Method for calculating ship safety risk level based on big data

Abstract

The invention discloses a method for calculating a ship safety risk level based on big data, and relates to the technical field of data processing, and the method comprises the steps: collecting ship state data and environment parameters in real time through a computer, and carrying out the multi-dimensional fusion and local uploading of the collected data parameters; monitoring and analyzing the integrated and uploaded data stream in real time, and automatically triggering a data acquisition self-adaptive adjustment mechanism when the system detects that the ship data is uploaded abnormally or interrupted; generating a statistical model of the environmental parameters, and associating the model with the ship performance data to identify the influence of the environmental factors on the safety performance and form quantitative evaluation of the influence of the environmental conditions on the safety performance; performing environment compensation on a safety characteristic curve of the ship by utilizing a machine learning model based on the quantitative evaluation of the calculated environmental condition on the safety influence; and based on the safety characteristic curve after environment compensation, calculating to obtain the comprehensive risk level of the ship. And the evaluation is compensated, so that the credibility of the evaluation is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for calculating ship safety risk levels based on big data. Background Art

[0002] As an important transportation industry, the shipping industry has a lot of production safety risks. Electrical failures of equipment and systems on ships, fuel leaks, gas leaks, boiler explosions, etc. may lead to fires and explosions. Ships may encounter bad weather, sea fog and other conditions during navigation, resulting in hull rupture, hull capsizing, sinking and other accidents. When a ship is sailing, if the navigation information and channel conditions are not correctly judged, a ship collision accident may occur. In addition, there are various complex safety risk situations. The safety risks of ships are mainly concentrated on their own factors and external environmental factors.

[0003] The existing assessment methods are insufficient in comprehensiveness of various situations, resulting in room for improvement in the accuracy and credibility of the assessment results. Summary of the invention

[0004] In order to solve the above technical problems, a method for calculating the ship safety risk level based on big data is provided. This technical solution solves the problem that the existing assessment method proposed in the above background technology is insufficient in comprehensiveness for various situations, resulting in room for improvement in the accuracy and credibility of the assessment results.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A method for calculating ship safety risk level based on big data, comprising:

[0007] The computer collects ship status data and environmental parameters in real time through the integrated environmental parameter perception module and ship monitoring module, and performs multi-dimensional fusion and local upload of the collected data parameters;

[0008] Real-time monitoring and analysis of the integrated uploaded data stream. When the system detects abnormalities in ship data upload or data interruption, it automatically triggers the data acquisition adaptive adjustment mechanism;

[0009] Generate statistical models of environmental parameters and correlate the models with ship performance data to identify the impact of environmental factors on safety performance and form a quantitative assessment of the impact of environmental conditions on safety;

[0010] Based on the quantitative assessment of the safety impact of calculated environmental conditions, a machine learning model is used to perform environmental compensation on the ship's safety characteristic curve;

[0011] Based on the safety characteristic curve after environmental compensation, the comprehensive risk level of the ship is calculated.

[0012] Preferably, the computer collects ship status data and environmental parameters in real time through an integrated environmental parameter perception module and a ship monitoring module, and multi-dimensionally integrates and locally uploads the collected data parameters, including the following steps:

[0013] An environmental parameter sensing device is set up in the computer to obtain ship data through a standardized communication protocol;

[0014] The analog signal output by the sensing device is converted into a digital signal through a high-precision analog-to-digital converter to achieve dynamic collection of environmental parameters and ship status parameters;

[0015] Transmit all collected digital signals to the central processing unit via a high-speed data bus or wireless network. During the data transmission process, each data packet contains a timestamp, a sensor identifier, a data value, and the current sensor status;

[0016] The time synchronization and data correction algorithms are used to synchronize and correct the collected data from different sensors in real time, and the corrected data are formatted and uploaded to the local computer server.

