Intelligent maintenance scheduling method, system and device for marine monitoring equipment and medium

Abstract

The application discloses a kind of marine monitoring equipment intelligent maintenance scheduling method, system, equipment and medium, belong to marine monitoring equipment operation and maintenance management technical field, comprising: S1, obtain environmental dynamic time series data and multidimensional operation and maintenance situation data;S2, environmental dynamic time series data and multidimensional operation and maintenance situation data are reconstructed to feature, and comprehensive data cube is obtained;S3, multidimensional weighted calculation is carried out to environmental parameter matrix, and the sea suitability score is obtained;S4, continuous future time point is aggregated and graded screening, and window period list is obtained;S5, space-time constraint matching is carried out to equipment state matrix, and feasible time window set is obtained;S6, by multi-objective path planning model with time window constraint, the minimum navigation mileage, fuel consumption and operation waiting time are solved, and operation instruction table is generated.The application carries out three-line coordination of time line, motion line planning line and economic line feedback, to have safe, efficient, controllable and have self-evolution ability.

Description

Technical Field

[0001] This invention belongs to the field of marine monitoring equipment operation and maintenance management technology, and particularly relates to an intelligent maintenance and scheduling method, system, equipment and medium for marine monitoring equipment. Background Technology

[0002] With the rapid development of marine observation technology, various marine monitoring equipment has become a key infrastructure for acquiring marine environmental data. These devices are exposed to harsh marine environments such as high salt spray, high humidity and heat, and strong winds and waves, making their sensors, power supply units, communication modules, and mooring systems highly susceptible to performance degradation, biofouling, corrosion, or energy depletion. To ensure data continuity and equipment reliability, regular on-site inspections, sensor calibration, battery replacement, biofouling removal, and emergency repairs are essential. While the industry has published relevant technical guidelines for the operation and maintenance management of marine environmental monitoring buoys, clarifying basic technical specifications such as maintenance procedures, operational requirements, data quality control, and record management for marine monitoring equipment, providing a basis for standardized operation and maintenance, these guidelines only focus on the execution of maintenance operations and do not address intelligent decision-making methods for maintenance scheduling. They fail to solve core scheduling issues such as predicting sea access windows, optimizing routes, and dynamically controlling costs.

[0003] In traditional technologies, most operation and maintenance units adopt a passive scheduling strategy that combines fixed-cycle inspections with emergency repairs after a failure. Specific methods include: relying on manual experience to determine the time window for departure based on rough weather forecasts; having the captain or operations manager manually plan the navigation sequence and work sequence between various equipment points using nautical charts; and only recording the total cost after each task without detailed cost aggregation and node analysis. These practices have been used for a long time, but lack the ability to perform refined modeling and intelligent decision-making regarding dynamic changes in sea conditions, equipment aging patterns, and the composition of operating costs.

[0004] However, existing maintenance scheduling methods have significant technical flaws. First, the selection of sea departure windows relies on manual experience and judgment, failing to comprehensively assess risks based on multiple parameters such as waves, wind speed, tidal range, visibility, and typhoon paths, resulting in numerous ineffective sea departures or dangerous operations. Second, the route planning of maintenance vessels is highly arbitrary, lacking optimization algorithms to support the navigation sequence, berthing time, and operation sequence between multiple maintenance points, often leading to problems such as circuitous routes, fuel waste, and excessively long waiting times. This results in low equipment coverage per sea departure and low overall operational efficiency. Third, maintenance cost control is rudimentary, only recording total costs without establishing a comprehensive cost aggregation model for each operation or individual piece of equipment. This makes it impossible to identify key cost overruns and difficult to evaluate the return on investment of different maintenance plans. Fourth, time planning, navigation routes, and economic costs are isolated, with data stored independently without a closed-loop feedback mechanism. Scheduling strategies cannot adaptively correct for historical execution deviations, leading to repeated problems and a lack of self-evolution capabilities in the system. Summary of the Invention

[0005] The purpose of this invention is to overcome at least one of the above-mentioned defects in the prior art and to provide an intelligent maintenance and scheduling method, system, equipment and medium for marine monitoring equipment, which can coordinate timeline prediction, movement planning and economic feedback to achieve safety, efficiency, controllability and self-evolution.

[0006] To achieve the above-mentioned objectives, the first objective of this invention is to provide an intelligent maintenance and scheduling method for marine monitoring equipment, comprising: S1. Acquire dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of marine monitoring equipment to be maintained; S2. Perform feature reconstruction on environmental dynamic time-series data and multi-dimensional operation and maintenance status data to obtain a comprehensive data cube; the comprehensive data cube includes an environmental parameter matrix, an equipment status matrix, and a maintenance task candidate set. S3. The environmental parameter matrix is ​​calculated using a multi-parameter weighted scoring model to obtain the seagoing suitability score for each future time point. S4. Aggregate consecutive future time points that meet the suitability score for going to sea into candidate time windows, and perform hierarchical filtering on the candidate time windows to obtain a list of window periods. S5. Based on the maintenance task candidate set and window list, perform spatiotemporal constraint matching on the device state matrix to obtain a feasible time window set. S6. Using the window period list and feasible time window set as rigid constraints, the multi-objective path planning model with time window constraints is used to minimize the navigation mileage, fuel consumption and operation waiting time, and generate the operation instruction table. The operation instruction table is used to represent the optimal navigation sequence, the expected arrival time of each marine monitoring equipment to be maintained and the operation start time.

[0007] Preferably, S2 includes: S21. Extract dynamic hydrological and meteorological parameters from environmental dynamic time series data, and align the dynamic hydrological and meteorological parameters according to timestamps to obtain an environmental parameter matrix; S22. Extract the equipment status characteristics of each marine monitoring equipment to be maintained from the multi-dimensional operation and maintenance status data, and aggregate and arrange the equipment status characteristics into a two-dimensional matrix according to the equipment identifier to obtain the equipment status matrix. S23. Perform multi-dimensional evaluation and calculation on the equipment status matrix to obtain the urgency index of each marine monitoring equipment to be maintained; S24. Marine monitoring equipment requiring maintenance whose urgency index is greater than the urgency threshold shall be considered as candidate maintenance equipment. S25. Based on candidate maintenance equipment, extract the spatial coordinates and expected operation time of the candidate maintenance equipment from the multi-dimensional operation and maintenance status data, and map the urgency index to the urgency weight. S26. Based on the equipment identifier, the spatial coordinates, expected operation duration and urgency weight are associated and combined to generate a set of maintenance task candidates.

[0008] Preferably, the maintenance task candidate set includes the spatial coordinates and home port location of each of the marine monitoring devices to be maintained; S5 includes: S51. Based on the spatial coordinates and the location of the home port, calculate the estimated sailing time from the home port to each marine monitoring equipment to be maintained, and extract the expected operation time of each marine monitoring equipment to be maintained from the equipment status matrix. S52. Iterate through each window in the window list, calculate the start time of each window and the sum of the estimated sailing time corresponding to the start time, and obtain the estimated arrival time of each window. S53. The window period in which the difference between the end time and the expected operation time is not less than the estimated arrival time shall be identified as the candidate window period for the marine monitoring equipment to be maintained. S54. Sort the candidate window periods of each marine monitoring equipment to be maintained according to the start time to obtain each feasible time window; S55. Aggregate the feasible time windows according to the device identifier to obtain a set of feasible time windows.

[0009] Preferably, the environmental parameters include significant wave height, maximum wave height, average wind speed, gust wind speed, tidal range, visibility, ocean current speed, and typhoon warning level; in S3, the expression for the suitability score for going to sea is:

[0010] In the formula, Indicates a future point in time. The suitability score for going to sea, Indicates the first The dynamic weighting coefficients of each environmental parameter, and satisfying , Indicates the first Piecewise linear penalty functions corresponding to each environmental parameter Indicates a future point in time. The The values ​​of each environmental parameter.

[0011] Preferred options also include: S7. Obtain the actual operation data generated by the marine monitoring equipment's operation instruction table, and collect and calculate the actual total cost of this task. S8. Compare and analyze the actual data of the operation and the preset plan data to obtain the execution deviation index; S9. Allocate the actual total cost according to the proportion of the voyage distance and the direct operating cost to obtain the actual cost of each marine monitoring device to be maintained; S10. If the actual cost of a single marine monitoring device exceeds the budgeted cost of the marine monitoring device to be maintained, the marine monitoring device to be maintained is marked as an overspending node, and a deviation feature vector is extracted based on the execution deviation index; the deviation feature vector is used to characterize the multidimensional deviation between the actual operation process and the preset plan. S11. Based on the deviation feature vector, update the dynamic weight vector in the multi-parameter weighted scoring model to obtain the updated multi-parameter weighted scoring model. S12. Based on the deviation feature vector, update the fuel consumption weight coefficient and relaxation penalty coefficient in the multi-objective path planning model to obtain the updated multi-objective path planning model. S13. Based on the actual cost per unit of the over-cost node, update the maintenance cycle of the marine monitoring equipment to be maintained corresponding to the over-cost node to obtain the updated maintenance cycle.

