A container ship wind load fast prediction method based on real-time image generation

Abstract

The present application relates to the technical field of ship, especially to a container ship wind load fast prediction method based on image real-time generation. It comprises the following steps: S1, collecting ship information, collecting wind speed and wind direction information near the ship; S2, taking a photo of the ship, then calling a generative three-dimensional large model to generate a three-dimensional model of the ship based on the taken photo of the ship; S3, adjusting the three-dimensional model of the ship generated in step S2 according to the ship information in step S1, then calculating the wind projection area of the ship according to the adjusted three-dimensional model of the ship and the wind direction information collected in step S1; S4, calculating the wind load of the ship according to the wind projection area obtained in step S3 and the wind speed and wind direction information collected in step S1; S5, outputting the wind load calculated in step S4. This method can solve the problems of poor timeliness, large calculation resource consumption and difficulty in meeting the real-time prediction demand of the pilot scene in the prior art.

Description

Technical Field

[0001] This invention relates to the field of shipbuilding technology, and in particular to a method for rapid forecasting of wind loads on container ships based on real-time image generation. Background Technology

[0002] Currently, ship pilotage technology is at a critical stage of transformation from traditional "experience-driven" to "data and rule-based intelligent integration." Accurate and efficient analysis of ship environmental loads has become the core technical support for intelligent pilotage operations. Container ships, with their massive superstructures and concentrated wind-exposed surfaces, are significantly affected by wind loads in port navigation environments. Simultaneously, these ships have large drafts when fully loaded, making them highly susceptible to the cumulative effects of shallow water during low-speed navigation and berthing / unberthing operations in port. This results in complex and variable ship maneuvering conditions, placing stringent technical requirements on real-time, accurate, and rapid analysis and forecasting of wind loads.

[0003] Currently, ship wind load analysis mainly relies on traditional techniques such as computational fluid dynamics and finite element simulation modeling. However, these methods have significant technical limitations and shortcomings in practical pilotage engineering applications. From an operational timeliness perspective, traditional simulation methods require detailed ship geometric modeling and high-density mesh generation for numerical iterative calculations. This process is cumbersome and time-consuming, failing to meet the real-time response requirements of pilotage operations and making it difficult to adapt to dynamically changing port navigation scenarios. From an engineering practicality perspective, these methods are highly dependent on the computing power of the computing hardware platform. As the number of ship model meshes increases, the computation time increases exponentially, resulting in high computational costs and significant implementation difficulties. This makes large-scale deployment in port pilotage operations impossible, hindering the dynamic monitoring and real-time prediction of ship wind load conditions.

[0004] In summary, the shortcomings in the timeliness and practicality of existing wind load calculation technologies severely restrict the safe decision-making and precise operation of intelligent pilotage. Therefore, developing a new technical method adapted to container ship operation scenarios and capable of real-time and rapid wind load calculation has significant engineering application value and practical significance for addressing the shortcomings of pilotage technology, realizing real-time perception of ship wind load status, and improving the safety and intelligence level of port pilotage operations. Summary of the Invention

[0005] The technical solution to be solved by the present invention is to provide a method for rapid forecasting of wind loads on container ships based on real-time image generation, so as to solve the problems of poor timeliness of ship wind load analysis, large consumption of computing resources, and difficulty in meeting the real-time forecasting requirements of pilotage scenarios in the existing technology.

[0006] The technical solution adopted in this invention is: a rapid forecasting method for wind loads on container ships based on real-time image generation, which includes the following steps: S1. Collect ship information, and simultaneously collect wind speed and direction information near the ship; S2. Take photos of the ship, and then call the generative 3D model to generate a 3D model of the ship based on the photos. S3. Adjust the ship 3D model generated in step S2 based on the ship information in step S1, and then calculate the windward projection area of ​​the ship based on the adjusted ship 3D model and the wind direction information collected in step S1. S4. Calculate the wind load on the ship based on the wind-received projected area obtained in step S3 and the wind speed and wind direction information collected in step S1. S5. Output the wind load calculated in step S4.

[0007] Preferably, all information collected in step S1 needs to be aggregated and timestamped through a unified edge gateway to form a spatiotemporally consistent input data stream.

[0008] Preferably, taking ship photos in step S2 includes taking multi-view photos of the ship, and after taking the photos, the photos need to be pre-processed, and then the three-dimensional model of the ship is generated based on the pre-processed multi-view ship photos by calling Tencent Hunyuan 3D model.

