloud instructions; the edge can directly trigger a first-level sealing reinforcement instruction with a response delay of ≤2s, ensuring the timeliness of fault handling. The cloud platform uses the cloud data processing center as the optimization hub. By receiving massive amounts of data synchronized from the edge, it performs big data analysis and diagnostic model parameter tuning to achieve global system performance optimization. As a communication bridge, the 5G relay unit uses signal relay amplification technology to extend the underwater 5G transmission distance to 50km, meeting the coverage needs of large-scale underwater projects. On the other hand, it supports data fragmentation transmission, solving the problem of time-consuming synchronization of large files and improving collaborative efficiency. This collaborative architecture not only ensures real-time handling capabilities under complex underwater conditions, but also improves the long-term reliability and adaptability of the system through continuous optimization in the cloud, achieving the dual advantages of fast local response and accurate global optimization.

[0030] In operation, this invention employs the following methods: The edge acquisition terminal integrates four types of sensors to simultaneously acquire four types of signals, including electrical signals and corrosion currents. Data integrity is ensured through CRC-32 verification; if verification fails, data is reacquired. Once the data meets the standards, distortion-free data is output. The preprocessing server purifies the data and maps it to the [0, 1] interval using db4 wavelet denoising, 3σ outlier removal, and normalization, outputting high-quality preprocessed data. The diagnostic server extracts features using an improved LSTM model and compares them with an underwater-specific fault feature library using cosine similarity to determine the fault type and its severity (light / medium / severe), outputting diagnostic results. The execution controller generates three levels of instructions based on the fault level, and after execution, acquires sealing pressure and position feedback signals; if the indicators are not met, a re-diagnosis is triggered. The waterproof storage node uses AES-256 encryption to store various types of data in partitions, forming a structured database. After compression using the improved LZ77 algorithm, the data is synchronized to the cloud to generate visual reports, and the cloud optimizes the feature library and updates it back.

[0031] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent connector management system for equipment, characterized in that: It includes an underwater data acquisition module, an anti-interference data preprocessing module, an underwater intelligent diagnostic module, a closed-loop decision execution module, an underwater distributed storage module, and an underwater cloud interaction module. These modules are connected in sequence to form a closed-loop data flow. The underwater data acquisition module acquires multi-dimensional physical signals from the smart connector, and after anti-interference processing, outputs distortion-free acquired data to the anti-interference data preprocessing module. The anti-interference data preprocessing module denoises and standardizes the data, and then outputs the anti-interference preprocessed data to the underwater intelligent diagnostic module. The underwater intelligent diagnostic module extracts features and compares them with underwater-specific fault features, and outputs diagnostic results containing fault type and level to the closed-loop decision execution module and the underwater distributed storage module. The closed-loop decision execution module generates and executes emergency commands, and outputs feedback signals to the anti-interference data preprocessing module. The underwater distributed storage module encrypts and partitions various types of data to form a structured database for the underwater cloud interaction module to call; The underwater cloud interaction module synchronizes data to the cloud and generates reports, and updates the fault feature database to the underwater intelligent diagnosis module.

2. The intelligent connector management system for equipment according to claim 1, characterized in that: The underwater data acquisition module includes an edge acquisition terminal, which integrates an electrochemical sensor, a fiber optic grating sealing sensor, an underwater low-frequency accelerometer, and an optical fiber transmission unit. The multi-dimensional physical signals include electrical signals, corrosion current signals, sealing pressure signals, and vibration signals, and the edge acquisition terminal acquires various signals simultaneously.

3. The intelligent connector management system for equipment according to claim 1, characterized in that: The anti-interference processing of the underwater data acquisition module includes synchronous acquisition and integrity verification. The integrity of the transmitted data is verified using a cyclic redundancy check (CRC) algorithm. The core model of the CRC algorithm is as follows: ； in, The original data is a polynomial. It generates a standard CRC-32 generator polynomial; the sending end calculates the CRC value during data transmission, and the receiving end verifies it.

4. The intelligent connector management system for equipment according to claim 1, characterized in that: The anti-interference data preprocessing module includes a preprocessing server. The denoising is performed using a wavelet thresholding algorithm. Wavelet thresholding employs a db4 wavelet basis 3-level decomposition, and the improved soft thresholding function is: ； in, These are wavelet coefficients. For the threshold, The standard deviation of noise; Outlier removal during the denoising process employs the 3σ criterion. The standardization process maps the data to the [0, 1] interval using a normalization algorithm. The mathematical model of this normalization algorithm is as follows: ； in, , These represent the minimum and maximum values ​​of the data, respectively.

5. The intelligent connector management system for equipment according to claim 1, characterized in that: The underwater intelligent diagnostic module includes a diagnostic server, which incorporates an improved LSTM neural network model and an underwater-specific fault feature library. The improved LSTM neural network model adds layer normalization and dropout regularization. The underwater-specific fault feature library includes standard feature vectors for three types of faults: sealing failure, seawater intrusion, and corrosion aging. Feature comparison uses cosine similarity calculation, with the following formula: ； in, The extracted fault feature vector, The standard vector in the fault feature library, This corresponds to sealing failure, seawater intrusion, and corrosion aging.

6. The intelligent connector management system for equipment according to claim 1, characterized in that: The closed-loop decision execution module includes an underwater sealed execution controller, which integrates an underwater acoustic communication module and a CAN bus interface. The emergency command is generated based on the fault level of the diagnostic results.

7. The intelligent connector management system for equipment according to claim 1, characterized in that: The underwater distributed storage module includes waterproof distributed storage nodes. These nodes encrypt data using an encryption algorithm and implement partitioned storage according to four areas: acquired data area, preprocessed data area, diagnostic result area, and feedback signal area. The encryption algorithm is as follows: ； in, A 256-bit key. Plain text data This is encrypted data.

8. The intelligent connector management system for equipment according to claim 1, characterized in that: The underwater cloud interaction module includes an underwater 5G relay cloud platform, which employs an improved LZ77 data compression algorithm with a compression ratio of: ； in, The length of the original data. The length is the compressed value; the visualization report includes fault occurrence time, fault type statistics, and key signal trend curves.

9. The intelligent connector management system for equipment according to claim 6, characterized in that: The feedback signals output by the closed-loop decision execution module include the real-time value of the sealing pressure and the position information of the underwater sealing type actuator. The emergency commands of the closed-loop decision execution module include a first-level sealing reinforcement command, a second-level parameter adjustment command, and a third-level emergency recovery command. After receiving the feedback signals, if the sealing pressure corresponding to the failure of the first-level sealing reinforcement command and the second-level parameter adjustment command is not restored to ≥0.8MPa, or if the emergency recovery operation is not executed in the third-level emergency recovery command, the underwater intelligent diagnostic module is triggered to re-execute the fault diagnosis process.

10. The intelligent connector management system for equipment according to claim 2, characterized in that: It also includes an underwater edge-cloud collaborative architecture, which includes an underwater edge cloud, a cloud platform, and a 5G relay unit. The underwater edge-cloud collaborative architecture takes the edge acquisition terminal as the data processing core and the cloud platform as the data optimization hub. The two achieve data interaction through the 5G relay unit.

Images