A ship roll risk assessment method based on AIS and ERA5 data fusion

Abstract

The application discloses a ship roll risk assessment method based on AIS and ERA5 data fusion, and belongs to the field of maritime safety and ship seakeeping analysis; in view of the problem that wave parameters and ship motion states are not accurately matched in existing roll risk analysis, the method comprises the following steps: fusing AIS trajectory and ERA5 wave reanalysis data, adopting space-time nearest neighbor pairing to construct a ship-wave sample and extracting input parameters; a simplified frequency domain roll response model is established based on an encounter angle frequency and a natural roll angle frequency, a roll response standard deviation and a short-time overrun probability are calculated; a frequency proximity function, a sea state intensity function and a wave direction function are defined, and a parameter roll amplification factor is constructed; the overrun probability and the amplification factor are fused to obtain a normalized roll risk index, and are mapped to a latitude and longitude grid to generate a global or regional roll risk distribution map. The application makes full use of ship dynamic and wave information, realizes efficient and robust roll risk assessment and visualization, and supports route planning and maritime early warning.

Description

Technical Field

[0001] This invention relates to the field of maritime safety and ship seakeeping analysis, and in particular to a method for assessing ship roll risk based on the fusion of AIS and ERA5 data. Background Technology

[0002] Ship rolling motion is one of the key factors affecting ship navigation safety, cargo securing, and crew comfort. In severe sea conditions, severe rolling can not only lead to loss of ship stability but also cause serious maritime accidents such as capsizing. With the International Maritime Organization's (IMO) mandatory promotion of the second-generation integrity stability criterion, the accurate assessment of rolling failure modes (especially parametric rolling and loss of pure stability) has become a research hotspot in the fields of ship engineering and maritime safety.

[0003] Existing technologies related to ship roll risk, dynamic stability, and maritime safety assessment can be mainly categorized as follows: dynamic stability vulnerability assessment methods based on ship stability theory and international rules; ship roll response analysis methods based on frequency domain seakeeping theory; and maritime risk analysis and spatial mapping technologies based on AIS trajectory data.

[0004] However, existing technologies still have the following shortcomings: First, SGISC-related methods tend to focus on the ship design stage or specific operating condition verification, usually analyzing selected ship types, loading states, and discrete operating conditions, making it difficult to reflect the frequency and spatial exposure level of dangerous rolling conditions encountered by ships during actual operation; Second, existing frequency domain roll analysis methods usually rely on relatively complete ship stability parameters and detailed hydrodynamic parameters, and are mostly conducted under limited standard sea states, headings, and speeds, making them difficult to directly apply to global-scale, large-sample, long-term operational risk statistics; Third, existing AIS risk studies mainly focus on navigation risks such as collisions and traffic conflicts, rarely combining AIS with wave reanalysis data, seakeeping response models, and parametric roll criteria for wave-induced dynamic stability risk assessment; Therefore, there is an urgent need for a ship roll risk assessment method that can integrate real operational trajectory data and marine environmental data, take into account both physical mechanisms and large-scale computational efficiency, and simultaneously output single-point risk results and spatial distribution results. Summary of the Invention

[0005] This invention provides a ship roll risk assessment method based on the fusion of AIS and ERA5 data, which overcomes the problems of existing technologies that are difficult to reflect the frequency and spatial exposure level of dangerous roll conditions encountered by ships in actual operation, are difficult to directly apply to global-scale, large-sample, long-term series operational risk statistics, and rarely combine AIS and ERA5 wave reanalysis data, seakeeping response models and parametric roll criteria for wave-induced dynamic stability risk assessment.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows: A method for assessing ship roll risk based on the fusion of AIS and ERA5 data includes: S1. Collect global AIS trajectory data and ERA5 wave reanalysis data, and form ship-wave pairing samples by combining time nearest neighbor and spatial nearest neighbor methods; then obtain input parameters through ship-wave pairing samples. S2. Calculate and obtain the encounter angle frequency and natural roll angle frequency based on the input parameters; construct a simplified frequency domain roll response model based on the encounter angle frequency and natural roll angle frequency; calculate the standard deviation of the roll response based on the simplified frequency domain roll response model, and then obtain the short-term over-limit probability of the roll angle. S3. Define the frequency proximity function, sea state intensity function, and wave direction function, and construct the parametric roll amplification factor; S4. The short-term overshoot probability of the roll angle is fused with the roll amplification factor to obtain the normalized roll risk index; the normalized roll risk index is mapped to a preset latitude and longitude grid to obtain the average roll risk value and sample size of the grid; based on the average roll risk value and sample size of the grid, a roll risk distribution map at the global or regional scale is obtained.

[0007] Furthermore, the expression for calculating and obtaining the encounter angular frequency is as follows:

[0008] In the formula, The encounter angular frequency; The wave angular frequency; The speed of the ship; Relative wave direction; Wave number; Wherein, the wave number The expression is:

[0009] In the formula, It is the acceleration due to gravity; The wave angular frequency The expression is:

[0010] In the formula, Waves represent cycles; The expression for calculating and obtaining the natural roll frequency is:

[0011] In the formula, The natural roll frequency; It is the natural rolling cycle, and ,in, This refers to the ship type coefficient determined based on the ship type. The width of the boat.

[0012] Furthermore, the simplified frequency domain roll response model is expressed as follows:

[0013] In the formula, For the roll response amplitude operator; It is the absolute value symbol; Amplitude factor; The damping ratio; It is the frequency ratio, and .

[0014] Furthermore, the expression for calculating the standard deviation of the roll response is:

[0015] In the formula, The standard deviation of the roll response; This is an estimate of the standard deviation of sea surface undulation determined based on significant wave height, and ,in, For significant wave height; The expression for calculating the short-term probability of exceeding the roll angle limit is:

[0016] In the formula, This represents the probability of short-term overshooting of the roll angle. This refers to the roll angle; Set a threshold for the roll angle; Let be the probability.

[0017] Furthermore, the expression for the yaw amplification factor is:

[0018] In the formula, The parameter is the roll amplification factor; This is the adjustment coefficient; It is a frequency approximation function; The parameter is the roll frequency ratio; The sea state intensity function; For significant wave height; For wave direction function; The expression for the frequency proximity function is as follows:

[0019] In the formula, These are frequency bandwidth control parameters; The expression for the roll frequency ratio parameter is: ; The expression for the sea state intensity function is:

[0020] In the formula, The effective wave height threshold; The expression for the wave direction function is:

[0021] In the formula, Relative wave direction; This is a directional sensitivity index.

