A method and system for integrated monitoring of ship auxiliary equipment suitable for high-speed ships

Abstract

The present invention provides a method and system for integrated monitoring of ship auxiliary equipment applicable to high-speed vessels. The present invention relates to the field of electrical digital data monitoring and is used for real-time monitoring and comprehensive analysis of key systems of high-speed vessels during travel. The method for integrated monitoring of ship auxiliary equipment applicable to high-speed vessels of the present invention specifically comprises: collecting multiple categories of information of high-speed vessels, wherein the multiple categories of information include stress data, temperature data, and operating parameters of fuel, hydraulic, and ventilation systems; inputting the stress data and temperature data into a stress monitoring model to obtain stress analysis results; obtaining status results of each subsystem based on analysis of the operating parameters of the fuel, hydraulic, and ventilation systems; and performing remote monitoring based on the stress analysis results and the status results of each subsystem. Through the above steps, the present invention not only overcomes the limitations of a single detection method, but also improves the monitoring capability and operability of high-speed vessels.

Description

Technical Field

[0001] The present invention relates to the field of electrical digital data monitoring, and in particular to a method and system for integrated monitoring of ship auxiliary equipment applicable to high-speed ships, for performing real-time monitoring and comprehensive analysis of key systems of high-speed ships during their travel. Background Art

[0002] With the rapid development of the global shipbuilding industry and the growing demand for maritime transportation, high-speed craft, as a type of ship with high efficiency and flexibility, have been widely used in maritime patrol, rescue, transportation, leisure and entertainment. Due to their high speed, excellent maneuverability and small size, high-speed craft occupy an important position in both military and civilian markets. At present, the design and manufacturing technology of modern high-speed craft are becoming more and more mature, and the hull materials are gradually developing in the direction of high strength and lightweight. At the same time, the power system and propulsion technology are also continuously optimized, which has greatly improved the operating ability of high-speed craft in complex sea conditions. However, due to the high speed and lightweight hull design of high-speed craft, they are subject to large power loads in harsh sea conditions. Safety issues have always been an important concern in the application field of high-speed craft.

[0003] In the operation of modern ships, auxiliary equipment plays a vital role in maintaining hull stability and ensuring navigation safety. Auxiliary equipment typically includes fuel systems, hydraulic systems, ventilation systems, fire protection systems, compressed air systems, daily water systems, and flooding alarm systems, covering multiple key aspects of ship operation. The operating status of these devices directly affects the performance and safety of the ship. Currently, monitoring of auxiliary equipment relies primarily on single-function monitoring devices. Sensors collect real-time data such as hydraulic pressure, fuel consumption, and ventilation duct temperature and humidity, and transmit this data to the ship's monitoring center. However, this decentralized monitoring approach is limited in its ability to comprehensively analyze complex situations, especially when equipment failures or other abnormalities occur, making it unable to effectively provide systematic solutions. Furthermore, while existing ship monitoring technologies can collect hull stress data, they have significant limitations in analyzing and responding to specific types of stress changes, particularly those caused by sudden external forces. For example, when a high-speed vessel is struck by an unidentified underwater object, accidentally grounded on a reef, or struck by violent waves in adverse sea conditions, the hull structure can be subjected to localized, transient, and significant stress shocks. However, currently used stress monitoring methods are limited to collecting stress values ​​at discrete, individual measurement points. These systems generally lack the ability to systematically detect, locate, and analyze the evolution of these sudden stress events across both time and space. Specifically, in the spatial dimension, it is difficult to accurately depict the exact location of stress concentrations caused by impact or grounding, the impact range, and the distribution and propagation of stress within the ship's structure. In the temporal dimension, it is also difficult to effectively capture the dynamic processes of the onset, development, and decay of stress impact events, and it is unable to distinguish between transient impacts and sustained high loads. This lack of analytical capability means that existing systems cannot provide operators with intuitive feedback based on temporal and spatial changes, such as the dynamic location of stress impact points and visualization of stress propagation. As a result, operators are often faced with a large number of isolated, single stress sensor readings or alarm messages based on simple thresholds. Faced with this fragmented, spatially and temporally unavailable data, operators struggle to quickly and accurately determine the root cause of abnormal stress readings, the specific location of potential damage within the hull, and the severity of the incident.

[0004] To address these issues, the present invention proposes an integrated monitoring method and system for high-speed vessel auxiliary equipment, designed to address the shortcomings of existing technologies. This system, centered around hull stress monitoring, integrates monitoring functions for multiple key auxiliary equipment, including the fuel system, hydraulic system, ventilation system, fire protection system, compressed air system, daily water system, and flooding alarm system. Through multi-sensor collaboration and data fusion technology, it enables comprehensive analysis of the vessel's operating status. Summary of the Invention

[0005] The present invention provides a method for integrated monitoring of ship auxiliary equipment applicable to high-speed ships, the method specifically comprising the following steps:

[0006] S1: Collecting multiple types of information of high-speed vessels, including stress data, temperature data, and operating parameters of fuel, hydraulic, and ventilation systems;

[0007] S2: Input stress data and temperature data into the stress monitoring model to obtain stress analysis results;

[0008] S3: Analyzing the operating parameters of the fuel, hydraulic, and ventilation systems to obtain status results of each subsystem;

[0009] S4: Remote monitoring based on stress analysis results and status results of each subsystem.

