Safety evaluation system after miter gate is impacted by ship

Abstract

The invention discloses a safety evaluation system after a miter gate is impacted by a ship. Based on sensor acquisition and an image acquisition module, actual measurement data such as vibration speed, acceleration response, a maximum inclination angle, maximum displacement at a door shaft column and a beveled joint column, water tightness and the like after the gate is collided are obtained. Based on results of a large number of numerical simulation working conditions, an atlas analysis module capable of visually displaying impact force and deformation damage conditions of the gate is obtained. And when partial measured data of the gate and parameters such as a deformation value obtained by map query are greater than or equal to a safety threshold value, giving an alarm and giving a safety level and a damage condition after the gate is collided. Through comparison of simulation data and actually measured data under the same working condition, whether the map query result is overestimated or underestimated can be further judged. The method can accurately and efficiently obtain the impact force and gate deformation after the miter gate is impacted by the ship through map query, and is especially suitable for occasions where a visual detection method is not suitable for being adopted, such as gate deformation detection when the underwater visibility is low.

Description

Technical Field

[0001] The present invention relates to the technical field of miter gate safety assessment, and more specifically, to a rapid safety assessment system for a miter gate after it is hit by a ship, and is also applicable to real-time safety assessment of a miter gate when it is not hit by a ship. Background Art

[0002] The miter gate has the advantages of small main beam bending moment, small deadweight and low cost. It is the most commonly used gate structure in high head and large span ship locks. However, with the continuous development of the shipping industry, the waterway is becoming more and more crowded. The miter gate is largely exposed to the risk of ship collision, which seriously threatens the watertightness and stability of the miter gate structure. In order to ensure the safe operation of the miter gate, a system that can evaluate the damage degree and risk level of the miter gate after being hit by a ship in real time is urgently needed.

[0003] The methods and technologies in this field have encountered at least the following challenges: First, the structure of the miter gate is complex and large in scale. If only sensors are deployed to obtain the real-time service status of the gate, the accuracy needs to be improved. Second, the damage degree of the miter gate after being hit by a ship is difficult to calculate quickly and efficiently in real time, and a safety assessment fast query map is urgently needed.

[0004] This patent combines the finite element simulation results of ships hitting miter gates to draw gate dynamic response maps under different impact scenarios. A comprehensive analysis of simulation results, image data and sensor measured parameters is performed to provide a solution for the rapid inspection and maintenance of miter gates through safety assessment and real-time warning. Summary of the invention

[0005] The present invention provides a real-time safety assessment system for a miter gate after being hit by a ship, which can accurately and efficiently obtain the impact force and gate deformation of the miter gate after being hit by a ship through graph query. It is particularly suitable for occasions where visual detection methods are not suitable, such as gate deformation detection when underwater visibility is low.

[0006] The technical solution of the present invention:

[0007] A safety assessment system for a miter gate after being hit by a ship, comprising a sensor acquisition module installed on the miter gate, an image acquisition module of the lock chamber, a data transmission module, a data processing and analysis module, a finite element simulation module for a ship hitting the miter gate, a spectrum analysis module, a safety assessment module and a real-time warning module;

[0008] The sensor acquisition module includes a velocity sensor, an acceleration sensor, a displacement sensor and an inclination sensor, which are used to monitor the dynamic response parameters of the miter gate after being hit by a ship in real time;

[0009] The image acquisition module monitors the service status of ships entering the lock area and the miter gates in real time;

[0010] The data transmission module transmits the sensor data and image data to the data processing and analysis module in real time;

[0011] The data processing and analysis module stores the data in real time, and obtains the maximum displacement, inclination, velocity and acceleration of the gate through sensor data analysis for safety assessment. The edge detection algorithm is used to monitor water leakage and provide real-time warning for the real-time image of the miter gate. The specific steps are as follows:

[0012] (1) The image of the miter gate is processed using the Canny edge detection algorithm to extract the outline of the miter gate; then, the features of the sealing area are extracted from the outline of the miter gate using the region segmentation method;

[0013] (2) Check the extracted sealing area for leaks, cracks or loose parts to determine the watertightness;

[0014] (3) When a poor seal or water leakage is detected in the miter gate, the system will automatically analyze the location, size and severity of the leakage, and immediately issue a warning message and provide the safety level of the gate operation;

[0015] The convolutional neural network is used to identify the ship speed, height and angle of the real-time transmission image data of the image acquisition module. The specific steps are as follows:

[0016] (1) Establish a known ship image dataset and mark the ship's speed, angle, and position. These labeled data will be used to train and test the convolutional neural network model.