[0012] The second objective of this invention is to provide an intelligent maintenance and scheduling system for marine monitoring equipment, comprising: The data acquisition module acquires dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of the marine monitoring equipment to be maintained; The feature reconstruction module performs feature reconstruction on the environmental dynamic time-series data and the multi-dimensional operation and maintenance status data to obtain a comprehensive data cube; the comprehensive data cube includes an environmental parameter matrix, an equipment status matrix, and a maintenance task candidate set. The suitability score calculation module performs multi-dimensional weighted calculation on the environmental parameter matrix through a multi-parameter weighted scoring model to obtain the seagoing suitability score corresponding to each future time point; The window period classification and screening module aggregates consecutive future time points that meet the sea-going suitability score threshold into candidate time windows, and performs classification and screening of candidate time windows to obtain a window period list. The spatiotemporal constraint matching module performs spatiotemporal constraint matching on the device state matrix based on the maintenance task candidate set and the window period list to obtain a set of feasible time windows; The instruction generation module uses the window period list and the set of feasible time windows as rigid constraints, and solves the optimization problem of minimizing navigation mileage, fuel consumption and operation waiting time through a multi-objective path planning model with time window constraints to generate an operation instruction table. The operation instruction table is used to represent the optimal navigation sequence, the expected arrival time of each marine monitoring equipment to be maintained and the operation start time.

[0013] Preferred options also include: The cost accounting module acquires the actual operation data generated by the marine monitoring equipment's operation instruction table, and collects and calculates the actual total cost of the task. The deviation index analysis module compares and analyzes the actual operation data and the preset plan data to obtain the execution deviation index; The actual cost allocation module allocates the actual total cost according to the proportion of the voyage distance and the direct operation cost to obtain the actual cost of each marine monitoring equipment to be maintained; The deviation feature extraction module marks the marine monitoring equipment to be maintained as an over-cost node if the actual cost of a single unit exceeds the budgeted cost of the marine monitoring equipment to be maintained. Based on the execution deviation index, the module extracts the deviation feature vector. The deviation feature vector is used to characterize the multi-dimensional deviation between the actual operation process and the preset plan. The weight update module of the multi-parameter weighted scoring model updates the dynamic weight vector in the multi-parameter weighted scoring model based on the deviation feature vector, so as to obtain the updated multi-parameter weighted scoring model. The multi-objective path planning model coefficient update module updates the fuel consumption weight coefficient and relaxation penalty coefficient in the multi-objective path planning model based on the deviation feature vector, and obtains the updated multi-objective path planning model. The maintenance cycle update module updates the maintenance cycle of the marine monitoring equipment to be maintained corresponding to the over-cost node based on the actual cost of a single unit of the over-cost node, thus obtaining the updated maintenance cycle.

[0014] A third objective of this invention is to provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the aforementioned intelligent maintenance and scheduling method for marine monitoring equipment.

[0015] A fourth objective of this invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned intelligent maintenance and scheduling method for marine monitoring equipment.

[0016] The fifth objective of this invention is to provide a computer program product, including a computer program that, when executed by a processor, implements the aforementioned intelligent maintenance and scheduling method for marine monitoring equipment.

[0017] The advantages and positive effects of this application are: This invention systematically integrates dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of marine monitoring equipment to be maintained. By reconstructing the features of the dynamic time-series environmental data and the multi-dimensional operation and maintenance status data of marine monitoring equipment to be maintained, the scattered environmental parameters, equipment status and maintenance task requirements are organized into a comprehensive data cube. This can formalize multi-source heterogeneous data into a structured and computable unified representation, laying a data foundation for time-series assessment and spatial optimization.

[0018] This invention uses a multi-parameter weighted scoring model to perform multi-dimensional weighted calculations on the environmental parameter matrix, which can quantify the overall suitability of sea operations at each future time point and transform dynamic changes in sea conditions into a comparable scoring sequence. By comparing the sea suitability score with a preset threshold and aggregating time points that continuously meet the conditions, a window list is obtained through hierarchical screening. This allows for the rigid screening of suitable and alternative sea operation windows from a time dimension, eliminating periods of severe sea conditions and reducing ineffective sea operations and safety risks from the source.

[0019] This invention performs spatiotemporal constraint matching on the equipment state matrix based on a window period list, generating a set of feasible time windows for each piece of equipment to be maintained. This accurately correlates time operability with equipment maintenance requirements, generating rigid constraints coupled in time and space. By using the window period list and the set of feasible time windows together as rigid constraints, a multi-objective path planning model with time window constraints is used to solve the optimization problem of minimizing flight mileage, fuel consumption, and operation waiting time, generating an operation instruction table. This allows for simultaneous optimization of route efficiency and time window compliance within an integrated framework, thereby achieving globally optimal scheduling with the shortest flight mileage, the least fuel consumption, and the least operation waiting time.

[0020] This invention integrates the traditionally fragmented time prediction, route planning, and cost constraints into an end-to-end intelligent decision-making process, reducing the rate of ineffective sea deployments and safety risks, improving the coverage of single-batch equipment and operation and maintenance efficiency, and providing a quantifiable, executable, and iterative technical solution for the maintenance and scheduling of marine monitoring equipment. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart of a first preferred embodiment of the present invention is shown; Figure 2 A flowchart of S2 in the first preferred embodiment of the present invention is shown; Figure 3 A flowchart of S5 in the first preferred embodiment of the present invention is shown; Figure 4 A system block diagram of a second preferred embodiment of the present invention is shown. Detailed Implementation

[0023] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0024] Please see Figure 1 A first preferred embodiment of an intelligent maintenance and scheduling method for marine monitoring equipment includes: S1. Acquire dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of marine monitoring equipment to be maintained.

[0025] Specifically, after acquiring the environmental dynamic time-series data of the target sea area and the multi-dimensional operation and maintenance status data of the marine monitoring equipment to be maintained, the maintenance and scheduling terminal can use the 3σ principle (Three Sigma Rule, three times the standard deviation principle) to detect and remove abnormal outliers in the environmental dynamic time-series data of the target sea area and the multi-dimensional operation and maintenance status data of the marine monitoring equipment to be maintained, so as to obtain the environmental dynamic time-series data and multi-dimensional operation and maintenance status data after removing outliers.

[0026] The maintenance and scheduling terminal can use linear interpolation or nearest neighbor filling to complete the missing data in the environmental dynamic time series data and multi-dimensional operation and maintenance status data after removing outliers. It can also convert the timestamps of all the environmental dynamic time series data and multi-dimensional operation and maintenance status data after completing the missing data into Coordinated Universal Time (UTC) and normalize the parameters of different dimensions to obtain the preprocessed environmental dynamic time series data and multi-dimensional operation and maintenance status data.

[0027] In the first preferred embodiment, parameters with different dimensions can be normalized using the following calculation formula:

[0028] In the formula, Indicates the first In the nth sample The normalized value of the term parameter, Indicates the first In the nth sample The original value of the parameter. Indicates the first The minimum value of the term parameter across all samples. Indicates the first The maximum value of the term parameter across all samples.

[0029] In the first preferred embodiment, the environmental dynamic time series data of the target sea area can be numerical forecast products released by the National Marine Environmental Forecasting Center, real-time observation data from coastal marine observation stations, marine meteorological data retrieved by satellite remote sensing, and field data collected by auxiliary observation buoys deployed in the target sea area.

[0030] In the first preferred embodiment, the multi-dimensional operation and maintenance status data of the marine monitoring equipment to be maintained may include, but is not limited to, the operating parameters uploaded in real time by the equipment edge computing unit, the maintenance data in the historical maintenance record database, the life curve data calibrated when the equipment left the factory, the sensor calibration cycle data, and the equipment fault alarm data.

[0031] S2. Perform feature reconstruction on environmental dynamic time-series data and multi-dimensional operation and maintenance status data to obtain a comprehensive data cube.

[0032] Specifically, the maintenance scheduling terminal can extract and align time-dimensional features from dynamic environmental time-series data, unifying hydrological and meteorological parameters from different sampling frequencies to the same time granularity, thus obtaining time-dimensional environmental features and spatial-dimensional equipment location features. The terminal can also aggregate equipment-dimensional features from multi-dimensional operational status data, integrating scattered equipment status information according to equipment identifiers to obtain task-dimensional maintenance requirement features. Finally, the terminal can perform correlation mapping between time-dimensional environmental features, spatial-dimensional equipment location features, and task-dimensional maintenance requirement features, generating a comprehensive data cube using a unified structured data organization format.

[0033] The comprehensive data cube can be used to represent a three-dimensional structured data organization form built around the time dimension, spatial dimension, and task dimension.

[0034] In a first preferred embodiment, the integrated data cube may include, but is not limited to, an environmental parameter matrix, a device status matrix, and a maintenance task candidate set.

[0035] S3. The environmental parameter matrix is ​​calculated using a multi-parameter weighted scoring model to obtain the seagoing suitability score corresponding to each future time point.