[0010] The stress analysis results include: the time series fusion features of the instantaneous stress data and the instantaneous stress data Input into the primary detection model to obtain the stress anomaly probability map , and using the reference frame feature map and time series fusion features and stress anomaly probability map Perform position correction and obtain stress analysis results.

[0011] Through the high-frequency sampling of fiber Bragg grating stress sensors, the stress values ​​of each sensor node can be continuously recorded over a period of time, and the corresponding time-series stress matrix can be generated. For each sensor node, a set of continuous stress value sequences is formed in the time dimension. Based on these time-series stress data, an instantaneous stress matrix reflecting the spatial stress distribution can be generated for each time point of a sensor layout grid. For the sensor layout grids on both sides of the bulkhead and the bottom of the ship, for the time point The instantaneous stress matrix is ​​expressed as: ,in, Indicates at a point in time When arranging the grid Row, No. The stress values ​​measured by the sensor nodes in the column. Inside, every moment Corresponding to an instantaneous stress matrix. By arranging these instantaneous stress matrices in chronological order, the time-series stress matrices of the left side of the bulkhead, the right side of the bulkhead, and the bottom of the ship are formed. Among them, the time-series stress matrix of any part is expressed as follows: .

[0012] In the collection of time-series temperature distribution, an array of temperature sensors is arranged on the inner surface of the left side of the bulkhead, the right side of the bulkhead and the bottom of the high-speed ship according to the spatial layout corresponding to the stress sensor grid. Specifically, if the stress sensor grid is m rows and n columns, then m rows and n columns of temperature sensors are also arranged at the same position on each surface to ensure that each temperature sensor node (x, y) is accurately matched or as close as possible to the corresponding stress sensor node (x, y) in space. The present invention uses a patch-type thermistor temperature sensor, which is firmly installed at the corresponding measuring point on the surface by mechanical fixing. All arranged temperature sensors are connected to the same data acquisition unit as the stress data. The data acquisition unit has multi-channel synchronous sampling capability and achieves strict time synchronization with the system for collecting instantaneous stress data. From arrive , the instantaneous temperature matrices collected at each moment are combined in chronological order to form the time series temperature matrices of the left side of the bulkhead, the right side of the bulkhead and any part of the bottom of the ship: ;

[0013] Input stress data and temperature data into the stress monitoring model to obtain stress analysis results including:

[0014] Will Instantaneous stress matrix and the instantaneous temperature matrix Combined into a multi-channel matrix in two-dimensional space As the current frame: , ， set up Multi-channel matrix of moments is the reference frame;

[0015] Initial position detection: reference frame and the current frame Perform feature extraction and combine it with the gradient change weight map Feature maps of different depths of the current frame Fusion to obtain multi-feature fusion ;

[0016]

[0017] in, and Respectively for and The extracted feature maps are and They are feature extraction and fusion processes respectively. The parameters in brackets are the input parameters of feature extraction or fusion process. d1 and d2 represent different depths. All current frames are traversed to obtain temporal fusion features. .

[0018] Among them, the gradient change weight map based on and The stress change and temperature change are obtained, including: obtaining the stress change matrix and temperature change matrix , normalize stress changes and temperature changes and , the normalized stress change and temperature change are weightedly fused to obtain the fusion change feature map , convert the fusion change feature map into a weight map through Softmax ;

[0019]

[0020] in, and represent the weights of stress and temperature respectively.

[0021] Time series fusion features Input to the primary detection model and output stress anomaly probability map The primary detection model uses a U-shaped network as the processing framework. In the encoding stage, GRU is used for feature extraction and 3D pooling is used for downsampling. The bottleneck layer uses a GRU module to connect the encoding and decoding stages. The decoding stage includes GRU sequence recovery and upsampling. The encoding and decoding stages correspond to skip connections.

[0022] Among them, the i-layer feature processing in the encoding stage is specifically as follows:

[0023]

[0024] in, represents the encoding of the kth hidden state of the i-th layer, and They are the input and output of the i-layer GRU respectively, and the input of the first layer is the temporal fusion feature , the input of the i-th layer is the output of the previous layer 3D pooling , For 3D pooling, , For the GRU module encoding stage, is the step size parameter, Represents kernel function parameters;

[0025] Bottleneck layer passes After processing, H bottle ;

[0026] The i-layer feature processing in the decoding stage is specifically as follows:

[0027]

[0028] in, and are the upsampling of layer i in the decoding stage and the GRU output of layer i in the encoding stage, respectively. and are the kth hidden state and output of the i-th layer in the decoding stage, For 3D upsampling, is the upsampled input, and the input of the 4th layer is the bottleneck layer output H bottle , Decoding stage of the GRU module;

[0029] The input and output of the upper GRU, the input of the first layer is the time series fusion feature , the input of the i-th layer is the output of the previous layer 3D pooling , It is 3D pooling.

[0030] The output layer aggregates the output of the first layer of the decoder in time, that is, takes the maximum value of the time dimension and activates it to obtain the stress anomaly probability map ;

[0031]

[0032] in, Indicates that for the third dimension, that is, the time dimension, the maximum value of each position is selected and aggregated. and They are maximum pooling and two-dimensional upsampling, and They are the feature maps after aggregation and upsampling respectively.