[0022] Furthermore, the expression for the normalized sway risk index is:

[0023] In the formula, This is a normalized oscillation risk index.

[0024] Furthermore, the normalized roll risk index is mapped to a preset latitude and longitude grid, and the point risk index within each grid is statistically analyzed to obtain the average roll risk value of the grid, expressed as:

[0025] In the formula, For the first Average roll risk value for each grid; For the first The number of paired samples within each grid; For the first in the grid Point risk index for each sample point; The number of samples; This is the index of the sample.

[0026] Beneficial effects: This invention provides a ship roll risk assessment method based on the fusion of AIS and ERA5 data. By constructing a simplified frequency domain roll response model through encounter angular frequency and natural roll angular frequency, it can quickly solve the roll response standard deviation and short-term over-limit probability. The simplified frequency domain roll response model significantly reduces the computational cost while retaining the core physical mechanism, and is particularly suitable for batch risk calculation of large-scale ship trajectory data covering the globe. It has good engineering practicality and real-time performance. By defining frequency proximity function, sea state intensity function and wave direction function, and constructing parametric roll amplification factor accordingly, it can effectively capture the conditions for the occurrence of this severe resonance phenomenon of parametric roll, and specifically enhance the standard deviation of conventional roll response. It overcomes the inherent defect of traditional frequency domain linear models that seriously underestimate roll amplitude under parametric roll conditions, making the risk assessment results closer to the extreme dangerous scenarios in actual navigation. By integrating the short-term over-limit probability with the roll amplification factor, a normalized roll risk index is formed. This index combines the probability of roll occurrence with the hazard amplification effect, and can quantitatively compare and rank the roll hazard levels in different sea areas and at different times under a unified dimension, providing a clear and intuitive decision-making basis for ship route planning and maritime supervision. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart illustrating the ship roll risk assessment method of the present invention. Figure 2 This is an overall flowchart of an embodiment of the present invention; Figure 3 This illustrates the relationship between the effective wave height and the average roll risk index for different ship types in this embodiment of the invention. Figure 4 This illustrates the relationship between the ratio of wave encounter period to natural roll period and the average roll risk index for different ship types in this embodiment of the invention. Figure 5 This is the distribution of the average roll risk index for different ship types under resonance and non-resonance conditions in the embodiments of the present invention; Figure 6 This invention illustrates the relationship between speed and average roll risk index for different ship types under conditions of transverse to oblique transverse waves, near resonance, and effective wave height greater than or equal to 3 meters. Figure 7 This invention illustrates the relationship between speed and risk for three ship types in a narrower effective wave height segment. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] This embodiment provides a method for assessing ship roll risk based on the fusion of AIS and ERA5 data, such as Figure 1 and Figure 2 As shown, it includes: S1. Collect global AIS trajectory data and ERA5 wave reanalysis data, and form ship-wave pairing samples by combining time nearest neighbor and spatial nearest neighbor methods; then obtain input parameters through ship-wave pairing samples. Specifically, the processing structure of step S1 includes: data acquisition, spatiotemporal registration, quality control, and input parameter generation; wherein, data acquisition is used to establish the sources of ship operating status data and wave environment data respectively; spatiotemporal registration is used to map the ship status under the same spatiotemporal conditions to the sea state; quality control is used to remove abnormal and missing samples that do not meet the requirements of subsequent modeling; input parameter generation is used to convert the registered ship-wave samples into unified variables required for subsequent roll response calculation; according to step S1, a unified ship-wave database for large-sample analysis on a global scale can be established, providing basic input for subsequent roll risk assessment.

[0031] Specifically, the AIS trajectory data includes ship identification information, timestamp, latitude and longitude position, speed, heading or track direction, ship width, and ship type information. The AIS trajectory data is used to characterize the ship's operating status during actual operation, including the ship's motion characteristics and ship type information at different times and locations, thereby providing ship-side input for subsequent ship-wave state registration and roll risk modeling. The ERA5 wave reanalysis data is a global ocean wave dataset provided by the European Centre for Medium-Range Weather Forecasts, which includes at least significant wave height, wave representative period, and wave dominant direction. The ERA5 data is used to characterize the wave environment of the sea area where the ship is located at the corresponding time, including sea state intensity, wave period characteristics, and wave propagation direction information, thereby providing environmental input for subsequent wave encounter angular frequency calculation, direction spectrum construction, and roll risk analysis.

[0032] Specifically, by combining temporal and spatial nearest neighbor methods, each AIS record is registered with ERA5 wave reanalysis data near the corresponding time and spatial location to form a ship-wave paired sample. After completing the temporal and spatial registration, the registration results are quality controlled to remove samples with abnormal positions, speeds, missing headings, missing ship types, and missing wave parameters. The combined temporal and spatial nearest neighbor method unifies the ship's operating status and the corresponding sea state status into the same paired record, thereby establishing a unified ship status-wave status dataset. This ship status-wave status dataset is used to support subsequent large-sample roll risk calculations and ensure the consistency of input samples in terms of time, space, and parameter integrity. The input parameters include ship speed, ship beam, ship type, significant wave height, wave representative period, wave dominant direction, ship heading, and relative wave direction. The relative wave direction is calculated from the difference between the ship heading and the wave dominant direction, converted to a range of 0° to 180°, to characterize the relative relationship between the ship heading and the wave direction under the current operating conditions. These input parameters serve as the basis for subsequent roll response modeling and risk calculation. Ship speed and relative wave direction are used for subsequent encounter angular frequency calculation; ship beam and ship type are used for natural roll period estimation; significant wave height, wave representative period, and wave dominant direction are used for wave environment characterization and subsequent direction spectrum construction. These input parameters can be understood as converting the registered original ship-wave record into a set of unified input variables that can be directly used by the subsequent roll response model.

[0033] S2. Calculate and obtain the encounter angle frequency and natural roll angle frequency based on the input parameters; construct a simplified frequency domain roll response model based on the encounter angle frequency and natural roll angle frequency; calculate the standard deviation of the roll response and the short-time over-limit probability of the roll angle based on the simplified frequency domain roll response model. Specifically, the processing structure of step S2 includes encounter angle frequency calculation, natural roll period estimation, simplified frequency domain roll response modeling, roll response statistics, and short-term roll angle exceedance probability calculation. Encounter angle frequency calculation characterizes the actual wave excitation frequency experienced by the ship under current speed and relative wave direction conditions. Natural roll period estimation characterizes the ship's intrinsic roll characteristics. Simplified frequency domain roll response modeling establishes an analytical relationship between wave input and roll response. Roll response statistics calculate the roll response variance or standard deviation using the directional wave spectrum and response amplitude operator. Short-term roll angle exceedance probability calculation transforms the response statistics into a probability representation of large roll under current operating conditions. Step S2 converts the ship-wave pairing sample obtained in step S1 into a basic roll risk probability, providing a basic input for subsequent parameter roll correction.

