Method and system for failure detection of dredging vessel equipment

Abstract

The application discloses a kind of dredging ship equipment fault detection method and system, the present application relates to the technical field of fault detection, determine multiple fault regions according to the identification of fault distribution map, determine the fault level of dredging ship equipment based on the fault content of each fault region, corresponding regional form and corresponding key position, determine the fault detection content of dredging ship equipment according to the fault level of dredging ship equipment, corresponding work task list and the fault event of each key position, improve the accuracy of the fault detection content of dredging ship equipment.Therefore, according to the detection of the dynamic review event, determine the main fault content and multiple branch fault content of dredging ship equipment, determine the fault maintenance event of dredging ship equipment based on the main fault content, multiple branch fault content and the work task list of dredging ship equipment, improve the accuracy of the fault maintenance event of dredging ship equipment.

Description

Technical Field

[0001] This invention relates to the technical field of fault detection, and in particular to a fault detection method and system for dredging vessel equipment. Background Technology

[0002] With the development of technology, dredging vessels have gradually been applied to people's lives. These are engineering vessels specifically designed for underwater earthwork excavation. During operation, dredging vessels further process underwater earthwork. Current technology collects vibration signals from the dredging vessel and determines the corresponding fault area based on these signals. However, this ignores the fault level and fault events in key components, affecting the accuracy of fault detection and maintenance. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a fault detection method and system for dredging vessel equipment.

[0004] This invention provides a fault detection method for dredging vessel equipment, comprising: Based on multiple images of dredging vessels and equipment in different dimensions and the overall shape of the dredging vessels and equipment, a distribution map of the dredging vessels and equipment is determined, and multiple key parts are identified based on the identification of the distribution map of the dredging vessels and equipment. Based on multiple working data of each key part, the corresponding part shape and the corresponding vibration signal, the fault event of the key part is determined, and the fault distribution map of the dredging vessel equipment is determined according to the location of each key part, the corresponding fault event and the current working status of the dredging vessel equipment. Multiple fault areas are identified based on the fault distribution map. The fault level of the dredging vessel equipment is determined based on the fault content, corresponding area shape, and corresponding key parts of each fault area. The fault detection content of the dredging vessel equipment is determined based on the fault level of the dredging vessel equipment, the corresponding work task list, and the fault events of each key part. Based on the fault detection content and the distribution map of dredging vessel equipment, the re-inspection area of ​​dredging vessel equipment is determined. According to the re-inspection area, the usage status of dredging vessel equipment and the corresponding current work tasks, dynamic re-inspection events are determined. The dynamic re-inspection events clarify the target area, data acquisition parameters, special analysis instructions and output alarm thresholds. Based on the detection of this dynamic review event, the main fault content and multiple branch fault content of the dredging vessel equipment are determined. Based on the main fault content, multiple branch fault content and the work task list of the dredging vessel equipment, the fault maintenance event of the dredging vessel equipment is determined. The fault maintenance event not only includes maintenance strategy, priority and suggested maintenance window, but also lists in detail the specific maintenance plan, required spare parts and resource list for the main fault content and all branch fault content.

[0005] This invention provides a fault detection system for dredging vessel equipment, which is applied to the aforementioned fault detection method for dredging vessel equipment. The fault detection system for dredging vessel equipment includes: The identification module is used to determine the distribution map of the dredging vessel based on multiple images of the dredging vessel in different dimensions and the overall shape of the dredging vessel, and to identify multiple key parts based on the identification of the distribution map of the dredging vessel. The fault distribution map module is used to determine the fault events of each key part based on multiple working data, corresponding part morphology and corresponding vibration signals. It determines the fault distribution map of the dredging vessel equipment based on the location of each key part, the corresponding fault events and the current working status of the dredging vessel equipment. The fault detection content module is used to identify multiple fault areas based on the fault distribution map, determine the fault level of the dredging vessel equipment based on the fault content of each fault area, the corresponding area shape and the corresponding key parts, and determine the fault detection content of the dredging vessel equipment based on the fault level of the dredging vessel equipment, the corresponding work task list and the fault events of each key part. The dynamic review event module is used to determine the review area of ​​dredging vessel equipment based on the fault detection content and the distribution map of dredging vessel equipment, and to determine the dynamic review event based on the review area, the usage status of dredging vessel equipment and the corresponding current work task; The fault maintenance event module is used to determine the main fault content and multiple branch fault content of the dredging vessel equipment based on the detection of the dynamic review event, and to determine the fault maintenance event of the dredging vessel equipment based on the main fault content, multiple branch fault content and the work task list of the dredging vessel equipment.

[0006] Compared with the prior art, the beneficial effects of the present invention are: In this embodiment of the invention, the method determines a distribution map of the dredging vessel based on multiple images of the dredging vessel in different dimensions and its overall shape. Multiple key components are identified based on this distribution map. Fault events of each key component are determined based on multiple operational data, corresponding component morphology, and corresponding vibration signals. A fault distribution map of the dredging vessel is then determined based on the location of each key component, its corresponding fault events, and the current operational status of the dredging vessel. Multiple fault areas are identified based on the fault distribution map. The fault level of the dredging vessel is determined based on the fault content of each fault area, its corresponding area morphology, and its corresponding key components. Finally, the fault detection content of the dredging vessel is determined based on its fault level, corresponding task list, and fault events of each key component. By introducing multiple key components, the fault distribution map of the dredging vessel is managed, incorporating a holistic consideration of the fault level, corresponding task list, and fault events of each key component, thus improving the accuracy of the fault detection content.

[0007] Therefore, based on the fault detection content and the distribution map of dredging vessel equipment, the re-inspection area of ​​dredging vessel equipment is determined. Dynamic re-inspection events are determined based on the re-inspection area, the usage status of the dredging vessel equipment, and the corresponding current work tasks. The main fault content and multiple branch fault content of the dredging vessel equipment are determined based on the detection of these dynamic re-inspection events. Fault maintenance events of the dredging vessel equipment are determined based on the main fault content, multiple branch fault content, and the work task list of the dredging vessel equipment. The introduction of dynamic re-inspection events allows for the control of the main fault content and multiple branch fault content, achieving a holistic consideration of the main fault content, multiple branch fault content, and the work task list of the dredging vessel equipment, thus improving the accuracy of fault maintenance events for dredging vessel equipment. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating the fault detection method for dredging vessel equipment in an embodiment of the present invention. Figure 2 This is a flowchart illustrating step S11 of the fault detection method for dredging vessel equipment in an embodiment of the present invention. Figure 3 This is a flowchart illustrating step S12 in the fault detection method for dredging vessel equipment in this embodiment of the invention. Figure 4 This is a flowchart illustrating step S13 in the fault detection method for dredging vessel equipment in this embodiment of the invention. Figure 5 This is a flowchart illustrating step S14 of the fault detection method for dredging vessel equipment in an embodiment of the present invention. Figure 6 This is a flowchart illustrating step S15 of the fault detection method for dredging vessel equipment in an embodiment of the present invention. Figure 7 This is a schematic diagram of the structural composition of the fault detection system for dredging vessel equipment in an embodiment of the present invention. Detailed Implementation

[0009] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0010] Please see Figures 1 to 7 A fault detection method for dredging vessel equipment, applied to fault detection scenarios; the fault detection method for dredging vessel equipment includes: Step S11: Based on multiple images of the dredging vessel equipment in different dimensions and the overall shape of the dredging vessel equipment, determine the distribution map of the dredging vessel equipment, and identify multiple key parts based on the identification of the distribution map of the dredging vessel equipment. Step S12: Based on multiple working data of each key part, the corresponding part shape and the corresponding vibration signal, determine the fault event of the key part, and determine the fault distribution map of the dredging vessel equipment according to the location of each key part, the corresponding fault event and the current working status of the dredging vessel equipment. Step S13: Based on the identification of the fault distribution map, identify multiple fault areas, determine the fault level of the dredging vessel equipment based on the fault content, corresponding area shape and corresponding key parts of each fault area, and determine the fault detection content of the dredging vessel equipment based on the fault level of the dredging vessel equipment, the corresponding work task list and the fault events of each key part. Step S14: Based on the fault detection content and the distribution map of dredging vessel equipment, determine the re-inspection area of ​​dredging vessel equipment, and determine the dynamic re-inspection event according to the re-inspection area, the usage status of dredging vessel equipment and the corresponding current work task; Step S15: Based on the detection of the dynamic review event, determine the main fault content and multiple branch fault content of the dredging vessel equipment, and determine the fault maintenance event of the dredging vessel equipment based on the main fault content, multiple branch fault content and the work task list of the dredging vessel equipment.