[0033] The input of the position correction process is the stress anomaly probability map , reference frame feature map and time series fusion features , for time series fusion features Perform temporal aggregation to obtain static feature maps ,calculate With the reference frame feature map The difference , based on the difference Generate channel attention weights ; Based on stress anomaly probability map Generating spatial attention weights , applying channel attention and spatial attention to temporal fusion features , get the temporal attention feature ;based on Generate precise location heatmaps of time series .

[0034] Specifically, The initial time series heat map is obtained through the four-layer position decoding network , where the first two layers of the network are composed of 3D convolution and ReLu activation function, and the last two layers of the network are composed of 3D transposed convolution and ReLu activation function. Softmax activation is applied independently at each time point to obtain a precise positioning heat map , is the length of the time series.

[0035] The ship's auxiliary equipment integrates the stress analysis results and the status results of each subsystem into a data packet, which contains a timestamp, ship identification number, current status results and a thermal map of the precise impact location. The data packet is sent to the preset shore-based remote monitoring center via the ship's onboard communication unit. The shore-based remote monitoring center server receives the data packet from the ship, stores the parsed structured information in the database, and triggers the real-time processing and visualization engine.

[0036] The present invention provides an integrated monitoring system for ship auxiliary equipment applicable to high-speed ships, the system comprising:

[0037] Data acquisition module: The data acquisition module collects multiple types of information of high-speed vessels, including stress data, temperature data, and operating parameters of fuel, hydraulic, and ventilation systems;

[0038] Stress analysis module: input stress data and temperature data into the stress monitoring model to obtain stress analysis results;

[0039] A subsystem analysis module, wherein the subsystem analysis module obtains status results of each subsystem based on the operating parameters of the fuel, hydraulic, and ventilation systems;

[0040] Remote control module: The remote control module performs remote monitoring based on stress analysis results and status results of each subsystem.

[0041] The stress analysis results include: the time series fusion features of the instantaneous stress data and the instantaneous stress data Input into the primary detection model to obtain the stress anomaly probability map , and using the reference frame feature map and time series fusion features and stress anomaly probability map Perform position correction and obtain stress analysis results.

[0042] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned integrated monitoring method for ship auxiliary equipment applicable to high-speed vessels is implemented.

[0043] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned integrated monitoring method for ship auxiliary equipment applicable to high-speed ships.

[0044] Compared to existing technologies, this paper proposes an integrated monitoring method and system for ship auxiliary equipment suitable for high-speed vessels. This two-stage approach combines the advantages of deep spatiotemporal pattern learning (stage one) and baseline difference-based attention focusing (stage two). Specifically, in scenarios such as high-speed vessels, which experience complex dynamic environments and numerous interfering signals, this method effectively processes complex background fluctuations using a U-Net / GRU framework. By comparing differences with the initial normal state and applying attention correction, it accurately captures and amplifies the signal characteristics of key abnormal events such as impacts and groundings. This enables more reliable and accurate stress anomaly detection and location, providing strong technical support for ensuring the structural safety of high-speed vessels. Furthermore, this method integrates the state analysis results of multiple subsystems, allowing for integrated supervision and control, thereby improving the monitoring capabilities and operability of high-speed vessels. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0046] Figure 1 It is a reference diagram of the components of each subsystem of the present invention;

[0047] Figure 2 Schematic diagram of a typical arrangement of the monitoring station of the present invention;

[0048] Figure 3 This is a control flow logic diagram of the automatic fuel transfer system of the present invention. DETAILED DESCRIPTION

[0049] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0050] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that, in the absence of conflict, the features in the following embodiments and embodiments can be combined with each other. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative work are within the scope of protection of this application.

[0051] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on this application, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspect described herein can be used to implement an apparatus and / or practice a method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this apparatus and / or practice this method.

[0052] Additionally, in the following description, specific details are provided to provide a thorough understanding of the examples, however, one skilled in the art will appreciate that the examples can be practiced without these specific details.

[0053] The embodiments of this specification provide a method for integrated monitoring of ship auxiliary equipment applicable to high-speed ships, which specifically includes the following steps:

[0054] S1: Collecting multiple types of information of high-speed vessels, including stress data, temperature data, and operating parameters of fuel, hydraulic, and ventilation systems;

[0055] S2: Input stress data and temperature data into the stress monitoring model to obtain stress analysis results;

[0056] S3: Analyzing the operating parameters of the fuel, hydraulic, and ventilation systems to obtain status results of each subsystem;

[0057] S4: Remote monitoring based on stress analysis results and status results of each subsystem.

[0058] The functional components of the monitoring system of the present invention are shown in Figure 1. This system is constructed as a centralized core unit responsible for unified management of all auxiliary equipment subsystems throughout the entire vessel. The system acquires key operating status parameters of each subsystem in real time through a built-in data acquisition interface and integrates these real-time parameters with pre-set control logic or operator instructions for comprehensive computational processing. The system also integrates a structural stress monitoring module, which collects structural stress data in real time through sensors located in key areas of the hull. Furthermore, the system is designed with a mutual backup and switching interface and logic with the power control system. In the event of an emergency failure in the main power control system, this system can seamlessly switch between key monitoring and control functions, ensuring the continued operation of core functions. Through this functional integration and data processing, the system enables centralized, real-time monitoring and remote control of all auxiliary equipment on the vessel. Combined with structural stress analysis, it comprehensively analyzes and understands the real-time operating status of each auxiliary subsystem and the hull structure under current navigation or operating conditions.