[0034] Preferably, the expression for calculating and obtaining the encounter angular frequency is:

[0035] In the formula, The encounter angular frequency; The wave angular frequency; The speed of the ship; Relative wave direction; Wave number; Wherein, the wave number The expression is:

[0036] In the formula, It is the acceleration due to gravity; The wave angular frequency The expression is:

[0037] In the formula, Waves represent cycles; The expression for calculating and obtaining the natural roll frequency is:

[0038] In the formula, The natural roll frequency; It is the natural rolling cycle, and ,in, This refers to the ship type coefficient determined based on the ship type. The width of the boat.

[0039] Specifically, based on the ship speed, relative wave direction, and wave period in the input parameters, the encounter angular frequency of the ship under the current operating conditions is calculated; the specific steps are as follows: S211. Calculate and obtain the wave angular frequency based on the wave representative period; S212. Under deep-water conditions, the wave number is calculated and obtained from the wave angular frequency and gravitational acceleration based on the wave dispersion relation. S213. Calculate and obtain the encounter angular frequency based on ship speed, relative wave direction, and wave number; The encounter angular frequency is used to characterize the actual wave excitation frequency experienced by the ship under the current speed and current relative wave direction conditions, and serves as the input for subsequent roll response modeling and resonance identification. S214. When it is necessary to analyze the relationship between the encounter period and the natural roll period, the encounter period can be calculated based on the encounter angular frequency, and the expression is:

[0040] In the formula, For encountering a cycle; The encounter period is a derived quantity obtained by converting the encounter angular frequency.

[0041] Specifically, the steps for calculating and obtaining the natural roll frequency are as follows: S221. When AIS data does not contain complete stability parameters and detailed hydrodynamic parameters, a simplified expression based on the ship's beam and ship type coefficient is used to estimate the natural roll period. Different ship types correspond to different ship type coefficients. In this embodiment, the ship type coefficient is set according to the ship type, such as 0.75 for passenger ships, 0.80 for cargo ships, and 0.85 for tankers. When the ship type cannot be effectively matched, the default value is 0.80. The natural roll period is used to characterize the ship's intrinsic roll characteristics and serves as a key parameter for resonance identification and roll response modeling. S222. Calculate the natural roll angular frequency based on the natural roll period; The natural roll angular frequency is used as the input for subsequent RAO modeling and parametric roll frequency ratio calculation. The calculation of the natural roll period enables an approximate expression of the ship's intrinsic roll characteristics in the absence of detailed stability parameters, allowing subsequent large-scale global analysis to retain its basic physical meaning.

[0042] Preferably, the simplified frequency domain roll response model is expressed as follows:

[0043] In the formula, For the roll response amplitude operator; It is the absolute value symbol; Amplitude factor; The damping ratio; It is the frequency ratio, and .

[0044] Specifically, the ship's rolling motion is considered as a single-degree-of-freedom linear oscillator, and a simplified frequency-domain rolling response model is established based on the encounter angular frequency and the natural rolling angular frequency. This simplified frequency-domain rolling response model uses the encounter angular frequency as the external wave excitation frequency, the natural rolling angular frequency to characterize the ship's intrinsic rolling characteristics, the damping ratio to characterize the rolling energy dissipation characteristics, and the amplitude factor to characterize the influence of wave direction on the rolling excitation intensity. The specific steps are as follows: S231. Define the frequency ratio to characterize the degree of closeness between the current excitation frequency and the ship's intrinsic frequency; S232. Based on the frequency ratio and damping ratio, construct a simplified frequency domain roll response model; The damping ratio, in subsequent application examples, refers to the damping ratio of passenger ships. The value is 0.08, representing the damping ratio of the cargo ship. The value is 0.05, representing the tanker damping ratio. The value is set to 0.04, and 0.05 for other ship types; The amplitude factor is used to characterize the sensitivity of the basic roll response to changes in wave direction. It belongs to the basic roll response direction effect and reflects the influence of different relative wave directions on the roll response amplitude under general wave excitation conditions, rather than being used specifically to characterize the prone direction of parametric roll. The structure of the simplified frequency domain roll response model can be understood as follows: the encounter angular frequency is used as the excitation input, the natural roll angular frequency and damping ratio are used as ship dynamics parameters, and the amplitude factor is used as the direction correction term to jointly determine the magnitude of the roll response amplitude operator.

[0045] Preferably, the expression for calculating the standard deviation of the roll response is:

[0046] In the formula, The standard deviation of the roll response; This is an estimate of the standard deviation of sea surface undulation determined based on significant wave height, and ,in, For significant wave height; Wherein, the standard deviation of sea surface undulation can be expressed as The effective wave height can be expressed as ,in, It is the zeroth order spectral moment; The expression for calculating the short-term probability of exceeding the roll angle limit is:

[0047] In the formula, This represents the probability of short-term overshooting of the roll angle. This refers to the roll angle; Set a threshold for the roll angle; Let be the probability.