[0009] Preferably, step S3 specifically includes the following steps: S31. Align the coordinate system of the 3D model so that the longitudinal axis of the hull coincides with the X-axis and the deck surface is parallel to the XOY plane. S32. Adjust the scale of the 3D model based on the ship information collected in step S1; S33. The transverse wind-receiving projected area, longitudinal wind-receiving projected area, and total wind-receiving projected area of ​​the ship are obtained by using the triangular patch projection summation method.

[0010] Preferably, step S33 specifically includes the following steps: S331. Discretize the three-dimensional model to obtain a triangular mesh model composed of several triangular facets, and record the three-dimensional spatial coordinates of the three vertices of each triangular facet. S332. Construct a projection reference plane based on the wind direction collected in step S1; S333. Traverse all triangular faces within the triangular mesh model, perform vertex projection operation on each triangular face, and obtain the projected triangle of each triangular face; S334. Calculate the total area covered by all projected triangles on the projection reference plane as the total wind-receiving projected area. Then, take the projected area in the range of 0° to 180° as the longitudinal wind-receiving projected area and the projected area in the range of 90° to 270° as the lateral wind-receiving projected area.

[0011] As a preferred option, step S334 also requires using a depth test method to remove the obscured triangular facets when calculating the windward projected area.

[0012] Preferably, the wind load is calculated in step S4 using the following formula: Where: F represents wind force, C h This is the height factor of the wind-receiving components; P is the wind pressure. ,in V 0 The wind speed collected in step S1; S is the wind-receiving area. Where A1 is the longitudinal wind-receiving projected area and A2 is the lateral wind-receiving projected area. The angle between the wind direction collected in step S1 and the longitudinal axis of the ship. ; Thus, the total wind load can be obtained as follows: ; The longitudinal wind load is: ; Lateral wind load is: .

[0013] Compared with the prior art, the present invention has the following advantages using the above method: I. High real-time performance: By rapidly constructing a 3D model of a ship through real-time image generation technology, combined with a simplified wind load physical calculation model (single-step calculation less than 0.5 seconds), it achieves second-level rapid forecasting of wind load, which can meet the stringent timeliness requirements of pilotage scenarios and has an order-of-magnitude efficiency improvement compared to traditional CFD methods (tens of minutes or even hours).

[0014] 2. Low computational resource consumption: This invention does not rely on high-density grid partitioning and large-scale parallel computing, and has low requirements for the computing platform's computing power. It can run stably on ordinary industrial control computers or edge computing devices, making it easy to deploy on pilot ships or port areas.

[0015] III. Accuracy and Reliability: Employing a projection area calculation method based on a 3D model, the windward area of ​​the ship under any wind direction is accurately obtained, avoiding errors caused by traditional empirical estimations. Through reasonable physical simplification, the relative error between the calculated wind load values ​​and CFD simulation results is controlled within an acceptable engineering range, ensuring the accuracy of the forecast.

[0016] IV. High adaptability: For container ships of different sizes, the corresponding model can be quickly generated simply by image acquisition, without the need to pre-establish a fine finite element mesh, and can flexibly adapt to changes in ship shape and different wind conditions.

[0017] V. Simple and easy to implement: This invention does not involve complex flow field solutions and iterative convergence processes. The algorithm structure is clear, easy to program and implement in engineering, and has good promotional value. Attached Figure Description

[0018] Figure 1 This is a technical roadmap of the present invention.

[0019] Figure 2 This is a schematic diagram of a ship photograph taken in Example 1.

[0020] Figure 3 This is a schematic diagram of a 3D model of a ship generated from photographs of the ship in Example 1.

[0021] Figure 4 This is a diagram illustrating how a ship is exposed to wind. Detailed Implementation

[0022] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0023] Example 1: A rapid forecasting method for wind loads on container ships based on real-time image generation, the overall technical approach is as follows: Figure 1 As shown, this embodiment uses the berthing process of the "Xinyingkou" vessel from the Meishan Terminal of Ningbo-Zhoushan Port Area on March 21, 2026 as a case study, and includes the following steps: Step 1: Multi-source data acquisition and preprocessing High-precision weather stations are deployed at key locations within the port area to collect meteorological parameters such as wind speed and direction in real time, with a data sampling frequency of no less than 1 Hz to ensure effective capture of transient wind fields. Simultaneously, AIS data of target container ships is acquired through the Automatic Identification System (AIS), and this AIS data includes ship type parameters. The aforementioned multi-source data is aggregated and timestamped through a unified edge gateway to form a spatiotemporally consistent input data stream, which is then uploaded to the system.