[0011] refer to Figure 2 In step S11, the specific steps are as follows: S111: Collect the current position of the dredging vessel equipment, determine the shooting space of the dredging vessel equipment based on the current position of the dredging vessel equipment and multiple surrounding cameras, and collect multiple images of the dredging vessel equipment in different dimensions based on the dynamic shooting of the shooting space. S112: Mark the overall shape of the dredging vessel equipment, determine multiple feature matching contents based on the overall shape of the dredging vessel equipment and multiple images of the dredging vessel equipment in different dimensions, and construct a distribution map of the dredging vessel equipment based on the synthesis of multiple feature matching contents. S113: Based on the identification of the distribution map of the dredging vessel equipment, multiple equipment parts are identified, and multiple key parts are identified according to the multiple equipment parts, their corresponding functions, and the usage scenarios of the dredging vessel equipment.

[0012] In the embodiments of this application, the current position of the dredging vessel is collected, and the shooting space of the dredging vessel is determined based on the current position of the dredging vessel and multiple surrounding cameras. Multiple images of the dredging vessel in different dimensions are collected according to the dynamic shooting of the shooting space. This approach takes into account both the current position of the dredging vessel and the overall consideration of multiple surrounding cameras, ensuring the accuracy of the shooting space of the dredging vessel.

[0013] At this time, the system collects latitude and longitude coordinates (such as 31.2345N, 121.4567E) in real time through GPS / BeiDou dual-mode positioning module. At the same time, it combines dead reckoning technology and uses ship speed (such as 8 knots) and heading angle (such as 45°) data to compensate for positioning deviations when satellite signals are weak. This multi-source positioning method can control the position error within 3 meters, which is particularly suitable for near-shore operating environments.

[0014] The system pre-builds a parameter database containing 12 fixed cameras in the port. Each camera records its installation location (e.g., dock coordinates), pan-tilt angle range (horizontal ±180° / vertical -30°~90°), and focal length parameters (5-50mm zoom). When it receives ship position data, the algorithm automatically calculates the spatial geometric relationship between the ship and each camera. For example, when the ship is located at coordinates (31.2345, 121.4567), the system determines that cameras A (distance 500m / azimuth 30°), B (distance 800m / azimuth 120°), and C (distance 600m / azimuth 270°) are within the effective shooting range. At the same time, it calculates the optimal focal length setting required for each camera based on the ship's dimensions (length 120m × width 22m).

[0015] The system simultaneously sends control commands to three available cameras: camera A adjusts to a 15° downward angle to capture the deck machinery status, camera B zooms to 35mm to focus on the mud pipe connection area on the side of the hull, and camera C executes panoramic mode to obtain an overall operational view. During the acquisition process, the system dynamically adjusts parameters through real-time image analysis. When it detects that the ship is rolling ±3° due to waves, it immediately compensates for the camera gimbal angle. The final image sequence includes: a close-up of the rake arm hydraulic system from the front view, details of the mud gate sealing surface from the side view, and the layout of deck equipment from the top view. These images are automatically stitched together to form a 3D point cloud model covering key equipment, providing a high-quality data foundation for the subsequent deep learning-based fault diagnosis module.

[0016] Furthermore, the overall shape of the dredging vessel equipment is marked, and multiple feature matching contents are determined based on the overall shape of the dredging vessel equipment and multiple images of the dredging vessel equipment in different dimensions. A distribution map of the dredging vessel equipment is constructed based on the synthesis of multiple feature matching contents, which takes into account both the overall shape of the dredging vessel equipment and the overall consideration of multiple images of the dredging vessel equipment in different dimensions, and ensures the accuracy of multiple feature matching contents.

[0017] At this point, the system retrieves its design CAD model and BOM list, and uses ontology technology to construct a knowledge graph. In this graph, the equipment is divided into primary units such as power modules and transmission modules, and further subdivided into secondary components such as Yaskawa variable frequency motors, ZF gearboxes, Warman mud pumps, and SKF bearing housings. The system not only records the model and dimensions of each component, but also clearly defines the spatial relationships (such as the motor being connected to the input end of the gearbox) and functional relationships (such as the motor driving the gearbox), forming a complete prior knowledge framework for the equipment.

[0018] The system performs in-depth analysis on the multi-dimensional images acquired in the S111 phase, employing a combination of top-down and bottom-up strategies. For example, in the front image of the unit, the algorithm identifies a cylindrical structure with heat dissipation fins through edge detection, and its contour features match the variable frequency motor model in the knowledge graph with a matching degree of up to 95%. In the side image, the system locates the box structure with an observation window and oil level gauge through feature point matching, confirming it as a ZF gearbox. Most importantly, the system performs multi-view geometric verification. It performs three-dimensional spatial back projection on the flanges and bolt groups identified in the front, side, and top images, respectively, to verify that they can accurately connect the motor, gearbox, and mud pump in space. This cross-verification mechanism improves the accuracy of component identification to over 99%, and all successfully matched components together constitute a reliable set of feature matching content.

[0019] Based on the calibration parameters of each camera, the system backprojects the component feature points in all images onto a unified three-dimensional coordinate system, generating a three-dimensional point cloud model containing millions of points. The system accurately assigns predefined semantic labels to this three-dimensional model; for example, it labels the motor's outline area in the point cloud as the motor drive end and the gearbox area as the input shaft bearing housing. The final generated dredging vessel equipment distribution map is a highly visualized interactive model. Operators can not only rotate 360 ​​degrees to view the overall structure of the unit, but also click on any component to instantly obtain its name, model, design parameters, and real-time status information extracted from the images (such as surface cleanliness: good, number of bolts: 8, no signs of loosening). This distribution map becomes a unified spatial benchmark for subsequent fault detection, condition assessment, and maintenance decisions, realizing a precise mapping from physical equipment to the digital world.

[0020] Therefore, based on the identification of the distribution map of the dredging vessel equipment, multiple equipment parts are determined. Based on the multiple equipment parts, their corresponding functions, and the usage scenarios of the dredging vessel equipment, multiple key parts are determined. This approach takes into account the overall consideration of multiple equipment parts, their corresponding functions, and the usage scenarios of the dredging vessel equipment, ensuring the accuracy of multiple key parts.

[0021] At this point, the system uses a 3D point cloud segmentation algorithm or a graph neural network (GNN) to traverse and segment the entire digital model, aggregating continuous regions belonging to the same physical object (such as all voxels marked as gearboxes) into an independent component instance. The system then automatically identifies and generates a detailed component list, including not only major modules such as the variable frequency motor, gearbox, and No. 1 mud pump, but also specific sub-components such as the motor drive end bearing housing, gearbox input shaft, and coupling. Each item in the list is accompanied by metadata such as its 3D spatial coordinates, geometric dimensions, and connection relationships, and is finally compared with the knowledge base to ensure the completeness and accuracy of the identification. For example, it confirms that the two bearing housings are correctly distinguished as the first bearing housing and the second bearing housing.

[0022] The system assigns a basic functional weight to each component based on its importance within the system. For example, the weight of the gearbox and mud pump bearing, which are the core of the power system, is set to 0.9, while the weight of transmission components such as couplings is 0.7. The system obtains current usage scenario information, such as workload, operation mode, and environmental parameters, from the ship's central control system in real time. In the scenario of dredging high-concentration clay on the Hangjun 401 with a load of 90%, the system dynamically adjusts the scenario impact factor of each component: the gearbox, which bears a huge load, has a factor of 1.6; the mud pump bearing, which faces a high risk of wear, has a factor of 1.5; and the motor bearing, which is under high load, has a factor of 1.3.

