Ship structure safety forecasting method based on artificial neural network

Abstract

The invention discloses a ship structure safety forecasting method based on an artificial neural network. The method comprises the following steps: collecting historical data; processing the collected historical data, performing time sequence alignment on the multi-source heterogeneous historical data, and normalizing the processed historical data; constructing an artificial neural network model; training an artificial neural network model by using historical data; and inputting wave, navigational speed and draught data collected in real time into the trained artificial neural network model, obtaining a stress result of the ship structure by the artificial neural network model, extracting stress circulation by using a rain flow counting method, performing fatigue accumulated damage calculation, and performing safety state evaluation according to the stress and the fatigue accumulated damage. According to the method, the real-time dynamic forecasting of the ship structure stress and the fatigue life is realized.

Description

Technical Field

[0001] The present invention relates to a safety prediction method, in particular to a ship structure safety prediction method based on an artificial neural network, and belongs to the technical field of intelligent ship navigation safety. Background Art

[0002] In recent years, the shipping industry has experienced rapid growth, with a trend toward larger and faster ships. Throughout the entire ship operation process, the safety of ship structures remains a core element in ensuring safe navigation, smooth cargo transportation, and the safety of life and property. The marine environment is complex and unpredictable, with multiple factors intertwined, including waves, currents, and meteorological conditions, significantly impacting ship structures. Traditional methods for ship structural stress monitoring and fatigue life assessment rely primarily on finite element analysis, empirical formulas, and periodic inspections, failing to achieve real-time dynamic forecasting and early warning. Faced with complex and volatile marine environments, these methods exhibit numerous limitations. These methods struggle to fully account for various complex factors and their dynamic changes, and are unable to fully tap into and utilize extensive historical data for in-depth analysis and learning. This results in a superficial understanding of the safety status of ship structures. Furthermore, significant deficiencies exist in the integration and utilization of real-time sea state data and future weather and sea state forecasts, resulting in assessment results that are both timeless and lack foresight. With the advancement of artificial intelligence technology, artificial neural networks have demonstrated powerful capabilities in nonlinear data processing. Therefore, integrating neural network algorithms to achieve real-time dynamic forecasting of ship structural stress and fatigue life has significant engineering application value. The development of a dynamic prediction system for ship structure safety that integrates multi-source data and neural network technology has become a key issue that needs to be urgently addressed in the shipping field and has extremely important practical significance and application value. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a ship structure safety prediction method based on artificial neural network, which can realize real-time dynamic prediction of ship structure stress and fatigue life.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is: A ship structure safety prediction method based on artificial neural network comprises the following steps: S1. Historical data collection; S2. Process the collected historical data, perform time series alignment on multi-source heterogeneous historical data, and normalize the processed historical data; S3, building an artificial neural network model; S4, using historical data to train the artificial neural network model; S5. The real-time collected wave, speed, and draft data are input into the trained artificial neural network model. The artificial neural network model obtains the stress results of the ship structure and uses the rain flow counting method to extract the stress cycle to calculate the fatigue cumulative damage, and performs a safety status assessment based on the stress and fatigue cumulative damage.

[0005] Furthermore, step S1 is specifically as follows: collecting stress in key areas of the hull through stress sensors at key positions of the ship, collecting sailing speed and heading through the ship navigation system, collecting ship loading information through the loader, and collecting wave information through the wave radar, where the wave information includes wave height, wave period and wave direction.

[0006] Furthermore, the processing of the collected historical data in step S2 is specifically as follows: performing preliminary screening on the collected historical data to eliminate data points with obvious errors or abnormalities, performing data cleaning on the wave information, ship loading information and stress collected by the wave radar, loading instrument and stress sensor to remove outliers and noise, and using interpolation to supplement data with missing values.

[0007] Furthermore, the time series alignment of the multi-source heterogeneous historical data in step S2 is specifically as follows: the collected historical data is multi-source heterogeneous data, and these multi-source heterogeneous data are counted on a time dimension in seconds. The navigation speed and heading collected by the ship's navigation system and the ship loading information collected by the loading instrument are second-level data and do not need to be processed. The stress collected by the stress sensor is millisecond-level, and redundant data is removed to retain second-level data. The wave information collected by the wave radar is minute-level data, and the processing method is as follows: Wave height follows Rayleigh distribution , Where h is the wave height, is the effective wave height per minute output by the wave radar, e is a natural constant, Through the inverse transform sampling method, the formula for generating the second-level wave height is: ,in ; The fluctuations in the wave direction in a short period of time follow a uniform distribution , , is the wave direction angle at a certain moment, is the random perturbation angle; The wave period changes very little within 1 minute, and the second-level wave period is set as , T is the minute wave period.

