Marine real-time database analysis platform

Abstract

The invention discloses a marine real-time database analysis platform, and relates to the technical field, and the system comprises the following components: a data acquisition and integration module, a fault prediction model depth embedding module, a navigation decision intelligent optimization module, and a user interface and interaction module. Through integrating the data acquisition and integration module and the fault prediction model deep embedding module, the platform can acquire and analyze the operation data of the ship key equipment in real time, accurately predict the equipment fault, and timely generate detailed preventive maintenance suggestions, so that the risk of navigation interruption caused by the equipment fault is reduced, and the maintenance efficiency is improved. And the cost of manual inspection and maintenance is greatly reduced, and the operation efficiency and safety of the ship are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ship operation and maintenance management and navigation decision optimization, and specifically to a marine real-time database analysis platform. Background Art

[0002] The marine real-time database analysis platform is an important invention in the field of ship operation and maintenance management and navigation decision optimization. With the rapid development of the global shipping industry, ships, as important carriers of maritime transportation, their safety and operational efficiency have received increasing attention.

[0003] In traditional technologies, there are many deficiencies in the deep embedding module of the fault prediction model. Traditional ship operation and maintenance management rely on manual inspections and experience judgments, making it difficult to accurately predict equipment failures and unable to generate detailed preventive maintenance suggestions based on the prediction results of the model, resulting in high maintenance costs and low navigation efficiency.

[0004] In summary, traditional technologies cannot meet the high requirements of the modern shipping industry for ship safety and operational efficiency in the marine real-time database analysis platform, especially in the deep embedding module of the fault prediction model. Therefore, it is particularly important to propose a marine real-time database analysis platform in the present invention. Summary of the Invention

[0005] The purpose of the present invention is to make up for the deficiencies of the existing technology and provide a marine real-time database analysis platform. This system, through the deep embedding module of the fault prediction model, not only improves the accuracy and timeliness of fault prediction but also can provide detailed preventive maintenance suggestions for the captain, thereby enhancing the safety and economy of the ship.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: A marine real-time database analysis platform, which includes a data collection and integration module, a deep embedding module of the fault prediction model, an intelligent optimization module for navigation decisions, and a user interface and interaction module;

[0007] The data collection and integration module: Real-time collects the operation data of each key equipment of the ship, ship position information, marine meteorological data, and route traffic flow data, integrates the historical data warehouse, stores and analyzes the historical data of the long-term operation of the ship, and provides rich training samples for the fault prediction model;

[0008] The fault prediction model deep embedding module: identify the key equipment fault types to be predicted, determine the specific content of the prediction, including fault time, location and type, collect the operation data of the ship's key equipment, the overall operation environment data of the ship, and the data of the interaction between equipment, perform preprocessing work such as data cleaning, denoising, and missing value handling, select features that have an important impact on the prediction target according to the characteristics of equipment faults and the characteristics of operation data, and extract features, including but not limited to time-domain features, frequency-domain features, and statistical features, to comprehensively reflect the operation status of the equipment;

[0009] According to the characteristics of the data and the prediction target, select a learning algorithm, consider factors such as the computational complexity, prediction accuracy, and interpretability of the algorithm, use the selected learning algorithm to build a ship equipment fault prediction model. The model should be able to comprehensively consider the overall operation environment of the ship and the interaction factors between equipment, use historical data as training samples, train the model, adjust the parameters of the model, use the validation data set to validate the trained model, evaluate its prediction performance, and make necessary adjustments and optimizations to the model according to the validation results;

[0010] Embed the trained model into the real-time data analysis system, input real-time data into the model to predict the possible fault time, location and type of key equipment, generate detailed preventive maintenance suggestions according to the prediction results of the model. The suggestions should include maintenance time, maintenance content, required spare parts and repair step information, collect new data, continuously update the training samples to improve the prediction accuracy of the model, continuously optimize and adjust the model according to the feedback in actual applications, and monitor the prediction results of the model in real time. When a possible fault is predicted, send an alarm signal in time. The alarm signal should be able to trigger the corresponding emergency response mechanism;

[0011] The navigation decision intelligent optimization module: integrate the marine weather forecasting system, obtain and analyze meteorological information in the next period of time in real time, combine the ship performance parameters and route traffic data, use the optimization algorithm to calculate the risks, fuel consumption and estimated arrival times of multiple potential routes, provide intelligent recommendations, display the optimal route plan for the captain, and allow dynamic adjustment of the navigation strategy according to real-time changes.