[0036] Specifically, the maintenance and scheduling terminal can invoke a pre-built multi-parameter weighted scoring model, traverse each time point in the environmental parameter matrix, and perform weighted fusion calculations on all environmental parameters corresponding to that time point. During the calculation, the multi-parameter weighted scoring model can map the original value of each environmental parameter to a unified score interval through a piecewise linear penalty function to obtain a standardized score. The multi-parameter weighted scoring model can then multiply the standardized score of each environmental parameter by the dynamic weight coefficient corresponding to the standardized score, and sum all products to obtain the seagoing suitability score for that time point.

[0037] In the first preferred embodiment, the multi-parameter weighted scoring model can be a weighted fusion model constructed based on a piecewise linear penalty function. The initial weights of the multi-parameter weighted scoring model can be calibrated according to the operational characteristics and historical operation and maintenance data of different sea areas.

[0038] In the first preferred embodiment, the sea suitability score corresponding to each future time point can be used as a quantitative value to characterize the overall safety level and feasibility of maritime operations at that time point. The higher the sea suitability score, the more suitable the sea conditions are for maintenance operations at that time point, and the lower the sea suitability score, the higher the operational risk.

[0039] S4. Aggregate consecutive future time points that meet the suitability score for going to sea into candidate time windows, and then perform hierarchical filtering on the candidate time windows to obtain a list of window periods.

[0040] Specifically, the maintenance and scheduling terminal can compare the seagoing suitability scores for all time points within the future preset time range calculated by S3 with the preset seagoing suitability thresholds point by point, and generate a Boolean comparison result sequence with timestamps as indexes, including two states: meeting the threshold conditions and not meeting the threshold conditions. The Boolean comparison result sequence can be used to record the preliminary judgment results of the operational feasibility at each future time point.

[0041] In the first preferred embodiment, the maintenance and scheduling terminal can traverse the Boolean comparison result sequence, identify all time intervals that are continuously marked as meeting the seagoing suitability threshold conditions, and aggregate all time points in each continuous time interval into candidate time windows, and record the start time, end time and duration attributes for each candidate time window.

[0042] In the first preferred embodiment, the maintenance scheduling terminal can construct a window quality assessment system based on the window duration, the average sea-going suitability score within the window, and the sea state fluctuation coefficient within the window. The maintenance scheduling terminal can quantitatively score and classify all candidate time windows: removing low-quality time windows whose duration is less than a preset minimum operation duration threshold, whose average sea-going suitability score is lower than a preset qualified score threshold, or whose sea state fluctuation coefficient exceeds a preset stability threshold; and arranging the remaining qualified candidate time windows in chronological order of their start times to generate a window period list.

[0043] In the first preferred embodiment, the seaworthiness threshold can be used to characterize the minimum sea state condition threshold at which offshore maintenance operations can be carried out safely. The seaworthiness threshold can be set differently according to the wind and wave resistance level of the maintenance vessel, the operational capabilities of the operators, and the operational requirements of different maintenance tasks.

[0044] In the first preferred embodiment, the candidate time window can be used to characterize a continuous time interval that initially meets the sea condition requirements for offshore operations.

[0045] In a first preferred embodiment, the window list can be used to characterize the set of all valid time windows that have been filtered and are available for scheduling maintenance work.

[0046] S5. Based on the maintenance task candidate set and window list, perform spatiotemporal constraint matching on the device state matrix to obtain a feasible time window set.

[0047] Specifically, the maintenance scheduling terminal can use each marine monitoring device to be maintained in the maintenance task candidate set as a matching object, use each time window in the window period list as a time constraint, and combine it with the operation requirement information in the equipment status matrix to perform spatiotemporal coupling constraint matching. The maintenance scheduling terminal can extract the spatial coordinates and home port reference position of each marine monitoring device to be maintained from the maintenance task candidate set, and calculate the estimated sailing time from the home port to the location of each device along the shortest great circle route by combining the rated navigation performance parameters of the maintenance vessel, the historical navigation resistance correction coefficient of the target sea area, and the average ocean current influence coefficient. The maintenance scheduling terminal can also extract the expected operation time corresponding to each marine monitoring device to be maintained from the equipment status matrix.

[0048] In a first preferred embodiment, the maintenance scheduling terminal can calculate the difference between the end time of each time window and the expected operation time corresponding to the end time of each time window for each combination of marine monitoring equipment to be maintained and each time window, thus obtaining the latest time that the maintenance vessel is allowed to arrive within the time window. If the estimated arrival time of the maintenance vessel, which departs on time at the start time of the time window, is not later than the aforementioned latest time, then the time window is determined to be a candidate window period for the marine monitoring equipment to be maintained. The maintenance scheduling terminal can arrange all candidate window periods corresponding to the same marine monitoring equipment to be maintained in ascending order of start time, remove redundant windows with completely overlapping times, and obtain the feasible time window for the equipment.

[0049] In the first preferred embodiment, the maintenance scheduling terminal can aggregate the feasible time windows of all marine monitoring equipment to be maintained according to the equipment identifier to obtain a set of feasible time windows.

[0050] In a first preferred embodiment, the feasible time window set can be used to characterize the set of all effective time windows in which maintenance operations can be carried out for each marine monitoring device to be maintained.

[0051] S6. Using the window list and feasible time window set as rigid constraints, solve the optimization problem of minimizing navigation mileage, fuel consumption and operation waiting time through a multi-objective path planning model with time window constraints, and generate an operation instruction table.

[0052] Specifically, the maintenance scheduling terminal can use the window list and feasible time window set as inviolable rigid constraints, inputting them into a multi-objective path planning model with time window constraints (Vehicle Routing Problem with Time Windows, VRPTW) for solution. The VRPTW model uses minimizing total voyage distance, total fuel consumption, and total job waiting time as multi-objective optimization functions, employing a heuristic algorithm for iterative search. Under the premise of satisfying all time and space constraints, it obtains the globally optimal voyage route and job sequence. After solving, the maintenance scheduling terminal can convert the globally optimal voyage route and job sequence into a standardized job instruction table.

[0053] In the first preferred embodiment, the objective function expression of the multi-objective path planning model can be:

[0054] In the formula, Indicates the comprehensive cost of multiple objectives. The weighting coefficients representing the voyage distance. The weighting coefficient representing fuel consumption. The weighting coefficient represents the job waiting time, and , Indicates the total number of nodes. Represents a node To the node The sailing distance, express Decision variables, when At that time, it indicates that the maintenance vessel is from the node. sail to the node Otherwise , The fuel consumption coefficient per unit distance traveled. This indicates the total number of marine monitoring devices requiring maintenance. Indicates the first Waiting time for maintenance of marine monitoring equipment in Taiwan.

[0055] In the first preferred embodiment, the multi-objective path planning model with time window constraints can be a multi-objective optimization solution model built based on genetic algorithm, simulated annealing algorithm or ant colony algorithm, which can quickly converge to an approximate optimal solution under complex constraints.

[0056] In a first preferred embodiment, the operation instruction table can be used to characterize the optimal navigation sequence, the expected arrival time of each marine monitoring device to be maintained, and the start time of the operation.

[0057] In the aforementioned intelligent maintenance and scheduling method for marine monitoring equipment, the maintenance and scheduling terminal systematically integrates dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of the marine monitoring equipment to be maintained. This provides a high-quality, aligned basic data source for subsequent scheduling decisions. Feature reconstruction is performed on the dynamic time-series environmental data and the multi-dimensional operation and maintenance status data of the marine monitoring equipment to be maintained, unifying the dispersed environmental parameters, equipment status, and maintenance task requirements into a comprehensive data cube. This formalizes multi-source heterogeneous data into a structured, computable, unified representation, laying a data foundation for time-series assessment and spatial optimization. A multi-parameter weighted scoring model is used to perform multi-dimensional weighted calculations on the environmental parameter matrix, quantifying the overall suitability for sea operations at each future time point and transforming dynamic sea state changes into a comparable scoring sequence. The sea operation suitability score is then compared with a preset threshold. This method aggregates consecutively satisfied time points and, through hierarchical filtering, generates a window list. This allows for the rigid selection of suitable and alternative sea operation windows from a time perspective, eliminating periods of severe sea conditions and reducing ineffective sea operations and safety risks at the source. Based on the window list, spatiotemporal constraints are matched to the equipment state matrix to generate a set of feasible time windows for each piece of equipment to be maintained. This precisely links time operability with equipment maintenance needs, generating rigid spatiotemporally coupled constraints. Using the window list and the set of feasible time windows together as rigid constraints, a multi-objective path planning model with time window constraints is used to solve the optimization problem of minimizing navigation mileage, fuel consumption, and operation waiting time, generating an operation instruction table. This allows for simultaneous optimization of route efficiency and time window compliance within an integrated framework, achieving globally optimal scheduling with the shortest navigation mileage, lowest fuel consumption, and minimized operation waiting time. This method integrates traditionally fragmented time prediction, route planning, and cost constraints into an end-to-end intelligent decision-making process, reducing ineffective sea operations and safety risks, improving single-batch equipment coverage and operational efficiency, and providing a quantifiable, executable, and iterative technical solution for the maintenance and scheduling of marine monitoring equipment.