[0024] Step 2: Reconstruction of the ship's 3D model based on real-time image generation The pilot departs from the pilotage base, boards the pilot boat, and proceeds to the designated boarding point outside the anchorage to begin pilotage operations. A photograph of the pilot vessel is taken before boarding; the photograph should look like this. Figure 2 As shown, the captured photos accurately reflect the ship's container capacity and load status. The images are uploaded to the system, which first preprocesses them using computer vision technology, including image denoising, enhancement, edge detection, and feature point extraction. Then, it calls the Tencent Hunyuan 3D Model API to generate a real-time 3D model of the ship based on a single image. Figure 3 As shown.

[0025] Step 3: Rapid Calculation of Ship's Windward Projected Area Align the model with the coordinate system so that the longitudinal axis of the hull coincides with the X-axis and the deck surface is parallel to the XOY plane. Based on the ship type parameter information collected in step one (including ship length and ship width; in this embodiment, the ship "Xinyingkou" has a length of 263 meters and a width of 32 meters), adjust the model scale and calculate the projected area of ​​the ship in the wind direction. The main method used is the triangular patch projection method, which includes the following steps. 1. Import the ship's 3D geometric model, discretize all the outer surfaces into a large number of non-overlapping triangular mesh units, and store the 3D coordinates of the three vertices of each triangular facet. 2. Based on the wind direction collected in step one, determine the projection direction and construct a projection plane perpendicular to the wind direction as the reference plane for area projection.

[0026] 3. For the coordinates of the three points of each triangular facet, perform a parallel projection transformation along the projection direction, and map the three vertices onto the projection plane to form a two-dimensional projected triangle; 4. Calculate the area of ​​the two-dimensional projected triangle; 5. Traverse all triangular faces, summarize all projected areas, and obtain the overall windward projected area of ​​the ship in that wind direction; 6. Traverse all wind directions and sum the projected areas of all wind directions from 0° to 180° as the longitudinal wind-receiving projected area A1; sum the projected areas of all wind directions from 90° to 270° as the lateral wind-receiving projected area A2. The calculated longitudinal windward projected area A1 of the "Xinyingkou" vessel is 947.4 m². 2 The lateral wind-receiving projected area A2 is 7685.4 m². 2 .

[0027] Step 4: Real-time calculation of wind load Step one allows for the real-time collection of the current average wind speed. V 0 and wind direction angle (relative to bow direction, such as) Figure 4(As shown), then based on the longitudinal wind-receiving projected area A1 and the lateral wind-receiving projected area A2 obtained in step three, the real-time wind load is calculated using the OCIMF (Oil Companies International Marine Forum) recommended formula, as follows: In the formula: F represents wind force, C h In this embodiment, C is the height factor of the wind-receiving component. h Take 1.2; P is the wind pressure. ; S represents the windward area, specifically the windward area of ​​the ship model in the direction of the incoming wind. X for: In the formula, A1 is the longitudinal wind-receiving projected area, and A2 is the lateral wind-receiving projected area. Wind direction angle ; This allows us to obtain the combined wind force F. X : ; The wind load (longitudinal wind load) on the model in the 0°-180° plane is as follows: ; The wind load (lateral wind load) on the model in the 90°-270° plane is as follows: .

[0028] Step 5: Outputting Forecast Results The wind load, longitudinal wind load, and lateral wind load calculated in step four are output in the form of structured data and can be transmitted in real time to the pilot terminal or pilotage assistance system through the data interface, so that the pilot can obtain the ship's wind load information in real time.

[0029] Example 2: The difference from Example 1 is that in Example 2, the pilot will take multiple photos of the ship from different perspectives, then preprocess the multiple photos, and then call the Tencent Hunyuan 3D Model API interface to generate a three-dimensional model of the ship in real time based on the multi-view photos.

[0030] Example 3: The difference from Embodiment 2 is that the multi-view photos of the ship in Embodiment 3 are taken by high-definition camera equipment deployed on pilot vessels or shore-based facilities in the port area. This results in photos with more diverse angles and higher shooting accuracy.

[0031] Example 4: The difference from Example 1 is that in Example 4, when calculating the windward projected area of ​​the ship, for areas with self-shading, a depth test method is used to remove the obscured triangular facets, retaining only the visible projected area; specifically, it includes the following steps: 1. For the determined wind direction collected in step one, establish a spatial depth coordinate system along the wind direction: Define the depth value as the distance from any point on the triangular facet to the projection starting point along the wind direction. The smaller the distance, the more the point faces the wind and is further forward. 2. Project all the vertices of the triangular facets along the wind direction onto a two-dimensional projection plane, and then divide the projection plane into pixels / small grids for sampling; 3. Sampling each patch individually: For the same sampling location on the projection plane, record the current minimum depth value; 4. If the depth of the current sampling point of the patch is greater than the existing minimum depth at that position, it means that the point is blocked by the preceding structure, and the projection of this part is invalid. The projection contribution of the patch is discarded. If the depth of the current sampling point of the patch is less than the existing minimum depth at that position, the patch is on the windward surface. Its projected area is retained, and the minimum depth at that position is updated. 5. After traversing all triangular faces, only the projected areas of all unremoved, windward faces are added together to obtain the unobstructed true windward projected area.