[0017] (2) Design a convolutional neural network structure consisting of convolutional layers, pooling layers, and fully connected layers to perform parameter recognition on ship images; perform preprocessing operations such as scaling, cropping, and grayscale on the images to prepare for input into the convolutional neural network model;

[0018] (3) Use the labeled image dataset to train the convolutional neural network model and optimize the model parameters through the back-propagation algorithm; continue to adjust and optimize the model;

[0019] (4) Test the trained convolutional neural network model using a new test dataset to evaluate the accuracy of the model in identifying ship parameters;

[0020] (5) Accurately identify the ship’s speed, angle, position and other parameters based on the convolutional neural network model; if it is identified that the ship may collide with the gate, an early warning is immediately issued to ensure that necessary control measures are taken in time to prevent the accident;

[0021] The finite element simulation module of ship impacting miter gates performs simulation analysis on ship impacting miter gates in different scenarios, and obtains the dynamic response and damage data of the miter gates for reference in safety assessment;

[0022] The spectrum analysis module uses the dynamic response data of the ship hitting the miter gate, taking the ship mass, speed, height and angle as independent variables, and the impact force and gate deformation value as dependent variables; uses a three-dimensional scatter plot to perform a preliminary visual analysis of the data distribution, and selects scatter fitting as the fitting model based on the distribution characteristics of the data; calculates the determination coefficient R through Python 2 To evaluate the quality of the fit, here are the detailed steps for the calculation:

[0023] (1) Calculate the total sum of squares (SST) of the differences between the data points and the mean:

[0024]

[0025] Among them, y i is the value of the data point, The average of the data points;

[0026] (2) Calculate the sum of squares of the difference between the data point and the model prediction value SSR:

[0027]

[0028] in, is the model prediction value, are the estimates of the fitted model parameters;

[0029] (3) Calculation of the coefficient of determination R 2 :

[0030]

[0031] R 2 The value of is between 0 and 1. A value close to 1 indicates that the model can predict the data well, and a value close to 0 indicates that the model does not provide a better prediction than the average. Using the selected model and the fitted parameters, a 3D bar graph is generated to show the changing pattern of impact force and gate deformation.

[0032] The trained convolutional neural network identifies the speed, height and angle of the ship at the time of collision, inputs these parameters and the mass of the ship into the system, and automatically and quickly outputs the current impact force and deformation value based on the 3D bar graph;

[0033] The overall damage degree of the gate is evaluated according to the permanent deformation energy ratio DR%, as shown in the following formula:

[0034]

[0035] Where E P is the kinetic energy dissipated by the plastic deformation of the gate, E F is the kinetic energy dissipated by the gate component fracture, E K It is the initial kinetic energy of the ship when no collision occurs; the damage degree and stress-strain cloud diagram of the miter gate can be efficiently and intuitively displayed through the 3D map interface;

[0036] Under the same impact conditions, when the damage degree reflected by the measured data is greater than that reflected by the simulated data, it is considered that the impact force and gate deformation obtained by the spectrum analysis are underestimated, and the actual damage is more serious than the quick estimate obtained by the query; otherwise, the actual damage is overestimated;

[0037] The safety assessment module evaluates the risk level and potential hazards of the miter gates after being hit. It divides the miter gates into three safety levels based on whether they meet the deformation specifications and normal operation after being hit by a ship, and proposes corresponding maintenance methods based on the damage description of each level of gates. When the measured data is greater than or equal to the safety threshold, the real-time early warning module will issue an alarm and give the safety level, damage status and maintenance suggestions of the gates after the collision. Finally, all data are transmitted to the cloud database for storage and display.