[0017] Preferably, the real-time monitoring and analysis of the integrated uploaded data stream, when the system detects abnormal data upload or data interruption of the ship data, automatically triggers the data acquisition adaptive adjustment mechanism, specifically includes:

[0018] Monitor the integrated uploaded data stream, identify the upload time of each data packet according to the ship timestamp, and continuously monitor the upload time interval between two consecutive data packets. If it is detected that the time interval exceeds the preset acceptable range, it is considered that the data packet upload interruption has occurred;

[0019] After detecting that the data packet upload is interrupted, check whether the protocol identifier in the data packet matches the set communication protocol. If the protocol identifier matches the communication protocol, check whether the length of the data packet complies with the provisions of the communication protocol. If the length of the data packet deviates from the predetermined range, it is determined to be a format error.

[0020] The CRC code is used to detect the integrity of the data in transmission. If the check code does not match, it is determined that the data packet is damaged or the data is wrong during the transmission process;

[0021] When a format error or checksum mismatch is detected, the current data transmission status between the computer and the ship is marked as abnormal, and the data acquisition adaptive adjustment mechanism is activated;

[0022] The data acquisition adaptive adjustment mechanism is to re-integrate and upload the data stream at the abnormal or data location.

[0023] Preferably, generating a statistical model of environmental parameters and associating the model with ship performance data to identify the impact of environmental factors on safety performance and form a quantitative assessment of the impact of environmental conditions on safety specifically includes:

[0024] Obtain the historical uploaded data of the current ship model in the local server of the computer, establish the historical environmental data set D(t), and obtain the safety performance data corresponding to the environmental parameter data for modeling and analysis of the compensation algorithm. The historical environmental data set is recorded in the form of a time series, specifically:

[0025] D(t)={env1(t),env2(t),…env n (t)};

[0026] Where, env n (t) is the time series representation of the nth environmental parameter;

[0027] Through the multivariate regression model, the environmental parameters are associated with the safety performance data to identify the impact of environmental factors on safety performance. The core formula of the model is as follows:

[0028] E charge (t)=β0+β1env1(t)+β2env2(t)+…+β n env n (t)+∈(t);

[0029] In the formula, E charge (t) is the quantitative index of safety performance at time t, ∈(t) is the calculation error term at time t, β0, β1, β2, …β n are the regression coefficients of environmental parameters, β0, β1, β2, …β n Indicates the weight of the impact of environmental parameters on safety performance, and is used to set the initial value of the compensation coefficient in environmental compensation calculation;

[0030] These coefficients are fitted through regression analysis methods such as the least squares method to quantitatively determine the impact of each environmental parameter on safety performance.

[0031] Preferably, the quantitative assessment of the safety impact of the calculated environmental conditions and the use of a machine learning model to perform environmental compensation on the safety characteristic curve of the ship specifically include:

[0032] Obtain a historical environmental dataset D(t), and use the historical environmental dataset as the machine model input, where the key input features include time series environmental parameters and time series ship performance data. The input feature set after feature processing is represented as X(t), and the complex relationship between the environment and ship performance is captured through feature selection and feature engineering.

[0033] A deep neural network model is selected. The basic structure of the model is designed as a multi-layer perceptron to process multi-dimensional environmental and performance data input. The model is trained based on the input feature set data. The correlation between environmental changes and safety feature changes is extracted layer by layer through the multi-layer perceptron to predict the optimal environmental safety level and ship safety level. The model parameters are optimized by minimizing the input loss function, where the loss function is specifically:

[0034]

[0035] In the formula, I real( t ) 、V real( t ) is the actual measured environmental safety level and ship safety level, I ML( t ) 、V ML( t ) is the model-predicted environmental safety level and ship safety level, and N is the total number of training data;

[0036] Use the gradient descent method to calculate the gradient of the loss function with respect to the network parameters, and update the parameters in the opposite direction of the gradient to reduce the prediction error;

[0037] The trained machine model is deployed to the local computer server to predict the optimal ship safety level and the optimal environmental safety level. Based on the prediction results of the machine learning model, environmental compensation is performed on the safety characteristic curve, specifically:

[0038]

[0039] In the formula, I adj ( t ) 、V adj ( t ) is the environmental safety level and ship safety level after environmental compensation adjustment, α n is the environmental safety level compensation coefficient of the nth environmental parameter, γ n It is the ship safety level compensation coefficient of the nth environmental parameter.

