A ship cabin volume measurement method and system based on digital twinning

Abstract

The application discloses a ship cabin volume measurement method and system based on digital twinning, relates to the field of ship cabin volume measurement and monitoring, and is suitable for liquid cargo ship cabin volume measurement and monitoring. Point cloud data of a cabin main body, additional structures and deducted structures is obtained through scanning, a digital twinning ship cabin volume system is established, a ship cabin volume table and a heeling and rolling correction table are calculated by a cabin volume calculation model and are input into the digital twinning ship cabin volume system, cabin volume information can be corrected in combination with the ship cabin volume table and the heeling and rolling correction table, the cabin volume information is displayed in real time, automatic calibration of the cabin volume information is realized, and the efficiency and accuracy of ship cabin volume measurement are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of ship cabin volume measurement and monitoring, in particular to a ship cabin volume measurement method and system based on digital twinning, which is suitable for liquid cargo ship cabin volume measurement and monitoring. BACKGROUND

[0002] The LNG market is facing an explosive development trend, and the demand for LNG transportation ships, LNG fuel ships and other LNG equipment is becoming increasingly strong. Ship cabin capacity is one of the important indicators of LNG ships. When the ship is delivered and the cargo is delivered, the ship cabin capacity is often the most concerned economic indicator of the ship owner. This indicator is particularly important for in-service ships, and accurate and effective measurement of the ship cabin capacity is an important guarantee for the core interests of the ship owner. At the same time, through accurate measurement of this indicator, important indicators such as the ship's voyage and carrying capacity can be determined, and reliable and effective performance data can be provided for the subsequent use and deployment of the ship. In addition, real-time monitoring and early warning of the ship cabin capacity during navigation also greatly improves the safety and reliability of the navigation process. SUMMARY

[0003] The present application mainly aims at the problem of low efficiency and inaccuracy of ship cabin volume measurement in the prior art, which relies on manual experience and high labor intensity, causing difficulty in measuring the ship cabin volume.

[0004] The technical scheme adopted by the present application is as follows:

[0005] The present application provides a ship cabin volume measurement method based on digital twinning, comprising the following steps:

[0006] Using a laser scanner to scan and obtain three-dimensional point cloud data of the ship cabin main body, additional structure and deducted structure, obtaining real-time stress data collected by the in-cabin sensor, correcting the point cloud data according to the real-time stress data, and optimizing the point cloud data;

[0007] According to the point cloud data, surface fitting is performed, a digital twinning ship cabin volume system is established, and a ship cabin capacity table and a heeling and rolling correction table are calculated by a cabin volume calculation model and input into the digital twinning ship cabin volume system;

[0008] The digital twinning ship cabin volume system collects physical cabin temperature, liquid level information and inclination angle change information, corrects the cabin volume information in combination with the ship cabin capacity table and the heeling and rolling correction table, and displays the cabin volume information in real time.

[0009] Further, the real-time stress data collected by the in-cabin sensor is obtained, and the point cloud data is corrected according to the real-time stress data, which comprises:

[0010] An infrared thermal imager is set up to acquire temperature data, and stress sensors are deployed in a gridded topology to form a strain sensing network, through which stress data is acquired.

[0011] The stress sensor position is matched with the corresponding position in the point cloud data to align the stress data with the point cloud data. The displacement caused by stress is calculated, and thermal expansion compensation is performed on the point cloud data based on the temperature data to correct the point cloud data.

[0012] Furthermore, the step of matching the stress sensor position with the corresponding position in the point cloud data to align the stress data with the point cloud data, calculating the displacement caused by stress, and performing thermal expansion compensation on the point cloud data based on the temperature data to correct the point cloud data includes:

[0013] pass , The point cloud data is corrected;

[0014] in, The mechanical deformation at the position of the stress sensor. The stress value is the value at the location of the stress sensor, and E is the Poisson's ratio of the material. This is due to the thermal expansion deformation at the location of the stress sensor. The coefficient of thermal expansion is This represents the change in temperature relative to a reference value. and The vector displacement of the stress sensor position is synthesized, the discrete displacement vector is interpolated onto the point cloud, and the displacement is decomposed into a unit direction vector to correct the point cloud data.

[0015] Further optimization processing of the point cloud data includes:

[0016] The acquired 3D point cloud data of the main body, additional structure and subtracted structure of the ship cabin are processed by removing point cloud noise, downsampling and point cloud completion defect data to obtain the point cloud data of the main body, additional structure and subtracted structure of the ship cabin.