[0038] Beneficial effects of the present invention:

[0039] (1) It realizes real-time assessment and early warning of the damage degree and risk level of the miter gate after being hit by a ship, providing data support for the rapid detection and maintenance of the miter gate.

[0040] (2) A quick query map of ship impact force and local deformation of miter gates and its analysis method are proposed. The 3D graphical interface can be used to quickly, intuitively and efficiently query the impact force, local deformation and damage of gates under different impact scenarios, avoiding time-consuming and laborious modeling and analysis. The safety level and recommendations of the miter gate after being hit are given through safety assessment, and the operator is reminded in the form of an automatic alarm so that necessary countermeasures can be taken in time to ensure the safety of personnel and the miter gate. It can provide support for the optimization design of gate anti-collision, and at the same time provide owners with a convenient and intuitive way to quickly refer to and predict the safe service of the gate.

[0041] (3) It can realize the all-weather "cloud-based monitoring and cloud alarm" operation mode, ensuring the safe and stable operation of the miter gate throughout its life cycle, and achieving low-cost, high-efficiency and high-precision safety assessment of the miter gate. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Schematic diagram of the safety assessment and early warning system after the miter gate is hit by a ship;

[0043] Figure 2 Finite element model diagram of miter gate and ship;

[0044] Figure 3 Energy conversion diagram of a ship hitting a miter gate;

[0045] Figure 4 Three-dimensional query map of impact force and local deformation response of miter gate;

[0046] Figure 5 Time history graph of ship impact force;

[0047] Figure 6 Energy consumption comparison chart of ships and miter gates; DETAILED DESCRIPTION

[0048] Taking the finite element model of the miter gate structure of a ship lock as an example, a real-time safety assessment system for ship impact with the miter gate is studied. The schematic diagram of the system is shown in the figure. Figure 1 As shown, the technical solution of the present invention is described in detail below in conjunction with the accompanying drawings:

[0049] 1. Miter gate leakage monitoring and early warning

[0050] (1) Obtaining the grayscale image of the miter gate

[0051] The miter gate image is obtained from the cameras installed in front and behind the miter gate, with an image resolution of 1280×720. The image is converted into a grayscale image while retaining the structural information.

[0052] (2) Extract the outline of the miter gate

[0053] Gaussian filtering is performed on the grayscale image to remove image noise; the Canny edge detection algorithm is used to extract the edge contour of the miter gate.

[0054] (3) Extract the sealing area of ​​the miter gate

[0055] The contour of the miter gate is segmented, and the features of the sealing area of ​​the miter gate are extracted using ROI (Region of Interest). The sealing area is locked using a boundary rectangle, and the sealing part is focused on based on the segmented area.

[0056] (4) Defect detection in the sealing area of ​​miter gates

[0057] Crack detection: Detect whether there are discontinuous areas or cracks on the edge. Analyze whether there are cracks by calculating the change in edge density. When the edge density is less than 0.3, it is determined that there is a crack.

[0058] Leak detection: The optical flow method is used to analyze whether there is an abnormal flow pattern in the sealed area. If the water flow pattern is detected to be inconsistent with the reference pattern with a good seal, it is marked as a leak.

[0059] (5) Analysis of miter gate leakage detection results

[0060] The detected leakage area is further analyzed; the optical flow vector field clustering method is used to mark the specific location of the leakage or crack; the safety level of the gate is evaluated according to the area A of the defect, and the severity can be divided into four levels of "normal", "minor", "medium" and "severe" through preset thresholds; among them, normal: A=0, minor: A<100px, medium: 100px≤R<500px, severe: A≥500px; px means pixel, that is, the area size of an area in the image, in pixels.

[0061] (6) Real-time early warning system

[0062] When abnormal sealing performance or water leakage is detected, the system will issue a real-time warning.