[0040] Preferably, the calculation of the comprehensive risk level of the ship based on the safety characteristic curve after environmental compensation specifically includes:

[0041] The environmental parameters collected in real time are multiplied and added with the degree of influence of the environmental parameters on safety performance to obtain the actual environmental safety level;

[0042] In the safety characteristic curve after environmental compensation, a point where the environmental safety level is equal to the actual environmental safety level is obtained as the target point;

[0043] The ship safety level at the target point is taken as the comprehensive risk level of the ship.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The computer collects ship status data and environmental parameters in real time, performs multi-dimensional fusion and local upload of the collected data parameters. Through multi-dimensional data fusion and real-time upload, the ship status and environmental parameters can be fully understood, which helps to evaluate the safety status more comprehensively and accurately. Real-time monitoring and analysis of the uploaded data stream after integration. When the system detects abnormal upload of ship data or data interruption, it automatically triggers the data acquisition adaptive adjustment mechanism. Generate a statistical model of environmental parameters and associate the model with ship performance data to identify the impact of environmental factors on safety performance and form a quantitative assessment of the impact of environmental conditions on safety. The impact of environmental factors on safety performance is identified through statistical models, providing a scientific basis for environmental compensation and improving safety accuracy. Based on the quantitative assessment of the calculated environmental conditions on safety, the machine learning model is used to perform environmental compensation on the safety characteristic curve of the ship. Based on the safety characteristic curve after environmental compensation, the assessment can be compensated, thereby improving the credibility of the assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of a method for calculating ship safety risk level based on big data according to the present invention;

[0047] Figure 2 The computer of the present invention collects ship status data and environmental parameters in real time through the integrated environmental parameter perception module and ship monitoring module, and performs multi-dimensional fusion and local uploading of the collected data parameters;

[0048] Figure 3 The present invention monitors and analyzes the uploaded data stream in real time and integrates it. When the system detects abnormal data upload or data interruption of the ship, it automatically triggers the data acquisition adaptive adjustment mechanism flow chart;

[0049] Figure 4 A statistical model for generating environmental parameters for the present invention, and associating the model with ship performance data to identify the impact of environmental factors on safety performance, forming a quantitative assessment flow chart of the impact of environmental conditions on safety;

[0050] Figure 5 A flow chart of the environmental compensation of the safety characteristic curve of a ship using a machine learning model based on the quantitative assessment of the safety impact of the calculated environmental conditions of the present invention;

[0051] Figure 6 The present invention is a flow chart for calculating the comprehensive risk level of a ship based on the safety characteristic curve after environmental compensation. DETAILED DESCRIPTION

[0052] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0053] Reference Figure 1 As shown, a method for calculating the ship safety risk level based on big data includes:

[0054] The computer collects ship status data and environmental parameters in real time through the integrated environmental parameter perception module and ship monitoring module, and performs multi-dimensional fusion and local upload of the collected data parameters;

[0055] Real-time monitoring and analysis of the integrated uploaded data stream. When the system detects abnormalities in ship data upload or data interruption, it automatically triggers the data acquisition adaptive adjustment mechanism;

[0056] Generate statistical models of environmental parameters and correlate the models with ship performance data to identify the impact of environmental factors on safety performance and form a quantitative assessment of the impact of environmental conditions on safety;

[0057] Based on the quantitative assessment of the safety impact of calculated environmental conditions, a machine learning model is used to perform environmental compensation on the ship's safety characteristic curve;

[0058] Based on the safety characteristic curve after environmental compensation, the comprehensive risk level of the ship is calculated.

[0059] In this scheme, the safety risk level is considered from two aspects, the environment and the ship itself, and the comprehensive risk level of the ship is calculated. In the calculation process, through the construction of a neural network model and through continuous training, finally, a machine learning model is formed. The machine learning model is used to perform environmental compensation on the safety characteristic curve of the ship. The safety characteristic curve is a curve that corresponds the environmental safety level caused by environmental parameters to the ship safety level. The environmental safety level is obtained by environmental parameters, and the environmental parameters include the internal environment of the ship and the environmental conditions outside the ship. Therefore, the ship safety level corresponding to the environmental safety level can be obtained through the safety characteristic curve.