[0012] Furthermore, during the data collection phase of the data collection and integration module, data sources including the operation, location, marine meteorology, and route traffic flow of key ship equipment need to be determined. Sensors, monitoring devices, and data communication modules are configured, and a collection protocol is established to obtain and preliminarily process data in real time. During the data integration phase, data cleaning, transformation, and standardization are carried out, and then the processed data is integrated into a data warehouse that supports efficient storage, retrieval, and analysis. Historical data is stored and its integrity and traceability are ensured. The data warehouse is updated and maintained regularly. During the data analysis and utilization phase, the integrated data is preprocessed, and a fault prediction model is trained using historical data. After verification and optimization, real-time data is used to monitor and analyze the ship's operating status to achieve timely early warning.

[0013] Furthermore, the formula weights of the learning algorithm selected by the deep embedding module of the fault prediction model are derived from the characteristics of the data and the prediction target: Paying attention to the interpretability of the algorithm, then ω3 may be set relatively large; being more concerned about the prediction accuracy, ω2 may be larger; having a high requirement for computational complexity, ω1 may account for a relatively large proportion. The selected learning algorithm provides a basis for algorithm selection to construct an accurate fault prediction model.

[0014] Furthermore, the formula for constructing a ship equipment fault prediction model using the selected learning algorithm by the deep embedding module of the fault prediction model is: where M represents the evaluation index, F acc represents the prediction accuracy, C represents the computational complexity of the algorithm, indicating that the lower the computational complexity, the greater the contribution to the model evaluation index M, E int is the interpretability strength of the model, α is the prediction accuracy weight coefficient, β is the computational complexity weight coefficient, γ is the interpretability weight coefficient. Hoping for a higher prediction accuracy, α may be set relatively large; considering the operating efficiency on marine equipment and being more concerned about the computational complexity, β may have a relatively large value; interpretability is important for understanding the model results and subsequent maintenance, γ may be increased. The computational complexity C, prediction accuracy F acc and interpretability E int characteristics are affected by the learning algorithm selected by the sovereign.

[0015] Furthermore, the formula for generating detailed preventive maintenance suggestions according to the prediction results of the model by the deep embedding module of the fault prediction model is: where P m represents the preventive maintenance suggestion index, T p represents the predicted fault occurrence time of the model, L p is the encoding of the predicted fault occurrence location, F t represents the predicted fault type, S iis a vector combination, where θ1, θ2, θ3, and μ i are all weight coefficients. In actual ship equipment maintenance, the determination of maintenance time is crucial for the overall maintenance plan. Then θ1 may be set relatively large because T p represents the predicted failure occurrence time by the model, which is closely related to the maintenance time. The location where the equipment fails has a great impact on arranging the maintenance team and preparing maintenance tools. Then θ2 may be large because L p is the predicted failure occurrence location code. For the failure type, the maintenance resources and difficulties required for different failure types vary greatly. Then θ3 may be set according to the importance and complexity of the failure type. Some factors in the vector combination S i play a key role in the accuracy or integrity of the maintenance advice. Then the corresponding μ i may be set relatively large.

[0016] Furthermore, the intelligent navigation decision optimization module, through the data integration and preprocessing stage, connects to the marine weather forecasting system to obtain real-time weather data, obtains performance parameters from the ship management system, and collects route traffic data from the traffic management system. Then it enters the real-time weather information acquisition and analysis stage, regularly updates the weather data, and performs preprocessing and analysis through statistical methods and machine learning models to identify the key factors affecting navigation, preparing for subsequent route calculation and optimization.