[0048] Specifically, the process for calculating the standard deviation of the roll response and the short-term probability of exceeding the roll angle limit is as follows: S241. Based on the obtained effective wave height, wave representative period and wave dominant direction, construct the directional wave spectrum under the current working condition using a preset directional spectrum model to characterize the distribution of wave energy in the frequency domain and directional domain under the current sea state. In this embodiment, the preset directional spectrum model may optionally be the JONSWAP spectrum model, the Pierson-Moskowitz (PM) spectrum model, or the Ochi-Hubble spectrum model; the JONSWAP spectrum model, the Pierson-Moskowitz (PM) spectrum model, and the Ochi-Hubble spectrum model are existing technologies and will not be described in detail here. S242. Based on frequency domain seakeeping theory, the directional wave spectrum is combined with the roll response amplitude operator to calculate the roll response variance, which is used to characterize the ship's roll response amplitude level under the current short-term sea state. The expression is:

[0049] In the formula, The standard deviation of the roll response; Directional spectrum; The dominant direction of the wave; To The differential; To The differential; S243. In large-scale global computing scenarios, in order to reduce the cost of point-by-point double integration calculation, a narrowband proxy approximation method is adopted based on the roll response variance to obtain the roll response standard deviation. The narrowband proxy approximation method is an engineering simplification algorithm used in ship seakeeping engineering to quickly estimate the standard deviation of the roll response. Its core idea is to approximate the complex directional wave spectrum double integral as the product of the spectral density at a single point of resonance frequency and the roll response amplitude operator, thereby significantly reducing the amount of computation. The calculation of the narrowband proxy approximation method is existing technology and will not be described in detail here. S244. Under the condition that the roll response satisfies the narrow-band Gaussian process assumption, the extreme value of the roll angle is approximated as following the Rayleigh distribution, and the short-term over-limit probability of the roll angle exceeding the preset threshold is calculated accordingly, that is, the short-term over-limit probability of the roll angle. The short-term roll angle exceedance probability is used to characterize the likelihood of a large roll occurring under the current sea state and current operating conditions; the higher the short-term roll angle exceedance probability, the higher the roll risk.

[0050] S3. Define the frequency proximity function, sea state intensity function, and wave direction function, and construct the parametric roll amplification factor; Preferably, the expression for the roll amplification factor is:

[0051] In the formula, The parameter is the roll amplification factor; This is the adjustment coefficient; It is a frequency approximation function; The parameter is the roll frequency ratio; The sea state intensity function; For significant wave height; For wave direction function; To avoid excessive strength that could mask the fundamental characterization of short-term overshoot probability of the roll angle, the overall amplification range of the parameter roll amplification factor is controlled, so that the parameter roll amplification factor is usually between 1 and 1.5. Based on the expression of the parametric roll amplification factor, the frequency proximity function is used to quantify the degree of proximity between the current operating condition and the parametric roll dangerous frequency condition; the closer the current encountered angular frequency is to the parametric roll 2:1 condition, the larger the value of the frequency proximity function; when it deviates from this condition, its value decreases. The width of the frequency proximity function is determined by the frequency bandwidth control parameter, which is used to limit the range of the critical frequency band of parameter roll and to characterize the degree of proximity of the current operating condition to the parameter roll 2:1 condition. The expression is:

[0052] In the formula, This is a frequency bandwidth control parameter used to control the width of the dangerous frequency band during pitching. In subsequent application examples, the frequency bandwidth control parameter... All values ​​are 0.2; The expression for the roll frequency ratio parameter is: ; The sea state intensity function is constructed based on the relationship between significant wave height and a reference significant wave height threshold. It characterizes whether the current sea state possesses the wave energy level to trigger severe rolling. In subsequent application examples, the significant wave height threshold is consistently set to 3.0m. When the significant wave height is low, the sea state intensity function takes a smaller value; as the significant wave height increases, its value gradually increases; when the significant wave height reaches or exceeds the reference significant wave height threshold, i.e., ... At that time, the sea state intensity function is set to 1; The expression for the sea state intensity function is:

[0053] In the formula, The effective wave height threshold; The wave direction function is constructed based on the relative wave direction and is used to characterize the influence of different wave directions on the susceptibility of parametric roll. When the wave is within the range of the up-wave, down-wave, or near-up-wave direction, the wave direction function takes a larger value. When the wave direction deviates from the above-mentioned range, the value of the wave direction function decreases. The wave direction function is only used for direction discrimination in the additional correction of parametric roll and is not directly equivalent to the distribution of high-risk directions in the final comprehensive roll risk under full-sample statistics. The expression for the wave direction function is:

[0054] In the formula, Relative wave direction; The directional sensitivity index is used to control the sensitivity of the wave direction function to the sway susceptibility of the parameter. In subsequent application examples, a value of 2 is preferred.

[0055] Specifically, the parameter roll amplification factor is used to correct the short-term probability of roll angle exceeding the limit, thereby improving the ability to identify dangerous parameter roll conditions. The structure of the parameter roll amplification factor is composed of an adjustment coefficient, a frequency proximity function, a sea state intensity function, and a wave direction function. The adjustment coefficient controls the overall amplitude of the additional amplification, and in the current embodiment, it is preferably set to 0.5. The frequency proximity function characterizes the degree to which the current operating condition approximates the parameter roll 2:1 condition. The sea state intensity function characterizes whether the current sea state has reached the wave energy level that triggers severe roll. The wave direction function characterizes whether the current relative wave direction is within the range of directions more likely to induce parameter roll. The wave direction function is used to characterize the directional susceptibility in the additional correction of parametric roll, which belongs to the directional effect of parametric roll correction and does not overlap with the amplitude factor; the wave direction function is used to identify the range of directions that are more likely to induce parametric roll when the frequency is close to the dangerous condition of parametric roll, and the amplitude factor is used to characterize the sensitivity of the basic roll response to changes in wave direction under general wave excitation. Therefore, steps S2 and S3 correspond to the basic roll mechanism and the parametric roll additional mechanism, respectively, and their combined effect forms the final comprehensive roll risk result; through the combination of the above parts, the parametric roll amplification factor is increased. It can simultaneously reflect the combined effects of frequency conditions, sea state conditions, and wave direction conditions on the hazard of parameter roll.

[0056] S4. The short-term overshoot probability of the roll angle is fused with the roll amplification factor to obtain the normalized roll risk index; the normalized roll risk index is mapped to a preset latitude and longitude grid to obtain the average roll risk value and sample size of the grid; based on the average roll risk value and sample size of the grid, a roll risk distribution map at the global or regional scale is obtained.

[0057] Specifically, based on the short-term overshoot probability of roll angle and the roll amplification factor, a normalized roll risk index is constructed, and the point risk results are mapped to a preset spatial grid for aggregation and statistics to obtain the grid average roll risk value and sample size, thereby forming the ship roll risk distribution results at the global or regional scale. The processing structure of step S4 includes: point risk fusion, risk normalization, spatial mapping, and grid statistics. Point risk fusion is used to uniformly express the basic roll risk and the additional effects of parametric roll. Risk normalization is used to ensure the comparability of risk values ​​between different samples. Spatial mapping is used to project the point risk results onto a unified latitude and longitude grid. Grid statistics are used to output the spatial average risk results and the corresponding number of samples. Through step S4, the risk results obtained based on a single ship-wave pairing sample can be further extended to spatial distribution results applicable to global or regional scale analysis.