[0060] Reference Figure 2 As shown, the computer collects ship status data and environmental parameters in real time through the integrated environmental parameter perception module and ship monitoring module, and multi-dimensional fusion and local uploading of the collected data parameters include the following steps:

[0061] An environmental parameter sensing device is set up in the computer to obtain ship data through a standardized communication protocol;

[0062] The analog signal output by the sensing device is converted into a digital signal through a high-precision analog-to-digital converter to achieve dynamic collection of environmental parameters and ship status parameters, while ensuring high accuracy and real-time data. The collection formula is as follows:

[0063]

[0064] Where V d (t) represents the digital ship safety level signal, V(t) is the collected original ship safety level signal, V ref For reference ship safety level, m represents the number of digits of ADC;

[0065] Transmit all collected digital signals to the central processing unit via a high-speed data bus or wireless network. During the data transmission process, each data packet contains a timestamp, a sensor identifier, a data value, and the current sensor status;

[0066] The time synchronization and data correction algorithms are used to synchronize and correct the collected data from different sensors in real time. The synchronization process is implemented through the time interpolation algorithm to ensure that the data of each sensor is aligned under the same time reference. The synchronization formula is as follows:

[0067]

[0068] In the formula, x(t) represents the interpolated signal at time t, x(t i ) and x(t i+1 ) is the original data at adjacent moments;

[0069] The correction part uses a filtering algorithm to remove noise and outliers in the data to ensure the accuracy of the collected data, and then uploads the corrected data to the computer local server after format integration.

[0070] Reference Figure 3 As shown, the data stream after integration and upload is monitored and analyzed in real time. When the system detects abnormal data upload or data interruption of the ship, the data acquisition adaptive adjustment mechanism is automatically triggered. Specifically, it includes:

[0071] Monitor the integrated uploaded data stream, identify the upload time of each data packet according to the ship timestamp, and continuously monitor the upload time interval between two consecutive data packets. If it is detected that the time interval exceeds the preset acceptable range, it is considered that the data packet upload interruption has occurred;

[0072] By calculating the time interval between consecutive data packets, we can detect whether there is data loss or delay. Assume that the timestamp of the i-th data packet is t i , the timestamp of the i+1th data packet is t i+1, the upload time interval between two consecutive data packets should satisfy Δt = t i+1 -t i ≤T s , where T s is the preset range of acceptable time intervals.

[0073] After detecting that the data packet upload is interrupted, check whether the protocol identifier in the data packet matches the set communication protocol. If the protocol identifier matches the communication protocol, check whether the length of the data packet complies with the provisions of the communication protocol. If the length of the data packet deviates from the predetermined range, it is determined to be a format error.

[0074] The CRC code is used to detect the integrity of the data in transmission. If the check code does not match, it is determined that the data packet is damaged or the data is wrong during the transmission process. The process is that the sender performs binary polynomial division on the data to generate the check code, and the receiver receives the data and verifies the data integrity with the same polynomial;

[0075] When a format error or checksum mismatch is detected, the current data transmission status between the computer and the ship is marked as abnormal, and the data acquisition adaptive adjustment mechanism is activated;

[0076] The data acquisition adaptive adjustment mechanism is to re-integrate and upload the data stream at the abnormal or data location.

[0077] During the transmission of ship data, there may be anomalies or interruptions, which need to be handled in a timely manner. Otherwise, it is easy to cause delays in real-time risk calculation.