[0063] Visual warning: The leaking area is marked with a red frame in the image and the warning level is displayed.

[0064] Sound and message warning: remind the operator through buzzer or text message.

[0065] Recommended actions: Based on the severity of the leak, repair suggestions are given (stop operation, routine inspection or emergency repair).

[0066] 2. Identify the speed, altitude and angle of the ship

[0067] (1) Construction of ship dataset

[0068] Collect ship image data: Obtain known ship image datasets, including ship motion images in multiple scenarios, covering parameter ranges of different speeds, angles, and altitudes; ensure that the dataset is diverse and covers complex environments such as daytime, nighttime, and rainy days.

[0069] Data annotation: Annotate each image with the following parameters: mark the actual speed of the ship (which can be obtained through GPS), mark the yaw angle of the ship, and mark the position and relative height of the ship in the image.

[0070] (2) Image data preprocessing

[0071] Image Scaling: Scale the input image to a fixed size (64×64) to normalize the input.

[0072] Image cropping: crop the area where the ship is located from the original image to reduce background interference.

[0073] Image grayscale: Convert to grayscale image to reduce the dimension of input data and reduce the computational complexity of the model.

[0074] (3) Convolutional Neural Network (CNN) Structure Design

[0075] Input layer: Input dimension: 64×64×1 (grayscale image).

[0076] Convolutional layer: extract image edge and shape features, activation function: ReLU, defined as: f(x) = max(0,x), x is the input value of the neuron; first layer: 32 filters, convolution kernel size 3×3, step size 1; second layer: 64 filters, convolution kernel size 3×33.

[0077] Pooling layer: Use maximum pooling to reduce the resolution and retain the main features; the pooling window size of each pooling layer is 2×22, and the step size is 2.

[0078] Fully connected layer: flatten the convolutional features into vectors; first layer: 128 neurons, ReLU activation; second layer: 64 neurons, ReLU activation.

[0079] Output layer: outputs 3 values ​​(speed, angle, height); activation function: linear activation, directly outputs continuous values.

[0080] (4) Model training

[0081] Loss function: Use mean square error (MSE) as the loss function to minimize the parameter identification error.

[0082] Training process: The labeled image dataset is divided into a training set and a validation set (80% for training and 20% for validation); the batch size is 64, the number of training rounds is 100, and the learning rate and regularization coefficient (regularization value is 10 -4 ).

[0083] (5) Model testing and evaluation

[0084] Test dataset: Use a completely independent test dataset to evaluate the model performance and ensure its generalization ability.

[0085] Model evaluation indicators: MSE is used to evaluate the numerical prediction error, and accuracy is used to determine whether the model can accurately classify ship speed, angle, and height.

[0086] (6) Collision Warning

[0087] Image input: Image frames are acquired from the camera in real time and input into the CNN model after the same preprocessing process as the training data.

[0088] Parameter identification: The model outputs the speed, angle and altitude of the ship.

[0089] Collision warning: Based on the identified parameters, the distance D between the ship and the gate is calculated; if D < 5m, an alarm is triggered; if D < 2m, emergency control measures are triggered; the operator is prompted to take necessary measures to adjust the ship's speed and direction, stop gate operation, and avoid potential collisions.