[0059] Figure 2 illustrates a typical layout of the monitoring console used in this invention. Typically installed within a vessel's central control room, the console physically integrates key hardware components, including a ruggedized computer for monitoring system operations and display, a high-resolution display for information display, and a human-computer interactive control panel with various switches, buttons, and indicator lights. The console's overall layout optimizes space and facilitates operator access. All information display and control command input related to auxiliary equipment monitoring are centralized here, enabling centralized monitoring and remote control of all auxiliary equipment on board.

[0060] The power supply system of the monitoring system of the present invention adopts a redundant power supply scheme, which is mainly connected to two independent marine AC power supplies, and the two AC power supplies are in a master-slave relationship with each other. The system has built-in power switching logic and devices, which can monitor the status of the main AC power supply in real time. Once the main power supply voltage is detected to be abnormal or fails, the system will automatically and uninterruptedly switch to the backup AC power supply for power supply to ensure the continuous operation of the monitoring system. In more extreme cases, that is, when both the main and backup AC power supplies lose power, the system is also equipped with two emergency DC power supplies as a last resort. At this time, the system will automatically switch to the emergency DC power supply, giving priority to ensuring the normal operation of core equipment such as reinforced computers, core controllers, key sensors and alarm devices. This multiple redundant power supply design ensures the continuity of data acquisition, processing and analysis functions, thereby ensuring the uninterrupted analysis and presentation of the real-time working status results of each subsystem.

[0061] The specific automation control subtasks of the monitoring system of the present invention, the control process logic of the automatic fuel transfer is as follows Figure 3As shown in Figure 1, this process automatically controls fuel transfers between fuel tanks based on a preset strategy. The process begins in Step 1, where the operator or system selects the current operating mode. In Step 2, the system determines whether it is set to "Automatic Mode." If "Manual Mode" is selected, the system enters the manual control interface and ultimately terminates (Step 10). If it is in "Automatic Mode," the system proceeds to Step 3, where it automatically issues a command to open the refueling valves in the service fuel tanks. Next, the system checks whether any of the oil extraction valves in standby fuel tanks 1 through 4 are open. If no valve is open, the "Automatic Mode" button flashes, prompting the operator to check and returning to Step 2 to wait for the conditions to be met. If a valve is open, the system proceeds to Step 5, where the "Automatic Mode" button remains lit, indicating that the conditions for automatic transfer are met. In Step 6, the system continuously monitors the liquid level in the service fuel tanks to determine whether a low or very low level warning has been triggered. If a low level warning has been triggered, the system proceeds to Step 7, where it automatically issues a command to start the transfer pump to initiate transfer, and then proceeds to Step 8. If the low-level alarm is not triggered, the system proceeds directly to step 8. In step 8, the system continues to monitor the daily fuel tank level to determine whether a high-level or over-high-level alarm has been triggered. If a high-level alarm is triggered, indicating that the daily fuel tank is full or nearly full, the system proceeds to step 9, where it automatically sends a command to stop the transfer pump. The system then returns to step 2 and restarts the monitoring cycle. If the high-level alarm is not triggered, transfer can continue or is unnecessary, and the system returns to step 2 to continue the monitoring cycle. This series of steps monitors the status of the involved level sensors, valve opening feedback, and the operating current, start / stop status, and other parameters of the transfer pump in real time.

[0062] In a specific embodiment, the fuel system of the present invention mainly collects key parameters such as fuel pressure, fuel temperature, fuel flow and fuel level. Specifically, a group of pressure sensors are arranged in the fuel supply pipeline to monitor the pressure fluctuations of the fuel in real time. The installation position of the pressure sensor is selected at the outlet of the fuel delivery pump to ensure that the acquired pressure data can accurately reflect the working conditions of the fuel supply system. At the same time, a fuel temperature sensor is designed in the fuel supply pipeline and installed at the key node of the fuel return pipeline to monitor the temperature rise of the fuel during the circulation process. The fuel flow is monitored by a high-precision turbine flowmeter or ultrasonic flowmeter, which is installed on the fuel delivery main pipeline to collect changes in the fuel supply amount in real time. In addition, in order to monitor the fuel storage status, a liquid level sensor is installed in the fuel storage tank, and an ultrasonic liquid level meter or a magnetostrictive liquid level meter is used to ensure that the collection of fuel level data is accurate and reliable.

[0063] The hydraulic system mainly collects parameters such as hydraulic oil pressure, temperature, flow rate, and oil contamination. The hydraulic system's pressure monitoring is achieved by installing pressure sensors at the hydraulic pump outlet and the hydraulic actuator inlet, respectively, to achieve real-time collection of hydraulic oil pressure. The hydraulic oil temperature is monitored by installing temperature sensors at the hydraulic oil tank outlet and return port to ensure that the temperature changes of the hydraulic oil as it flows through the entire system can be fully understood. The hydraulic oil flow is monitored using an electromagnetic flowmeter or vortex flowmeter, installed on the main oil line at the hydraulic pump outlet, to monitor changes in the hydraulic system's oil supply in real time. An online oil contamination detection sensor is installed in the hydraulic oil tank return port or main circuit. Using capacitance detection technology, the content of particulate matter or water in the hydraulic oil is monitored in real time to ensure that the cleanliness of the hydraulic system meets operational requirements.