[0017] Furthermore, the intelligent navigation decision optimization module calculates the route risk, fuel consumption, and estimated arrival time, generates multiple potential route plans, comprehensively considers the weather and traffic data to evaluate the risk, calculates the estimated fuel consumption and arrival time. Based on these calculations, it intelligently recommends the optimal route plan to the captain and displays the detailed information in a graphical manner. The captain can dynamically adjust the navigation strategy according to the real-time changes, provide the adjusted plan and evaluation, collect actual data during the navigation process, continuously optimize the algorithm and model, and make improvements according to the captain's feedback to enhance the user experience.

[0018] Furthermore, the system also includes a user interface and interaction module. According to understanding the needs and usage habits of the captain and management personnel for equipment failure warnings, maintenance advice, and navigation optimization plans, it develops the interface using a front-end technology stack and conducts testing and optimization. The implementation of the interaction function includes integrating equipment failure warnings and maintenance advice, displaying the navigation optimization plan, implementing remote monitoring and instant notification, and providing decision confirmation and instruction issuance functions. It collects feedback through user surveys and interviews and continuously optimizes and iterates the user interface and interaction module.

[0019] Compared with the prior art, the marine real-time database analysis platform has the following beneficial effects:

[0020] 1. By integrating a data collection and integration module and a deep embedding module for fault prediction models, the platform can collect and analyze the operation data of key ship equipment in real time, accurately predict equipment failures, and generate detailed preventive maintenance suggestions in a timely manner. This not only reduces the risk of navigation interruption caused by equipment failures but also significantly reduces the costs of manual inspections and maintenance, improving the operational efficiency and safety of the ship.

[0021] 2. In terms of optimizing navigation decisions, the platform integrates a marine weather forecasting system, which can obtain and analyze meteorological information for a period of time in the future in real time. Combining the ship's performance parameters and route traffic data, the platform can use optimization algorithms to calculate the risks, fuel consumption, and estimated arrival times of multiple potential routes, providing the captain with an intelligent recommended optimal route plan. This not only helps the captain dynamically adjust the navigation strategy according to real-time changes to reduce navigation risks and save fuel consumption but also displays detailed information in a graphical manner, enhancing the intuitiveness and convenience of decision-making. In addition, the platform allows the collection of actual data during navigation to continuously optimize the algorithms and models and make improvements based on the captain's feedback, thereby continuously enhancing the user experience and navigation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings described below are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0023] Figure 1 Operation flow chart of the marine real-time database analysis platform. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0025] Embodiment 1

[0026] This embodiment describes the process of the marine real-time database analysis platform.

[0027] Data including the engine, generator, and navigation equipment is obtained through GPS or other positioning systems. Data including wind speed, wind direction, wave height, and seawater temperature can be obtained through meteorological satellites, ocean observation stations, or dedicated meteorological sensors. Data is obtained through AIS or other traffic monitoring means. Sensors and monitoring equipment are installed on the ship to ensure that the required data can be collected in real time. A data communication module is configured to transmit the data to the central processing system. The time interval, data format, and transmission protocol for data collection are formulated to ensure the consistency and accuracy of data collection. According to the protocol, data is collected from each data source in real time, and the collected data is preliminarily processed;

[0028] The collected data is cleaned to remove duplicate, incorrect, or invalid data. Missing data is filled or interpolated. Data from different sources is converted into a unified format and standard to ensure consistency and comparability during data integration. The cleaned and converted data is integrated into a unified data warehouse. The data warehouse should be able to support efficient data storage, retrieval, and analysis. The historical data of long-term operation is stored in the data warehouse to ensure the integrity and traceability of historical data. The data in the data warehouse is updated regularly to reflect the latest ship operation status and ocean meteorological conditions. The data warehouse is maintained to ensure the accuracy and reliability of the data. The integrated data is further preprocessed to provide high-quality input data for subsequent data analysis and fault prediction models. Using historical data as training samples, a fault prediction model is constructed. Through model training, the accuracy and reliability of fault prediction are improved. The trained model is verified to evaluate its prediction performance. According to the verification results, the model is optimized and adjusted to improve its prediction effect. Using the integrated real-time data, the ship operation status is monitored and analyzed in real time. When abnormal situations are detected, warning signals are sent in a timely manner to ensure the safe operation of the ship.