[0090] 3. Finite element model and analysis of miter gate and ship

[0091] The miter gate is 34m wide and 38.42m high. The angle between the left and right gate leaves and the vertical direction of the lock chamber is 22°. The number of grids is 345,214. The ship is a 5,000-ton bulbous structure with a length of 90m, a width of 15.55m, a depth of 5.95m, and a maximum draft of 5.22m. The model is as follows: Figure 2 As shown. The dynamic response of the ship hitting the miter gate was simulated and verified using Abaqus finite element analysis software. The Linux system was used to call two AMD9654 CPUs with a total of 192 cores for parallel computing, which can significantly improve the calculation speed. The S4R four-node surface thin shell element was used. The initial mesh convergence test determined that the mesh size of the impact area was 150mm, which met the best accuracy requirements, and the mesh size of the remaining areas was 250mm. The Johnson-Cook damage model was used to define the steel of the miter gate; the displacement constraints in the X and Y directions were set for the miter column and the gate axis column, the displacement constraints in the X and Y directions were set at the top pivot, and the displacement constraints in the X, Y, and Z directions were set for the bottom pivot. The prestress of the back tie rod of the miter gate adopts the implicit and explicit coupling analysis method. The implicit analysis of the prestress is used to obtain the shape and stress state after pre-deformation, which is then used as the initial state of the explicit dynamic analysis. In order to make the simulation results accurate, the mass scaling function is not used. The impact contact between the ship and the gate adopts the penalty contact mode, allowing a small sliding between the contact surfaces. The friction coefficient for the tangential behavior is 0.3; the contact stiffness for the normal behavior is K = 80000GPa. The energy conversion of the system during the collision between the ship and the gate is as follows: Figure 3 As shown, the proportion of pseudo-strain energy to internal energy is only 1.4%. It is generally required that the pseudo-strain energy of the simulation results be controlled within 5% of the internal energy; therefore, from the perspective of energy balance, it is proved that the grid size and the accuracy of the numerical simulation results are satisfactory.

[0092] 4. Sensor and camera placement

[0093] Displacement sensors are installed at the upper, middle and lower parts of the gate shaft column and the miter column of the miter gate, the acceleration sensor is installed in the middle of the miter column of the miter gate, the inclination sensor is installed at the top pivot position of the miter gate, the speed sensor is installed at the top of the miter column, and the high-resolution cameras with night vision function are installed above the gate walls inside and outside the lock chamber. The collected sensor data and image data are uploaded to the cloud for storage, connected to the cloud server through the Internet, to achieve real-time data transmission and access, and provide data support for subsequent data processing analysis and safety assessment, and realize all-weather intelligent identification and early warning of ships passing through the lock.

[0094] 5. Quick query of gate impact force and deformation based on graph

[0095] 180 different impact scenarios are defined, including different impact tonnage, impact speed, impact angle, impact position and other parameters. Different impact scenarios are automatically generated using Python scripts, and corresponding inp files are generated. Python scripts are used to call Abaqus command to batch run multiple inp files, which can conveniently and efficiently obtain the dynamic response results of the miter gate under different impact scenarios. Based on 180 simulated impact data, a three-dimensional query map is constructed using the surface interpolation fitting method, such as Figure 4 As shown. Mapping the impact force and local deformation values ​​of the gate at different speeds and tonnages onto a three-dimensional surface can intuitively and efficiently display the overall severity of the gate after the collision. Given the mass of the ship, the speed, angle and impact position of the ship identified by image recognition can be matched with the three-dimensional map to directly obtain the required impact force and deformation, and display the damage and safety level of the current miter gate without time-consuming modeling and analysis, solving the problem of the difficulty in quickly and accurately evaluating the service status of the gate after the ship hits it.

[0096] 6. Safety classification of miter gates after being hit by ships

[0097] According to whether the miter gate meets the design standards and normal operation requirements after being hit by a ship, with reference to the "Design Code for Steel Gates of Water Conservancy and Hydropower Projects" (SL74-2019) and the "Design Code for Steel Structures" (GB 50017-2017), the miter gate after collision is divided into three safety levels and corresponding maintenance recommendations. The specific classification standards are as follows:

[0098] Level I: When all gate operation indicators meet the design standards, it is defined as slight damage. The overall appearance of the gate is intact, the load-bearing and traffic capacity remain unchanged, there are no cracks, the water tightness is intact, and there are slight deformations and scratches at the impact site; the maximum local deformation of the steel structure panel does not exceed 25mm, and the maximum deflection of the miter gate does not exceed 45mm; the overall stability is intact and can be opened and closed normally; it can be navigated normally after routine durability maintenance.

