Method for predicting maximum displacement of dredging ship gauge pile

Abstract

The invention discloses a method for predicting the maximum displacement of a dredging ship gauge pile. The method comprises the following steps: constructing a trolley-gauge pile-soil body coupling numerical model based on ABAQUS software; establishing a cubic polynomial regression model of the maximum displacement of the gauge pile with respect to the three influence factors by applying a response surface method; data samples are stochastic in a regression model parameter range, and a training set, a verification set and a prediction set are divided; constructing a BP neural network model, and training and testing the BP neural network model by using the training set and the verification set to obtain an optimized prediction model; and inputting influence factors of the prediction set into the trained BP neural network for prediction, performing dual verification on a prediction result by using a relative error and a determination coefficient, and outputting the prediction result when the prediction result meets a verification standard. According to the method, high-precision real-time prediction of the displacement of the gauge pile is realized, and an important decision basis is provided for safe construction of a dredging ship.

Description

Technical Field

[0001] This invention relates to the field of dredging vessels, and more particularly to a method for predicting the maximum displacement of positioning piles on dredging vessels. Background Technology

[0002] Dredging vessels, as a high-tech type of ship, are the core equipment for dredging projects. Their operational stability and the reliability of positioning piles are key issues in ensuring construction safety and efficiency. The positioning piles are subjected to complex forces due to the complex marine environment and numerous loads encountered by dredging vessels. Therefore, positioning piles are prone to buckling failure or breakage, significantly impacting normal construction efficiency and safety. It is necessary to study and analyze their stress and deformation.

[0003] During dredging operations, the positioning pile system is the core support for the stable operation of the vessel. Deeply embedded in the riverbed or seabed, the positioning piles provide a stable fulcrum for the hull, enabling the dredger to operate precisely under complex hydrogeological conditions. However, the operating environment is complex and variable; the impact of water flow, siltation, and the movement of the hull all combine to exert complex external forces on the positioning piles. These forces can cause displacement of the positioning piles. Once the displacement exceeds the safe range, the stability of the hull will be severely threatened, potentially leading to major safety accidents such as positioning pile breakage and hull capsizing. This would not only cause huge economic losses but also pose a serious threat to the surrounding water environment and the safety of the workers. Researching the technology of positioning piles for cutter suction dredgers not only helps solve practical problems in marine dredging engineering but also promotes the development of related technologies. Through in-depth theoretical research and technological innovation, the design and construction process of positioning piles can be continuously optimized, promoting the advancement of marine engineering technology and providing technical support for the efficient operation of marine engineering equipment. Traditional methods for predicting positioning pile displacement mainly involve establishing complex numerical calculation models and then calculating boundary conditions. Although this method is simple, the calculations are complex and tedious, and the calculations must be recalculated every time the influencing factors are changed.

[0004] MATLAB is a mathematical software that can be used in fields such as deep learning. Algorithms can be implemented, and corresponding data processing and visualization can be performed in MATLAB, providing a good display effect for prediction results.

[0005] Backpropagation (BP) neural network prediction models are mainly used in fields such as image recognition and natural language processing. However, in the field of force analysis for positioning piles in dredging vessel construction, there is currently no research on the prediction of relevant parameters. Existing research on dredging vessel prediction also lacks studies on constructing relationships between physical indicators and corresponding parameters by influencing physical indices. Other parameter prediction methods mostly rely on experimental data for training and prediction, resulting in small sample sizes and a lack of universality. BP neural networks are more advantageous in handling nonlinear relationships. Linear regression models exhibit large prediction errors when facing complex nonlinear data, while BP neural networks, through their multi-layered structure and nonlinear activation functions, can better adapt to the complexity of the data.

[0006] Considering the above factors, it is necessary to develop a system that can quickly predict the corresponding parameters, enabling the system to predict specified parameters for different physical indices, and providing guidance for prevention and application in practical engineering. Summary of the Invention

[0007] This invention provides a method for predicting the maximum displacement of positioning piles for dredging vessels, overcoming the problems of complex and cumbersome calculations and large prediction errors in traditional positioning pile prediction methods when faced with complex nonlinear data.

[0008] To achieve the above objectives, the technical solution of the present invention is as follows:

[0009] A method for predicting the maximum displacement of positioning piles for dredging vessels includes:

[0010] S1. A coupled numerical model of trolley-positioning pile-soil is constructed based on ABAQUS software to calculate the maximum displacement of the positioning pile under various influencing factors, including the bridge deflection angle, the ship deflection angle, and the wind, wave and current angles.