[0058] In the first preferred embodiment, such as Figure 3 As shown, a flowchart illustrating the process of generating a set of feasible time windows is provided. S5 may include: S51. Based on spatial coordinates and home port location, calculate the estimated sailing time from the home port to each marine monitoring equipment to be maintained, and extract the expected operation time of each marine monitoring equipment to be maintained from the equipment status matrix.

[0059] Optionally, the maintenance task candidate set may include, but is not limited to, the spatial coordinates and home port locations of each marine monitoring device to be maintained.

[0060] For example, the maintenance scheduling terminal can extract the geospatial coordinates of each marine monitoring device to be maintained and the baseline geographical coordinates of the preset home port from a pre-generated set of maintenance task candidates. Based on the rated navigation performance parameters of the maintenance vessel, combined with the historical navigation resistance correction coefficient and the average ocean current influence coefficient of the target sea area, it calculates the estimated sailing time from the home port to the location of each device to be maintained along the shortest great circle route. The maintenance scheduling terminal can retrieve the expected operation time corresponding to the unique identifier of each marine monitoring device from the device status matrix. The expected operation time can be pre-calibrated and stored according to the device type, the category of the maintenance task to be performed, and the statistical time consumption of similar historical operations.

[0061] S52. Iterate through each window in the window list, calculate the start time of each window and the sum of the estimated sailing time corresponding to the start time, and obtain the estimated arrival time of each window.

[0062] For example, the maintenance scheduling terminal can traverse each candidate time window in the window period list in chronological order of start time. For each marine monitoring device to be maintained, the start time of the current window period and the estimated sailing time corresponding to the marine monitoring device to be maintained are timestamped together to obtain the estimated arrival time of the maintenance vessel after it departs from the home port on time at the start time of the window period and arrives at the location of the marine monitoring device to be maintained. During the traversal, the maintenance scheduling terminal can generate a unique association identifier for each marine monitoring device-window period combination to be maintained, and bind and store the estimated arrival time corresponding to the association identifier.

[0063] In the first preferred embodiment, the estimated arrival time can be used to characterize the expected time when the maintenance vessel departs from the home port and arrives at the designated location of the marine monitoring equipment to be maintained, under ideal operating conditions without unexpected delays.

[0064] S53. The window period in which the difference between the end time and the expected operation time is not less than the estimated arrival time shall be identified as the candidate window period for the marine monitoring equipment to be maintained.

[0065] For example, the maintenance scheduling terminal can calculate the difference between the end time of the current window and the expected operation time of the marine monitoring equipment to be maintained for each combination of marine monitoring equipment and window period, thus obtaining the latest allowed arrival time of the maintenance vessel within the window period. The maintenance scheduling terminal can compare the estimated arrival time and the latest allowed arrival time of the marine monitoring equipment to be maintained within the window period. If the estimated arrival time is less than or equal to the latest allowed arrival time, it is determined that the window period can meet the complete operation time requirement of the equipment, and the window period is marked as a candidate window period for the marine monitoring equipment to be maintained.

[0066] In the first preferred embodiment, the candidate window period can be used to characterize the time interval that can provide a complete operational time guarantee for the marine monitoring equipment to be maintained, based on meeting the sea state suitability requirements.

[0067] S54. Sort the candidate window periods of each marine monitoring equipment to be maintained according to the start time to obtain each feasible time window.

[0068] For example, the maintenance scheduling terminal can use the device's unique identifier as an index to retrieve all candidate time windows corresponding to each marine monitoring device to be maintained, and extract the start time, end time, and mean sea state suitability score attributes of each candidate time window. The maintenance scheduling terminal can arrange all candidate time windows in ascending order of start time, and after removing redundant windows with completely overlapping times, generate a feasible time window for the marine monitoring device to be maintained.

[0069] In the first preferred embodiment, the feasible time window can be used to characterize an ordered set of all effective time intervals in which maintenance operations can be carried out for a single marine monitoring device to be maintained, reflecting the time feasibility sequence of maintenance operations for the marine monitoring device to be maintained within a future preset time range.

[0070] S55. Aggregate the feasible time windows according to the device identifier to obtain a set of feasible time windows.

[0071] For example, the maintenance scheduling terminal can assign a globally unique device identifier to each marine monitoring device to be maintained. The terminal can then associate and bind each feasible time window corresponding to the device to be maintained with that device identifier, constructing a key-value pair structured data storage structure indexed by the device identifier. The terminal can then integrate the associated data of all devices to be maintained to generate a set of feasible time windows.

[0072] In the first preferred embodiment, the maintenance scheduling terminal generates a set of feasible time windows for each marine monitoring device to be maintained by matching the time constraints in the window period list with the spatial location and operation duration of each device. This accurately associates the suitable sea condition time window with the equipment maintenance requirements, providing a rigid and computable spatiotemporal constraint basis for subsequent multi-objective path planning.

[0073] In the first preferred embodiment, please refer to Figure 2 S2 may include: S21. Extract dynamic hydrological and meteorological parameters from environmental dynamic time series data, and align the dynamic hydrological and meteorological parameters according to timestamps to obtain an environmental parameter matrix.

[0074] For example, the maintenance and dispatch terminal can filter out dynamic hydro-meteorological parameters directly related to maritime operation safety from preprocessed environmental dynamic time-series data, and use linear interpolation to unify all dynamic hydro-meteorological parameters to the same time granularity. It then performs point-by-point timestamp alignment based on Coordinated Universal Time (UTC) to eliminate time discrepancies between different data sources. After alignment, the maintenance and dispatch terminal can construct a two-dimensional array with time points as rows and hydro-meteorological parameters as columns to obtain a structured environmental parameter matrix.

[0075] In the first preferred embodiment, dynamic hydro-meteorological parameters can be used to characterize the marine hydrological and meteorological elements of the target sea area that change dynamically over time.

[0076] In the first preferred embodiment, the environmental parameter matrix can be used to characterize a structured set of all core hydrological and meteorological parameters corresponding to each time point within a future preset time range.

[0077] S22. Extract the equipment status characteristics of each marine monitoring device to be maintained from the multi-dimensional operation and maintenance status data, and aggregate and arrange the equipment status characteristics into a two-dimensional matrix according to the equipment identifier to obtain the equipment status matrix.

[0078] For example, the maintenance scheduling terminal can extract multi-dimensional features that comprehensively reflect the health status and maintenance needs of equipment from preprocessed multi-dimensional operation and maintenance status data. The maintenance scheduling terminal can use the globally unique equipment identifier as an index to aggregate all status features of the same marine monitoring equipment to be maintained into a row, and arrange different status features as columns to generate a two-dimensional structured equipment status matrix. Each element in the equipment status matrix corresponds to a specific status feature value of the marine monitoring equipment to be maintained.

[0079] In the first preferred embodiment, the multi-dimensional features may include, but are not limited to, the cumulative running time of the equipment, the degree of sensor accuracy attenuation, the remaining battery capacity, the signal strength of the communication module, the wear status of the mooring system, the frequency of historical faults, and the time of the last maintenance completion.

[0080] S23. Perform multi-dimensional evaluation and calculation on the equipment status matrix to obtain the urgency index of each marine monitoring equipment to be maintained.

[0081] For example, the maintenance scheduling terminal can use the entropy weight method combined with expert experience correction to determine the weight coefficients corresponding to each state feature in the equipment state matrix. The weight coefficients are used to reflect the degree of influence of different state features on the equipment failure risk. The maintenance scheduling terminal can perform weighted summation of all state feature values ​​corresponding to each marine monitoring equipment to be maintained to obtain the urgency index of the marine monitoring equipment to be maintained.

[0082] In the first preferred embodiment, the expression for the urgency index can be:

[0083] In the formula, Indicates the first The urgency index of Taiwan's marine monitoring equipment awaiting maintenance This represents the total number of dimensions representing the device status characteristics. Indicates the first The comprehensive weighting coefficient corresponding to each equipment status characteristic is obtained by weighted fusion of objective weights calculated by the entropy weight method and subjective weights calibrated by expert experience. Indicates the first Taiwan awaits maintenance of marine monitoring equipment. The normalized values ​​of the equipment status characteristics.

[0084] Optionally, the urgency index can be used to characterize the urgency of the marine monitoring equipment to be maintained requiring on-site maintenance. The higher the urgency index, the greater the risk of equipment failure, indicating that maintenance work should be prioritized.

[0085] S24. Marine monitoring equipment with an urgency index greater than the urgency threshold is selected as candidate maintenance equipment.

[0086] For example, the maintenance scheduling terminal can traverse each row of data in the device status matrix, compare the urgency index of each marine monitoring device to be maintained with the pre-stored urgency threshold, filter out marine monitoring devices whose urgency index exceeds the urgency threshold, and mark marine monitoring devices whose urgency index exceeds the preset urgency threshold as candidate maintenance devices.

[0087] In the first preferred embodiment, the urgency threshold can be a pre-set critical value based on the operation and maintenance resource configuration capability, the equipment importance level, and historical fault statistics, and can be dynamically adjusted according to actual operation and maintenance needs.