[0029] Identify the key equipment fault types that need to be predicted, and determine the specific content of the prediction, including the fault time, location, and type. Collect the operation data of the ship's key equipment, the overall ship operation environment data, and the data of the interaction between equipment. Perform preprocessing work such as data cleaning, denoising, and missing value processing on the data. According to the characteristics of equipment faults and the characteristics of operation data, select the features that have an important impact on the prediction target, and extract features, including but not limited to time-domain features, frequency-domain features, and statistical features, to comprehensively reflect the operation status of the equipment;

[0030] According to the characteristics of the data and the prediction target, select a learning algorithm, where S represents the selection score, C represents the computational complexity, indicating that the lower the computational complexity, the greater the positive contribution to the selection score, P represents the prediction accuracy, and I represents the interpretability, It indicates that the higher the interpretability, the greater the positive contribution to the selection score. ω1, ω2, and ω3 represent weight coefficients. If we focus on the interpretability of the algorithm, then ω3 may be set relatively large. If we pay more attention to the prediction accuracy, ω2 may be larger. If there are high requirements for the computational complexity, ω1 may account for a relatively large proportion. Considering the factors of the algorithm's computational complexity, prediction accuracy, and interpretability, using the selected learning algorithm, a ship equipment fault prediction model is constructed, M = α × F acc + β × where M represents the evaluation index, F acc represents the prediction accuracy, C represents the computational complexity of the algorithm, It indicates that the lower the computational complexity, the greater the contribution to the model evaluation index M, E int is the interpretability strength of the model, α is the prediction accuracy weight coefficient, β is the computational complexity weight coefficient, γ is the interpretability weight coefficient. If we hope the model has higher prediction accuracy, α may be set relatively large. If we need to consider the operating efficiency on marine equipment and pay more attention to the computational complexity, β may have a relatively large value. Interpretability is very important for understanding the model results and subsequent maintenance, so γ may be increased. The model should be able to comprehensively consider the overall operating environment of the ship and the interaction factors between equipment. Using historical data as training samples, the model is trained, the parameters of the model are adjusted, and a validation data set is used to validate the trained model and evaluate its prediction performance. According to the validation results, the model is adjusted and optimized as necessary;

[0031] The trained model is embedded in the real-time data analysis system, and real-time data is used as the input to the model to predict the possible fault time, location, and type of key equipment. According to the prediction results of the model, detailed preventive maintenance suggestions are generated, where P m represents the preventive maintenance suggestion index, T p represents the fault occurrence time predicted by the model, L p is the encoding of the predicted fault occurrence location, F t represents the predicted fault type, S i is a vector combination, and θ1, θ2, θ3, μ i are all weight coefficients. In actual ship equipment maintenance, the determination of the maintenance time is crucial for the overall maintenance plan. Then θ1 may be set relatively large because T p represents the fault occurrence time predicted by the model, which is closely related to the maintenance time. The location of the equipment fault has a greater impact on arranging the maintenance team and preparing maintenance tools. Then θ2 may be relatively large because L pIt is the encoding of the predicted fault occurrence location. For fault types, the maintenance resources and difficulties required for different fault types vary greatly. Then, θ3 may be set according to the importance and complexity of the fault type. Some factors in the vector combination S i play a key role in the accuracy or integrity of the maintenance suggestions. Then, the corresponding μ i may be set to be relatively large. The suggestions should include maintenance time, maintenance content, required spare parts, and repair step information. Collect new data and continuously update the training samples to improve the prediction accuracy of the model. According to the feedback in actual applications, continuously optimize and adjust the model, and monitor the prediction results of the model in real time. When a possible fault is predicted, send an alarm signal in a timely manner. The alarm signal should be able to trigger the corresponding emergency response mechanism.