[0078] Reference Figure 4 As shown in the figure, a statistical model of environmental parameters is generated and the model is associated with the ship performance data to identify the impact of environmental factors on safety performance and form a quantitative assessment of the impact of environmental conditions on safety. Specifically, it includes:

[0079] Obtain the historical uploaded data of the current ship model in the local server of the computer, establish the historical environmental data set D(t), and obtain the safety performance data corresponding to the environmental parameter data for modeling and analysis of the compensation algorithm. The historical environmental data set is recorded in the form of a time series, specifically:

[0080] D(t)={env1(t),env2(t),…env n (t)};

[0081] Where, env n (t) is the time series representation of the nth environmental parameter;

[0082] Through the multivariate regression model, the environmental parameters are associated with the safety performance data to identify the impact of environmental factors on safety performance. The core formula of the model is as follows:

[0083] E charge (t)=β0+β1env1(t)+β2env2(t)+…+β n env n (t)+∈(t);

[0084] In the formula, E charge (t) is the quantitative index of safety performance at time t, ∈(t) is the calculation error term at time t, β0, β1, β2, …β n are the regression coefficients of environmental parameters, β0, β1, β2, …β n Indicates the weight of the impact of environmental parameters on safety performance, and is used to set the initial value of the compensation coefficient in environmental compensation calculation;

[0085] These coefficients are fitted through regression analysis methods such as the least squares method to quantitatively determine the impact of each environmental parameter on safety performance.

[0086] Among them, the quantifiable index value of safety is comprehensively evaluated based on the ship's model specification standards, safety performance and actual safety requirements, and flexible assessment methods such as weighted calculation can be adopted.

[0087] Based on the analysis results of the statistical model, the impact of environmental factors on safety performance is quantitatively evaluated. For example, excessive temperature can cause fire, and temperature env C (t) Safety performance E charge The influence of (t) is shown by the regression coefficient β C It shows that if the temperature rises, the safety performance predicted by the model will change according to β C The value decreases, and the result is used to further optimize the security strategy to compensate for the impact of environmental factors.

[0088] Reference Figure 5 As shown in the figure, based on the quantitative assessment of the impact of the calculated environmental conditions on safety, the environmental compensation of the safety characteristic curve of the ship using the machine learning model specifically includes:

[0089] Obtain the historical environmental data set D(t), and use the historical environmental data set as the machine model input, where the key input features include time series environmental parameters and time series ship performance data. The input feature set after feature processing is represented as X(t). The complex relationship between the environment and ship performance is captured through feature selection and feature engineering. The output can be expressed as: I ML (t),V ML (t) = f(X(t)), where f is the prediction function, which is modeled using a deep neural network to express the complex nonlinear relationship between input features and generate high-precision security parameter predictions;

[0090] A deep neural network model is selected. The basic structure of the model is designed as a multi-layer perceptron to process multi-dimensional environmental and performance data input. The model is trained based on the input feature set data. The correlation between environmental changes and safety feature changes is extracted layer by layer through the multi-layer perceptron to predict the optimal environmental safety level and ship safety level. The model parameters are optimized by minimizing the input loss function, where the loss function is specifically:

[0091]

[0092] In the formula, I real( t ) 、V real( t ) is the actual measured environmental safety level and ship safety level, I ML( t ) 、V ML( t ) is the model-predicted environmental safety level and ship safety level, and N is the total number of training data;

[0093] Use the gradient descent method to calculate the gradient of the loss function with respect to the network parameters, and update the parameters in the opposite direction of the gradient to reduce the prediction error;

[0094] The loss function is mainly used to evaluate the training of the deep neural network model. When the parameters of the deep neural network model satisfy the minimum loss function, the training ends. At the same time, the trained deep neural network model is used as a machine learning model.

[0095] The environmental safety level and the ship safety level are both obtained by inputting the ship status data and environmental parameters into the machine learning model. The environmental safety level is obtained by the environmental parameters, and the ship safety level is obtained by the ship status data. The environmental safety level and the ship safety level at the same time are matched as coordinates to form a safety characteristic curve, and the safety characteristic curve is compensated, thereby eliminating errors;

[0096] The trained machine model is deployed to the local computer server to predict the optimal ship safety level and the optimal environmental safety level. Based on the prediction results of the machine learning model, environmental compensation is performed on the safety characteristic curve, specifically:

[0097]

[0098] In the formula, I adj ( t ) 、V adj ( t ) is the environmental safety level and ship safety level after environmental compensation adjustment, α nis the environmental safety level compensation coefficient of the nth environmental parameter, γ n It is the ship safety level compensation coefficient of the nth environmental parameter.