[0088] In a first preferred embodiment, candidate maintenance equipment can be used to characterize the set of marine monitoring equipment that require priority on-site maintenance operations during the current scheduling cycle.

[0089] S25. Based on candidate maintenance equipment, extract the spatial coordinates and expected operation time of the candidate maintenance equipment from the multi-dimensional operation and maintenance status data, and map the urgency index into urgency weight.

[0090] For example, the maintenance scheduling terminal can use the unique identifier of the candidate maintenance equipment as the search key to extract the geospatial coordinates of the deployment location of each candidate equipment and the expected operation time corresponding to the maintenance task type from the preprocessed multi-dimensional operation and maintenance status data. The maintenance scheduling terminal can use a linear normalized mapping function to transform the urgency index of all candidate maintenance equipment into a unified numerical range to obtain the urgency weight corresponding to each candidate maintenance equipment.

[0091] S26. Based on the equipment identifier, the spatial coordinates, expected operation duration and urgency weight are associated and combined to generate a set of maintenance task candidates.

[0092] For example, the maintenance scheduling terminal can assign a unique task identifier to each candidate maintenance device and associate and bind the device identifier, geospatial coordinates, expected operation duration, and urgency weight one by one to obtain an independent maintenance task entry. The maintenance scheduling terminal can integrate all maintenance task entries to construct a structured set of maintenance task candidates.

[0093] In a first preferred embodiment, the maintenance task candidate set can be stored in the form of key-value pairs, with the device identifier as the key and the task information containing spatial coordinates, expected operation duration and urgency weight as the value.

[0094] In the first preferred embodiment, the maintenance task candidate set can be used to characterize the core information set of all maintenance tasks to be executed in the current scheduling cycle. The maintenance task candidate set may include, but is not limited to, all basic task data required for subsequent spatiotemporal constraint matching and path planning.

[0095] In the first preferred embodiment, the maintenance scheduling terminal extracts and constructs an environmental parameter matrix from dynamic time-series environmental data and an equipment status matrix from multi-dimensional operation and maintenance status data. Then, it calculates an urgency index to screen candidate maintenance equipment and generates a maintenance task candidate set by associating and combining spatial coordinates, expected operation duration, and urgency weight. This transforms multi-source heterogeneous environmental and equipment data into a structured, computable, and unified data organization form, providing a complete and standardized input basis for subsequent overseas suitability scoring and spatiotemporal constraint matching.

[0096] In the first preferred embodiment, the expression for the suitability score for going to sea can be:

[0097] In the formula, Indicates a future point in time. The suitability score for going to sea, Indicates the first The dynamic weighting coefficients of each environmental parameter, and satisfying , Indicates the first Piecewise linear penalty functions corresponding to each environmental parameter Indicates a future point in time. The The values ​​of environmental parameters, which may include, but are not limited to, significant wave height, maximum wave height, average wind speed, gust wind speed, tidal range, visibility, ocean current speed, and typhoon warning level.

[0098] For example, the environmental parameters can be in the following order: significant wave height, maximum wave height, average wind speed, gust wind speed, tidal range, visibility, ocean current speed, and typhoon warning level.

[0099] For example, the maintenance scheduling terminal can extract the original values ​​of significant wave height, maximum wave height, average wind speed, gust wind speed, tidal range, visibility, ocean current speed, and typhoon warning level for each time point within a preset future time range from the environmental parameter matrix by timestamp index. Then, it sequentially loads a predefined piecewise linear penalty function corresponding to each environmental parameter. The predefined piecewise linear penalty function can be calibrated based on the safety regulations for offshore operations and the performance parameters of the maintenance vessel. The predefined piecewise linear penalty function maps environmental parameters with different numerical ranges to a unified scoring interval. When the environmental parameters are within the safe operating range, a high score is output, and when they exceed the critical threshold, a zero score is output.

[0100] In the first preferred embodiment, the maintenance and scheduling terminal can load the currently effective dynamic weight coefficients, perform a weighted summation operation for each time point, multiply the standardized scores of each parameter by their corresponding weights, and then sum them to obtain the seagoing suitability score for that time point. The maintenance and scheduling terminal can arrange the seagoing suitability scores of all time points in chronological order to generate a continuous score sequence.

[0101] In the first preferred embodiment, the maintenance and scheduling terminal maps the effective wave height, maximum wave height, average wind speed, gust wind speed, tidal range, visibility, ocean current speed, and typhoon warning level into standardized scores through piecewise linear penalty functions, and then weights and sums them with dynamic weight coefficients to generate a series of seagoing suitability scores for each future time point. This quantifies the multi-dimensional and non-linearly coupled sea conditions into unified and comparable operational feasibility indicators, providing an objective and accurate quantitative decision-making basis for subsequent window selection.

[0102] In the first preferred embodiment, after generating the job instruction table, it may further include: S7. Obtain the actual operation data generated by the marine monitoring equipment's operation instruction table, and collect and calculate the actual total cost of this task.

[0103] For example, the maintenance dispatch terminal can collect real-time data on the entire operation process of this maintenance task. After the real-time operation data is collected, the maintenance dispatch terminal can classify and collect the real-time operation data according to cost categories, and separately calculate the fuel consumption, ship depreciation and port charges generated during the navigation phase, the labor hours, equipment consumables, tool wear and tear and additional costs generated during the operation phase, and add up all the cost categories to obtain the actual total cost of this task.

[0104] In the first preferred embodiment, the real-time operation data can be used to characterize the actual operating status and resource consumption of the maintenance task from the ship's departure to its return to port, and is the basic data source for cost accounting and deviation analysis.

[0105] In the first preferred embodiment, the actual total cost of this task can be the sum of all direct and indirect costs incurred during the execution of this maintenance task.

[0106] S8. Compare and analyze the actual data of the operation and the preset plan data to obtain the execution deviation index.

[0107] For example, the maintenance scheduling terminal can retrieve the preset plan data corresponding to this task. The preset plan data may include, but is not limited to, the planned voyage mileage, planned voyage time, planned fuel consumption, planned operating time of each piece of equipment, and planned consumable usage. The maintenance scheduling terminal can compare the actual operation data and the preset plan data item by item, calculate the absolute deviation value and relative deviation rate of each parameter, generate multi-dimensional execution deviation indicators, and quantify and classify the degree of deviation, distinguishing between minor deviation, general deviation, and major deviation.

[0108] In the first preferred embodiment, the execution deviation index can be a set of quantitative parameters used to characterize the degree of deviation between the actual operation process and the preset plan. The execution deviation index may include, but is not limited to, navigation deviation, operation deviation, cost deviation and time deviation.

[0109] S9. Allocate the actual total cost according to the proportion of the voyage distance and the direct operating cost to obtain the actual cost of each marine monitoring device to be maintained.

[0110] For example, the maintenance dispatch terminal can break down the actual total cost of this mission into two parts: navigation-related costs and direct operating costs. The maintenance dispatch terminal can allocate navigation-related costs according to the proportion of the distance between each piece of equipment to be maintained to the total navigation mileage of this mission, calculating the navigation cost to be borne by each piece of equipment. Direct operating costs, including consumable costs and specific labor costs specific to each piece of equipment, can be directly attributed to the corresponding marine monitoring equipment to be maintained. The maintenance dispatch terminal can then add the navigation costs and direct operating costs allocated to each piece of marine monitoring equipment to be maintained to obtain the actual cost per piece of marine monitoring equipment to be maintained.

[0111] In the first preferred embodiment, the expression for the voyage cost can be:

[0112] In the formula, Indicates the first The navigation costs that Taiwan should share in the maintenance of its marine monitoring equipment. This indicates the total distance of the voyage related to the marine monitoring equipment requiring maintenance during this mission. This indicates the total voyage distance for this mission. This represents the total navigation-related costs for this mission.

[0113] In the first preferred embodiment, the actual cost per unit can be the total cost corresponding to the maintenance operation of a single marine monitoring device. The actual cost per unit may include, but is not limited to, the shared public navigation costs and the dedicated direct operating costs.

[0114] S10. If the actual cost of a single marine monitoring device exceeds the budgeted cost of the marine monitoring device to be maintained, the marine monitoring device to be maintained is marked as an over-cost node, and the deviation feature vector is extracted based on the execution deviation index.

[0115] For example, the maintenance scheduling terminal can iterate through the actual cost of each individual marine monitoring device to be maintained and compare it one by one with the pre-stored maintenance budget cost for each device, filtering out the marine monitoring devices whose actual cost exceeds the budget cost. The maintenance scheduling terminal can mark the marine monitoring devices whose actual cost exceeds the budget cost as over-cost nodes. For each over-cost node, the maintenance scheduling terminal can extract deviation data related to the maintenance process of the marine monitoring device from multi-dimensional execution deviation indicators, and arrange these deviation data in a preset dimension order to construct the deviation feature vector corresponding to the over-cost node.

[0116] In the first preferred embodiment, the deviation data may include, but is not limited to, deviations in arrival time, operation duration, fuel consumption, and deviations between actual and predicted sea state values.