[0032] This method yields a more accurate windward projected area.

[0033] Example 5: The difference from Example 1 is that when calculating the windward projected area of ​​the ship, cylindrical components such as the lowering legs are calculated separately. Based on their geometric parameters, diameter D and height h exposed above the free surface, the windward projected area is obtained as D*h.

[0034] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

[0035] For those skilled in the art, various changes and modifications will undoubtedly be apparent after reading the above description. Therefore, the appended claims should be construed as covering all changes and modifications that encompass the true intent and scope of the invention. Any and all equivalent scope and content within the scope of the claims should be considered to remain within the intent and scope of the invention.

Claims

1. A method for rapid forecasting of wind loads on container ships based on real-time image generation, characterized in that, It includes the following steps: S1. Collect ship information, and simultaneously collect wind speed and direction information near the ship; S2. Take photos of the ship, and then call the generative 3D model to generate a 3D model of the ship based on the photos. S3. Adjust the ship 3D model generated in step S2 based on the ship information in step S1, and then calculate the windward projection area of ​​the ship based on the adjusted ship 3D model and the wind direction information collected in step S1. S4. Calculate the wind load on the ship based on the wind-received projected area obtained in step S3 and the wind speed and wind direction information collected in step S1. S5. Output the wind load calculated in step S4.

2. The method for rapid forecasting of wind loads on container ships based on real-time image generation according to claim 1, characterized in that: All information collected in step S1 needs to be aggregated and timestamped through a unified edge gateway to form a spatiotemporally consistent input data stream.

3. The method for rapid forecasting of wind loads on container ships based on real-time image generation according to claim 1, characterized in that: Step S2 involves taking photos of the ship from multiple perspectives. After taking the photos, they need to be preprocessed. Then, a three-dimensional model of the ship is generated based on the preprocessed multi-view photos by calling Tencent Hunyuan 3D Model.

4. The method for rapid forecasting of wind loads on container ships based on real-time image generation according to claim 3, characterized in that, Step S3 specifically includes the following steps: S31. Align the coordinate system of the 3D model so that the longitudinal axis of the hull coincides with the X-axis and the deck surface is parallel to the XOY plane. S32. Adjust the scale of the 3D model based on the ship information collected in step S1; S33. The transverse wind-receiving projected area, longitudinal wind-receiving projected area, and total wind-receiving projected area of ​​the ship are obtained by using the triangular patch projection summation method.

5. The method for rapid forecasting of wind loads on container ships based on real-time image generation according to claim 4, characterized in that: Step S33 specifically includes the following steps: S331. Discretize the three-dimensional model to obtain a triangular mesh model composed of several triangular facets, and record the three-dimensional spatial coordinates of the three vertices of each triangular facet. S332. Construct a projection reference plane based on the wind direction collected in step S1; S333. Traverse all triangular faces within the triangular mesh model, perform vertex projection operation on each triangular face, and obtain the projected triangle of each triangular face; S334. Calculate the total area covered by all projected triangles on the projection reference plane as the total wind-receiving projected area. Then, take the projected area in the range of 0° to 180° as the longitudinal wind-receiving projected area and the projected area in the range of 90° to 270° as the lateral wind-receiving projected area.

6. The method for rapid forecasting of wind loads on container ships based on real-time image generation according to claim 5, characterized in that: In step S334, when calculating the windward projected area, a depth test method is also needed to remove the obscured triangular facets.

7. The method for rapid forecasting of wind loads on container ships based on real-time image generation according to claim 4, characterized in that: The wind load is calculated in step S4 using the following formula: Where: F represents wind force, C h This is the height factor of the wind-receiving components; P is the wind pressure. ,in V 0 The wind speed collected in step S1; S is the wind-receiving area. Where A1 is the longitudinal wind-receiving projected area and A2 is the lateral wind-receiving projected area. The angle between the wind direction collected in step S1 and the longitudinal axis of the ship. ; Thus, the total wind load can be obtained as follows: ; The longitudinal wind load is: ; Lateral wind load is: .

Images