[0058] Preferably, the short-term overshoot probability of the roll angle is fused with the roll amplification factor to construct a normalized roll risk index; the normalized roll risk index is used to uniformly quantify the roll risk of a single vessel under specific time, position, speed, wave direction and sea state conditions. The expression for the normalized sway risk index is:

[0059] In the formula, To normalize the sway risk index; By normalizing the index, the point risk index is limited to the range of 0 to 1, so as to facilitate a unified comparison and subsequent spatial statistical analysis of risk results under different ship types, sea states, speeds and wave directions. The normalized roll risk index inherits the basic characterization function of the short-term roll angle exceedance probability for roll risk under general sea conditions. At the same time, it enhances the ability to identify resonance-prone conditions through the parametric roll amplification factor, so that the point risk result can simultaneously reflect the basic roll level and the additional danger of parametric roll.

[0060] Preferably, the normalized roll risk index is mapped to a preset 1°×1° latitude and longitude grid, and the risk index of each point within the grid is statistically analyzed to obtain the average roll risk value of the grid, expressed as:

[0061] In the formula, For the first Average roll risk value for each grid; For the first The number of paired samples within each grid; For the first in the grid Point risk index for each sample point; The number of samples; For the index of the sample; The grid average roll risk value is used to characterize the overall exposure level of ship roll risk within the spatial cell, while the sample size is used to characterize the reliability of the grid statistical results. The number of samples is retained as a reliability indicator of the corresponding grid statistical results; when it is lower than a preset threshold, the corresponding grid can be removed, marked with low confidence, or not included in subsequent spatial output, so as to avoid low sample areas interfering with the spatial risk distribution results. Specifically, the grid average roll risk value and the corresponding number of samples are used to generate roll risk distribution maps, risk statistics tables, and conditional grouping analysis results at the global or regional scale. This enables the conversion from point risk results to spatial risk representation, which can not only identify the roll hazard level under individual sample conditions, but also identify the spatial distribution characteristics of high-risk sea areas, key routes, and high-exposure areas.

[0062] In this embodiment, global ship AIS trajectory data from 2022 and ERA5 wave reanalysis data from the same period are selected as input data, and steps S1 to S4 are executed. Specifically, step S1 involves acquiring ship operation data and wave environment data, and completing the spatiotemporal registration of AIS data and ERA5 data, as well as the extraction of ship parameters and sea state parameters. Step S2 involves calculating the encounter angle frequency, estimating the natural roll period, modeling the simplified frequency domain roll response, calculating the roll response variance, and calculating the short-term overshoot probability of the roll angle, thereby obtaining the basic roll risk probability. Step S3 involves constructing a parametric roll amplification factor to correct for the additional effects under parametric roll-prone conditions. Step S4 involves generating a normalized roll risk index and mapping the point risk index to a 1°×1° latitude and longitude grid for aggregation and statistics to obtain the global-scale ship roll risk distribution results. The results show that global ship roll risk exhibits significant spatial heterogeneity. High-risk areas are mainly distributed along major deep-sea shipping routes and storm paths, with the North Atlantic being the most prominent high-risk area, located approximately in the 30°N-60°N shipping belt between the east coast of North America and Northwest Europe. A continuous high-risk belt also forms along the great circle shipping route from East Asia to the west coast of North America in the North Pacific. High-risk areas in the Southern Hemisphere are mainly distributed in strips along routes near South Africa, Australia, New Zealand, and the Drake Passage. In contrast, the average roll risk in tropical and subtropical waters is generally lower. Analysis of global grid statistics shows that, assuming a minimum of 20 samples, the effective number of grids after merging all ship types is 33,919; the median average roll risk per grid is... The 95th percentile is The 99th percentile is The maximum value reached 0.446; exceeding The proportion of high-risk grids is approximately 1.20%; The above results indicate that the average roll risk is generally low in most sea areas around the world, but high risk values ​​are concentrated in a few deep-sea shipping lanes that are affected by storms and have dense traffic. The ship roll risk assessment method in this embodiment can continuously assess, spatially identify and quantitatively count the ship roll risk in different sea areas based on the 2022 global AIS and ERA5 paired samples, and can provide a basis for identifying high roll exposure shipping lanes around the world, screening key routes and subsequent ship type subdivision analysis.

[0063] In the second embodiment, passenger ships, cargo ships and tankers from the 2022 global ship AIS trajectory data and ERA5 wave reanalysis data from the same period are selected as input data. Steps S1 to S4 are executed to form independent point risk indices and grid average roll risk results for different ship types. Specifically, step S1 constructs AIS-ERA5 paired samples for different ship types and extracts corresponding ship parameters and sea state parameters; step S2 completes the calculation of encounter angle frequency, estimation of natural roll period, roll response modeling, calculation of roll response variance, and calculation of short-term roll angle exceedance probability for each ship type, thereby obtaining the basic roll risk probability for each ship type; step S3 constructs the corresponding parametric roll amplification factor to correct the additional effects under parametric roll-prone conditions; and step S4 generates the normalized roll risk index for each ship type and performs grid aggregation statistics on the risk results, thereby forming a roll risk distribution map and statistical indicators differentiated by ship type; The comparative results show that there are significant differences in the roll risk distribution among different ship types. Passenger ships have the sparsest spatial coverage and the lowest overall median risk. High risks are mainly concentrated on a few transoceanic routes and high-latitude sections, with relatively high average roll risks observed near the Drake Passage at the southern tip of South America, on transoceanic passenger / cruise routes in the North Atlantic, and on some sections of the North Sea-Norwegian Sea. Apart from the aforementioned localized areas, the risk distribution of passenger ships is generally scattered and localized. Cargo ships have the widest spatial coverage, distributed along all major global trade routes. High risks are mainly distributed in a continuous band along the mid-to-high latitude shipping lanes in the North Atlantic, the North Pacific routes, and some mid-to-high latitude sections in the Southern Hemisphere. Their risk distribution is more even, reflecting the extensive exposure of cargo ships on major global trade routes. Tankers exhibit a more obvious high-risk tail characteristic. Their high-risk areas, in addition to the mid-to-high latitude shipping lanes in the North Atlantic and the Norwegian Sea-Barents Sea entrance, form an approximately latitudinal high-risk band between the South Atlantic, the Southern Indian Ocean, and the South Pacific. Higher average risk values ​​are observed in some mid-to-high latitude sections and high-energy choke points in the Southern Hemisphere. A quantitative comparison of the grid statistics results for the three types of ships shows that, assuming a sample size of at least 20, the effective number of grids for passenger ships is 16,428, and the median average roll risk per grid is [missing data]. The 95th percentile is The 99th percentile is The maximum value is 0.430, exceeding... The proportion of high-risk grids is approximately 1.26%; the number of effective grids for cargo ships is 32,494, with a median of [missing data]. The 95th percentile is The 99th percentile is The maximum value is 0.469, and the proportion of high-risk grids is approximately 1.12%; the number of effective tanker grids is 29,794, with a median of [missing value]. The 95th percentile is The 99th percentile is The maximum value reached 0.824, and the proportion of high-risk grids was approximately 2.53%. The above results indicate that although the median risk of cargo ships is higher than that of tankers, tankers are significantly higher than passenger ships and cargo ships in terms of high quantile risk, maximum risk value, and proportion of high-risk grids, indicating that tankers are more concentrated and intensely exposed on a few high-risk routes. Comprehensive analysis shows that passenger ships exhibit a risk characteristic of "low overall risk, high local risk," meaning that the overall risk is relatively low, but localized high risks may occur on a few high-risk routes; cargo ships exhibit a risk characteristic of "wide coverage and even distribution," meaning that risks are widely distributed across major global trade routes; and oil tankers exhibit a risk characteristic of "heavy at the stern and prominent hotspots," meaning that high-risk values ​​are more concentrated in some mid-latitude sections of the Southern Hemisphere and key high-risk chokepoint areas. The ship roll risk assessment method in this embodiment can not only conduct differentiated roll risk assessments for different ship types, but also identify the spatial distribution characteristics of high-risk sections for each ship type, and provide a basis for differentiated management of ship types, setting differentiated early warning thresholds for ship types, and identifying key routes.