[0099] Here the environmental compensation coefficient α n With γ n The initial values ​​are set based on a regression model based on historical data. During the safety process, the environmental compensation coefficients are not fixed, and the system dynamically adjusts these coefficients according to real-time environmental changes. For example, under extreme weather conditions (such as sudden temperature changes or sharp increases in humidity), the system can speed up the adjustment of environmental compensation coefficients.

[0100] Through continuous learning of the machine learning model, the system will record the relationship between environmental parameters and safety performance during each safety process and continuously update the compensation coefficient library. After a long period of safe operation, the system can gradually optimize the compensation coefficient through the self-learning mechanism, making it more accurate and efficient in long-term operation.

[0101] Reference Figure 6 As shown in the figure, based on the safety characteristic curve after environmental compensation, the comprehensive risk level of the ship is calculated to include:

[0102] The environmental parameters collected in real time are multiplied and added with the degree of influence of the environmental parameters on safety performance to obtain the actual environmental safety level;

[0103] In the safety characteristic curve after environmental compensation, a point where the environmental safety level is equal to the actual environmental safety level is obtained as the target point;

[0104] Furthermore, the present solution also proposes a method storage medium for calculating the ship safety risk level based on big data, on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned method for calculating the ship safety risk level based on big data is executed.

[0105] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0106] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for calculating ship safety risk level based on big data, characterized in that: include: The computer collects ship status data and environmental parameters in real time through the integrated environmental parameter perception module and ship monitoring module, and performs multi-dimensional fusion and local upload of the collected data parameters; Real-time monitoring and analysis of the integrated uploaded data stream. When the system detects abnormalities in ship data upload or data interruption, it automatically triggers the data acquisition adaptive adjustment mechanism; Generate statistical models of environmental parameters and correlate the models with ship performance data to identify the impact of environmental factors on safety performance and form a quantitative assessment of the impact of environmental conditions on safety; Based on the quantitative assessment of the safety impact of calculated environmental conditions, a machine learning model is used to perform environmental compensation on the ship's safety characteristic curve; Based on the safety characteristic curve after environmental compensation, the comprehensive risk level of the ship is calculated.

2. According to the method for calculating the ship safety risk level based on big data in claim 1, it is characterized in that: The computer collects ship status data and environmental parameters in real time through the integrated environmental parameter perception module and ship monitoring module, and multi-dimensionally integrates and locally uploads the collected data parameters, including the following steps: An environmental parameter sensing device is set up in the computer to obtain ship data through a standardized communication protocol; The analog signal output by the sensing device is converted into a digital signal through a high-precision analog-to-digital converter to achieve dynamic collection of environmental parameters and ship status parameters; Transmit all collected digital signals to the central processing unit via a high-speed data bus or wireless network. During the data transmission process, each data packet contains a timestamp, a sensor identifier, a data value, and the current sensor status; The time synchronization and data correction algorithms are used to synchronize and correct the collected data from different sensors in real time, and the corrected data are formatted and uploaded to the local computer server.

3. The method for calculating ship safety risk level based on big data according to claim 2 is characterized in that: The real-time monitoring and analysis of the integrated uploaded data stream, when the system detects abnormal data upload or data interruption, automatically triggers the data acquisition adaptive adjustment mechanism, specifically including: Monitor the integrated uploaded data stream, identify the upload time of each data packet according to the ship timestamp, and continuously monitor the upload time interval between two consecutive data packets. If it is detected that the time interval exceeds the preset acceptable range, it is considered that the data packet upload interruption has occurred; After detecting that the data packet upload is interrupted, check whether the protocol identifier in the data packet matches the set communication protocol. If the protocol identifier matches the communication protocol, check whether the length of the data packet complies with the provisions of the communication protocol. If the length of the data packet deviates from the predetermined range, it is determined to be a format error. The CRC code is used to detect the integrity of the data in transmission. If the check code does not match, it is determined that the data packet is damaged or the data is wrong during the transmission process; When a format error or checksum mismatch is detected, the current data transmission status between the computer and the ship is marked as abnormal, and the data acquisition adaptive adjustment mechanism is activated; The data acquisition adaptive adjustment mechanism is to re-integrate and upload the data stream at the abnormal or data location.