[0017] The point cloud data is subjected to point cloud filtering, point cloud data stitching, and point cloud registration processing. The point cloud filtering employs a bilateral filtering method, which includes:

[0018] Let a point in the point cloud be... , for A point within the neighborhood, Points after bilateral filtering The position is calculated as follows:

[0019] ;

[0020] ;

[0021] ;

[0022] wherein, is a spatial weight function, is a value weight function, is a spatial standard deviation parameter for controlling the size of the geometric neighborhood, is an attribute standard deviation parameter for controlling the feature similarity tolerance, is a feature function of a point, containing the normal vector, curvature or coordinate of the point.

[0023] Further, the acquisition of real-time stress data collected by the cabin sensor, and the correction of the point cloud data according to the real-time stress data further comprises:

[0024] Setting the liquid level meter matrix to obtain the liquid level surface data, and setting the pressure sensor to obtain the pressure data;

[0025] According to the liquid level surface data, the point cloud data is deformed and compensated, and according to the pressure data, the point cloud data is pressure compensated.

[0026] Further, the surface fitting according to the point cloud data, and the establishment of the digital twin ship cabin system further comprise:

[0027] Obtaining the three-dimensional point cloud data of the ship cabin main body, the additional structure and the deducted structure after optimization processing;

[0028] The three-dimensional point cloud data of the ship cabin main body and the additional structure are spliced to form spliced point cloud data;

[0029] According to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure, the point cloud data is fitted using a B-spline surface fitting algorithm, and a B-spline surface is fitted.

[0030] Further, the fitting of the point cloud data using the B-spline surface fitting algorithm according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure, and the fitting of the B-spline surface further comprise:

[0031] The spliced point cloud data and the three-dimensional point cloud data of the deducted structure are taken as input point clouds, the input point clouds are parameterized, node vectors are determined, control points are solved, surfaces are generated and error analysis is performed, and it is determined whether the accuracy requirement is met;

[0032] If yes, the B-spline surface is output, otherwise the input point clouds are re-parameterized.

[0033] Further, when fitting the point cloud data using the B-spline surface fitting algorithm according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure, a cubic B-spline curve is selected for fitting, and the parameter expression of the cubic B-spline curve is as follows:

[0034] ;

[0035] wherein, , , , are any four discrete points in a plane, P(t) is the coordinate of the fitted curve, and the component form of the above formula is as follows:

[0036] ;

[0037] wherein: , .

[0038] Further, when the B-spline surface is fitted, the B-spline surface is formed by multiple B-spline curves in the u and v directions, u and v are two independent parameters defined in the parameter domain, a control grid is formed by (m+1)×(n+1) control points, and the B-spline surface is fitted, and the equation of the B-spline surface is as follows:

[0039] ;

[0040] wherein is the control point set (i=0,1…m;j=0,1…n), m is the number of control points of the surface in the u direction minus 1, n is the number of control points of the surface in the v direction minus 1, and are B-spline surface basis functions.

[0041] Further, the ship's cabin capacity table and the correction table for rolling and listing calculated by the cabin capacity calculation model include:

[0042] The cabin capacity calculation model is set to calculate the cabin volume using the surface reconstruction method, and the B-spline surface fitted by the cabin capacity calculation model according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure is obtained by a fitting algorithm to obtain a main capacity table and a deducted structure capacity table with a minimum unit of 1 cm;

[0043] Under the condition of room temperature 20℃, a rolling ball boundary with a radius of α is used to extract non-convex surfaces, and a uniformly distributed grid is generated, the ship's cabin volume result is obtained by superposition calculation, and the ship's cabin capacity table is output, and the rolling and listing correction table is output synchronously.

[0044] Further, the digital twin ship cabin volume system collects physical cabin temperature, liquid level information, and inclination angle change information, and corrects the cabin volume information in combination with the ship cabin volume table and the heeling correction table, which includes:

[0045] determining whether the physical cabin temperature of the ship cabin is room temperature 20℃, and if the ship cabin is not at 20℃, correcting the ship cabin volume based on a temperature calibration formula, wherein the temperature calibration formula is:

[0046] ;

[0047] wherein: is the volume after temperature correction, in cubic meters; is the volume shown in the cabin volume table, in cubic meters; is the average temperature of the cabin wall, in degrees Celsius; is the linear expansion coefficient of the cabin material, in degrees Celsius per one, and the sample is taken =0.000012 / ℃.