[0023] After calculating the criticality score using the formula = basic functional weight × scenario impact factor, the gearbox scored 1.44, the mud pump bearing scored 1.35, and the motor bearing scored 1.105. The system used a threshold of 1.0 for filtering and finally output a dynamic list of critical components: [gearbox, mud pump bearing, motor bearing]. This list guides the subsequent monitoring system to prioritize high-precision sensors and analysis resources on these weakest and most critical links, ensuring accurate and efficient fault early warning under complex working conditions.

[0024] refer to Figure 3 In step S12, the specific steps are as follows: S121: In multiple key parts, multiple working data of the key parts are determined based on the detection of each key part, so as to collect multiple working data of each key part. At the same time, the corresponding part shape is determined based on the shape recognition of each key part. S122: Collect vibration signals from various key components, determine the corresponding vibration data combination based on the tracing of the vibration signals, and determine the fault event of the key component based on the vibration data combination of each key component, multiple working data of each key component, and the corresponding component morphology; at the same time, collect the current working status of the dredging vessel equipment, and determine the fault distribution map of the dredging vessel equipment based on the location of each key component, the corresponding fault event, and the current working status of the dredging vessel equipment.

[0025] In the embodiments of this application, multiple working data of a key part are determined based on the detection of each key part, so as to collect multiple working data of each key part. At the same time, the corresponding part shape is determined based on the shape recognition of each key part, which is compatible with the overall consideration of the detection of each key part and ensures the accuracy of multiple working data of the key parts.

[0026] At this point, based on the list of key components identified by S113, the system queries the built-in equipment knowledge base and automatically associates each component with the data type of work that needs to be monitored. For example, the system will associate the gearbox with lubricating oil temperature and pressure, and the motor bearing with bearing temperature and motor stator current. The system subscribes to and collects these data streams from the ship's PLC or DCS in real time through industrial bus protocols such as ModbusTCP, ensuring that all parameters are accurately aligned with the vibration signal in terms of timestamps, providing reliable data support for subsequent correlation analysis.

[0027] The system utilizes the equipment distribution map generated by S112 to accurately define the two-dimensional region of interest (ROI) corresponding to each key component in the multi-dimensional image. Advanced vision algorithms are applied to each ROI: the geometry and pose of the component are measured through edge detection and contour analysis, and minute deformations or displacements are identified by comparison with a benchmark model; semantic segmentation algorithms based on convolutional neural networks are used to analyze the surface texture of the component to identify abnormal morphologies such as cracks, corrosion, oil stains, and liquid leaks; and object detection algorithms are used to identify bolts and other connectors to determine whether they are missing or loose. All analysis results are structured into a dataset describing the current physical state of each key component, forming a comprehensive visual status report.

[0028] Specifically, the Hangjun 401 trailing suction hopper dredger was introduced. In phase S113, the key components were identified as [gearbox, mud pump bearing, and motor bearing]. After querying the knowledge base, the system configured data acquisition tasks for these three key components: the gearbox required data acquisition of lubricating oil temperature and pressure; the mud pump bearing and motor bearing required data acquisition of their respective temperatures; and the motor bearing also required data acquisition of three-phase current. At the current time T, the system successfully acquired data from the ship's PLC via the Modbus TCP protocol: the gearbox lubricating oil temperature was 85.2℃ (overheated), and the pressure was 2.1MPa (normal); the mud pump bearing temperature was 72.5℃ (normal); the motor bearing temperature was 68.1℃ (normal); and there was a slight imbalance in the motor's three-phase current (phase A 322A, phase B 315A, phase C 319A).

[0029] The system automatically located three key areas in the real-time images and performed feature analysis. For the gearbox, the surface texture analysis algorithm detected a dark infiltrated area on its bottom, and the morphological recognition model classified it as an oil stain with a 92% confidence level. Geometric analysis showed that the gearbox body had no tilting deformation. For the mud pump and motor bearing housing, the target detection algorithm confirmed that all fixing bolts were in place and without displacement, and the surface texture analysis results were all clean. The system generated a morphological report for the gearbox: {Geometric Morphology: Normal, Surface Anomaly: [{Type: Oil Stain, Location: Near the Bottom Oil Seal, Confidence Level: 92%}]}. Thus, at time point T, the system prepared a complete data package containing working data and part morphology for each key part.

[0030] Furthermore, vibration signals from various key components are collected, and corresponding vibration data combinations are determined based on the tracing of these vibration signals. Fault events for each key component are then identified based on these vibration data combinations, multiple operational data points for each key component, and the corresponding component morphology. Simultaneously, the current operational status of the dredging vessel is collected. A fault distribution map of the dredging vessel is determined based on the location of each key component, the corresponding fault events, and the current operational status of the dredging vessel. This comprehensive approach, considering the location of each key component, the corresponding fault events, and the current operational status of the dredging vessel, ensures the accuracy of the fault distribution map.

[0031] At this point, the system synchronously acquires the original vibration time-domain signal at a high sampling rate using IEPE-type accelerometers installed on key parts. The acquisition host or edge computing module processes the original signal in real time to generate a multi-dimensional vibration data combination. This combination is not a single data set, but a feature vector set containing time-domain features (such as root mean square and kurtosis), frequency-domain features (rotation frequency, gear meshing frequency, bearing failure frequency, etc. obtained through FFT), envelope spectrum features (sensitive to early impact failures), and order tracking features (eliminating the influence of speed fluctuations). This constitutes the core basis for subsequent diagnosis.

[0032] The system uses a multimodal fusion model to integrate all available data and make accurate fault judgments for each key component. An advanced fusion diagnostic model (such as a multi-input deep learning network based on an attention mechanism) takes three types of data as input: a combination of vibration data as the core diagnostic basis, working data that provides the operating condition context, and part morphology that provides visual verification and corroboration. The model performs intelligent diagnosis by cross-validating evidence from different sources. For example, when vibration characteristics point to gear damage, working data shows that the gearbox temperature exceeds the standard, and morphological recognition finds oil stains at the bottom, the model will determine with high confidence that the fault event is: gearbox gear damage (moderate), accompanied by decreased lubrication performance and potential oil seal leakage.

[0033] The system obtains the current working status (such as heavy-load excavation) from the ship's mission management system in real time, providing key working condition background for fault assessment; the system uses the fault events diagnosed in each key part as semantic tags and accurately attaches them to the corresponding spatial coordinate points in the three-dimensional equipment distribution map generated by S112; the final generated fault distribution map is dynamically rendered, conveying rich information through color coding (such as red for severe and yellow for warning), icon labels and dynamic effects (such as flashing frequency to indicate urgency), and superimposing the current working status on the map, providing the operator with a clear fault situation map.

[0034] Specifically, in stage S121, working data and part morphology of [gearbox, mud pump bearing, motor bearing] have been collected, and it is currently in a heavy-load excavation state; the system collects raw signals from the vibration sensor of the gearbox and generates vibration data in real time, showing that the amplitude of the 2nd and 3rd harmonics of the gear meshing frequency exceeds the normal threshold by 25%, and obvious frequency sidebands appear on both sides, which are typical characteristics of gear damage.

[0035] The system inputs the vibration data of the gearbox, the working data collected by S121 (temperature 85.2℃), and the part morphology (oil stains at the bottom) into the fusion diagnostic model. After calculation, the model outputs the fault events: gearbox gear pitting / stripping (moderate), decreased lubrication performance, suspected oil seal leakage, with a confidence level of up to 96%; while all data of the mud pump bearing and motor bearing are within the normal range and are judged to be fault-free.

[0036] After acquiring the current working status of the heavy-duty excavation, the system begins spatial rendering on the 3D equipment distribution map: the gearbox component model is rendered in yellow (warning) and labeled with the fault: gear damage (moderate); at the same time, the location of oil stains is marked at the bottom of the gearbox model with a flashing blue water droplet icon; other components remain green; this dynamic fault distribution map seen by the operator on the monitoring screen allows them to grasp the overall situation instantly, providing an accurate and intuitive decision-making basis for the next step of fault level assessment and maintenance strategy formulation.