[0099] Level II: When the gate operation indicators basically meet the design standards, it is defined as moderate damage. The main beam in the gate impact area has slight bending deformation, the load-bearing and traffic capacity is reduced to a certain extent, non-penetrating cracks appear, and slight water seepage occurs; the maximum local deformation of the steel structure panel does not exceed 35mm, and the maximum deflection of the miter gate does not exceed 52mm; the overall stability meets the requirements of the specification and can be opened and closed normally; after minor repairs to restore the accuracy of the components or replace parts, normal navigation can be carried out.

[0100] Level III: When the gate operation index does not meet the design standard, it is defined as severe damage. Obvious local deformation occurs in the gate impact area, the maximum local deformation of the steel structure panel exceeds 35mm, and the maximum deflection of the miter gate exceeds 52mm; the main beam undergoes bending and torsion deformation, the deflection exceeds the specified value, through cracks appear, and serious water seepage occurs; the load-bearing and traffic capacity are greatly reduced, the overall stability is reduced, the opening and closing are blocked, and major repairs are required.

[0101] Based on the above classification, the safety level of the miter gate after being hit can be described qualitatively and quantitatively. Combined with the graph analysis, the damage of the miter gate after being hit can be intuitively displayed, which is conducive to the rapid and accurate assessment of the service status of the damaged gate.

[0102] 7. Safety assessment and real-time warning

[0103] Evaluate whether the structural deformation, stability, water tightness and ship impact force of the gate after being hit meet the safety standards: the impact force is set according to the design requirements and bearing capacity of the miter gate, the deformation value is determined according to the structural material and deformation limit of the miter gate, the stability is set by analyzing the overturning and sliding of the miter gate, and the water tightness is set according to the leakage standard; the integration of the real-time warning module and other modules can realize the comprehensive analysis and evaluation of the real-time data of the miter gate, and can display the impact force time history curve of the miter gate in real time.

[0104] like Figure 5 As shown in the figure, IT1-IT5 represent the impact force of a 1000-5000 ton ship when it hits the gate head-on. Figure 6 As shown, the comparison of ship, miter gate main beam and overall energy consumption under 12 collision positions is shown. By displaying the gate's safety level, stability, watertightness and other data in real time, the structural strength and stability of the miter gate can be intuitively understood. These data can be used to judge the service status of the miter gate and provide necessary safety recommendations and measures. Once these parameters exceed or equal to the set threshold, it means there is a safety risk; at this time, the real-time early warning module issues an alarm and gives the safety level, damage and maintenance recommendations after the gate is hit. Finally, all data is transmitted to the cloud database for storage and display; relevant departments can access and analyze the data at any time to understand the operating status and operation and maintenance history of the miter gate.

[0105] The present invention realizes low-cost, high-efficiency and high-precision real-time safety assessment and early warning of ship collisions with miter gates. It integrates the common analysis results of real-time sensor monitoring data, real-time image information and finite element model analysis, and gives the overall safety level, damage status and maintenance suggestions of the miter gates. The system is also suitable for real-time safety assessment and early warning of gates when they are not hit by ships.