[0064] Ventilation system data collection primarily focuses on wind speed, wind pressure, temperature and humidity, and air quality parameters within the air ducts. Wind speed monitoring in the ventilation system involves installing impeller-type wind speed sensors at key points in the supply and exhaust ducts, providing real-time monitoring of wind speed changes. Pressure collection involves deploying pressure sensors at the fan outlet, duct branch nodes, and terminal vents to ensure comprehensive monitoring of the ventilation system's wind pressure distribution. Temperature and humidity data collection involves installing temperature and humidity sensors at the fan outlet to monitor the supply air's temperature and humidity in real time. Additional sensors are deployed at branch nodes to monitor the temperature and humidity distribution within the ship's ventilation system.

[0065] Data collected from each system is converted from analog sensor output signals to analog signals. After amplification, filtering, and analog-to-digital conversion by the signal conditioning module, it is transmitted to the data acquisition terminal. The data acquisition terminal uses a multi-channel data acquisition card to uniformly collect signals from each sensor and preprocess the data, including noise reduction, calibration, and outlier removal. This preprocessed data is then transmitted to the main control server at the ship monitoring center via a wireless communication module.

[0066] In order to fully reflect the stress distribution on both sides of the bulkhead and the bottom of the ship, the present invention arranges fiber grating stress sensors at the following key locations on both sides of the bulkhead: the upper area of ​​the bulkhead, i.e., near the deck connection location, the middle area, i.e., near the intersection of the bulkhead ribs, and the lower area, i.e., near the transition area of ​​the bottom. At the same time, in order to capture the stress distribution changes of the bulkhead along the longitudinal direction, multiple fiber grating sensor nodes are evenly arranged along the longitudinal direction of the bulkhead to form a longitudinal stress data acquisition matrix. At the same time, the present invention arranges sensors at the following locations in the bottom area: the middle and end of the keel, the intersection of the crossbeams of the bottom plate, and key locations of the bottom outer plate. In order to capture the stress distribution characteristics of the bottom, a uniform sensor grid is also established in the bottom area, and sensor nodes are arranged in the transverse and longitudinal directions to form a stress acquisition matrix covering the entire bottom.

[0067] Fiber Bragg grating stress sensors are installed using an adhesive fixation method. Specifically, at each sensor installation point, a surface treatment agent is first used to remove impurities such as oil and oxide layers, and the installation area is polished to ensure a smooth and clean surface. The fiber Bragg grating sensor is attached to the designated location using high-strength epoxy resin adhesive, and a fiber protection groove is set along the wiring direction of the sensor fiber. To reduce vibration interference on the sensor signal, for sensors installed on both sides of the bulkhead, the fiber optic cable is laid along the direction of the bulkhead ribs; for sensors installed in the bottom area, the fiber optic cable is laid along the direction of the keel, and a fiber optic buffer protection device is added at the intersection of the keel and the crossbeam.

[0068] Each FBG sensor is connected via an optical fiber to a fiber interrogator, which collects real-time signals indicating changes in the FBG wavelength at each sensor node. FBG sensors, based on the Bragg grating effect, measure changes in the reflected wavelength of the FBG to calculate strain data at the corresponding location. To ensure accurate and synchronized data acquisition, the fiber interrogator utilizes a high sampling rate and a data acquisition unit for multi-channel simultaneous acquisition, enabling simultaneous data collection from multiple sensor nodes on both sides of the bulkhead and the bottom of the ship.

[0069] Through the high-frequency sampling of fiber Bragg grating stress sensors, the stress values ​​of each sensor node can be continuously recorded over a period of time, and the corresponding time-series stress matrix can be generated. For each sensor node, a set of continuous stress value sequences is formed in the time dimension. Based on these time-series stress data, an instantaneous stress matrix reflecting the spatial stress distribution can be generated for each time point of a sensor layout grid. For the sensor layout grids on both sides of the bulkhead and the bottom of the ship, for the time point The instantaneous stress matrix is ​​expressed as: ,in, Indicates at a point in time When arranging the grid Row, No. The stress values ​​measured by the sensor nodes in the column. Inside, every moment Corresponding to an instantaneous stress matrix. By arranging these instantaneous stress matrices in chronological order, the time-series stress matrices of the left side of the bulkhead, the right side of the bulkhead, and the bottom of the ship are formed. Among them, the time-series stress matrix of any part is expressed as follows: .