[0037] refer to Figure 4 In step S13, the specific steps are as follows: S131: In the fault distribution map, multiple fault areas are identified based on the fault distribution map, and the fault content of each fault area is marked; at the same time, the regional morphology of each fault area is collected, and the first fault coefficient is determined according to the fault content of each fault area and the corresponding key parts. S132: Determine the second fault coefficient based on the regional morphology and corresponding key parts of each fault area, and determine the fault level of the dredging vessel equipment according to the mapping relationship between the first fault coefficient, the second fault coefficient and the fault level. S133: Collect the work task list of dredging vessel equipment, determine multiple sub-work tasks based on the detection of the work task list of dredging vessel equipment, determine the first sub-fault detection content according to the multiple sub-work tasks and the fault level of dredging vessel equipment, determine the second sub-fault detection content according to the multiple sub-work tasks and the fault events of each key part, and determine the fault detection content of dredging vessel equipment based on the first sub-fault detection content and the second sub-fault detection content.

[0038] In the embodiments of this application, multiple fault regions are identified based on the fault distribution map, and the fault content of each fault region is marked. At the same time, the regional morphology of each fault region is collected, and a first fault coefficient is determined according to the fault content of each fault region and the corresponding key parts. This takes into account the overall consideration of the fault content of each fault region and the corresponding key parts, and ensures the accuracy of the first fault coefficient.

[0039] At this point, the system uses image analysis or graph traversal algorithms to scan the entire fault distribution map and identify all components or spatial areas that are assigned abnormal states (such as yellow warnings or red alarms) or have fault labels as independent fault areas. The system reads the fault event semantic labels generated in step S12 on each fault area and organizes them into a structured list. Each item contains a unique area ID, component name, spatial coordinates, and a specific description of the fault content, laying the data foundation for subsequent quantitative evaluation.

[0040] The system will re-invoke the part morphology data generated for the critical part in stage S121, but this time the collection is more targeted, focusing on morphological evidence strongly related to the fault content. For example, for gear damage faults, it will focus on collecting data such as oil stains and deformation on the gearbox surface. The system calculates the first fault coefficient through a built-in evaluation model. This coefficient is mainly determined by two dimensions: one is the inherent danger of the fault type (represented by the danger weight preset by the expert knowledge base, such as 0.8 for gear damage and 0.4 for poor lubrication); the other is the severity of the fault (quantifying descriptions such as moderate and severe into scores, such as 0.6 for moderate). The system calculates a value between 0 and 1 through a weighting function (such as multiplying the two). The higher this first fault coefficient, the more dangerous the physical properties of the fault itself are.

[0041] Specifically, the fault distribution map generated in stage S12 shows that the gearbox is marked with a yellow warning, and the fault content is gear damage (moderate), decreased lubrication performance, and suspected oil seal leakage; the motor is marked with a blue information prompt, and the fault content is rotor imbalance (minor).

[0042] After scanning the fault distribution map, the system identified and marked two fault areas. Fault area 1 (FA-001) was recorded as follows: the component name is gearbox, the spatial coordinates are [x1,y1,z1], the fault content is gear damage (moderate), decreased lubrication performance, and suspected oil seal leakage. Fault area 2 (FA-002) was recorded as follows: the component name is motor, the spatial coordinates are [x2,y2,z2], and the fault content is rotor imbalance (minor).

[0043] The system begins calculating the first fault coefficient for each region. For fault region 1 (gearbox), the system collects morphological evidence of oil stains at its bottom, which is highly correlated with the fault content. According to the model, the risk weight for the fault type "gear damage" is 0.8, and the score for "moderate severity" is 0.6. The system calculates its first fault coefficient as 0.8 × 0.6 = 0.48. For fault region 2 (motor), its appearance is not obviously abnormal, the risk weight for the fault type "rotor imbalance" is 0.5, and the score for "slight severity" is 0.3. Its first fault coefficient is 0.5 × 0.3 = 0.15. The system assigns a quantified first fault coefficient to all fault regions. This preliminary quantification result will be passed to S132 for use in conjunction with other factors to make the final fault level assessment.

[0044] Furthermore, a second fault coefficient is determined based on the regional morphology and corresponding key components of each fault area. The fault level of the dredging vessel equipment is determined according to the mapping relationship between the first fault coefficient, the second fault coefficient, and the fault level. This approach takes into account the overall consideration of the mapping relationship between the first fault coefficient, the second fault coefficient, and the fault level, ensuring the accuracy of the fault level of the dredging vessel equipment.

[0045] At this point, the system performs in-depth analysis of the morphology of the area collected in S131. Based on whether the fault has caused visible leakage, structural damage, or connection failure, a morphological impact score is given; for example, a slight oil stain is 0.1, while obvious dripping is 0.4. At the same time, the system retrieves information from stage S113 and assigns a critical weight to the components corresponding to the fault area. The weight of core power components (such as the main gearbox) is 1.0, while that of general auxiliary components is 0.4. The system combines the two through a fusion model (such as: second fault coefficient = morphological impact score + (1 - morphological impact score) × critical weight) to obtain the second fault coefficient. This coefficient cleverly reflects that even if the morphological manifestation of the fault itself is not serious, if it occurs in an extremely critical part, its overall risk will still be very high.

[0046] The system employs a weighted fusion approach, combining a primary fault coefficient representing inherent risk and a secondary fault coefficient representing external risk to calculate a comprehensive fault score (e.g., comprehensive score = 0.4 × primary coefficient + 0.6 × secondary coefficient, with external risk having a higher weight). The system then compares this comprehensive score with a built-in mapping rule, which is typically a piecewise function. For example, a comprehensive score greater than 0.75 indicates a Level 1 fault (urgent), while a score between 0.50 and 0.75 indicates a Level 2 fault (important). The final output fault level is a clear and standardized conclusion that directly determines the priority and urgency of subsequent maintenance responses.

[0047] Specifically, in stage S131, the first fault coefficients have been calculated for the two fault areas: 0.48 for the gearbox (FA-001) and 0.15 for the motor (FA-002). The system then begins to determine the second fault coefficients. For fault area 1 (gearbox), the system analysis shows obvious oil stains at its bottom, with a shape influence score of 0.3. At the same time, as a core transmission component, the gearbox has a criticality weight of 1.0. Therefore, its second fault coefficient is calculated as: 0.3 + (1 - 0.3) × 1.0 = 1.0. For fault area 2 (motor), its appearance shows no obvious abnormalities, with a shape influence score of 0.0. However, as a core power component, the motor also has a criticality weight of 1.0, and its second fault coefficient is calculated as: 0.0 + (1 - 0.0) × 1.0 = 1.0.

[0048] Using weights w1=0.4 and w2=0.6, the overall fault score was calculated as follows: the gearbox (FA-001) had an overall score of 0.4×0.48+0.6×1.0=0.792; the motor (FA-002) had an overall score of 0.4×0.15+0.6×1.0=0.66. According to the mapping rules, the gearbox's overall score of 0.792 is greater than 0.75, and it is rated as a Level 1 fault (urgent); the motor's overall score of 0.66 falls within the range of (0.50, 0.75), and it is rated as a Level 2 fault (important). This final classification clearly indicates that the gearbox problem requires the highest priority emergency response, providing a clear and unambiguous action guide for the crew and shore-based management personnel.

[0049] Therefore, a work task list for dredging vessels is collected. Based on the detection of the work task list, multiple sub-tasks are determined. The first level of sub-fault detection content is determined based on these sub-tasks and the fault level of the dredging vessel equipment. The second level of sub-fault detection content is determined based on these sub-tasks and fault events in various key components. Finally, the overall fault detection content of the dredging vessel equipment is determined based on both the first and second level sub-fault detection content, ensuring the accuracy of the fault detection content. Furthermore, multiple key components are introduced to control the fault distribution map of the dredging vessel equipment, further improving the accuracy of the fault detection content by considering the fault level, corresponding work task list, and fault events in various key components.