[0032] Connect and integrate the marine meteorological forecasting system to ensure real-time acquisition of meteorological data. The meteorological data includes, but is not limited to, wind speed, wind direction, wave height, visibility, temperature, humidity, and precipitation probability. Obtain the performance parameters of the current ship from the ship management system or database. The performance parameters include the ship's speed, fuel consumption rate, load capacity, and stability. Obtain the traffic data on the route from the traffic management system or third-party data sources. The traffic data includes channel congestion conditions, the positions and speeds of other ships, and potential dangerous areas. Regularly obtain the latest meteorological data from the marine meteorological forecasting system to ensure the accuracy and timeliness of the data. Preprocess and analyze the obtained meteorological data to identify the key meteorological factors that may affect navigation. Use statistical methods and machine learning models to predict and trend-analyze the meteorological data;

[0033] Generate multiple potential route plans based on the starting and ending points of the ship. Consider factors such as channel restrictions, water depth, and safety distance. Combine meteorological data and traffic data to evaluate the risks of each route. The risks include meteorological risks and traffic risks. Calculate the estimated fuel consumption of each route according to the performance parameters of the ship and the route length. Consider the fuel consumption changes under different meteorological conditions. Calculate the estimated arrival time of each route according to the ship's speed and route length. Consider the impact of meteorological conditions and traffic conditions on the navigation speed. Comprehensively consider the route risks, fuel consumption, and estimated arrival time, and use an optimization algorithm to calculate the comprehensive score of each route. According to the comprehensive score, recommend the optimal route plan to the captain. Display the optimal route plan to the captain in a graphical manner, providing detailed route information, including risk points, fuel consumption distribution, and estimated arrival time. Allow the captain to dynamically adjust the navigation strategy according to real-time changes, and provide the adjusted route plan and the corresponding risk assessment, fuel consumption prediction, and estimated arrival time;

[0034] During the voyage, the navigation data of the ship is collected in real time, including actual fuel consumption, actual arrival time, and actual meteorological and traffic conditions encountered. Based on the collected navigation data, the optimization algorithms and models are continuously improved and optimized to enhance the accuracy and reliability of the algorithms. The feedback from the captain is collected to understand the user experience and requirements for the intelligent optimization module. According to the user feedback, the module is improved and optimized to enhance the user experience.

[0035] Conduct in-depth communication with the captain and management personnel, the target users, to understand their needs for equipment failure warnings, maintenance suggestions, and navigation optimization plans. Analyze the users' usage habits, work processes, and expected interaction methods. Based on the results of the requirements analysis, design the prototype of the user interface, including layout, color, icons, and font elements, to ensure that the interface is intuitive and easy to use, the information is presented clearly, and the operation logic is reasonable. Develop the user interface using the front-end technology stack, conduct internal testing and user testing, collect feedback, and optimize the interface design. Integrate the equipment failure warning and maintenance suggestion background data into the user interface.

[0036] Design an intuitive way to display warning information, such as using a red warning light and a flashing icon to attract the user's attention. Provide detailed maintenance suggestions, including maintenance steps and required spare parts. Integrate the navigation optimization plan, including the optimal route, estimated fuel consumption, and estimated arrival time information. Use charts and maps, visualization tools to display the navigation optimization plan to help the user understand better. Implement the function that allows the captain or management personnel to remotely monitor the ship's status through a mobile device or an on-board terminal. Design an instant notification mechanism to notify the user in a timely manner when there is a device failure warning or a navigation optimization plan update. Provide an interface for the user to confirm the decision to ensure that the user makes a decision after fully understanding the information. Implement the function of issuing commands to allow the user to issue maintenance instructions and adjust the route through the interface.