[0070] In the collection of time-series temperature distribution, an array of temperature sensors is arranged on the inner surface of the left side of the bulkhead, the right side of the bulkhead and the bottom of the high-speed ship according to the spatial layout corresponding to the stress sensor grid. Specifically, if the stress sensor grid is m rows and n columns, then m rows and n columns of temperature sensors are also arranged at the same position on each surface to ensure that each temperature sensor node (x, y) is accurately matched or as close as possible to the corresponding stress sensor node (x, y) in space. The present invention uses a patch-type thermistor temperature sensor, which is firmly installed at the corresponding measuring point on the surface by mechanical fixing. All arranged temperature sensors are connected to the same data acquisition unit as the stress data. The data acquisition unit has multi-channel synchronous sampling capability and achieves strict time synchronization with the system for collecting instantaneous stress data. From arrive , the instantaneous temperature matrices collected at each moment are combined in chronological order to form the time series temperature matrices of the left side of the bulkhead, the right side of the bulkhead and any part of the bottom of the ship: ;

[0071] Input stress data and temperature data into the stress monitoring model to obtain stress analysis results including:

[0072] Will Instantaneous stress matrix and the instantaneous temperature matrix Combined into a multi-channel matrix in two-dimensional space As the current frame: , ， set up Multi-channel matrix of moments is the reference frame;

[0073] When a high-speed vessel is underway, its hull structure is not only subjected to wave impact and its own vibrations, but also to temperature fluctuations and the resulting thermal stresses caused by factors such as frictional heating due to speed, ambient water temperature fluctuations, and heat dissipation from nearby equipment. Analyzing stress data alone makes it difficult to identify the root cause of stress changes. Specifically, localized high stresses may be caused by external impact or thermal expansion and contraction due to a sharp temperature gradient. By incorporating a temperature time-series matrix as a parallel channel into the analysis, the present invention can obtain the temperature context corresponding to the stress change. When a significant stress change is detected, combined with synchronized temperature data, its nature can be more accurately determined. If a sudden stress increase is accompanied by a significant local temperature anomaly, it is highly likely due to a collision, grounding, or abnormal contact with another object. If the stress change is associated with large, slow temperature changes, it is related to thermal stress or temperature-dependent changes in material properties. Furthermore, material mechanical properties, such as yield strength and elastic modulus, are also affected by temperature. The same stress value has different implications for structural safety at different temperatures. Incorporating temperature data allows for more precise assessment of stress levels. Furthermore, by analyzing the coupling patterns between stress and temperature in the spatiotemporal dimensions, a more robust anomaly detection model can be constructed. Under normal navigation conditions, stress and temperature fluctuations exhibit an inherent correlation pattern; however, this correlation pattern is disrupted when structural damage, equipment failure, or special external events occur. Therefore, integrating temperature information not only provides richer criteria for identifying the source and nature of stress anomalies, but also improves detection accuracy and reliability through multi-physical quantity cross-validation, reducing the potential for misjudgments or missed detections based solely on stress data, thereby providing more comprehensive and accurate data support for structural health monitoring and operational safety assurance of high-speed vessels.

[0074] Initial position detection: reference frame and the current frame Perform feature extraction and combine it with the gradient change weight map Feature maps of different depths of the current frame Fusion to obtain multi-feature fusion ;

[0075]

[0076] in, and Respectively for and The extracted feature maps are and They are feature extraction and fusion processes respectively. The parameters in brackets are the input parameters of feature extraction or fusion process. d1 and d2 represent different depths. All current frames are traversed to obtain temporal fusion features. .

[0077] Among them, the gradient change weight map based on and The stress change and temperature change are obtained, including: obtaining the stress change matrix and temperature change matrix , normalize stress changes and temperature changes and , the normalized stress change and temperature change are weightedly fused to obtain the fusion change feature map , convert the fusion change feature map into a weight map through Softmax ;

[0078]

[0079] in, and represent the weights of stress and temperature respectively.

[0080] Time series fusion features Input to the primary detection model and output stress anomaly probability map The primary detection model uses a U-shaped network as the processing framework. In the encoding stage, GRU is used for feature extraction and 3D pooling is used for downsampling. The bottleneck layer uses a GRU module to connect the encoding and decoding stages. The decoding stage includes GRU sequence recovery and upsampling. The encoding and decoding stages correspond to skip connections.

[0081] Among them, the i-layer feature processing in the encoding stage is specifically as follows:

[0082]

[0083] in, represents the encoding of the kth hidden state of the i-th layer, and They are the input and output of the i-layer GRU respectively, and the input of the first layer is the temporal fusion feature , the input of the i-th layer is the output of the previous layer 3D pooling , For 3D pooling, , For the GRU module encoding stage, is the step size parameter, Represents kernel function parameters;

[0084] Bottleneck layer passes After processing, H bottle ;

[0085] The i-layer feature processing in the decoding stage is specifically as follows:

[0086]

[0087] in, and are the upsampling of layer i in the decoding stage and the GRU output of layer i in the encoding stage, respectively. and are the kth hidden state and output of the i-th layer in the decoding stage, For 3D upsampling, is the upsampled input, and the input of the fourth layer is the bottleneck layer output H bottle , Decoding stage of the GRU module;

[0088] The input and output of the upper GRU, the input of the first layer is the time series fusion feature , the input of the i-th layer is the output of the previous layer 3D pooling , It is 3D pooling.

[0089] The output layer aggregates the output of the first layer of the decoder in time, that is, takes the maximum value of the time dimension and activates it to obtain the stress anomaly probability map ;

[0090]

[0091] in, Indicates that for the third dimension, that is, the time dimension, the maximum value of each position is selected and aggregated. and They are maximum pooling and two-dimensional upsampling, and They are the feature maps after aggregation and upsampling respectively.