[0050] At this time, the system obtains the current macro-level work task list in real time from the ship's navigation or construction management system through the API interface, such as completing the dredging operation of the N7 channel of the Yangtze River Estuary within 24 hours. The system uses natural language processing and task decomposition engine to break down this high-level task into a series of sub-work tasks directly related to the equipment's operating status, such as parsing it as: [High-power navigation to the N7 channel, continuous heavy-load dredging operation], providing a precise operational context for the subsequent formulation of targeted detection strategies.

[0051] The system has a built-in fault level-response strategy knowledge base. Based on the highest fault level determined by S132 and combined with the current sub-task, the first level of detection content is generated. For example, for a level 1 fault (urgent), regardless of the current task, the first level of content is usually to immediately stop the current task and execute the safety shutdown procedure. For a level 2 fault (important), the content may be to assess task compatibility. If compatible, the system will reduce the load and strengthen monitoring.

[0052] The system has a built-in fault event-detection checklist knowledge base, which associates standard detection methods with each specific fault event. For example, for a gearbox gear damage event, the system will generate specific instructions such as collecting oil samples for ferrography analysis and using an endoscope to inspect the gear meshing surface; for a motor rotor imbalance event, it will generate steps such as measuring the rotor imbalance when the machine is stopped.

[0053] The system prioritizes the first layer of content as the overall guideline, followed by the second layer of content as a specific execution checklist. It also prioritizes the test items based on urgency and relevance. Finally, it outputs a structured electronic work order, which includes overall instructions, a specific test checklist (including methods, standards, and tools), and safety tips, seamlessly transforming the diagnostic results into clear instructions that the crew can immediately execute.

[0054] Specifically, in phase S132, it has been determined that the gearbox is a Level 1 fault (urgent) and the motor is a Level 2 fault (important). The current task is to complete the dredging operation of the N7 channel of the Yangtze River Estuary within 24 hours.

[0055] The system identified the highest fault level as Level 1 (emergency) and generated the first level of content accordingly: immediately suspend dredging operations, start the backup pump unit (if available), execute the main mud pump unit safety shutdown procedure, isolate the gearbox, and prepare for port maintenance; the system generated detection items for the two fault events respectively: for gearbox gear damage, generate the collection of oil samples for testing, preparation of endoscope, and recording of vibration data; for motor rotor imbalance, generate the measurement of rotor radial runout after shutdown.

[0056] The system integrates the above information and generates a final instruction, which is then sent to the crew's terminal. The instruction begins with "Level 1 Fault Alarm," lists general instructions such as immediately stopping operations and safely shutting down the machine, and includes a detailed checklist of specific tests. The first item is for the gearbox, the second for the motor, and each item contains clear operating steps. This work order received by the crew ensures the safety of the current operation and provides necessary data collection guidance for subsequent precise maintenance, perfectly transforming technical diagnosis into efficient on-site action.

[0057] refer to Figure 5 In step S14, the specific steps are as follows: S141: Collect a distribution map of dredging vessel equipment, identify the distribution areas of multiple key parts based on the distribution map of dredging vessel equipment, and determine the re-inspection area of ​​dredging vessel equipment according to the distribution areas of multiple key parts, the corresponding part functions and the fault detection content. S142: Collect multiple usage data of dredging vessel equipment, determine the usage status of dredging vessel equipment based on the multiple usage data of dredging vessel equipment, and determine the content of the first dynamic review based on the review area and the usage status of dredging vessel equipment; S143: Obtain the current working task of the dredging vessel equipment, determine the second dynamic review content based on the current working task of the dredging vessel equipment and the review area, and determine the dynamic review event based on the first and second dynamic review content.

[0058] In the embodiments of this application, a distribution map of dredging vessel equipment is collected, and the distribution areas of multiple key parts are determined based on the identification of the distribution map of dredging vessel equipment. The re-inspection area of ​​dredging vessel equipment is determined according to the distribution area of ​​multiple key parts, the corresponding part functions and the fault detection content. This takes into account the overall consideration of the distribution area of ​​multiple key parts, the corresponding part functions and the fault detection content, and ensures the accuracy of the re-inspection area of ​​dredging vessel equipment.

[0059] At this point, the system retrieves the 3D distribution map of the equipment with complete semantic information generated in stage S112 as a spatial reference. Based on the list of key parts determined in stage S113, the system highlights the key parts on the distribution map and automatically generates a precise distribution area by analyzing the 3D model of each key part. This area is usually a 3D bounding box or a tightly wrapped convex hull, which clearly defines the precise boundary and volume of the component in space, providing a quantitative basis for subsequent spatial planning.

[0060] All critical components diagnosed as fault events in S12 are directly and unconditionally designated as review areas. More intelligently, the system uses an associated extension step: it analyzes fault-free critical components and, through built-in equipment function topology diagrams and fault impact path predictions, determines whether they need to be included in the review scope. For example, if the gearbox is a vibration fault area, then the couplings and motor drive-end bearings located on the vibration transmission path, even if they are fault-free, will have their areas extended to review areas. Finally, the system integrates all directly designated and associated extended distribution areas to form one or more final review area sets with review priorities, achieving proactive risk control.

[0061] Specifically, in stage S13, the gearbox was identified as a Level 1 fault (urgent) and the motor as a Level 2 fault (important). The list of critical components included [gearbox, motor, mud pump, coupling, coupling B]. The system retrieved the 3D distribution map of the equipment and automatically generated precise 3D boundaries for each critical component: a bounding box covering the entire gearbox housing, a bounding box covering the motor stator and the bearing housings at both ends, and a cylindrical bounding box precisely enclosing the coupling, etc. The spatial location of all critical components was clearly defined.

[0062] The system begins to identify the review areas; the distribution areas of the gearbox, which has a first-level fault, and the motor, which has a second-level fault, are directly designated as review areas RA-001 and RA-002, and are given the highest and second-highest priorities, respectively; correlation expansion is performed: the system analysis shows that the gear damage in the gearbox is a typical mechanical vibration fault, and by querying the functional topology map, it is found that the vibration is directly transmitted to the motor through the coupling.

[0063] Therefore, even though the coupling itself is not faulty, as a critical path for fault propagation, its distribution area is intelligently expanded into the review area RA-003, with a priority set to medium. The system integrates and outputs three review areas: [RA-001 (gearbox), RA-002 (motor), RA-003 (coupling)]. The delineation of this review area provides a precise spatial target for subsequent dynamic monitoring, effectively preventing the spread and omission of risks caused by fault propagation.

[0064] Furthermore, multiple usage data of dredging vessels and equipment are collected, and the usage status of the dredging vessels and equipment is determined based on these multiple usage data. The content of the first dynamic review is determined according to the review area and the usage status of the dredging vessels and equipment, which takes into account the overall consideration of the review area and the usage status of the dredging vessels and equipment, and ensures the accuracy of the content of the first dynamic review.

[0065] At this time, the system collects a series of multi-dimensional usage data in real time through the ship's onboard sensor network and control system, including the main motor power and output torque reflecting the load, the engine room temperature and ship roll angle reflecting the environment, and the engine speed reflecting operating parameters. The system integrates these multi-dimensional data into a qualitative usage state description, such as standby, steady-state operation, high load or severe sea conditions, through a state assessment model based on a rule engine or fuzzy logic, providing a precise operating condition background for formulating dynamic review strategies.

[0066] The system has a built-in usage status-review strategy mapping table. Once the current usage status is determined, the system will match the corresponding review strategy for all review areas and generate specific first-level dynamic review content. For example, when the status is high load and heavy load, the strategy will require the highest density of monitoring, which may include increasing the sampling frequency of vibration sensors to 200%, shortening the data acquisition interval, and enabling high-frequency impact pulse monitoring. When the status is severe sea state, the strategy will focus on data validity, which may include enabling adaptive noise filtering algorithms to suppress interference and marking affected data segments to reduce the false alarm rate.

[0067] Specifically, S141 has identified the inspection areas as [RA-001 (gearbox), RA-002 (motor), RA-003 (coupling)], and the vessel is currently conducting full-power dredging operations; the system has collected real-time data showing that the main motor power is 98% and the output torque is 102%, far exceeding the 85% threshold; based on this data, the condition assessment model quickly determines that the current operating condition is high load and heavy load.