[0064] In the third embodiment, based on the ship-wave pairing samples and point risk index results obtained in steps S1 to S4, analysis is performed on three types of ships: passenger ships, cargo ships, and tankers, and the analysis is based on the significant wave height. The samples were grouped to analyze the variation of the average roll risk index with sea state intensity; Specifically, in step S1, AIS data and ERA5 wave reanalysis data are acquired and registered to form a ship-wave paired sample and extract the effective wave height. Sea state parameters are used; step S2 completes the calculation of encounter angle frequency, estimation of natural roll period, roll response modeling, calculation of roll response variance, and calculation of short-term roll angle exceedance probability to obtain the basic roll risk probability; step S3 constructs a parametric roll amplification factor to correct for additional effects under parametric roll-prone conditions; and step S4 generates the normalized roll risk index for each sample. and according to different The interval is used to statistically compare the point risk indices, such as Figure 3 As shown, the variable values ​​and grouping conditions are shown in Table 1. Table 1. Variable values ​​and grouping conditions for examples of the impact of sea state intensity on roll risk.

[0065] Statistical results show that for passenger ships, cargo ships, and oil tankers, the risk of rolling is generally low in low sea states, but increases with the significant wave height. The increase shows a clear non-linear growth trend, but the growth threshold and growth rate differ among different ship types; for passenger ships, when Below approximately 3 meters, the average roll risk is close to zero; in the range of approximately 3 to 6 meters, the risk begins to rise significantly; when Beyond 6m to 7m, the risk increases rapidly, reaching a high level in some groupings. For cargo ships, the risk gradually increases from about 2m, with a relatively stable increase under moderate sea states, until more obvious high-risk fluctuations appear under higher wave heights. For oil tankers, the overall risk is low and changes slowly over a wide range of wave heights, but it increases significantly under higher wave heights and adverse dynamic conditions, demonstrating a strong threshold effect. Furthermore, from Figure 3 According to the corresponding statistical patterns, the significant wave height It has a significant effect on controlling roll risk, but its effect varies significantly depending on the type of ship. Passenger ships are more sensitive to changes in medium and high sea states, cargo ships show a continuous increasing trend over a wider range of wave heights, while oil tankers are relatively stable in most medium and low sea states, but may experience a sudden increase in risk under high wave height conditions. This indicates that wave height alone cannot fully reflect roll risk, and further identification is needed by combining ship type characteristics with subsequent factors such as period ratio, wave direction, and speed. Therefore, the effective wave height can be determined. As a preliminary screening indicator for sway risk; preferably, it can be As a low alert condition, As a medium alert condition, As a high alert condition; after entering moderate or above sea state, a more detailed judgment is made by combining the ratio of wave encounter period to natural roll period and relative wave direction; the ship roll risk assessment method in this embodiment can extract sea state threshold information that can be used for operational prediction from large sample statistical results, and can provide a basis for graded early warning during maritime navigation.

[0066] In the fourth embodiment, based on the point risk index results obtained in steps S1 to S4, analysis is performed on three types of ships: passenger ships, cargo ships, and tankers, analyzing the ratio of wave encounter period to natural roll period. The impact on roll risk is assessed through the following steps: Step S1 involves acquiring and registering AIS data with ERA5 wave reanalysis data to form ship-wave paired samples and extracting ship parameters and sea state parameters; Step S2 involves calculating the encounter angle frequency, estimating the natural roll period, modeling the roll response, calculating the roll response variance, and calculating the short-term roll angle exceedance probability to obtain the basic roll risk probability; Step S3 involves constructing a parametric roll amplification factor to correct for additional effects under parametric roll-prone conditions; and Step S4 involves generating the normalized roll risk index for each sample. and based on The values ​​of are used to group and statistically analyze the samples, comparing the average sway risk level under different period intervals, such as... Figure 4 As shown in Table 2, the variable values ​​and grouping conditions are as follows; Table 2. Variable values ​​and grouping conditions for examples of period ratio resonance sensitivity analysis

[0067] Analysis results show that the average rolling risk is within the range for passenger ships, cargo ships, and oil tankers. The peak value is reached within a narrow band close to 1, indicating that the roll risk is significantly amplified when waves encounter a period close to the ship's natural roll period; however, the risk decreases rapidly after deviating from this range. For passenger ships, the average roll risk is within... When the value is less than approximately 0.8, it approaches zero, and a significant peak appears near approximately 1. A secondary peak also exists in the range of approximately 1.4–1.6, reflecting the coexistence of synchronous roll and parametric roll amplification phenomena. For cargo ships, the risk curve exhibits a single and relatively obvious main peak in the range slightly above 1, with the peak risk being about an order of magnitude higher than the level far from the resonance zone. For oil tankers, the risk peak is sharper and more concentrated, with the peak essentially focused on… Nearby; furthermore, it can The resonance warning zone is set as the operating condition of the ship. When the ship's operating condition enters this zone, the rolling risk may increase significantly even if the effective wave height is at a moderate level. The ship rolling risk assessment method in this embodiment can identify high-risk operating conditions dominated by resonance conditions and can provide a basis for subsequent detailed risk judgment based on relative wave direction and speed.