[0048] Further, the method further includes:

[0049] collecting liquid level changes in the ship cabin, constructing a CNN-LSTM (Convolutional Neural Network-Long Short Term Memory Network) hybrid model to capture the spatio-temporal feature information of the liquid level changes, inputting the model into the LSTM model to build a time series prediction model, and outputting the leakage risk result from the last hidden layer of the LSTM model to prompt the leakage risk;

[0050] collecting ship cabin stress data history, establishing a 3D-CNN model, fusing spatio-temporal information through three-dimensional convolution, retaining the correlation between spatial topology and spatio-temporal information sequence, and outputting a fatigue accumulation index for fatigue damage prediction.

[0051] The application also provides a ship cabin volume measurement system based on digital twinning, which is used to implement the ship cabin volume measurement method based on digital twinning described above, and the system includes a multi-source data acquisition module, a core fitting algorithm module, and a digital twinning application module.

[0052] The multi-source data acquisition module includes a laser radar scanning unit, a radar liquid level meter matrix, a strain sensing network, an infrared thermal imaging unit, and a ship body inclination angle measurement unit deployed in the ship cabin body, which is used to record real-time monitoring data of the physical ship cabin and share the data to the core fitting algorithm module and the digital twinning application module.

[0053] The core fitting algorithm module is used to fuse point cloud dynamic data, perform surface fitting calculation and cabin volume calculation, and correct the cabin volume information in combination with the real-time monitoring data collected by the multi-source data acquisition module.

[0054] The digital twin application module includes a visual cabin volume panel, a leakage early warning platform and a fatigue damage prediction platform, the visual cabin volume panel is used to display the cabin volume information in real time, the leakage early warning platform is used to display the leakage risk result, and the fatigue damage prediction platform is used to display the fatigue accumulation index.

[0055] The application provides a ship cabin volume measurement method and system based on digital twinning, point cloud data of a ship cabin main body, additional structure and deducted structure is obtained by scanning, a digital twin ship cabin volume system is established, a ship cabin volume table and a heeling correction table are calculated by a cabin volume calculation model and input into the digital twin ship cabin volume system, the cabin volume information can be corrected in combination with the ship cabin volume table and the heeling correction table, and the cabin volume information is displayed in real time, automatic calibration of the cabin volume information is realized, and the ship cabin volume measurement efficiency and accuracy are improved. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0057] Figure 1 is a step diagram of a ship cabin volume measurement method based on digital twinning of the present application;

[0058] Figure 2 is a hardware structure diagram of a ship cabin volume measurement system based on digital twinning of the present application;

[0059] Figure 3 is a flowchart of B-spline surface fitting in the present application;

[0060] Figure 4 is a structure diagram of CNN-LSTM model in the present application;

[0061] Figure 5 is a flowchart of a ship cabin volume calculation method in the present application. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0063] As shown in Figure 1 , the present application provides a ship cabin volume measurement method based on digital twinning, comprising the following steps:

[0064] Step S1, using a laser scanner to scan to obtain three-dimensional point cloud data of the ship cabin body, additional structure and deduction structure, obtaining real-time stress data collected by the cabin sensor, correcting the point cloud data according to the real-time stress data, and optimizing the point cloud data;

[0065] Step S2, according to the point cloud data, surface fitting is carried out, a digital twin ship cabin system is established, and the ship cabin capacity table and the heeling correction table are calculated by the cabin capacity calculation model and input into the digital twin ship cabin system;

[0066] Step S3, the digital twin ship cabin system collects physical cabin temperature, liquid level information and inclination angle change information, corrects the cabin capacity information in combination with the ship cabin capacity table and the heeling correction table, and displays the cabin capacity information in real time.

[0067] The application obtains the point cloud data of the cabin body, additional structure and deduction structure by scanning, establishes a digital twin ship cabin system, calculates the ship cabin capacity table and the heeling correction table by the cabin capacity calculation model and inputs the digital twin ship cabin system, can correct the cabin capacity information in combination with the ship cabin capacity table and the heeling correction table, and display the cabin capacity information in real time, realize the automatic calibration of the cabin capacity information, and improve the efficiency and accuracy of the ship cabin volume measurement.

[0068] Further, the real-time stress data collected by the cabin sensor is obtained, and the point cloud data is corrected according to the real-time stress data, including:

[0069] An infrared thermal imager is arranged to obtain temperature data, and a stress sensor is arranged in a grid topology to form a strain sensing network, and stress data is obtained through the strain sensing network;

[0070] The stress sensor position is matched with the corresponding position of the point cloud data, the stress data is aligned with the point cloud data, the displacement caused by stress is calculated, the point cloud data is compensated for thermal expansion according to the temperature data, and the point cloud data is corrected.