[0008] Furthermore, the normalization of the processed historical data in step S2 is specifically to normalize the historical data using the formula Normalization, where X is historical data, Xmin is the minimum value in historical data, X max This is the maximum value in historical data.

[0009] Furthermore, step S3 is specifically to construct a BP neural network as a basic prediction model. The basic prediction model includes an input layer, a hidden layer and an output layer. The number of neurons in the input layer is consistent with the number of historical data features. The hidden layer is set to one or more layers. The number of neurons in the hidden layer is determined through experimental debugging. The activation function adopts ReLU or sigmoid function. The number of neurons in the output layer is 1 and the ship structure stress is output.

[0010] Furthermore, step S4 is specifically as follows: using the collected historical data to train the artificial neural network model, dividing the historical data into 80% training set and 20% validation set, during the training process, monitoring the loss function and accuracy of the artificial neural network model in real time, and dynamically adjusting the parameters of the artificial neural network model according to the training results, so that the loss function of the artificial neural network model is gradually reduced until it is minimized.

[0011] Furthermore, step S5 is specifically as follows: inputting the wave information, speed, heading and ship loading information collected and processed in real time into the trained artificial neural network model, the artificial neural network model quickly calculates the ship structure stress and uses the rain flow counting method to extract the stress cycle, and calculates the fatigue cumulative damage. The system performs a safety status assessment in real time based on the predicted ship structure stress and fatigue cumulative damage, or predicts the stress and fatigue life of key areas of the ship in the next 24 hours in combination with the future wave information of the weather forecast. The system judges the predicted results of the safety status based on the preset safety threshold. If the predicted result shows that it is in a safe state, real-time monitoring will continue. If the predicted result reaches a warning or dangerous state, the system will immediately issue an alarm to the crew through an audible and visual alarm to remind the crew to take effective measures in time.

[0012] Compared with the existing technology, the present invention has the following advantages and effects: the present invention provides a ship structure safety prediction method based on artificial neural network, which comprehensively considers various complex sea conditions and their dynamic changes during ship operation, and uses advanced artificial neural network algorithms to fully utilize and mine a large amount of historical data, conduct in-depth analysis and learning, and integrate sea condition data and ship data in real time, so that the ship structure safety assessment results are timely. In addition, by accessing future meteorological and sea condition forecast data, the ship structure safety assessment results are forward-looking, ensuring the safety of ship structure operation in actual sea conditions and greatly extending the service life of the ship; by processing the ship's historical data and unifying multi-source heterogeneous data, the multi-source heterogeneous ship data can be used for neural network model training and prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a flow chart of a method for predicting ship structure safety based on artificial neural network.

[0014] Figure 2 It is a schematic diagram of the present invention that counts these multi-source heterogeneous data into a time dimension in seconds.

[0015] Figure 3 It is a structural diagram of the artificial neural network model of the present invention.

[0016] Figure 4 It is a flow chart of the real-time stress and fatigue damage prediction of the present invention.

[0017] Figure 5 It is a flow chart of the present invention for predicting stress and fatigue damage in the next 24 hours. DETAILED DESCRIPTION

[0018] In order to elaborate on the technical solutions adopted by the present invention to achieve the predetermined technical purpose, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments, and the technical means or technical features in the embodiments of the present invention can be replaced without creative work. The present invention will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0019] like Figure 1 As shown, a ship structure safety prediction method based on artificial neural network of the present invention comprises the following steps: S1. Historical data collection.

[0020] Using sensors onboard the vessel, multi-source historical data is collected for neural network training and subsequent real-time forecasting. Specifically, stress sensors located in key areas of the ship collect stress data, the navigation system collects speed and heading, the loader collects loading information, and the wave radar collects wave information, including wave height, wave period, and wave direction. This information can be used for neural network training and real-time forecasting.

[0021] The system also integrates wave information from professional meteorological forecasting agencies for the next 24 hours, including wave height, wave period, wave direction, and other data. This data forms the basis for subsequent neural network model training, analysis, and prediction, and its accuracy and completeness directly impact the reliability of safety forecasts.

[0022] S2. Process the collected historical data, perform time series alignment on multi-source heterogeneous historical data, and normalize the processed historical data.

[0023] The collected historical data is initially screened to remove data points with obvious errors or anomalies. Wave information, ship loading information, and stress data collected by wave radar, loading instruments, and stress sensors are cleaned to remove outliers and noise. For example, by setting reasonable wave height, draft, and stress ranges, data points that clearly exceed the normal range are filtered out. For wave height data, if the wave height measured at a certain moment far exceeds the historical wave height data for the sea area and differs significantly from the wave height data at surrounding times, it can be determined as an outlier and removed. For data with missing values, interpolation is used to supplement them to ensure data integrity and accuracy. Subsequently, advanced filtering algorithms are used to remove noise interference.