[0037] Collect the user feedback on the user interface and interaction module through user surveys and interviews, pay attention to the user satisfaction and improvement suggestions. According to the user feedback, continuously optimize the user interface design, improve the interaction method, and regularly update the function module to ensure that the user interface and interaction module always meet the user requirements.

[0038] Embodiment 2

[0039] This embodiment describes an ocean-going cargo ship that is performing a long-term voyage mission from Asia to Europe. To ensure navigation safety and improve equipment maintenance efficiency, the cargo ship adopts a marine real-time database analysis platform.

[0040] During the voyage, the marine real-time database analysis platform collects the operation data of key equipment such as the main engine, auxiliary engine, and navigation system of the ship in real time and accurately through carefully configured sensors and monitoring devices. At the same time, the platform also obtains the ship's position information, marine meteorological data, and route traffic flow data in real time. After preliminary processing and cleaning, these data are efficiently integrated into a data warehouse that supports rapid retrieval and analysis, providing a solid foundation for subsequent data analysis and utilization.

[0041] The platform uses rich historical data and, through learning algorithms, trains an accurate ship equipment fault prediction model. Among them, M represents the evaluation index, F acc represents the prediction accuracy, C represents the computational complexity of the algorithm. It means that the lower the computational complexity, the greater the contribution to the model evaluation index M. E int is the interpretability strength of the model, α is the prediction accuracy weight coefficient, β is the computational complexity weight coefficient, and γ is the interpretability weight coefficient. If a higher prediction accuracy is desired for the model, α may be set relatively large; considering the operating efficiency on marine equipment and being more concerned about the computational complexity, β may have a relatively large value; interpretability is important for understanding the model results and subsequent maintenance, so γ may be increased. During a voyage, the model successfully predicted that there might be a potential fault risk in the main engine cooling system. Immediately based on the prediction result, the platform generated detailed preventive maintenance suggestions. Among them, P m represents the preventive maintenance advice index, T p represents the predicted fault occurrence time by the model, L p is the encoded predicted fault occurrence location, F t represents the predicted fault type, S i is a vector combination, and θ1, θ2, θ3, μ i are all weight coefficients. In actual ship equipment maintenance, determining the maintenance time is crucial for the overall maintenance plan. Then θ1 may be set relatively large because T p represents the predicted fault occurrence time by the model, which is closely related to the maintenance time. The location where the equipment fails has a greater impact on arranging the maintenance team and preparing maintenance tools. Then θ2 may be relatively large because L p is the encoded predicted fault occurrence location. For the fault type, the required maintenance resources and difficulty vary greatly for different fault types. Then θ3 may be set according to the importance and complexity of the fault type. Some factors in the vector combination S i play a key role in the accuracy or completeness of the maintenance advice. Then the corresponding μ iIt may be set relatively large, including the best maintenance time, specific maintenance content, required spare parts, and detailed repair steps. After receiving these suggestions, the captain quickly arranged for a professional maintenance team to conduct maintenance, thus effectively avoiding potential failures and ensuring the safe operation of the ship.

[0042] During the voyage, the platform obtains and analyzes marine weather forecast data in real time. At the same time, by combining the performance parameters of the ship and the route traffic data, it provides the captain with a detailed analysis of the risks, fuel consumption, and estimated arrival time of multiple potential routes. According to the optimal route plan recommended by the platform and combined with the real-time navigation situation, the captain flexibly adjusts the navigation strategy, which not only effectively reduces the navigation cost but also significantly shortens the navigation time and improves the overall navigation efficiency.

[0043] The captain and management personnel can access the user interface of the marine real-time database analysis platform anytime and anywhere through mobile devices or on-board terminals. The interface is designed to be intuitive and easy to use, and can clearly display important information such as equipment failure warnings, maintenance suggestions, and navigation optimization plans. At the same time, the platform also supports remote monitoring functions. The captain and management personnel can understand the operating status of the ship in real time, receive instant notifications, and make decision confirmations and issue instructions. This efficient interaction method greatly improves the work efficiency of the captain and management personnel and further ensures the safety of navigation.