[0071] Further, the stress sensor position is matched with the corresponding position of the point cloud data, the stress data is aligned with the point cloud data, the displacement caused by stress is calculated, the point cloud data is compensated for thermal expansion according to the temperature data, and the point cloud data is corrected, including:

[0072] The point cloud data is corrected by ,

[0073] Wherein, is the mechanical deformation of the stress sensor position,​ is the stress value at the stress sensor position, E is the Poisson's ratio of the material, is the thermal expansion deformation at the stress sensor position, is the thermal expansion coefficient, is the change of temperature relative to the reference value, and the vector displacement at the stress sensor position is synthesized, the discrete displacement vector is interpolated to the point cloud, and the displacement is decomposed to the unit direction vector to correct the point cloud data.

[0074] Further, the optimization processing of the point cloud data comprises:

[0075] The three-dimensional point cloud data of the ship cabin main body, additional structure and deducted structure is processed to remove point cloud noise, reduce sampling, and complete defect data processing to obtain the point cloud data of the ship cabin main body, additional structure and deducted structure;

[0076] The point cloud data is processed by point cloud filtering, point cloud data splicing and point cloud registration, wherein the point cloud filtering adopts a bilateral filtering method, and the bilateral filtering method comprises:

[0077] Let a point in the point cloud be , is a point in the neighborhood, , the position of the point after bilateral filtering is calculated as follows:

[0078] ;

[0079] ;

[0080] ;

[0081] wherein, is a spatial weight function, is a value weight function, is a spatial standard deviation parameter for controlling the size of the geometric neighborhood, is an attribute standard deviation parameter for controlling the feature similarity tolerance, is a point feature function, including the normal vector, curvature or coordinate of the point, denotes the point coordinate, i.e. .

[0082] Further, the real-time stress data collected by the in-cabin sensor is obtained, and the point cloud data is corrected according to the real-time stress data, which further comprises:

[0083] A liquid level meter matrix is set to obtain liquid level surface data, and a pressure sensor is set to obtain pressure data;

[0084] According to the liquid level data, the point cloud data is morphologically compensated, and according to the pressure data, the point cloud data is pressure compensated.

[0085] According to the liquid level data, the point cloud data is morphologically compensated, and according to the pressure data, the point cloud data is pressure compensated.

[0086] According to the liquid level data, the point cloud data is morphologically compensated, and according to the pressure data, the point cloud data is pressure compensated. Figure 3 According to the point cloud data, a curved surface fitting is performed to establish a digital twin ship cabin system, as shown in the figure.

[0087] The three-dimensional point cloud data of the ship cabin main body, the additional structure and the deducted structure after optimization processing is obtained.

[0088] The three-dimensional point cloud data of the ship cabin main body and the additional structure is spliced to form spliced point cloud data.

[0089] According to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure, a B-spline surface fitting algorithm is used to fit the point cloud data, and a B-spline surface is formed.

[0090] Further, according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure, a B-spline surface fitting algorithm is used to fit the point cloud data, and a B-spline surface is formed.

[0091] The spliced point cloud data and the three-dimensional point cloud data of the deducted structure are taken as input point clouds, and the input point clouds are parameterized, node vector determination, control point solving, surface generation and error analysis are performed, and whether the accuracy requirement is met is judged.

[0092] If yes, the B-spline surface is output, otherwise the input point cloud is re-parameterized.

[0093] Further, when fitting the point cloud data using the B-spline surface fitting algorithm according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure, a cubic B-spline curve is selected for fitting, and the parameter expression of the cubic B-spline curve is as follows:

[0094] ;

[0095] wherein, , , , are any four discrete points in a plane, P(t) is the coordinate of the fitted curve, and the component form of the above formula is as follows:

[0096]

[0097] wherein: , .

[0098] Further, when the B-spline surface is fitted, the B-spline surface is formed by multiple B-spline curves in the u and v directions, u and v are two independent parameters defined in the parameter domain, and a control grid is formed by (m+1)×(n+1) control points. The B-spline surface is fitted, and the equation of the B-spline surface is as follows:

[0099] ;

[0100] wherein is a control point set (i=0,1…m;j=0,1…n), m is the number of control points of the surface in the u direction minus 1, n is the number of control points of the surface in the v direction minus 1, and are B-spline surface basis functions.