[0068] Based on the high-load and heavy-duty state, the system matched the highest priority review strategy to all review areas and generated the first level of dynamic review content for execution. The instruction required: immediately increasing the sampling frequency of all vibration sensors in the gearbox, motor, and coupling areas from 50kHz to 100kHz; shortening the data acquisition and processing interval from 10 seconds to 2 seconds; enabling high-frequency impact pulse monitoring for the gearbox and coupling to capture early impacts; and simultaneously opening a dedicated window for these areas on the main monitoring interface to display the trend graphs of key vibration indicators in real time.

[0069] Therefore, by obtaining the current working task of the dredging vessel and equipment, determining the second level of dynamic review content based on the current working task of the dredging vessel and equipment and the review area, and determining the dynamic review event based on the first level of dynamic review content and the second level of dynamic review content, the overall consideration of the first level of dynamic review content and the second level of dynamic review content is taken into account, ensuring the accuracy of the dynamic review event.

[0070] At this point, the system obtains the most granular current work task being executed in real time from the ship's construction or control system via API interface, such as dredging operations or the mud pump starting up. The system cross-analyzes this specific task with the functional attributes of the review area defined in S141 to generate task-oriented review instructions. For example, if the current task is maneuvering, the review content for the transmission system review area will focus on torque modulation sidebands and transient impact components; while for the mud pump review area in a steady-state dredging task, the focus may be on the correlation analysis between vibration and outlet pressure, thus making the review more targeted.

[0071] The system intelligently merges the state-based review strategy (first level) and the task-based review focus (second level) to form a final, complete, and immediately executable dynamic review event. The system integrates the first-level content generated in S142 (macro-level instructions on how to review, such as sampling frequency) and the second-level content generated in S143.1 (specific focus on what to review). The first-level content forms the framework of the review, and the second-level content fills the framework. The system formats the merged content into a structured dynamic review event containing all necessary parameters. The dynamic review event clearly defines the target area, data acquisition parameters, special analysis instructions, and output alarm thresholds, and can be directly sent to the monitoring and analysis modules for execution.

[0072] Specifically, S141 has identified the re-inspection areas as [RA-001 (gearbox), RA-002 (motor), RA-003 (coupling)], and S142 has generated the first level of content, such as increasing the sampling rate, based on the high load and heavy load status. The ship is currently performing a maneuver from the dredging area to the dumping area.

[0073] The system obtains that the current task is maneuvering and generates a second dynamic review based on this. The system analysis shows that during maneuvering, the gearbox and coupling will be subjected to frequent torque changes, so the focus of the review is the torque modulation sideband and transient impact component in the vibration signal. As the power source, the speed and load of the motor will fluctuate drastically, so the focus of the review becomes the analysis of current harmonics and speed stability.

[0074] The system integrates the high-intensity monitoring framework of S142 and the task-oriented analysis focus of S143.1 to generate and issue a complete dynamic review event command. This command not only includes execution parameters such as a sampling frequency of 100kHz and a sampling interval of 2 seconds, but also specifies the specific analysis tasks: for gearboxes and couplings, it calculates the amplitude of the torque modulation sideband in real time and captures transient impacts during speed changes; for motors, it activates the current-vibration joint analysis module. The system ultimately generates a highly intelligent and dynamic review command, which perfectly matches the monitoring activities with the actual operation and potential risk points of the ship.

[0075] refer to Figure 6 In step S15, the specific steps are as follows: S151: Collect the dynamic review event, determine the set of fault contents of the dredging vessel equipment based on the identification of the dynamic review event, and determine the main fault contents and multiple branch fault contents of the dredging vessel equipment according to the set of fault contents, the overall shape of the dredging vessel equipment and the corresponding current working status. S152: Obtain the work task list of dredging vessel equipment, and determine the first fault maintenance coefficient based on the work task list of dredging vessel equipment and the main fault content. S153: Determine the second fault maintenance coefficient based on the work task list of the dredging vessel equipment and multiple branch fault contents, and determine the fault maintenance events of the dredging vessel equipment based on the mapping relationship between the first fault maintenance coefficient, the second fault maintenance coefficient and the fault maintenance events.

[0076] In the embodiments of this application, the dynamic review event is collected, and the set of fault contents of the dredging vessel equipment is determined based on the identification of the dynamic review event. The main fault contents and multiple branch fault contents of the dredging vessel equipment are determined according to the set of fault contents, the overall shape of the dredging vessel equipment and the corresponding current working state. This approach takes into account the overall consideration of the set of fault contents, the overall shape of the dredging vessel equipment and the corresponding current working state, and ensures the accuracy of the main fault contents and multiple branch fault contents of the dredging vessel equipment.

[0077] At this point, the system fully collects all outputs generated by the dynamic review events executed in stage S14, including high-fidelity data collected under high load or specific tasks, special analysis reports generated for specific tasks, and short-term trend data of key indicators. The system integrates and cross-validates this in-depth information from the dynamic review with the preliminary diagnostic results of stage S12. By comparing the data changes before and after the review, the system confirms, corrects, or enriches the initial judgment, ultimately forming a more accurate and comprehensive set of fault information than that of stage S12.

[0078] Before making a decision, the system retrieves the overall morphology (such as oil stains or cracks on the equipment's exterior) and current operating status (such as high load or heavy load) as contextual information to assist in causal inference. The system uses an inference engine based on expert rules and causal graphs to identify the primary fault from the fault content set. The criteria for identification include fundamentality (it is the root cause of other faults), severity (usually having the highest fault level), and impact breadth (it has a negative impact on multiple related components). All faults in the fault content set that are not identified as primary faults are classified as branch faults and their relationship with the primary fault is further analyzed, such as direct consequences or concurrent issues.

[0079] Specifically, the dynamic review of S14 (under high load and heavy-duty navigation missions) has been completed, and a large amount of key information has been added; the system has collected the complete output of S14, including a special analysis report showing that the amplitude of the second harmonic sideband of the gearbox meshing frequency increased sharply by 40% during maneuvering navigation, an SPM report showing that the impact pulse value of the gearbox output bearing has entered the red danger zone, and morphological data confirming that the oil stain area at the bottom of the gearbox has expanded by about 50% within 2 hours; the system integrates this new evidence with the preliminary diagnosis of S12 to form an updated and more detailed set of fault contents: {gearbox gear damage (moderate, rapidly deteriorating), early pitting corrosion of the gearbox output bearing (moderate), severe leakage of gearbox oil seal, motor rotor imbalance (minor)}.

[0080] The system utilizes contextual information for corroboration: the high load condition explains why a moderate damage deteriorates rapidly, while the large amount of oil stains at the bottom of the gearbox provides strong visual evidence for the causal chain of gear damage leading to oil seal failure. The system's causal reasoning engine analyzes the set: motor imbalance is independent; bearing pitting and oil seal leakage are both downstream problems of gear damage. Therefore, the system ultimately determines the main fault as gearbox gear damage (moderate, rapidly deteriorating). At the same time, the system categorizes other items in the set as branch faults, namely early pitting of the gearbox output bearing (moderate), severe oil seal leakage in the gearbox, and slight motor rotor imbalance. Through this in-depth analysis, the system successfully organizes a complex set of faults into a clear main-branch structure, providing a crucial logical foundation for subsequently developing precise maintenance strategies.

[0081] Furthermore, the work task list of the dredging vessel equipment is obtained, and the first fault maintenance coefficient is determined based on the work task list and the main fault content of the dredging vessel equipment. This takes into account the overall consideration of the work task list and the main fault content of the dredging vessel equipment, and ensures the accuracy of the first fault maintenance coefficient. At this point, the system uses standardized API interfaces to obtain data in real time from multiple business systems such as the ship's navigation planning system and construction management system, forming a structured list of work tasks. This list is not a single task, but includes multiple task items such as completing the dredging of the main channel of the N7 waterway. Each task item is accompanied by rich attributes, including task type, deadline, contract amount, and even late payment penalties, providing key business background for subsequent quantitative evaluation.