[0092] The input of the position correction process is the stress anomaly probability map obtained by the initial positioning , reference frame feature map and time series fusion features ;

[0093] Time series fusion features Perform temporal aggregation to obtain static feature maps ;

[0094] calculate With the reference frame feature map The difference : ;

[0095] Based on the difference Generate channel attention weights ;

[0096] Based on stress anomaly probability map Generating spatial attention weights ;

[0097] Apply channel attention and spatial attention to temporal fusion features : , get the temporal attention feature ;

[0098] In order to capture the dynamic changes of the impact position over time, based on Generate precise location heatmaps of time series .

[0099] Specifically, The initial time series heat map is obtained through the four-layer position decoding network , where the first two layers of the network are composed of 3D convolution and ReLu activation function, and the last two layers of the network are composed of 3D transposed convolution and ReLu activation function. Softmax activation is applied independently at each time point to obtain a precise positioning heat map , is the length of the time series.

[0100] The present invention adopts a two-stage method combining initial detection and positioning with positioning correction to detect stress anomalies of high-speed ships. In the first stage, a deep learning network structure based on the U-Net framework embedded with the GRU module is used to directly process the time series data that integrates stress and temperature information. The network design of the present invention makes full use of the fact that the U-Net architecture itself is good at capturing multi-scale spatial features through the encoder-decoder structure and jump connections, and has a natural advantage in identifying local abnormal areas in the stress / temperature field; and the GRU modules embedded in each layer effectively process the dependencies and dynamic evolution laws of the data in the time dimension. This enables the first-stage network to fully learn the complex spatiotemporal coupling pattern of the normal stress and temperature field of high-speed ships under complex navigation conditions, so that when processing new time series data, it can quickly and preliminarily identify areas that may deviate from the normal pattern, and generate a probability map indicating the possibility of anomalies, complete preliminary detection, and provide key candidate targets for subsequent analysis.

[0101] Relying solely on the results of the first stage may result in errors in the recognition results. For example, the positioning of the probability map is difficult to fully distinguish between violent normal load fluctuations and real structural abnormal events. For this reason, the second stage of positioning correction is introduced. The core of this stage is to use the difference matrix between the current time series data (stress and temperature) and a specific reference frame to construct an attention network, and use it to correct the time series data. The first frame of data represents the normal state of the hull structure under operational loads, including the initial temperature distribution and possible prestress. Using this as a benchmark to calculate the difference between the data at all subsequent moments and it can very intuitively quantify and highlight all changes caused by environmental changes and potential abnormal events. Based on the comparison method with the initial stable state, it is simpler, more stable and easier to implement than trying to define a normal reference baseline that changes dynamically over time.

[0102] The attention network constructed based on this difference matrix can focus on the spatiotemporal regions and feature channels with the most significant changes and the most likely associations with abnormal events. The attention mechanism is trained to identify regions in the difference matrix that exhibit specific patterns—namely, large values, steep gradients, and durations consistent with impact characteristics—and assign higher weights to these regions. When correcting the original time series data, sharp, localized difference signals truly caused by sudden events such as impacts and groundings are significantly amplified, while smaller, more gradual differences or those consistent with normal operating load fluctuations, such as those caused by normal navigation vibrations, are suppressed. This correction of the time series data based on the difference and attention mechanisms significantly improves the signal-to-noise ratio, making the characteristics of abnormal events more prominent. Using this corrected data for final location analysis significantly improves the accuracy of abnormal event location, reduces potential bias in initial location analysis, and facilitates a more accurate assessment of the true scope and severity of the event.

[0103] This two-stage approach combines the advantages of deep spatiotemporal pattern learning (stage one) and baseline difference-based attention focusing (stage two). Especially in scenarios like high-speed vessels, which face complex dynamic environments and numerous interfering signals, this approach effectively handles complex background fluctuations through the U-Net / GRU framework. Furthermore, by comparing differences with the initial normal state and applying attention correction, it accurately captures and amplifies the signal characteristics of key abnormal events such as impacts and groundings. This enables more reliable and accurate stress anomaly detection and location, providing strong technical support for ensuring the structural safety of high-speed vessels.

[0104] In this embodiment, the integrated monitoring system for ship auxiliary equipment deployed on board completes the acquisition of status parameters of subsystems such as fuel, hydraulics, and ventilation, and can process and derive the current status results: normal, warning, fault; or specific operating parameters such as pressure, flow, liquid level, and valve position status; at the same time, the stress monitoring model within the system completes the processing of stress sensor data and generates a thermal map containing the precise location of the impact within the time window of the event. .

[0105] The vessel's auxiliary equipment then packages these stress analysis results and subsystem status data into a single data packet containing a timestamp, vessel identification number, current status results, and a heat map accurately locating the impact. When no abnormal events are detected, only a general status summary is transmitted at a low frequency. However, if an abnormal event is detected or its severity exceeds a preset threshold, a high-priority transmission is triggered, containing the complete data and relevant subsystem status. The data is compressed and encrypted before transmission and sent to a pre-determined shore-based remote monitoring center via the ship's onboard communication unit.

[0106] The shore-based remote monitoring center is equipped with a communication receiving gateway and data processing server. The server receives data packets from the ship, decrypts, decompresses, and verifies the data, and stores the parsed structured information in a database, triggering the real-time processing and visualization engine.