[0101] As shown in Figure 5 , the ship's cabin capacity table and the correction table for listing and heeling are calculated by the cabin capacity calculation model, and the cabin capacity calculation model is set to calculate the cabin volume using the surface reconstruction method. The B-spline surface fitted by the cabin capacity calculation model according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure is obtained by a fitting algorithm to obtain a main capacity table and a deducted structure capacity table with a minimum unit of 1 cm;

[0102] The cabin capacity calculation model is set to calculate the cabin volume using the surface reconstruction method. The B-spline surface fitted by the cabin capacity calculation model according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure is obtained by a fitting algorithm to obtain a main capacity table and a deducted structure capacity table with a minimum unit of 1 cm;

[0103] Under the condition of room temperature 20℃, the non-convex surface is extracted using a rolling ball boundary with a radius of α, and a uniformly distributed grid is generated. The ship's cabin volume result is obtained by superposition calculation and output, and the listing and heeling correction table is output synchronously.

[0104] It can be understood that the calculation of the cabin volume operation is carried out at room temperature 20℃, first, the non-convex surface is extracted using the rolling ball boundary with radius α, and a uniformly distributed grid is generated, the ship cabin volume result is calculated and output using the mesh.get_volume() function of Open3D in python, and the correction table of the heeling and listing is output synchronously. Subsequently, based on the ship cabin volume result at room temperature 20℃, the cabin capacity information is corrected.

[0105] Further, the digital twin ship cabin volume system collects physical cabin temperature, liquid level information, and inclination angle change information, and corrects the cabin capacity information in combination with the ship cabin capacity table and the heeling and listing correction table, including:

[0106] Determine whether the physical cabin temperature of the ship cabin is room temperature 20℃, if the ship cabin is not at 20℃, correct the ship cabin volume based on the temperature calibration formula, the temperature calibration formula is:

[0107] ;

[0108] Wherein: V is the capacity after temperature correction, unit: cubic meters; V0 is the capacity shown in the cabin capacity table, unit: cubic meters; T is the average temperature of the cabin wall, unit: Celsius; α is the linear expansion coefficient of the cabin material, unit: 1 / ℃, sample =0.000012 / ℃.

[0109] Wherein, the correction of the ship cabin capacity according to the heeling angle and the listing angle includes: the liquid level height of the ship cabin at a certain angle of inclination is measured by the radar liquid level meter, then the height correction amount Δh at this inclination is:

[0110] Δh = H0-H’;

[0111] Wherein, H0 is the height (m) measured by the radar liquid level meter under the normal floating state of the ship; H' is the height measured by the radar liquid level meter under the inclined state of the ship (including heeling and listing).

[0112] The ship cabin normal position volume table is a corresponding table of the actual liquid level height and the volume of the cabin content under the liquid level under the normal position state of the ship; the ship cabin heeling correction table is a corresponding table of the actual liquid level height and the height correction value under the heeling angle; the ship listing correction table is a corresponding table of the actual liquid level height and the height correction value under the bow stern draft difference.

[0113] According to the point cloud fitting result and the cabin capacity calculation model, the right position volume table, the correction table of transverse inclination and the correction table of lateral inclination of the ship cabin are calculated and integrated with the digital twin ship cabin capacity system. When the ship is actually sailing, the actual liquid level height, the transverse inclination angle and the bow stern draft difference data in the cabin are collected through the radar liquid level meter and various sensors. The digital twin ship cabin capacity system obtains the corresponding height correction amount according to the correction table, and then the liquid volume in the cabin at this time is obtained by searching the right position volume table according to the corrected height.

[0114] As shown in Figure 1 , the method further comprises:

[0115] S4, constructing an AI model according to the liquid level change in the ship cabin, the stress data history and the infrared thermal imaging result, and prompting the leakage risk and the fatigue damage prediction.

[0116] Specifically, please refer to Figure 4 , Figure 4 is a CNN-LSTM (Convolutional Neural Network-Long Short Term Memory Network) hybrid model, wherein CNN is a convolutional neural network, and LSTM is a long short term memory network.

[0117] In other words, the method further comprises:

[0118] Collecting the liquid level change in the ship cabin, constructing a CNN-LSTM (Convolutional Neural Network-Long Short Term Memory Network) hybrid model, capturing the spatio-temporal feature information of the liquid level change, inputting the LSTM model to build a time series prediction model, and outputting the leakage risk result from the last hidden layer of the LSTM model to prompt the leakage risk.

[0119] Collecting the stress data history of the ship cabin, establishing a 3D-CNN model, fusing spatio-temporal information through three-dimensional convolution, retaining the correlation between spatial topology and spatio-temporal information sequence, outputting a fatigue accumulation index, and predicting fatigue damage.