[0044] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A marine real-time database analysis platform, characterized in that, The system includes a data acquisition and integration module, a deep embedding module for fault prediction models, an intelligent optimization module for navigation decision-making, and a user interface and interaction module; The data acquisition and integration module: Real-time collects the operation data of key ship equipment, ship position information, marine meteorological data, and route traffic flow data, integrates the historical data warehouse, stores and analyzes the historical data of long-term ship operation, and provides rich training samples for the fault prediction model; The deep embedding module for fault prediction models: Identifies the types of key equipment faults to be predicted, determines the specific content of the prediction, including fault time, location, and type, collects the operation data of key ship equipment, the overall ship operation environment data, and the data of the interaction between equipment, performs preprocessing work such as data cleaning, denoising, and missing value handling on the data, selects features that have an important impact on the prediction target according to the characteristics of equipment faults and the characteristics of operation data, and extracts features, including but not limited to time-domain features, frequency-domain features, and statistical features, to comprehensively reflect the operation state of the equipment; Select a learning algorithm according to the characteristics of the data and the prediction target. Among them, S represents the selection score, and C represents the computational complexity. It means that the lower the computational complexity, the greater the positive contribution to the selection score. P represents the prediction accuracy, and I represents the interpretability. It means that the higher the interpretability, the greater the positive contribution to the selection score. ω1, ω2, and ω3 represent the weight coefficients. Considering the computational complexity, prediction accuracy, and interpretability factors of the algorithm, use the selected learning algorithm to construct a ship equipment fault prediction model. The model should be able to comprehensively consider the overall operating environment of the ship and the interaction factors between equipment. Use historical data as training samples to train the model, adjust the parameters of the model, use the validation dataset to validate the trained model, evaluate its prediction performance, and make necessary adjustments and optimizations to the model according to the validation results. Embeds the trained model into the real-time data analysis system, inputs the model with real-time data, predicts the possible fault time, location, and type of key equipment, generates detailed preventive maintenance suggestions according to the prediction results of the model, and the suggestions should include maintenance time, maintenance content, required spare parts, and repair step information, collects new data, continuously updates the training samples to improve the prediction accuracy of the model, continuously optimizes and adjusts the model according to the feedback in actual applications, monitors the prediction results of the model in real time, and when a possible fault is predicted, issues an alarm signal in a timely manner, and the alarm signal should be able to trigger the corresponding emergency response mechanism; The intelligent optimization module for navigation decision-making: Integrates the marine meteorological forecasting system, obtains and analyzes the meteorological information in the next period of time in real time, combines the ship performance parameters and route traffic data, calculates the risks, fuel consumption, and estimated arrival times of multiple potential routes using optimization algorithms, provides intelligent recommendations, displays the optimal route plan for the captain, and allows dynamic adjustment of the navigation strategy according to real-time changes.

2. The marine real-time database analysis platform according to claim 1, wherein, In the data acquisition stage of the data acquisition and integration module, it is necessary to determine the data sources including the operation, position, marine meteorology, and route traffic flow of key ship equipment, configure sensors, monitoring equipment, and data communication modules, and establish a collection protocol to obtain and preliminarily process data in real time. In the data integration stage, data cleaning, transformation, and standardization are carried out, and then the processed data is integrated into a data warehouse that supports efficient storage, retrieval, and analysis. The data warehouse is updated and maintained regularly. In the data analysis and utilization stage, the integrated data is preprocessed, and the historical data is used to train the fault prediction model. After verification and optimization, the real-time data is used to monitor and analyze the ship operation state to achieve timely early warning.