[0117] In a first preferred embodiment, the deviation feature vector can be used to characterize the multidimensional deviation between the actual operation process and the preset plan.

[0118] S11. Based on the deviation feature vector, update the dynamic weight vector in the multi-parameter weighted scoring model to obtain the updated multi-parameter weighted scoring model.

[0119] For example, the maintenance scheduling terminal can parse the dimensional data related to sea state prediction deviation in the deviation feature vector and calculate the contribution of the prediction deviation of each environmental parameter to the cost overrun or time delay of this operation. The maintenance scheduling terminal can use gradient descent to iteratively update the dynamic weight vector in the multi-parameter weighted scoring model. The maintenance scheduling terminal can positively adjust the weight coefficients corresponding to environmental parameters whose contribution is higher than a preset contribution threshold, and negatively adjust the weight coefficients corresponding to environmental parameters whose contribution is lower than the preset contribution threshold, keeping the sum of all weight coefficients at 1 during the update process. After the update is completed, the new dynamic weight vector overwrites the original weight vector, resulting in the updated multi-parameter weighted scoring model.

[0120] In the first preferred embodiment, the expression for the updated dynamic weight coefficient can be:

[0121] In the formula, Indicates the updated number The dynamic weighting coefficients of the environmental parameters. This represents the dynamic weight coefficients before the update. This represents the preset learning rate, used to control the step size of weight updates. The loss function is a weighted average of the cost and time deviations of the current operation. This represents the gradient of the loss function with respect to the dynamic weight coefficients of the k-th environmental parameter.

[0122] S12. Based on the deviation feature vector, update the fuel consumption weight coefficient and relaxation penalty coefficient in the multi-objective path planning model to obtain the updated multi-objective path planning model.

[0123] For example, the maintenance scheduling terminal can extract relevant data on fuel consumption deviation and job waiting time deviation from the deviation feature vector, and calculate the relative deviation rate between actual fuel consumption and planned value and the relative deviation rate between actual job waiting time and planned value based on the relevant data on fuel consumption deviation and job waiting time deviation.

[0124] In the first preferred embodiment, the maintenance and scheduling terminal can adjust the fuel consumption weight coefficient in the multi-objective path planning model according to the calculated relative deviation rate between the actual fuel consumption and the planned value. The larger the fuel consumption deviation rate, the higher the corresponding fuel consumption weight coefficient.

[0125] In the first preferred embodiment, the maintenance scheduling terminal can adjust the relaxation penalty coefficient based on the relative deviation rate between the actual operation waiting time and the planned value. The larger the relative deviation rate between the actual operation waiting time and the planned value, the higher the penalty coefficient. After the update is completed, the updated multi-objective path planning model is obtained.

[0126] In the first preferred embodiment, the expression for the updated fuel consumption weighting coefficient can be:

[0127] In the formula, This represents the updated fuel consumption weighting coefficient. The weighting coefficient representing fuel consumption. This indicates the preset adjustment coefficient. This represents the relative deviation rate of fuel consumption, which is the difference between actual fuel consumption and planned fuel consumption divided by the planned fuel consumption. This is an abbreviation for fuel, used to identify the fuel consumption dimension corresponding to this relative deviation rate.

[0128] S13. Based on the actual cost per unit of the over-cost node, update the maintenance cycle of the marine monitoring equipment to be maintained corresponding to the over-cost node to obtain the updated maintenance cycle.

[0129] For example, the maintenance scheduling terminal can calculate the ratio of the actual cost per unit of the over-cost node to the corresponding budgeted cost, thus obtaining the cost overrun ratio. The maintenance scheduling terminal can then combine the historical maintenance cost data and current health status of the marine monitoring equipment to be maintained, employing a linear adjustment strategy to correct the original maintenance cycle. The higher the cost overrun ratio, the longer the maintenance cycle should be, and upper and lower thresholds for the maintenance cycle can be set. After the update, the maintenance scheduling terminal can write the updated maintenance cycle into the operation and maintenance file of the marine monitoring equipment to be maintained, serving as the basis for scheduling maintenance tasks in the next scheduling cycle.

[0130] In the first preferred embodiment, the maintenance scheduling terminal obtains the actual total cost by collecting real-time work execution data and allocates it to the actual cost of a single unit. Then, it compares the budgeted cost to mark over-cost nodes, extracts deviation feature vectors, and performs closed-loop adaptive updates on the dynamic weight vector in the multi-parameter weighted scoring model, the fuel consumption weight coefficient and relaxation penalty coefficient in the multi-objective path planning model, and the maintenance cycle of over-cost nodes. This feeds the actual execution deviations back to the key parameters of the scheduling decision model, thereby achieving continuous optimization and self-evolution of the scheduling strategy.

[0131] In the aforementioned intelligent maintenance and scheduling method, system, equipment, and medium for marine monitoring equipment, a comprehensive data cube is constructed by integrating dynamic environmental time-series data and multi-dimensional equipment operation and maintenance status data. A multi-parameter weighted scoring model quantifies sea state parameters into a seagoing suitability score sequence, which is then graded and filtered to obtain a window period list. Based on this window period list, a set of feasible time windows for each device is generated. A multi-objective path planning model with time window constraints is used to solve the optimization problem of minimizing navigation mileage, fuel consumption, and operation waiting time, outputting an operation instruction table. Furthermore, actual total costs are collected and allocated to individual units using real-time operation data. Over-cost nodes are marked, and deviation feature vectors are extracted. The dynamic weight vector, fuel consumption weight coefficient, relaxation penalty coefficient, and maintenance cycle are then updated in a closed-loop adaptive manner. This technical solution replaces the traditional fixed-cycle inspection and manual experience-based scheduling mode, achieving a three-line collaborative closed loop of timeline prediction, route planning, and economic feedback. This effectively reduces the rate of ineffective seagoing operations and safety risks, shortens navigation mileage, and improves equipment coverage. It also enables comprehensive cost accounting and automatic early warning, giving the scheduling strategy continuous self-evolution capabilities.

[0132] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0133] Based on the same inventive concept, this application also provides a system for implementing the intelligent maintenance and scheduling method for marine monitoring equipment described above. The solution provided by this system is similar to the implementation scheme described in the above method. Therefore, the specific limitations of one or more embodiments of the intelligent maintenance and scheduling system for marine monitoring equipment provided below can be found in the limitations of the intelligent maintenance and scheduling method for marine monitoring equipment described above, and will not be repeated here.

[0134] Specific application examples: Four 1-meter ecological monitoring buoys deployed at different locations in a certain sea area by the State Oceanic Administration were selected as the implementation targets. The four 1-meter ecological monitoring buoys of the State Oceanic Administration were named F1, F2, F3 and F4 respectively. The measurement and forecast period was from May 1, 2025 to May 5, 2025 (time t1 to t5). The environmental dynamic time series data were obtained from the monthly marine environmental report of the sea area, the actual situation of the coastal observation station, the inversion of Fengyun satellite remote sensing, and the marine numerical forecast products. The multi-dimensional operation and maintenance data were obtained from the real-time operating parameters of the edge acquisition unit of the buoy body, the historical maintenance log of the Fisheries Bureau, the equipment factory life calibration curve, and the platform fault alarm record.

[0135] Step 1: Raw data preprocessing (3σ outlier removal + missing data imputation + Min-Max normalization) Raw data sampling: Environmental dynamic time-series data includes surface water temperature, seawater salinity, sea area wind speed, and significant wave height; multi-dimensional operation and maintenance status data includes buoy battery voltage and solar power load power consumption.

[0136] In the original sampling data, F1 showed an abnormal temperature jump of 30.2℃ on May 3. It was identified as an outlier due to a momentary fault of the equipment sensor and removed using the 3σ three-standard deviation criterion. The missing data at this moment was filled in using the nearest neighbor mean filling method.

[0137] Unified operation and maintenance data: F1: Battery average voltage 12.3V, power consumption normalization value 0.89; 55 days past the last maintenance cycle, remaining service life 17%, sensor drift alarm 4 times in the last 7 days, marked as high priority emergency maintenance. F2 / F3 / F4: Power consumption normalization values ​​are 0.31, 0.35, and 0.33 respectively, and voltage parameters are within the factory health threshold range. They are all classified as monthly routine inspection tasks.

[0138] Step 2: 3D feature reconstruction to construct a comprehensive data cube Time dimension alignment: Satellite remote sensing samples once a day, buoy in-situ hydrological data twice a day, and shore-based meteorological stations sample every 6 hours; using the UTC calendar day as the granularity, the weighted average of multiple daily observation data is used to generate daily environmental characteristics. Taking time t1 and buoy F1 as an example, the environmental characteristics of the day are obtained by fusing multi-source data: surface water temperature 13.1℃, salinity 30.2‰, nearshore average wind speed 3.4m / s, and significant wave height 0.65m.

[0139] Spatial dimension binding points: The geographical locations of the four buoys are based on measured nautical charts: Points F1 to F44 belong to the same fishery monitoring area, and are bound to the area number and coastal geographic coordinates to form spatial feature labels.