[0120] Among them, collecting the liquid level change in the ship cabin, constructing a CNN-LSTM (Convolutional Neural Network-Long Short Term Memory Network) hybrid model, capturing the spatio-temporal feature information of the liquid level change, inputting the LSTM model to build a time series prediction model, and outputting the leakage risk result from the last hidden layer of the LSTM model to prompt the leakage risk, comprising:

[0121] Deploy multiple radar liquid level meters in each liquid tank of the ship, synchronously collect and record the ship attitude data (including roll, pitch and tank liquid level height) under the time coordinate; based on Kalman filtering, the collected ship attitude data is denoised and abnormal values are removed, the feature data liquid level change rate dh and the liquid level difference fh of adjacent cabins are established, and the time series ship attitude database is constructed; the database is used to train the CNN-LSTM network, which is composed of two convolutional layers with 64 convolutional kernels and a long short-term memory network LSTM, and the loss function is composed of standard binary cross entropy and focal loss function (focalloss); after training convergence, the leakage risk prediction model is obtained and deployed in the digital twin system, the system collects the ship attitude data in real time and inputs the prediction model, the risk coefficient at this time is output by the last full connection layer of the model, and the purpose of prompting the abnormal leakage risk in the cabin is achieved.

[0122] A stress sensor matrix is arranged on the ship cabin, and the sensor density is increased in the key structure area, the sensor data w, the corresponding point coordinates (x, y, z) and the time information t are collected, a five-dimensional data set (t, x, y, z, w) is constructed, and the data is denoised and abnormal data points are removed; the database is used to train the 3D-CNN network, which is mainly composed of 3D convolutional layer, 3D maximum pooling layer, LSTM layer and full connection layer, the spatial features and time features are extracted by 3D convolution and 1D time sequence convolution, and the sensitivity of the model to time is enhanced by LSTM, and the linear function is used to accumulate the fatigue damage; the loss function uses standard binary cross entropy, the fatigue damage prediction model is trained and deployed in the digital twin system; during the navigation of the ship, the system collects the stress sensor data and inputs the fatigue damage prediction model to obtain the predicted future fatigue damage and prompt the service life of the system.

[0123] By inputting the ship tank capacity table into the digital twin ship tank capacity system, automatic calibration is realized in combination with real-time temperature and pressure, during the navigation of the ship, the digital twin ship tank capacity system realizes real-time visualization of the ship tank capacity according to the actual temperature, pressure, liquid level and other information in the cabin combined with the tank capacity table information, records the change of liquid level and tank capacity, and prompts the leakage risk. Its function is to construct an AI model according to the liquid level change in the ship cabin, the history of stress data and the results of the infrared thermal imager, and to prompt the leakage risk and fatigue damage prediction.

[0124] As shown in Figure 2 The application also provides a ship tank capacity measurement system based on digital twinning, which is used to realize the ship tank capacity measurement method based on digital twinning described above, and the system comprises a multi-source data acquisition module, a core fitting algorithm module and a digital twin application module.

[0125] The multi-source data acquisition module includes a laser radar scanning unit, a radar liquid level meter matrix, a strain sensing network, an infrared thermal imaging unit, and a hull inclination angle measuring unit arranged in the cabin of the ship, for recording real-time monitoring data of the physical cabin of the ship and sharing to the core fitting algorithm module and the digital twin application module;

[0126] The core fitting algorithm module is used for fusing point cloud dynamic data, performing surface fitting calculation and cabin volume calculation, and correcting cabin volume information in combination with real-time monitoring data collected by the multi-source data acquisition module;

[0127] The digital twin application module includes a visual cabin volume board, a leakage warning platform, and a fatigue damage prediction platform. The visual cabin volume board is used for real-time display of the cabin volume information, the leakage warning platform is used for display of leakage risk results, and the fatigue damage prediction platform is used for display of fatigue accumulation index.

[0128] The specific limitations of the ship cabin volume measurement system based on digital twinning can refer to the limitations of the ship cabin volume measurement method based on digital twinning in the above, which will not be repeated here. Each module in the above ship cabin volume measurement system based on digital twinning can be realized by software, hardware, and combinations thereof, in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0129] The ship cabin volume measurement method and system based on digital twinning provided by the present application can obtain point cloud data of the cabin main body, additional structure, and deducted structure by scanning, establish a digital twin ship cabin volume system, calculate a ship cabin volume table and a heeling correction table by a cabin volume calculation model, and input the digital twin ship cabin volume system. The cabin volume information can be corrected in combination with the ship cabin volume table and the heeling correction table, and the cabin volume information can be displayed in real time, realizing automatic calibration of the cabin volume information and improving the efficiency and accuracy of ship cabin volume measurement.