[0082] The system analyzes the potential impact of the primary failure on various tasks, assesses the risk of task interruption and task degradation, and quantifies the potential economic losses, including direct maintenance costs and indirect losses caused by breach of contract, missed opportunities, etc. The system integrates all factors through a weighted function (e.g., first failure maintenance coefficient = w1 × primary failure technical score + w2 × task impact score + w3 × economic loss score), where the weight representing business impact is usually higher, and finally calculates a coefficient between 0 and 1. This coefficient is a comprehensive reflection of the technical severity and business urgency.

[0083] Specifically, S151 has identified the main fault as gearbox gear damage (moderate, rapidly deteriorating), and the system now needs to assess the urgency of its maintenance; the system has obtained the current work task list from the construction management system: one task is to complete the dredging of the last 500 meters of the main channel of N7 within 8 hours, and this task has the highest priority.

[0084] Since the main failure is a core component of the main propulsion system, the system determines that its mission interruption risk is extremely high, and the mission impact score is rated as 0.95; the economic loss score is rated as 0.9; combined with the main failure technical score of 0.9 (from the first-level failure of S132), and setting a model with higher business impact weight (w1=0.2, w2=0.4, w3=0.4), the system calculates the maintenance coefficient of the first failure as: (0.2×0.9)+(0.4×0.95)+(0.4×0.9)=0.92. This coefficient of 0.92 clearly tells the decision-making system that although it is a moderate damage from a purely technical point of view, considering its fatal impact on the critical mission, the urgency of its maintenance has reached the highest level. This coefficient will directly drive the subsequent steps to make the decision to immediately shut down and repair.

[0085] Therefore, a second fault maintenance coefficient is determined based on the work task list and multiple branch fault contents of the dredging vessel equipment. Fault maintenance events of the dredging vessel equipment are determined based on the mapping relationship between the first fault maintenance coefficient, the second fault maintenance coefficient, and fault maintenance events. This approach takes into account the overall consideration of the mapping relationship between the first fault maintenance coefficient, the second fault maintenance coefficient, and fault maintenance events, ensuring the accuracy of fault maintenance events for the dredging vessel equipment. At the same time, dynamic review events are introduced to control the main fault contents and multiple branch fault contents. This approach achieves an overall consideration of the main fault contents, multiple branch fault contents, and the work task list of the dredging vessel equipment, thereby improving the accuracy of fault maintenance events for the dredging vessel equipment.

[0086] At this point, the system iterates through all branch faults and calculates a comprehensive risk score by weighted summation of their levels to reflect the overall severity of the secondary problems. The system performs a core maintenance complexity assessment, analyzing the correlation between all branch faults and the primary fault in terms of physical space and maintenance process, such as whether they are located on the same equipment or whether additional disassembly steps are required during maintenance. The system combines comprehensive risk and maintenance complexity through a fusion model to calculate the second fault maintenance coefficient. A high second coefficient means that even if the primary fault is not urgent, it is still a good time to perform maintenance, considering that several minor problems can be resolved together at low cost.

[0087] The system has a built-in maintenance decision matrix that maps different maintenance strategies based on different combinations of two coefficients: a high first coefficient indicates the highest priority emergency action; a medium / low first coefficient but a high second coefficient indicates planned centralized maintenance; and both coefficients are low, indicating monitoring should be postponed. Once a decision is made, the system generates a detailed, structured fault maintenance event. The fault maintenance event not only includes the maintenance strategy, priority, and suggested maintenance window, but also details the specific maintenance plan, required spare parts, and resource list for the main fault and all branch faults, which can be directly issued to the maintenance team for execution. Specifically, S151 has identified the main fault as gearbox gear damage, the secondary faults as motor rotor imbalance (minor) and gearbox oil seal leakage (moderate), and S152 has calculated the maintenance factor for the first fault to be 0.92 (extremely high).

[0088] The system begins calculating the second fault maintenance factor; it weights motor imbalance (score 0.3) and oil seal leakage (score 0.6) to obtain a comprehensive risk score of 0.45; the system assesses the maintenance complexity and finds that the motor, gearbox, and oil seal are all on the same power chain, and repairing the gearbox requires disassembling the motor, and replacing the oil seal is also a standard procedure, so the process coupling is extremely high, and the maintenance complexity score is rated as 0.9; the system calculates the second fault maintenance factor to be 0.675, indicating that this is a very suitable time to handle maintenance together.

[0089] Since the maintenance coefficient for the first fault is as high as 0.92, the highest-level rule in the decision matrix is ​​triggered, and the system directly decides to take immediate action. The system generates a detailed P0-level emergency maintenance work order, which not only clarifies the overall strategy of immediately suspending operations and preparing for port repair, but also provides solutions: for the main fault, it proposes to apply for a spare gearbox or carry out a major overhaul at the factory; for the branch fault, it suggests replacing the entire set of seals during the gearbox overhaul and using the disassembly window to perform dynamic balancing verification on the motor.

[0090] Please see Figure 7 , Figure 7This is a schematic diagram of the structural composition of the fault detection system for dredging vessel equipment in an embodiment of the present invention; the fault detection system for dredging vessel equipment includes: The identification module 21 is used to determine the distribution map of the dredging vessel equipment based on multiple images of the dredging vessel equipment in different dimensions and the overall shape of the dredging vessel equipment, and to determine multiple key parts based on the identification of the distribution map of the dredging vessel equipment. The fault distribution map module 22 is used to determine the fault events of each key part based on multiple working data, corresponding part morphology and corresponding vibration signals, and to determine the fault distribution map of the dredging vessel equipment according to the part location of each key part, the corresponding fault events and the current working status of the dredging vessel equipment. The fault detection content module 23 is used to identify multiple fault areas based on the fault distribution map, determine the fault level of the dredging vessel equipment based on the fault content of each fault area, the corresponding area shape and the corresponding key parts, and determine the fault detection content of the dredging vessel equipment based on the fault level of the dredging vessel equipment, the corresponding work task list and the fault events of each key part. The dynamic review event module 24 is used to determine the review area of ​​the dredging vessel equipment based on the fault detection content and the distribution map of the dredging vessel equipment, and to determine the dynamic review event based on the review area, the usage status of the dredging vessel equipment and the corresponding current work task. The fault maintenance event module 25 is used to determine the main fault content and multiple branch fault content of the dredging vessel equipment based on the detection of the dynamic review event, and to determine the fault maintenance event of the dredging vessel equipment based on the main fault content, multiple branch fault content and the work task list of the dredging vessel equipment.

[0091] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A method for fault detection of dredging vessel equipment, characterized in that, include: Based on multiple images of dredging vessels and equipment in different dimensions and the overall shape of the dredging vessels and equipment, a distribution map of the dredging vessels and equipment is determined, and multiple key parts are identified based on the identification of the distribution map of the dredging vessels and equipment. Based on multiple working data, corresponding part shapes, and corresponding vibration signals of each key part, the fault events of the key part are determined. The fault distribution map of the dredging vessel is determined according to the location of each key part, the corresponding fault events, and the current working status of the dredging vessel equipment. The fault distribution map is dynamically rendered, conveying rich information through color coding, icon labels, and dynamic effects, and the current working status is superimposed on the map. Multiple fault areas are identified based on the fault distribution map. The fault level of the dredging vessel equipment is determined based on the fault content, corresponding area shape, and corresponding key parts of each fault area. The fault detection content of the dredging vessel equipment is determined based on the fault level of the dredging vessel equipment, the corresponding work task list, and the fault events of each key part. Based on the fault detection content and the distribution map of dredging vessel equipment, the re-inspection area of ​​dredging vessel equipment is determined. According to the re-inspection area, the usage status of dredging vessel equipment and the corresponding current work tasks, dynamic re-inspection events are determined. The dynamic re-inspection events clarify the target area, data acquisition parameters, special analysis instructions and output alarm thresholds. Based on the detection of this dynamic review event, the main fault content and multiple branch fault content of the dredging vessel equipment are determined. Based on the main fault content, multiple branch fault content and the work task list of the dredging vessel equipment, the fault maintenance event of the dredging vessel equipment is determined. The fault maintenance event not only includes maintenance strategy, priority and suggested maintenance window, but also lists in detail the specific maintenance plan, required spare parts and resource list for the main fault content and all branch fault content.