[0140] Task Dimension Aggregation: F1 aggregates (emergency fault repair) tasks, F2 / F3 / F4 aggregates (monthly routine inspection) tasks.

[0141] Construct a three-dimensional integrated data cube: X-axis (time t1~t5), Y-axis (4 spatial points), Z-axis (two types of tasks: maintenance / inspection); t1 is day 1, t2 is day 2, t3 is day 3, t4 is day 4, t5 is day 5. Single grid storage example: [Time: 2025.5.1, Space: F1, Task: Emergency maintenance, Environment {13.1℃, 30.2‰, Wind speed 3.4m / s}, Operation and maintenance {12.3V, Power consumption normalized 0.89, Remaining lifespan 17%}].

[0142] Step 3: Calculate the seagoing suitability score daily using multi-parameter weighting. Four environmental factors were selected: water temperature, sea surface wind speed, significant wave height, and surface current velocity. The weights were determined based on historical fisheries operation and maintenance data: water temperature weight = 0.25, wind speed weight = 0.40, wave height weight = 0.25, and ocean current weight = 0.10. The scoring formula for this embodiment, based on four parameters, is as follows:

[0143] Daily score calculation over 5 days: The calculation score for the first day was 0.98, the calculation score for the second day was 0.91, the calculation score for the third day was 0.39, the calculation score for the fourth day was 0.86, and the calculation score for the fifth day was 0.77. On the third day, the calculated score was affected by the passage of cold air, strong winds and increased waves (wind speed 7.9 m / s, wave height 1.55 m), exceeding the parameters and resulting in a score of 0.39, which is below the qualified line for going to sea. Therefore, it was prohibited to go to sea for maintenance on that day.

[0144] Step 4: Threshold-based screening of candidate maintenance time windows The preset seagoing suitability qualification threshold S(t) is ≥0.6, and the minimum continuous operation time per window is ≥8h; Boolean sequence determination: [t1 qualified, t2 qualified, t3 unqualified, t4 qualified, t5 qualified]; Two sets of candidate windows are obtained by aggregating consecutive time periods: W1 priority window: May 1st to May 2nd, a continuous 48 hours, with an average window score of 0.945; W2 backup window: May 4th to May 5th, 48 consecutive hours, average window score 0.815; Both window durations meet the minimum operation time limit. The final window period list is generated by sorting by start time: [W1, W2], where W1 is the preferred time window period and W2 is the emergency backup window period.

[0145] Step 5: Spatiotemporal matching + path optimization to generate standardized work instructions. Home Port: Calculate the voyage distance and one-way travel time at each point based on the great circle route: Home port → F1: 31 nautical miles, sailing time 3.4 hours; F1 → F2: 18 nautical miles, sailing time 1.9 hours; F2 → F3: 33 nautical miles, sailing time 3.6 hours; F3 → F4: 12 nautical miles, sailing time 1.3 hours; F4 returning to home port: 36 nautical miles, sailing time 3.8 hours. Work hours: 6.5 hours for emergency maintenance of F1 unit, and 3 hours each for routine inspection of F2 / F3 / F4 units; Multi-objective planning weights: range weight = 0.5, fuel consumption = 0.3, waiting time = 0.2; Substituting the VRPTW model with time window for optimization, the optimal route for sea operations on W1 (May 1st) is: Home port → F1 → F2 → F3 → F4 → return to home port, the total voyage is 130 nautical miles, and the entire process of operation and navigation can be completed within the May 1st window, with no cross-day standby loss.

[0146] Table 1 is the final work instruction table.

[0147] May 4th to May 5th will serve as the emergency backup window W2. If W1 experiences sudden and short-term severe sea conditions that prevent departure, the entire maintenance plan will be postponed to the backup window.

[0148] Second preferred embodiment, such as Figure 4 As shown, an intelligent maintenance and scheduling system for marine monitoring equipment includes: The data acquisition module can be used to acquire dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of marine monitoring equipment to be maintained; The feature reconstruction module can be used to reconstruct features from dynamic time-series environmental data and multi-dimensional operation and maintenance status data to obtain a comprehensive data cube; the comprehensive data cube includes an environmental parameter matrix, an equipment status matrix, and a maintenance task candidate set; The suitability score calculation module can be used to perform multi-dimensional weighted calculation of the environmental parameter matrix through a multi-parameter weighted scoring model to obtain the seagoing suitability score corresponding to each future time point; The window period classification and filtering module can be used to aggregate consecutive future time points that meet the sea availability threshold conditions into candidate time windows, and to classify and filter the candidate time windows to obtain a window period list. The spatiotemporal constraint matching module can perform spatiotemporal constraint matching on the device state matrix based on the maintenance task candidate set and window period list to obtain a set of feasible time windows; The instruction generation module can be used to solve the optimization problem of minimizing navigation mileage, fuel consumption and operation waiting time by using the window period list and feasible time window set as rigid constraints through a multi-objective path planning model with time window constraints, and generate an operation instruction table. The operation instruction table is used to represent the optimal navigation sequence, the expected arrival time of each marine monitoring equipment to be maintained and the operation start time.

[0149] In a second preferred embodiment, the spatiotemporal constraint matching module includes: The expected operation time extraction unit can be used to calculate the estimated sailing time from the home port to each marine monitoring equipment to be maintained based on spatial coordinates and home port location, and extract the expected operation time of each marine monitoring equipment to be maintained from the equipment status matrix. The estimated arrival time calculation unit can be used to traverse each window in the window list, calculate the start time of each window and the sum of the estimated flight time corresponding to the start time, and obtain the estimated arrival time of each window. The candidate window period determination unit can be used to determine the window period in which the difference between the end time and the expected operation time is not less than the estimated arrival time as a candidate window period for the marine monitoring equipment to be maintained. The feasible time window generation unit can be used to sort the candidate window periods of each marine monitoring equipment to be maintained according to the start time to obtain each feasible time window; The feasible time window aggregation unit can be used to aggregate each feasible time window according to the device identifier to obtain a feasible time window set.

[0150] In a second preferred embodiment, the feature reconstruction module includes: The environmental parameter matrix construction unit can be used to extract dynamic hydrological and meteorological parameters from dynamic environmental time-series data, and align the dynamic hydrological and meteorological parameters according to timestamps to obtain the environmental parameter matrix; The equipment status matrix construction unit can be used to extract the equipment status characteristics of each marine monitoring equipment to be maintained from multi-dimensional operation and maintenance status data, and aggregate and arrange the equipment status characteristics into a two-dimensional matrix according to the equipment identifier to obtain the equipment status matrix. The urgency index calculation unit can be used to perform multi-dimensional evaluation calculations on the equipment status matrix to obtain the urgency index of each marine monitoring equipment to be maintained. The candidate maintenance equipment screening unit can be used to select marine monitoring equipment to be maintained that has an urgency index greater than the urgency threshold as candidate maintenance equipment. The weight mapping unit can be used to extract the spatial coordinates and expected operation time of candidate maintenance equipment from multi-dimensional operation and maintenance status data based on candidate maintenance equipment, and map the urgency index into urgency weight. The maintenance task candidate set generation unit can be used to generate a maintenance task candidate set by associating and combining spatial coordinates, expected operation duration and urgency weight based on equipment identifier.

[0151] In the second preferred embodiment, the suitability score calculation module includes: The expression for the suitability score for going to sea is:

[0152] In the formula, Indicates a future point in time. The suitability score for going to sea, Indicates the first The dynamic weighting coefficients of each environmental parameter, and satisfying , Indicates the first Piecewise linear penalty functions corresponding to each environmental parameter Indicates a future point in time. The The values ​​of several environmental parameters, including significant wave height, maximum wave height, average wind speed, gust wind speed, tidal range, visibility, ocean current speed, and typhoon warning level.

[0153] In a second preferred embodiment, the system further includes: The cost accounting module can be used to obtain the actual operation data generated by the operation instruction table of marine monitoring equipment, and to collect and calculate the actual total cost of the task. The deviation index analysis module can be used to compare and analyze the actual operation data and the preset plan data to obtain the execution deviation index; The actual cost allocation module can be used to allocate the actual total cost according to the proportion of the voyage distance and the direct operation cost to obtain the actual cost of each marine monitoring device to be maintained; The deviation feature extraction module can be used to mark the marine monitoring equipment to be maintained as an over-cost node if the actual cost of a single unit exceeds the budget cost of the marine monitoring equipment to be maintained, and extract the deviation feature vector based on the execution deviation index; wherein, the deviation feature vector is used to characterize the multi-dimensional deviation state between the actual operation process and the preset plan; The weight update module for the multi-parameter weighted scoring model can be used to update the dynamic weight vector in the multi-parameter weighted scoring model based on the deviation feature vector, so as to obtain the updated multi-parameter weighted scoring model. The multi-objective path planning model coefficient update module can be used to update the fuel consumption weight coefficient and relaxation penalty coefficient in the multi-objective path planning model based on the deviation feature vector, so as to obtain the updated multi-objective path planning model. The maintenance cycle update module can be used to update the maintenance cycle of the marine monitoring equipment to be maintained corresponding to the over-cost node based on the actual cost of a single unit at the over-cost node, thus obtaining the updated maintenance cycle.