[0024] The collected historical data is multi-source heterogeneous data with asynchronous frequencies. In order to train the model on the same time dimension, it is necessary to align the time series of the multi-source heterogeneous ship data. Figure 2 As shown, these multi-source heterogeneous data are counted on a time dimension of seconds. The navigation speed and heading collected by the ship's navigation system and the ship loading information collected by the loading instrument are second-level data and do not require processing. The stress collected by the stress sensor is millisecond-level data. The redundant data is removed and the second-level data is retained. The wave information collected by the wave radar is minute-level data, and its processing method is as follows: Wave height follows Rayleigh distribution , Where h is the wave height, is the effective wave height per minute output by the wave radar, e is a natural constant, Through the inverse transform sampling method, the formula for generating the second-level wave height is: ,in ; The effective wave height needs to be verified every minute, and the effective wave height is defined as the average of the maximum wave heights in the first 1 / 3.

[0025] The fluctuations in the wave direction in a short period of time follow a uniform distribution , , is the wave direction angle at a certain moment, is the random perturbation angle; The wave period changes very little within 1 minute, and the second-level wave period is set as , T is the minute wave period.

[0026] The historical data is converted into Normalization, where X is historical data, X min is the minimum value in historical data, X maxThis is the maximum value in historical data.

[0027] S3. Build an artificial neural network model.

[0028] like Figure 3 As shown in the figure, a BP neural network is constructed as a basic forecasting model. The basic forecasting model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is consistent with the number of historical data features. The input data features include wave data (wave height, wave period, wave direction), ship loading information (i.e. draft), speed, and stress. The hidden layer is set to one or more layers, and the number of hidden layer neurons is determined through experimental debugging. For example, the preset number of input layers is , is the number of input nodes, is the number of output nodes, The activation function uses ReLU or sigmoid function to achieve nonlinear transformation of input data. The number of neurons in the output layer is 1 and it outputs the ship structure stress.

[0029] S4. Use historical data to train the artificial neural network model.

[0030] The artificial neural network model is trained using the collected historical data, which is divided into an 80% training set and a 20% validation set. During the training process, the loss function and accuracy of the artificial neural network model are monitored in real time, and the parameters of the artificial neural network model are dynamically adjusted according to the training results, so that the loss function of the artificial neural network model is gradually reduced until it is minimized.

[0031] S5. The real-time collected wave, speed, and draft data are input into the trained artificial neural network model. The artificial neural network model obtains the stress results of the ship structure and uses the rain flow counting method to extract the stress cycle to calculate the fatigue cumulative damage, and performs a safety status assessment based on the stress and fatigue cumulative damage.

[0032] like Figure 4 As shown in the figure, the wave information, speed, heading and ship loading information collected and processed in real time are input into the trained artificial neural network model. The artificial neural network model quickly calculates the ship structure stress and uses the rain flow counting method to extract the stress cycle and calculate the fatigue cumulative damage. The system conducts a safety status assessment based on the predicted ship structure stress and fatigue cumulative damage in real time. The system judges the predicted results of the safety status based on the preset safety threshold. If the predicted result shows that it is in a safe state, real-time monitoring will continue. If the predicted result reaches a warning or dangerous state, the system will immediately send an alarm to the crew through an audible and visual alarm to remind the crew to take effective measures in time.

[0033] like Figure 5As shown, combined with future wave information, the artificial neural network model can predict stress and fatigue life in key areas of the ship for the next 24 hours, providing crew members with a trend analysis of the ship's navigation safety risks over the next period of time. The structural safety forecast for the scheduled routes for the next 24 hours is evaluated. If weather conditions significantly impact structural strength, route optimization can be performed to select the optimal route with the lowest stress.

[0034] The present invention provides a ship structure safety prediction method based on artificial neural network, which comprehensively considers various complex sea conditions and their dynamic changes during ship operation, uses advanced artificial neural network algorithms to fully utilize and mine a large amount of historical data, conducts in-depth analysis and learning, and integrates sea condition data and ship data in real time, so that the ship structure safety assessment results are timely. In addition, by accessing future meteorological and sea condition forecast data, the ship structure safety assessment results are forward-looking, ensuring the safety of ship structure operation in actual sea conditions and greatly extending the service life of the ship; by processing the ship's historical data and unifying multi-source heterogeneous data, the multi-source heterogeneous ship data can be used for neural network model training and prediction.