2. The fault detection method for dredging vessel equipment according to claim 1, characterized in that, The distribution map of the dredging vessel equipment is determined based on multiple images of the equipment in different dimensions and the overall shape of the equipment. Based on the identification of this distribution map, several key components are determined, including: The current location of the dredging vessel is collected. Based on the current location of the dredging vessel and multiple surrounding cameras, the shooting space of the dredging vessel is determined. Multiple images of the dredging vessel in different dimensions are collected based on the dynamic shooting of this shooting space. The overall shape of the dredging vessel equipment is marked. Based on the overall shape of the dredging vessel equipment and multiple images of the dredging vessel equipment in different dimensions, multiple feature matching contents are determined. Based on the synthesis of multiple feature matching contents, a distribution map of the dredging vessel equipment is constructed. Based on the identification of the distribution map of the dredging vessel equipment, multiple equipment locations were determined, and multiple key locations were determined according to the multiple equipment locations, their corresponding functions, and the usage scenarios of the dredging vessel equipment.

3. The fault detection method for dredging vessel equipment according to claim 1, characterized in that, The method involves determining the fault event of a key component based on multiple operational data, corresponding component morphology, and corresponding vibration signals. A fault distribution map of the dredging vessel is then determined based on the location of each key component, the corresponding fault event, and the current operational status of the dredging vessel equipment. This includes: In multiple key parts, multiple working data of each key part are determined based on the detection of each key part, so as to collect multiple working data of each key part. At the same time, the corresponding part shape is determined based on the shape recognition of each key part. Vibration signals from various key components are collected, and corresponding vibration data combinations are determined based on the tracing of these vibration signals. Fault events of key components are determined based on the vibration data combinations, multiple working data of each key component, and the corresponding component morphology. Simultaneously, the current working status of the dredging vessel equipment is collected, and a fault distribution map of the dredging vessel equipment is determined based on the location of each key component, the corresponding fault events, and the current working status of the dredging vessel equipment.

4. The fault detection method for dredging vessel equipment according to claim 1, characterized in that, The process involves identifying multiple fault zones based on the fault distribution map, determining the fault level of the dredging vessel equipment based on the fault content, corresponding regional morphology, and corresponding key components of each fault zone, and determining the fault detection content of the dredging vessel equipment based on the fault level, corresponding work task list, and fault events of each key component, including: In the fault distribution map, multiple fault areas are identified based on the fault distribution map, and the fault content of each fault area is marked; at the same time, the regional morphology of each fault area is collected, and the first fault coefficient is determined according to the fault content of each fault area and the corresponding key parts. The second fault coefficient is determined based on the regional morphology of each fault area and the corresponding key parts. The fault level of the dredging vessel equipment is determined according to the mapping relationship between the first fault coefficient, the second fault coefficient and the fault level.

5. The fault detection method for dredging vessel equipment according to claim 4, characterized in that, The process of identifying multiple fault areas based on the fault distribution map, determining the fault level of the dredging vessel equipment based on the fault content, corresponding area morphology, and corresponding key components of each fault area, and determining the fault detection content of the dredging vessel equipment based on the fault level, corresponding work task list, and fault events of each key component, also includes: Collect the work task list of dredging vessel equipment, determine multiple sub-work tasks based on the detection of the work task list of dredging vessel equipment, determine the first level of sub-fault detection content based on the multiple sub-work tasks and the fault level of dredging vessel equipment, determine the second level of sub-fault detection content based on the multiple sub-work tasks and the fault events of each key part, and determine the fault detection content of dredging vessel equipment based on the first level of sub-fault detection content and the second level of sub-fault detection content.

6. The fault detection method for dredging vessel equipment according to claim 1, characterized in that, The process involves determining the re-inspection area for dredging vessels and equipment based on the fault detection content and the distribution map of the dredging vessels and equipment. Dynamic re-inspection events are then determined based on this re-inspection area, the usage status of the dredging vessels and equipment, and the corresponding current work tasks. These events include: A distribution map of dredging vessel equipment is collected. Based on the identification of the distribution map of dredging vessel equipment, the distribution areas of multiple key parts are determined. Based on the distribution areas of multiple key parts, the corresponding functions of the parts, and the fault detection content, the re-inspection area of ​​dredging vessel equipment is determined.

7. The fault detection method for dredging vessel equipment according to claim 6, characterized in that, The process of determining the re-inspection area for dredging vessels and equipment based on the fault detection content and the distribution map of the dredging vessels and equipment, and determining dynamic re-inspection events based on the re-inspection area, the usage status of the dredging vessels and equipment, and the corresponding current work tasks, also includes: Collect multiple usage data of dredging vessels and equipment, determine the usage status of dredging vessels and equipment based on the multiple usage data of dredging vessels and equipment, and determine the content of the first dynamic review based on the review area and the usage status of dredging vessels and equipment; Obtain the current working task of the dredging vessel and equipment, determine the second level of dynamic review content based on the current working task of the dredging vessel and equipment and the review area, and determine the dynamic review event based on the first level of dynamic review content and the second level of dynamic review content.

8. The fault detection method for dredging vessel equipment according to claim 1, characterized in that, The process involves determining the main fault content and multiple branch fault content of the dredging vessel equipment based on the detection of the dynamic review event, and determining the fault maintenance event of the dredging vessel equipment based on the main fault content, multiple branch fault content, and the work task list of the dredging vessel equipment, including: The dynamic review event is collected, and the set of fault contents of the dredging vessel equipment is determined based on the identification of the dynamic review event. The main fault contents and multiple branch fault contents of the dredging vessel equipment are determined according to the set of fault contents, the overall shape of the dredging vessel equipment and the corresponding current working status.

9. The fault detection method for dredging vessel equipment according to claim 8, characterized in that, The process of determining the main fault content and multiple branch fault content of the dredging vessel equipment based on the detection of the dynamic review event, and determining the fault maintenance event of the dredging vessel equipment based on the main fault content, multiple branch fault content, and the work task list of the dredging vessel equipment, further includes: Obtain the work task list of the dredging vessel equipment, and determine the first fault maintenance coefficient based on the work task list and the main fault content of the dredging vessel equipment. The second fault maintenance coefficient is determined based on the work task list of the dredging vessel equipment and multiple branch fault contents. The fault maintenance events of the dredging vessel equipment are determined based on the mapping relationship between the first fault maintenance coefficient, the second fault maintenance coefficient and the fault maintenance events.

10. A fault detection system for dredging vessel equipment, characterized in that, The fault detection system for dredging vessel equipment is applied to the fault detection method for dredging vessel equipment as described in any one of claims 1-9, and the fault detection system for dredging vessel equipment includes: The identification module is used to determine the distribution map of the dredging vessel based on multiple images of the dredging vessel in different dimensions and the overall shape of the dredging vessel, and to identify multiple key parts based on the identification of the distribution map of the dredging vessel. The fault distribution map module is used to determine the fault events of each key part based on multiple working data, corresponding part morphology and corresponding vibration signals. It determines the fault distribution map of the dredging vessel equipment based on the location of each key part, the corresponding fault events and the current working status of the dredging vessel equipment. The fault detection content module is used to identify multiple fault areas based on the fault distribution map, determine the fault level of the dredging vessel equipment based on the fault content of each fault area, the corresponding area shape and the corresponding key parts, and determine the fault detection content of the dredging vessel equipment based on the fault level of the dredging vessel equipment, the corresponding work task list and the fault events of each key part. The dynamic review event module is used to determine the review area of ​​dredging vessel equipment based on the fault detection content and the distribution map of dredging vessel equipment, and to determine the dynamic review event based on the review area, the usage status of dredging vessel equipment and the corresponding current work task; The fault maintenance event module is used to determine the main fault content and multiple branch fault content of the dredging vessel equipment based on the detection of the dynamic review event, and to determine the fault maintenance event of the dredging vessel equipment based on the main fault content, multiple branch fault content and the work task list of the dredging vessel equipment.

Images