[0154] In a third preferred embodiment, a computer device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described intelligent maintenance and scheduling method for marine monitoring equipment.

[0155] In a fourth preferred embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the above-described intelligent maintenance and scheduling method for marine monitoring equipment.

[0156] In a fifth preferred embodiment, a computer program product includes a computer program that, when executed by a processor, implements the above-described intelligent maintenance and scheduling method for marine monitoring equipment.

[0157] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented, in whole or in part, as a computer program product, the computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line, or wireless (e.g., infrared, wireless, microwave, etc.) means). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0158] The above description is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for intelligent maintenance and scheduling of marine monitoring equipment, characterized in that, include: S1. Acquire dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of marine monitoring equipment to be maintained; S2. Perform feature reconstruction on environmental dynamic time-series data and multi-dimensional operation and maintenance status data to obtain a comprehensive data cube; the comprehensive data cube includes an environmental parameter matrix, an equipment status matrix, and a maintenance task candidate set. S3. The environmental parameter matrix is ​​calculated using a multi-parameter weighted scoring model to obtain the seagoing suitability score for each future time point. S4. Aggregate consecutive future time points that meet the suitability score for going to sea into candidate time windows, and perform hierarchical filtering on the candidate time windows to obtain a list of window periods. S5. Based on the maintenance task candidate set and window list, perform spatiotemporal constraint matching on the device state matrix to obtain a feasible time window set. S6. Using the window period list and feasible time window set as rigid constraints, the multi-objective path planning model with time window constraints is used to minimize the navigation mileage, fuel consumption and operation waiting time, and generate the operation instruction table. The operation instruction table is used to represent the optimal navigation sequence, the expected arrival time of each marine monitoring equipment to be maintained and the operation start time.

2. The intelligent maintenance and scheduling method for marine monitoring equipment according to claim 1, characterized in that, S2 include: S21. Extract dynamic hydrological and meteorological parameters from environmental dynamic time series data, and align the dynamic hydrological and meteorological parameters according to timestamps to obtain an environmental parameter matrix; S22. Extract the equipment status characteristics of each marine monitoring equipment to be maintained from the multi-dimensional operation and maintenance status data, and aggregate and arrange the equipment status characteristics into a two-dimensional matrix according to the equipment identifier to obtain the equipment status matrix. S23. Perform multi-dimensional evaluation and calculation on the equipment status matrix to obtain the urgency index of each marine monitoring equipment to be maintained; S24. Marine monitoring equipment requiring maintenance whose urgency index is greater than the urgency threshold shall be considered as candidate maintenance equipment. S25. Based on candidate maintenance equipment, extract the spatial coordinates and expected operation time of the candidate maintenance equipment from the multi-dimensional operation and maintenance status data, and map the urgency index to the urgency weight. S26. Based on the equipment identifier, the spatial coordinates, expected operation duration and urgency weight are associated and combined to generate a set of maintenance task candidates.

3. The intelligent maintenance and scheduling method for marine monitoring equipment according to claim 1 or 2, characterized in that, The maintenance task candidate set includes the spatial coordinates and home port locations of each of the marine monitoring devices to be maintained; S5 includes: S51. Based on the spatial coordinates and the location of the home port, calculate the estimated sailing time from the home port to each marine monitoring equipment to be maintained, and extract the expected operation time of each marine monitoring equipment to be maintained from the equipment status matrix. S52. Iterate through each window in the window list, calculate the start time of each window and the sum of the estimated sailing time corresponding to the start time, and obtain the estimated arrival time of each window. S53. The window period in which the difference between the end time and the expected operation time is not less than the estimated arrival time shall be identified as the candidate window period for the marine monitoring equipment to be maintained. S54. Sort the candidate window periods of each marine monitoring equipment to be maintained according to the start time to obtain each feasible time window; S55. Aggregate the feasible time windows according to the device identifier to obtain a set of feasible time windows.

4. The intelligent maintenance and scheduling method for marine monitoring equipment according to claim 1, characterized in that, Environmental parameters include significant wave height, maximum wave height, average wind speed, gust wind speed, tidal range, visibility, ocean current speed, and typhoon warning level; in S3, the expression for the suitability score for going to sea is: In the formula, Indicates a future point in time. The suitability score for going to sea, Indicates the first The dynamic weighting coefficients of each environmental parameter, and satisfying , Indicates the first Piecewise linear penalty functions corresponding to each environmental parameter Indicates a future point in time. The The values ​​of each environmental parameter.

5. The intelligent maintenance and scheduling method for marine monitoring equipment according to claim 1, characterized in that, Also includes: S7. Obtain the actual operation data generated by the marine monitoring equipment's operation instruction table, and collect and calculate the actual total cost of this task. S8. Compare and analyze the actual data of the operation and the preset plan data to obtain the execution deviation index; S9. Allocate the actual total cost according to the proportion of the voyage distance and the direct operating cost to obtain the actual cost of each marine monitoring device to be maintained; S10. If the actual cost of a single marine monitoring device exceeds the budgeted cost of the marine monitoring device to be maintained, the marine monitoring device to be maintained is marked as an over-cost node, and the deviation feature vector is extracted based on the execution deviation index. deviation Feature vectors are used to characterize the multidimensional deviation between the actual operation process and the preset plan; S11. Based on the deviation feature vector, update the dynamic weight vector in the multi-parameter weighted scoring model to obtain the updated multi-parameter weighted scoring model. S12. Based on the deviation feature vector, update the fuel consumption weight coefficient and relaxation penalty coefficient in the multi-objective path planning model to obtain the updated multi-objective path planning model. S13. Based on the actual cost per unit of the over-cost node, update the maintenance cycle of the marine monitoring equipment to be maintained corresponding to the over-cost node to obtain the updated maintenance cycle.

6. An intelligent maintenance and scheduling system for marine monitoring equipment, characterized in that, include: The data acquisition module acquires dynamic time-series environmental data of the target sea area and multi-dimensional operation and maintenance status data of the marine monitoring equipment to be maintained; The feature reconstruction module performs feature reconstruction on the environmental dynamic time-series data and the multi-dimensional operation and maintenance status data to obtain a comprehensive data cube; the comprehensive data cube includes an environmental parameter matrix, an equipment status matrix, and a maintenance task candidate set. The suitability score calculation module performs multi-dimensional weighted calculation on the environmental parameter matrix through a multi-parameter weighted scoring model to obtain the seagoing suitability score corresponding to each future time point; The window period classification and screening module aggregates consecutive future time points that meet the sea-going suitability score threshold into candidate time windows, and performs classification and screening of candidate time windows to obtain a window period list. The spatiotemporal constraint matching module performs spatiotemporal constraint matching on the device state matrix based on the maintenance task candidate set and the window period list to obtain a set of feasible time windows; The instruction generation module uses the window period list and the set of feasible time windows as rigid constraints, and solves the optimization problem of minimizing navigation mileage, fuel consumption and operation waiting time through a multi-objective path planning model with time window constraints to generate an operation instruction table. The operation instruction table is used to represent the optimal navigation sequence, the expected arrival time of each marine monitoring equipment to be maintained and the operation start time.

7. The intelligent maintenance and scheduling system for marine monitoring equipment according to claim 6, characterized in that, Also includes: The cost accounting module acquires the actual operation data generated by the marine monitoring equipment's operation instruction table, and collects and calculates the actual total cost of the task. The deviation index analysis module compares and analyzes the actual operation data and the preset plan data to obtain the execution deviation index; The actual cost allocation module allocates the actual total cost according to the proportion of the voyage distance and the direct operation cost to obtain the actual cost of each marine monitoring equipment to be maintained; The deviation feature extraction module marks the marine monitoring equipment to be maintained as an over-cost node if the actual cost of a single unit exceeds the budget cost of the marine monitoring equipment to be maintained, and extracts the deviation feature vector based on the execution deviation index. deviation Feature vectors are used to characterize the multidimensional deviation between the actual operation process and the preset plan; The weight update module of the multi-parameter weighted scoring model updates the dynamic weight vector in the multi-parameter weighted scoring model based on the deviation feature vector, so as to obtain the updated multi-parameter weighted scoring model. The multi-objective path planning model coefficient update module updates the fuel consumption weight coefficient and relaxation penalty coefficient in the multi-objective path planning model based on the deviation feature vector, and obtains the updated multi-objective path planning model. The maintenance cycle update module updates the maintenance cycle of the marine monitoring equipment to be maintained corresponding to the over-cost node based on the actual cost of a single unit of the over-cost node, thus obtaining the updated maintenance cycle.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the intelligent maintenance and scheduling method for marine monitoring equipment as described in any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the intelligent maintenance and scheduling method for marine monitoring equipment as described in any one of claims 1-5.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the intelligent maintenance and scheduling method for marine monitoring equipment as described in any one of claims 1-5.

Images