[0130] The above is only the preferred embodiment of the present application. It should be noted that for ordinary skilled persons in the technical field, several improvements and modifications can be made without departing from the technical principles of the present application, and these improvements and modifications should be considered as the protection scope of the present application.

Claims

1. A method for ship's hold volume measurement based on digital twinning, characterized in that, The method comprises the following steps: acquiring three-dimensional point cloud data of a ship cabin body, additional structure and deduction structure by using a laser scanner, acquiring real-time stress data collected by a cabin sensor, correcting the point cloud data according to the real-time stress data, and optimizing the point cloud data; performing surface fitting according to the point cloud data, establishing a digital twin ship cabin system, calculating a ship cabin capacity table and a heeling correction table by using a cabin capacity calculation model, and inputting the digital twin ship cabin system; the digital twin ship cabin system collects physical cabin temperature, liquid level information and inclination angle change information, corrects cabin capacity information in combination with the ship cabin capacity table and the heeling correction table, and displays the cabin capacity information in real time; wherein the acquiring of the real-time stress data collected by the cabin sensor and the correcting of the point cloud data according to the real-time stress data comprise: setting an infrared thermal imager to acquire temperature data and setting a stress sensor to be deployed in a grid topology to form a strain sensing network, and acquiring stress data through the strain sensing network; matching the position of the stress sensor with the corresponding position of the point cloud data, aligning the stress data with the point cloud data, calculating displacement caused by stress, and performing thermal expansion compensation on the point cloud data according to the temperature data to correct the point cloud data; wherein the matching of the position of the stress sensor with the corresponding position of the point cloud data, the aligning of the stress data with the point cloud data, the calculation of displacement caused by stress, and the thermal expansion compensation on the point cloud data according to the temperature data to correct the point cloud data comprise: By , correcting the point cloud data; wherein, is a mechanical deformation quantity at the stress sensor position, is a stress value at the stress sensor position, E is a Poisson's ratio of the material, is a thermal expansion deformation at the stress sensor position, is a thermal expansion coefficient, is a change in temperature relative to a reference value, and a vector displacement at the stress sensor position is synthesized, the discrete displacement vector is interpolated to the point cloud, and the displacement is decomposed to a unit direction vector to correct the point cloud data.

2. The digital twin based ship hatch volume measurement method according to claim 1, characterized in that, the optimization of the point cloud data comprises: performing point cloud noise removal, downsampling, point cloud defect data processing on the acquired three-dimensional point cloud data of the ship cabin body, additional structure and deduction structure to obtain point cloud data of the ship cabin body, additional structure and deduction structure; performing point cloud filtering, point cloud data splicing and point cloud registration processing on the point cloud data, wherein the point cloud filtering adopts a bilateral filtering method, and the bilateral filtering method comprises: Let a point in the point cloud be , is a point in the neighborhood, , the position of the point after bilateral filtering is calculated as follows: ； ； ； wherein, is a spatial weight function, is a value weight function, is a spatial standard deviation parameter for controlling the size of the geometric neighborhood, is an attribute standard deviation parameter for controlling the feature similarity tolerance, is a point feature function containing the normal vector, curvature or the point coordinates of the point.

3. The digital twin based ship hatch volume measurement method according to claim 1, characterized in that, the acquiring of the real-time stress data collected by the cabin sensor and the correcting of the point cloud data according to the real-time stress data further comprise: setting a liquid level gauge matrix to acquire liquid level data and setting a pressure sensor to acquire pressure data; performing deformation compensation on the point cloud data according to the liquid level data and performing pressure compensation on the point cloud data according to the pressure data.

4. The digital twin based ship hatch volume measurement method according to claim 1, characterized in that, the surface fitting according to the point cloud data and the establishment of the digital twin ship cabin system comprise: acquiring three-dimensional point cloud data of the ship cabin body, additional structure and deduction structure after optimization processing; performing point cloud splicing on the three-dimensional point cloud data of the ship cabin body and the additional structure to form spliced point cloud data; performing fitting on the point cloud data by using a B-spline surface fitting algorithm according to the spliced point cloud data and the three-dimensional point cloud data of the deduction structure, respectively, to form a B-spline surface.

5. The digital twin based ship hatch volume measurement method according to claim 4, characterized in that, the performing of fitting on the point cloud data by using a B-spline surface fitting algorithm according to the spliced point cloud data and the three-dimensional point cloud data of the deduction structure, respectively, to form a B-spline surface comprises: The spliced point cloud data and the three-dimensional point cloud data of the deducted structure are taken as input point clouds, and the input point clouds are subjected to parameterization processing, node vector determination, control point solving, surface generation and error analysis to determine whether the accuracy requirement is met; If yes, a B-spline surface is output, otherwise the input point clouds are subjected to parameterization processing again.