Claims

1. A safety assessment system for a miter gate after being hit by a ship, characterized in that: It includes a sensor acquisition module installed on the miter gate, an image acquisition module in the lock chamber, a data transmission module, a data processing and analysis module, a finite element simulation module for ships impacting the miter gate, a spectrum analysis module, a safety assessment module and a real-time warning module; The sensor acquisition module includes a velocity sensor, an acceleration sensor, a displacement sensor and an inclination sensor, which are used to monitor the dynamic response parameters of the miter gate after being hit by a ship in real time; The image acquisition module monitors the service status of ships entering the lock area and the miter gates in real time; The data transmission module transmits the sensor data and image data to the data processing and analysis module in real time; The data processing and analysis module stores the data in real time, and obtains the maximum displacement, inclination, velocity and acceleration of the gate through sensor data analysis for safety assessment. The edge detection algorithm is used to monitor water leakage and provide real-time warning for the real-time image of the miter gate. The specific steps are as follows: (1) The image of the miter gate is processed using the Canny edge detection algorithm to extract the outline of the miter gate; then, the features of the sealing area are extracted from the outline of the miter gate using the region segmentation method; (2) Check the extracted sealing area for leaks, cracks or loose parts to determine the watertightness; (3) When a poor seal or water leakage is detected in the miter gate, the system will automatically analyze the location, size and severity of the leakage, and immediately issue a warning message and provide the safety level of the gate operation; The convolutional neural network is used to identify the ship speed, height and angle of the real-time transmission image data of the image acquisition module. The specific steps are as follows: (1) Establish a known ship image dataset and mark the ship's speed, angle, and position. These labeled data will be used to train and test the convolutional neural network model. (2) Design a convolutional neural network structure consisting of convolutional layers, pooling layers, and fully connected layers to perform parameter recognition on ship images; perform preprocessing operations such as scaling, cropping, and grayscale on the images to prepare for input into the convolutional neural network model; (3) Use the labeled image dataset to train the convolutional neural network model and optimize the model parameters through the back-propagation algorithm; continue to adjust and optimize the model; (4) Test the trained convolutional neural network model using a new test dataset to evaluate the accuracy of the model in identifying ship parameters; (5) Accurately identify the ship’s speed, angle, position and other parameters based on the convolutional neural network model; if it is identified that the ship may collide with the gate, an early warning is immediately issued to ensure that necessary control measures are taken in time to prevent the accident; The finite element simulation module of ship impacting miter gates performs simulation analysis on ship impacting miter gates in different scenarios, and obtains the dynamic response and damage data of the miter gates for reference in safety assessment; The spectrum analysis module uses the dynamic response data of the ship hitting the miter gate, taking the ship mass, speed, height and angle as independent variables, and the impact force and gate deformation value as dependent variables; uses a three-dimensional scatter plot to perform a preliminary visual analysis of the data distribution, and selects scatter fitting as the fitting model based on the distribution characteristics of the data; calculates the determination coefficient R through Python 2 To evaluate the quality of the fit, here are the detailed steps for the calculation: (1) Calculate the total sum of squares (SST) of the differences between the data points and the mean: Among them, y i is the value of the data point, The average of the data points; (2) Calculate the sum of squares of the difference between the data point and the model prediction value SSR: in, is the model prediction value, are the estimates of the fitted model parameters; (3) Calculation of the coefficient of determination R 2 : R 2 The value of is between 0 and 1. A value close to 1 indicates that the model can predict the data well, and a value close to 0 indicates that the model does not provide a better prediction than the average. Using the selected model and the fitted parameters, a 3D bar graph is generated to show the changing pattern of impact force and gate deformation. The trained convolutional neural network identifies the speed, height and angle of the ship at the time of collision, inputs these parameters and the mass of the ship into the system, and automatically and quickly outputs the current impact force and deformation value based on the 3D bar graph; The overall damage degree of the gate is evaluated according to the permanent deformation energy ratio DR%, as shown in the following formula: Where E P is the kinetic energy dissipated by the plastic deformation of the gate, E F is the kinetic energy dissipated by the gate component fracture, E K It is the initial kinetic energy of the ship when no collision occurs; the damage degree and stress-strain cloud diagram of the miter gate can be efficiently and intuitively displayed through the 3D map interface; Under the same impact conditions, when the damage degree reflected by the measured data is greater than that reflected by the simulated data, it is considered that the impact force and gate deformation obtained by the spectrum analysis are underestimated, and the actual damage is more serious than the quick estimate obtained by the query; otherwise, the actual damage is overestimated; The safety assessment module evaluates the risk level and potential hazards of the miter gates after being hit. It divides the miter gates into three safety levels based on whether they meet the deformation specifications and normal operation after being hit by a ship, and proposes corresponding maintenance methods based on the damage description of each level of gates. When the measured data is greater than or equal to the safety threshold, the real-time early warning module will issue an alarm and give the safety level, damage status and maintenance suggestions of the gates after the collision. Finally, all data are transmitted to the cloud database for storage and display.

Images