[0068] In the fifth embodiment, based on the sample risk results obtained from steps S1 to S4, analysis is conducted on three types of vessels: passenger ships, cargo ships, and tankers. The relative wave direction is further examined under resonance and non-resonance conditions. The impact on roll risk varies; specifically, AIS data and ERA5 wave reanalysis data are acquired and registered through step S1 to form a ship-wave paired sample and extract the relative wave direction. Significant wave height The process involves considering wave period, ship heading, and ship parameters. Step S2 calculates the encounter angle frequency, estimates the natural roll period, models the roll response, calculates the roll response variance, and calculates the short-term roll angle exceedance probability, thus obtaining the basic roll risk probability. Step S3 constructs a parametric roll amplification factor to correct for additional effects under parametric roll-prone conditions. Step S4 generates the normalized roll risk index R for each sample. Based on this, the ratio of wave encounter period to natural roll period is used to determine the risk level. Grouping the samples will satisfy The samples are defined as resonance condition samples, and the remaining samples are defined as non-resonance condition samples, and different relative wave directions are considered separately. Statistical analysis was performed on the average sway risk index within the interval, and the results are as follows: Figure 5 As shown, the variable values ​​and grouping conditions are shown in Table 3. Table 3. Variable values ​​and grouping conditions for examples of the impact of relative wave direction on roll risk under resonance and non-resonance conditions.

[0069] Overall, roll risk is lower near the head and tail of the waves, but increases significantly near the cross and oblique cross waves. It should be noted that the relative wave direction-risk distribution here reflects the combined statistical result of the basic roll response direction effect in step S2 and the parametric roll additional correction in step S3, and is not equivalent to the individual value trend of the wave direction function in step S3. The high-risk range for different ship types is roughly located in... Within the scope, including cargo ships and oil tankers Nearby areas reach a higher risk level, with passenger ships exhibiting a relatively wider high-risk range; furthermore, under resonance conditions, the relative wave direction sensitivity of the three types of vessels is significantly enhanced; passenger ships in A significant risk peak has emerged nearby; cargo ships are... Nearby risks escalated rapidly and reached their peak; tankers were... The interval exhibits a higher peak level; in contrast, under non-resonance conditions, the roll risk curves for all ship types are generally flatter, with a significantly reduced amplitude of changes in relative wave direction, indicating that when the wave encounter period is far from the ship's natural roll period, simply changing the wave direction has a relatively limited impact on roll risk; the ship roll risk assessment method in this embodiment can not only identify the distribution pattern of roll risk under different relative wave direction conditions, but also reveal the amplification effect of resonance conditions on wave direction sensitivity, as well as the combined effect of the basic roll response direction effect and the parametric roll additional correction in the process of comprehensive risk formation; in actual operation, when the ship is in When the resonance warning range is reached, priority should be given to avoiding it. The course should be adjusted from a transverse to a diagonal transverse course, and combined with the relationship between the wave encounter cycle and the natural roll cycle, the relative wave direction and sea conditions, to a lower-risk course to reduce roll risk. Under non-resonance conditions, the course adjustment has a relatively weak effect on improving roll risk, and the focus of operations can be shifted to other sea condition risk control.

[0070] In the sixth embodiment, based on the sample risk results obtained from steps S1 to S4, the analysis is conducted on passenger ships, cargo ships, and tankers respectively, and the speed under different sea state conditions is further examined. Impact on roll risk; specifically, in step S1, ship AIS data and ERA5 wave reanalysis data are acquired and registered to form ship-wave paired samples and extract speed. Significant wave height Relative wave direction The process involves: 1) determining the wave period and ship parameters; 2) calculating the encounter angle frequency, estimating the natural roll period, modeling the roll response, calculating the roll response variance, and calculating the short-term overshoot probability of the roll angle, thus obtaining the basic roll risk probability; 3) constructing a parametric roll amplification factor to correct for additional effects under parametric roll-prone conditions; and 4) generating the normalized roll risk index for each sample. Based on this, samples with transverse to oblique transverse waves, close to resonance, and whose significant wave height meets the preset threshold conditions are preferentially selected. Statistical analysis is then performed on the average roll risk index under different speed ranges, such as... Figure 6 As shown; preferably, the sample satisfies , and It should be noted that the sample selection criteria reflect the combined high-risk operating conditions resulting from the combined effect of the basic roll response direction effect in step S2 and the additional parameter roll correction in step S3, rather than being determined solely by the wave direction function in step S3. Furthermore, the samples can be grouped according to different effective wave height ranges to compare the speed-risk relationship under different sea state intensities, such as... Figure 7 As shown in Table 4, the variable values ​​and filtering conditions are as follows; Table 4. Variable values ​​and selection criteria for examples of the impact of ship speed on roll risk under different sea states.