3. The marine real-time database analysis platform according to claim 1, wherein The formula weights of the learning algorithm selected by the fault prediction model deep embedding module are derived from the characteristics of the data and the prediction target: if emphasis is placed on the interpretability of the algorithm, then ω3 may be set relatively large; if more attention is paid to the prediction accuracy, ω2 may be larger; if there are high requirements for computational complexity, ω1 may account for a relatively large proportion. The selected learning algorithm provides a basis for algorithm selection to build an accurate fault prediction model.

4. The marine real-time database analysis platform according to claim 1, characterized in that, The fault prediction model depth embedding module uses the selected learning algorithm to construct the formula of the ship equipment fault prediction model as follows: where M represents the evaluation index, F acc represents the prediction accuracy, C represents the computational complexity of the algorithm, indicating that the lower the computational complexity, the greater the contribution to the model evaluation index M, E int is the interpretability strength of the model, α is the prediction accuracy weight coefficient, β is the computational complexity weight coefficient, γ is the interpretability weight coefficient. If a higher prediction accuracy of the model is desired, α may be set to a larger value. Considering the operating efficiency on marine equipment and being more concerned about the computational complexity, β may have a larger value. Interpretability is important for understanding the model results and subsequent maintenance, so γ may be increased. The computational complexity C, prediction accuracy F acc and interpretability E int characteristics are affected by the learning algorithm selected by the sovereignty.

5. The marine real-time database analysis platform according to claim 1, wherein The formula for the failure prediction model depth embedding module to generate detailed preventive maintenance suggestions based on the model's prediction results is as follows: where P m represents the preventive maintenance suggestion index, T p represents the predicted failure occurrence time of the model, L p is the predicted failure occurrence location code, F t represents the predicted failure type, S i is a vector combination, and θ1, θ2, θ3, μ i are all weight coefficients. In the actual maintenance of ship equipment, the determination of the maintenance time is crucial for the overall maintenance plan. Then θ1 may be set relatively large because T p represents the predicted failure occurrence time of the model, which is closely related to the maintenance time. The location of equipment failure has a great impact on arranging the maintenance team and preparing maintenance tools. Then θ2 may be relatively large because L p is the predicted failure occurrence location code. For different failure types, the required maintenance resources and difficulties vary greatly. Then θ3 may be set according to the importance and complexity of the failure type. Some factors in the vector combination S i play a key role in the accuracy or integrity of the maintenance suggestions. Then the corresponding μ i may be set relatively large.

6. The marine real-time database analysis platform according to claim 1, wherein The navigation decision intelligent optimization module, through the data integration and preprocessing stage, connects to the marine meteorological forecasting system to obtain real-time meteorological data, obtains performance parameters from the ship management system, and collects route traffic data from the traffic management system. It enters the real-time meteorological information acquisition and analysis stage, regularly updates the meteorological data, and performs preprocessing and analysis through statistical methods and machine learning models to identify the key factors affecting navigation, preparing for subsequent route calculation and optimization.

7. The marine real-time database analysis platform according to claim 6, wherein The navigation decision intelligent optimization module calculates the route risk, fuel consumption, and estimated arrival time, generates multiple potential route plans, comprehensively considers meteorological and traffic data to evaluate risks, calculates the estimated fuel consumption and arrival time. Based on these calculations, it intelligently recommends the optimal route plan to the captain and displays detailed information in a graphical manner. The captain can dynamically adjust the navigation strategy according to real-time changes, provide the adjusted plan and evaluation, collect actual data during the navigation process, continuously optimize the algorithm and model, and make improvements based on the captain's feedback.

8. The marine real-time database analysis platform according to claim 1, characterized in that, The system also includes a user interface and interaction module. According to the needs and usage habits of the captain and management personnel for equipment fault warnings, maintenance suggestions, and navigation optimization plans, it develops the interface using a front-end technology stack and conducts testing and optimization. The realization of the interaction function includes integrating equipment fault warnings and maintenance suggestions, displaying navigation optimization plans, implementing remote monitoring and instant notifications, and providing decision confirmation and instruction issuance functions. It collects feedback through user surveys and interviews and continuously optimizes and iterates the user interface and interaction module.

Images