[0107] The present invention provides an integrated monitoring system for ship auxiliary equipment applicable to high-speed ships, the system comprising:

[0108] Data acquisition module: The data acquisition module collects multiple types of information of high-speed vessels, including stress data, temperature data, and operating parameters of fuel, hydraulic, and ventilation systems;

[0109] Stress analysis module: input stress data and temperature data into the stress monitoring model to obtain stress analysis results;

[0110] A subsystem analysis module, wherein the subsystem analysis module obtains status results of each subsystem based on the operating parameters of the fuel, hydraulic, and ventilation systems;

[0111] Remote control module: The remote control module performs remote monitoring based on stress analysis results and status results of each subsystem.

[0112] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned integrated monitoring method for ship auxiliary equipment applicable to high-speed vessels is implemented.

[0113] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned integrated monitoring method for ship auxiliary equipment applicable to high-speed ships.

[0114] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0115] In this specification, the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the embodiments described later, the description is relatively simple, and the relevant parts can be referred to the partial description of the previous embodiments.

[0116] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for integrated monitoring of ship auxiliary equipment applicable to high-speed ships, characterized in that: The method comprises: S1: Collecting multiple types of information of high-speed vessels, including stress data, temperature data, and operating parameters of fuel, hydraulic, and ventilation systems; S2: Input stress data and temperature data into the stress monitoring model to obtain stress analysis results; S3: Analyzing the operating parameters of the fuel, hydraulic, and ventilation systems to obtain status results of each subsystem; S4: Remote monitoring based on stress analysis results and status results of each subsystem; The stress analysis results include: the time series fusion features of the instantaneous stress data and the instantaneous stress data Input into the primary detection model to obtain the stress anomaly probability map , and using the reference frame feature map and time series fusion features and stress anomaly probability map Perform position correction to obtain stress analysis results; Will Instantaneous stress matrix and the instantaneous temperature matrix Combined into a multi-channel matrix in two-dimensional space As the current frame: , ， set up Multi-channel matrix of moments is the reference frame; Reference frame and the current frame Perform feature extraction and combine it with the gradient change weight map Feature maps of different depths of the current frame Fusion to obtain multi-feature fusion ; in, and Respectively for and The extracted feature maps are and They are feature extraction and fusion processes respectively. The parameters in brackets are the input parameters of feature extraction or fusion process. d1 and d2 represent different depths. All current frames are traversed to obtain temporal fusion features. .

2. The integrated monitoring method for ship auxiliary equipment applicable to high-speed vessels according to claim 1, characterized in that: Gradient change weight map based on and The stress change and temperature change are obtained to obtain the stress change matrix and temperature change matrix , normalize stress changes and temperature changes and , the normalized stress change and temperature change are weightedly fused to obtain the fusion change feature map , convert the fusion change feature map into a weight map through Softmax ; in, and represent the weights of stress and temperature respectively.

3. The integrated monitoring method for ship auxiliary equipment applicable to high-speed vessels according to claim 2, characterized in that: The primary detection model uses a U-shaped network as the processing framework. GRU is used for feature extraction in the encoding stage and 3D pooling is used for downsampling. The bottleneck layer uses the GRU module to connect the encoding stage and the decoding stage. The decoding stage includes the use of GRU to restore the sequence and upsampling. The encoding stage and the decoding stage correspond to jump connections.

4. The integrated monitoring method for ship auxiliary equipment applicable to high-speed ships according to claim 3, characterized in that: The input of the position correction process is the stress anomaly probability map , reference frame feature map and time series fusion features , for time series fusion features Perform temporal aggregation to obtain static feature maps ,calculate With the reference frame feature map The difference , based on the difference Generate channel attention weights ; Based on stress anomaly probability map Generating spatial attention weights , applying channel attention and spatial attention to temporal fusion features , get the temporal attention feature ;based on Generate precise location heatmaps of time series .

5. The integrated monitoring method for ship auxiliary equipment applicable to high-speed ships according to claim 1, characterized in that: The ship's auxiliary equipment integrates the stress analysis results and the status results of each subsystem into a data packet, which contains a timestamp, ship identification number, current status results and a thermal map of the precise impact location. The data packet is sent to the preset shore-based remote monitoring center via the ship's onboard communication unit. The shore-based remote monitoring center server receives the data packet from the ship, stores the parsed structured information in the database, and triggers the real-time processing and visualization engine.

6. An integrated monitoring system for ship auxiliary equipment suitable for high-speed ships, characterized in that: The system is used to execute the integrated monitoring method for ship auxiliary equipment applicable to high-speed ships according to any one of claims 1 to 5, and the system comprises: Data acquisition module: The data acquisition module collects multiple types of information of high-speed vessels, including stress data, temperature data, and operating parameters of fuel, hydraulic, and ventilation systems; Stress analysis module: input stress data and temperature data into the stress monitoring model to obtain stress analysis results; A subsystem analysis module, wherein the subsystem analysis module obtains status results of each subsystem based on the operating parameters of the fuel, hydraulic, and ventilation systems; Remote control module: The remote control module performs remote monitoring based on stress analysis results and status results of each subsystem; The stress analysis results include: the time series fusion features of the instantaneous stress data and the instantaneous stress data Input into the primary detection model to obtain the stress anomaly probability map , and using the reference frame feature map and time series fusion features and stress anomaly probability map Perform position correction and obtain stress analysis results.

Images