[0071] Statistical results show that in the transverse to oblique transverse wave, close to resonance and Under these conditions, passenger ships, cargo ships, and tankers all exhibit significant speed sensitivity, and the roll risk shows a non-monotonic relationship with speed; for passenger ships, the risk increases from low speeds, reaching a peak around 10 km / h. A relatively wide peak range appears within 15 knots, followed by a decrease at higher speeds; for cargo ships, the trend is similar to that of passenger ships, but the peaks are higher and more distinct, also concentrated around 10 knots. The range is 15 knots; for tankers, the overall risk level is the highest, with high risk even at lower speeds, and around 9 knots. The risk peaks around 12 knots and remains high over a wide speed range. This result indicates that there is a typical high-risk speed band under conditions of cross waves to diagonal cross waves and near resonance. It should be noted that the high-risk speed range mentioned here corresponds to the combined effect of the basic roll response directional effect and the parametric roll additional correction, and is not determined by a single directional correction term. like Figure 7 As shown, under the premise of keeping the conditions of transverse to oblique transverse waves and near resonance unchanged, the grouped analysis according to different effective wave height ranges shows that, at any given speed, the roll risk increases with the effective wave height. The systemic increase occurs as 2m < When the depth is less than 3 meters, the overall risk of the three types of vessels is relatively low; when the depth is less than 4 meters... At speeds <6m, the risk increases significantly, and clearly defined high-risk speed zones appear in all cases; when At speeds ≥6m, the risk remains high over a wider range of speeds, and the high-risk speed band expands, weakening the effect of simply adjusting speed to reduce risk. This result indicates that the stronger the sea state, the more limited the ability of speed to adjust risk. The ship roll risk assessment method in this embodiment can not only identify the changing patterns of roll risk under different speed conditions, but also reveal the modulating effect of sea state intensity on the speed-risk relationship; in actual operation, when the ship is in cross waves to oblique cross waves, close to resonance and In working conditions ≥3m, priority should be given to avoiding areas around 10m. The high-risk speed zone of 15 knots can be used to change the encounter situation by adjusting the speed; and when When the speed is ≥6m, speed adjustment alone is usually insufficient to significantly reduce the risk of rolling. Risk avoidance should be further carried out by combining measures such as course adjustment, detouring, storm avoidance, or fleet-level scheduling. It should be noted that the speed adjustment recommendations are applicable to the comprehensive high-risk operating condition identification results corresponding to this example, and together with the aforementioned relative wave direction and period ratio conditions, they constitute the basis for operational decision-making.

[0072] The present invention has the following beneficial effects: This invention presents a ship roll risk assessment method based on the fusion of AIS and ERA5 data. By constructing a simplified frequency domain roll response model using the encounter angular frequency and the natural roll angular frequency, it can quickly solve the roll response standard deviation and short-term over-limit probability. The simplified frequency domain roll response model significantly reduces the computational cost while retaining the core physical mechanism, and is particularly suitable for batch risk calculation of large-scale ship trajectory data covering the globe, with good engineering practicality and real-time performance. By defining frequency proximity function, sea state intensity function and wave direction function, and constructing parametric roll amplification factor accordingly, it can effectively capture the conditions for the occurrence of this severe resonance phenomenon of parametric roll, and specifically enhance the standard deviation of conventional roll response. It overcomes the inherent defect of traditional frequency domain linear models that seriously underestimate roll amplitude under parametric roll conditions, making the risk assessment results closer to the extreme dangerous scenarios in actual navigation. By integrating the short-term over-limit probability with the roll amplification factor, a normalized roll risk index is formed. This index combines the probability of roll occurrence with the hazard amplification effect, and can quantitatively compare and rank the roll hazard levels in different sea areas and at different times under a unified dimension, providing a clear and intuitive decision-making basis for ship route planning and maritime supervision.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for assessing ship roll risk based on the fusion of AIS and ERA5 data, characterized in that, include: S1. Collect global AIS trajectory data and ERA5 wave reanalysis data, and form ship-wave pairing samples by combining time nearest neighbor and spatial nearest neighbor methods; Input parameters are obtained through ship-wave pairing samples; S2. Calculate and obtain the encounter angle frequency and natural roll angle frequency based on the input parameters; construct a simplified frequency domain roll response model based on the encounter angle frequency and natural roll angle frequency; calculate the standard deviation of the roll response based on the simplified frequency domain roll response model, and then obtain the short-term over-limit probability of the roll angle. S3. Define the frequency proximity function, sea state intensity function, and wave direction function, and construct the parametric roll amplification factor; S4. The short-term overshoot probability of the roll angle is fused with the roll amplification factor to obtain the normalized roll risk index. The normalized roll risk index is mapped to a preset latitude and longitude grid to obtain the average roll risk value and sample size of the grid. Based on the average roll risk value and sample size of the grid, a roll risk distribution map at the global or regional scale is obtained.

2. The method for assessing ship roll risk based on AIS and ERA5 data fusion as described in claim 1, characterized in that, The expression for calculating and obtaining the encounter angular frequency is: In the formula, The encounter angular frequency; The wave angular frequency; The speed of the ship; Relative wave direction; Wave number; Wherein, the wave number The expression is: In the formula, It is the acceleration due to gravity; The wave angular frequency The expression is: In the formula, Waves represent cycles; The expression for calculating and obtaining the natural roll frequency is: In the formula, The natural roll frequency; It is the natural rolling cycle, and ,in, This refers to the ship type coefficient determined based on the ship type. The width of the boat.

3. The method for assessing ship roll risk based on AIS and ERA5 data fusion according to claim 2, characterized in that, The simplified frequency domain roll response model is expressed as follows: In the formula, For the roll response amplitude operator; It is the absolute value symbol; Amplitude factor; The damping ratio; It is the frequency ratio, and .

4. The method for assessing ship roll risk based on AIS and ERA5 data fusion as described in claim 3, characterized in that, The expression for calculating the standard deviation of the roll response is: In the formula, The standard deviation of the roll response; This is an estimate of the standard deviation of sea surface undulation determined based on significant wave height, and ,in, For significant wave height; The expression for calculating the short-term probability of exceeding the roll angle limit is: In the formula, This represents the probability of short-term overshooting of the roll angle. This refers to the roll angle; Set a threshold for the roll angle; Let be the probability.

5. The ship roll risk assessment method based on AIS and ERA5 data fusion according to claim 4, characterized in that, The expression for the roll amplification factor is: In the formula, The parameter is the roll amplification factor; This is the adjustment coefficient; It is a frequency approximation function; The parameter is the roll frequency ratio; The sea state intensity function; For significant wave height; For wave direction function; The expression for the frequency proximity function is as follows: In the formula, These are frequency bandwidth control parameters; The expression for the roll frequency ratio parameter is: ； The expression for the sea state intensity function is: In the formula, The effective wave height threshold; The expression for the wave direction function is: In the formula, Relative wave direction; This is a directional sensitivity index.

6. The method for assessing ship roll risk based on AIS and ERA5 data fusion according to claim 5, characterized in that, The expression for the normalized sway risk index is: In the formula, This is a normalized oscillation risk index.

7. The method for assessing ship roll risk based on AIS and ERA5 data fusion as described in claim 6, characterized in that, The normalized roll risk index is mapped to a preset latitude and longitude grid. The risk index of each point in the grid is statistically analyzed to obtain the average roll risk value of the grid, expressed as: In the formula, For the first Average roll risk value for each grid; For the first The number of paired samples within each grid; For the first in the grid Point risk index for each sample point; The number of samples; This is the index of the sample.

Images