4. The method for calculating ship safety risk level based on big data according to claim 3 is characterized in that: The generating of the statistical model of environmental parameters and associating the model with the ship performance data to identify the impact of environmental factors on safety performance and form a quantitative assessment of the impact of environmental conditions on safety specifically includes: Obtain the historical uploaded data of the current ship model in the local server of the computer, establish the historical environmental data set D(t), and obtain the safety performance data corresponding to the environmental parameter data for modeling and analysis of the compensation algorithm. The historical environmental data set is recorded in the form of a time series, specifically: D(t)={env1(t),env2(t),…env n (t)}; Where, env n (t) is the time series representation of the nth environmental parameter; Through the multivariate regression model, the environmental parameters are associated with the safety performance data to identify the impact of environmental factors on safety performance. The core formula of the model is as follows: E charge (t)=β0+β1env1(t)+β2env2(t)+…+β n environment n (t)+∈(t); In the formula, E charge (t) is the quantitative index of safety performance at time t, ∈(t) is the calculation error term at time t, β0, β1, β2, …β n are the regression coefficients of environmental parameters, β0, β1, β2, …β n Indicates the weight of the impact of environmental parameters on safety performance, and is used to set the initial value of the compensation coefficient in environmental compensation calculation; These coefficients are fitted through regression analysis methods such as the least squares method to quantitatively determine the impact of each environmental parameter on safety performance.

5. The method for calculating ship safety risk level based on big data according to claim 4 is characterized in that: The quantitative assessment of the safety impact of the calculated environmental conditions and the use of a machine learning model to perform environmental compensation on the safety characteristic curve of the ship specifically include: Obtain a historical environmental dataset D(t), and use the historical environmental dataset as the machine model input, where the key input features include time series environmental parameters and time series ship performance data. The input feature set after feature processing is represented as X(t), and the complex relationship between the environment and ship performance is captured through feature selection and feature engineering. A deep neural network model is selected. The basic structure of the model is designed as a multi-layer perceptron to process multi-dimensional environmental and performance data input. The model is trained based on the input feature set data. The correlation between environmental changes and safety feature changes is extracted layer by layer through the multi-layer perceptron to predict the optimal environmental safety level and ship safety level. The model parameters are optimized by minimizing the input loss function, where the loss function is specifically: In the formula, I real (t), V real (t) is the actual measured environmental safety level and ship safety level, I ML (t), V ML (t) is the model-predicted environmental safety level and ship safety level, and N is the total number of training data; Use the gradient descent method to calculate the gradient of the loss function with respect to the network parameters, and update the parameters in the opposite direction of the gradient to reduce the prediction error; The trained machine model is deployed to the local computer server to predict the optimal ship safety level and the optimal environmental safety level. Based on the prediction results of the machine learning model, environmental compensation is performed on the safety characteristic curve, specifically: In the formula, I adj (t), V adj (t) is the environmental safety level and ship safety level after environmental compensation adjustment, α n is the environmental safety level compensation coefficient of the nth environmental parameter, γ n It is the ship safety level compensation coefficient of the nth environmental parameter. The safety characteristic curve is the curve formed by the coordinates of the paired environmental safety level and the ship safety level.

6. The method for calculating ship safety risk level based on big data according to claim 5 is characterized in that: The safety characteristic curve after environmental compensation is used to calculate the comprehensive risk level of the ship, which specifically includes: The environmental parameters collected in real time are multiplied and added with the degree of influence of the environmental parameters on safety performance to obtain the actual environmental safety level; In the safety characteristic curve after environmental compensation, a point where the environmental safety level is equal to the actual environmental safety level is obtained as the target point; The ship safety level at the target point is taken as the comprehensive risk level of the ship.

Images