[0035] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any technician familiar with the present profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical content disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent replacement and improvement of the above embodiments made according to the technical essence of the present invention, within the spirit and principles of the present invention, without departing from the content of the technical solution of the present invention, shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A ship structure safety prediction method based on artificial neural network, characterized in that The following steps are involved: S1. Historical data collection; S2. Process the collected historical data, perform time series alignment on multi-source heterogeneous historical data, and normalize the processed historical data; S3, building an artificial neural network model; S4, using historical data to train the artificial neural network model; S5. The real-time collected wave, speed, and draft data are input into the trained artificial neural network model. The artificial neural network model obtains the stress results of the ship structure and uses the rain flow counting method to extract the stress cycle to calculate the fatigue cumulative damage, and performs a safety status assessment based on the stress and fatigue cumulative damage.

2. A ship structure safety prediction method based on artificial neural network according to claim 1, characterized in that: Specifically, step S1 includes collecting stress in key areas of the hull through stress sensors at key locations on the ship, collecting sailing speed and heading through the ship navigation system, collecting ship loading information through the loading instrument, and collecting wave information through the wave radar, where the wave information includes wave height, wave period, and wave direction.

3. The ship structure safety prediction method based on artificial neural network according to claim 1, characterized in that: The processing of the collected historical data in step S2 specifically includes: performing preliminary screening on the collected historical data to eliminate data points with obvious errors or abnormalities, performing data cleaning on the wave information, ship loading information and stress collected by the wave radar, loading instrument and stress sensor to remove outliers and noise, and using interpolation to supplement data with missing values.

4. The ship structure safety prediction method based on artificial neural network according to claim 1, characterized in that: In step S2, the time series alignment of the multi-source heterogeneous historical data is specifically performed as follows: the collected historical data is multi-source heterogeneous data, and these multi-source heterogeneous data are counted on a time dimension in seconds. The navigation speed and heading collected by the ship's navigation system and the ship loading information collected by the loading instrument are second-level data and do not need to be processed. The stress collected by the stress sensor is millisecond-level, and redundant data is removed to retain the second-level data. The wave information collected by the wave radar is minute-level data, and the processing method is as follows: Wave height follows Rayleigh distribution , Where h is the wave height, is the effective wave height per minute output by the wave radar, e is a natural constant, Through the inverse transform sampling method, the formula for generating the second-level wave height is: ,in ; The fluctuations in the wave direction in a short period of time follow a uniform distribution , , is the wave direction angle at a certain moment, is the random perturbation angle; The wave period changes very little within 1 minute, and the second-level wave period is set as , T is the minute wave period.

5. The ship structure safety prediction method based on artificial neural network according to claim 1 is characterized in that: The specific step S2 of normalizing the processed historical data is to use the formula Normalization, where X is historical data, X min is the minimum value in historical data, X max This is the maximum value in historical data.

6. The ship structure safety prediction method based on artificial neural network according to claim 1, characterized in that: Specifically, step S3 is to construct a BP neural network as a basic prediction model. The basic prediction model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is consistent with the number of historical data features. The hidden layer is set to one or more layers. The number of neurons in the hidden layer is determined through experimental debugging. The activation function adopts ReLU or sigmoid function. The number of neurons in the output layer is 1 and outputs the ship structure stress.

7. The ship structure safety prediction method based on artificial neural network according to claim 1, characterized in that: Specifically, step S4 is to use the collected historical data to train the artificial neural network model, divide the historical data into an 80% training set and a 20% validation set, monitor the loss function and accuracy of the artificial neural network model in real time during the training process, and dynamically adjust the parameters of the artificial neural network model according to the training results, so that the loss function of the artificial neural network model is gradually reduced until it is minimized.

8. The ship structure safety prediction method based on artificial neural network according to claim 1, characterized in that: Specifically, step S5 includes inputting the wave information, speed, heading and ship loading information collected and processed in real time into a trained artificial neural network model. The artificial neural network model quickly calculates the ship structure stress and uses the rain flow counting method to extract the stress cycle and calculate the fatigue cumulative damage. The system performs a safety status assessment in real time based on the predicted ship structure stress and fatigue cumulative damage, or predicts the stress and fatigue life of key areas of the ship in the next 24 hours in combination with future wave information from the weather forecast. The system judges the predicted results of the safety status based on a preset safety threshold. If the predicted results show that the ship is in a safe state, real-time monitoring is continued. If the predicted results reach a warning or dangerous state, the system immediately issues an alarm to the crew through an audible and visual alarm to remind the crew to take effective measures in a timely manner.

Images