6. The digital twin based ship hatch volume measurement method according to claim 4, characterized in that, When the B-spline surface fitting algorithm is used to fit the point cloud data according to the spliced point cloud data and the three-dimensional point cloud data of the deducted structure, a cubic B-spline curve is selected for fitting, and the parameter expression of the cubic B-spline curve is as follows: ； wherein , , , are any four discrete points in the plane, P(t) is the coordinate of the fitted curve, and the components of the above equation are: P(t) = P0+ P1t+ P2t2+ P3t3 ； wherein: , .

7. The digital twin based ship's hold volumetric measurement method according to claim 6, characterized in that, When the B-spline surface is fitted, the B-spline surface is formed by multiple B-spline curves in the u and v directions, u and v are two independent parameters defined in the parameter domain, and a control grid is formed by (m+1)×(n+1) control points. The B-spline surface is fitted, and the equation of the B-spline surface is: ； wherein is a set of control points (i=0,1…m;j=0,1…n), m is the number of control points of the surface in the u direction minus 1, n is the number of control points of the surface in the v direction minus 1, and is a B-spline surface basis function.

8. The digital twin based ship hatch volume measurement method according to claim 4, characterized in that, The ship's tank capacity table and the correction table for rolling and tilting calculated by the tank capacity calculation model include: The tank capacity calculation model is set to calculate the volume of the tank body using the surface reconstruction method. The B-spline surface formed by fitting the spliced point cloud data and the three-dimensional point cloud data of the deducted structure according to the tank capacity calculation model is obtained by a fitting algorithm to obtain a main capacity table and a deducted structure capacity table with a minimum unit of 1 cm; Under the condition of room temperature 20℃, a non-convex surface is extracted using a rolling ball boundary with a radius of α, and a uniformly distributed grid is generated. The ship's tank volume result is obtained by superposition calculation and output, and the rolling and tilting correction table is output synchronously.

9. The digital twin based ship hatch volume measurement method according to claim 8, characterized in that, The digital twin ship tank capacity system collects physical tank temperature, liquid level information, and inclination angle change information, and corrects the tank capacity information in combination with the ship tank capacity table and the rolling and tilting correction table, which includes: Determine whether the physical tank temperature of the ship tank is room temperature 20℃. If the ship tank is not at 20℃, correct the ship tank volume based on the temperature calibration formula, which is: ； Where: is the temperature corrected volume in cubic meters; is the volume shown in the tank table in cubic meters; is the average temperature of the tank walls in degrees Celsius; is the linear expansion coefficient of the tank material in degrees Celsius per one, the sample is taken from the tank table = 0.000012 / °C.

10. The digital twin based ship hatch volume measurement method according to claim 8, characterized in that, The method further includes: Collecting the liquid level change in the ship tank, constructing a CNN-LSTM hybrid model to capture the spatio-temporal feature information of the liquid level change, inputting the model into the LSTM model to build a time series prediction model, and outputting the leakage risk result from the last hidden layer of the LSTM model to prompt the leakage risk; Collecting the history of ship tank stress data, establishing a 3D-CNN model, and fusing spatio-temporal information through three-dimensional convolution to retain the correlation between spatial topology and spatio-temporal information sequence, and outputting a fatigue accumulation index for fatigue damage prediction.

11. A digital-twin-based ship's hold volumetric measurement system, characterized in that, The system for implementing the digital twin-based ship tank volume measurement method of any one of claims 1 to 10 includes a multi-source data acquisition module, a core fitting algorithm module, and a digital twin application module. The multi-source data acquisition module includes a laser radar scanning unit, a radar liquid level meter matrix, a strain sensing network, an infrared thermal imaging unit, and a ship body inclination angle measurement unit deployed in the ship tank body, which is used to record real-time monitoring data of the physical ship tank and share the data to the core fitting algorithm module and the digital twin application module. The core fitting algorithm module is used for fusing point cloud dynamic data, carrying out surface fitting calculation and cabin volume calculation, and correcting cabin volume information in combination with real-time monitoring data collected by the multi-source data acquisition module; The digital twin application module includes a visual cabin volume board, a leakage early warning platform and a fatigue damage prediction platform. The visual cabin volume board is used for real-time display of the cabin volume information, the leakage early warning platform is used for display of leakage risk results, and the fatigue damage prediction platform is used for display of fatigue accumulation indexes.

Images