[0011] S2. Based on the response surface methodology, establish a cubic polynomial regression model of the maximum displacement of the positioning pile with respect to the bridge deflection angle, the ship deflection angle, and the wind, wave, and current angle;

[0012] S3. Randomly generate multiple sets of data samples within the range of the input parameters of the cubic polynomial regression model. The data sample sets include: maximum displacement of the positioning pile, bridge deflection angle, ship deflection angle, and wind, wave, and current angle. Divide the data sample sets into training set, validation set, and prediction set.

[0013] S4. Construct a BP neural network model, train and test the BP neural network model using the training set and validation set, and obtain a trained BP neural network model;

[0014] S5. Input the bridge deflection angle, ship deflection angle and wind, wave and current angle from the prediction set into the trained BP neural network and output the predicted value; calculate the relative error and coefficient of determination between the predicted value and the measured value of the maximum displacement of the positioning pile in the prediction set. If the calculated relative error and coefficient of determination meet the designed threshold, the prediction is deemed valid and the predicted value is output as the result; otherwise, the BP neural network is retrained.

[0015] Furthermore, in the coupled numerical model of the trolley-positioning pile-soil, the Mohr-Coulomb model is used to simulate the constitutive structure of the soil; by refining the mesh in the pile head area and using a mesh division method with uniform mesh density in the pile body and the trolley, uniform stress simulation is achieved.

[0016] Furthermore, the implementation steps for establishing a cubic polynomial regression model of the maximum displacement of the positioning pile with respect to the bridge girder deflection angle, the ship hull deflection angle, and the wind, wave, and current angle based on the response surface method include:

[0017] S21. The bridge deflection angle, hull deflection angle and wind, wave and current angle are used as input variables, and the range of values ​​for the input variables is limited.

[0018] S22. Generate an input set of combinations of input variables through experimental design;

[0019] S23. Calculate the maximum displacement of the positioning pile corresponding to the input set using the coupled numerical model of the trolley-positioning pile-soil.

[0020] S24. Combining the input set and the maximum displacement of the positioning pile, and using a cubic polynomial model, construct the mapping relationship between the bridge deflection angle, the ship deflection angle, and the wind, wave and current angle and the maximum displacement of the positioning pile, i.e., a cubic polynomial regression model.

[0021] Furthermore, the expression for constructing the cubic polynomial regression model is as follows:

[0022]

[0023] In the formula, Y is the maximum displacement of the positioning pile; A is the bridge deflection angle; B is the hull deflection angle; and C is the wind, wave and current angle.

[0024] Furthermore, the relative error is the deviation between the predicted value output by the BP neural network model and the measured value of the predicted maximum displacement of the concentrated positioning piles;

[0025] The coefficient of determination is the goodness of linear fit between the predicted value output by the BP neural network model and the measured value of the maximum displacement of the positioning piles in the prediction set. It is calculated based on the goodness of linear fit formula and is used to quantify the correlation strength between the predicted value and the measured value.

[0026] The present invention has the following beneficial effects:

[0027] This invention combines the Mohr-Coulomb pile-soil model, response surface methodology, and BP neural network to construct a predictive model for the maximum displacement of positioning piles in dredging projects, filling a technological gap in parametric prediction in this field. It establishes a quantitative relationship between bridge deflection angle, hull deflection angle, wind, wave, and current angle, and displacement through a cubic polynomial regression model, replacing the traditional random data input method. A dual verification mechanism (relative error and coefficient of determination) ensures that the prediction results meet engineering accuracy requirements. The rapid and reliable prediction results of this invention can provide a basis for adjusting the operational attitude of dredging vessels, guiding relevant decision-makers to make more informed decisions and ensuring the smooth progress of engineering projects. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a schematic diagram of the displacement prediction method of the present invention;

[0030] Figure 2 This invention provides a numerical model of the vehicle-positioning pile-soil coupling and various load angles.

[0031] Figure 3 This is an overall assembly drawing of the vehicle-positioning pile-soil coupled numerical model of the present invention;

[0032] Figure 4 The displacement of the positioning piles is determined by the different bridge frame deflections and different ship hull deflections of this invention.

[0033] Figure 5 The displacement of the positioning piles is determined by the different wind, wave and current angles and different bridge deflection angles of this invention.

[0034] Figure 6 The displacement of the positioning piles is determined by the different wind, wave, and current angles and different ship hull deflections of this invention.

[0035] Figure 7 The response surface and contour plot of the influence of bridge deflection and hull deflection on the displacement of positioning piles in this invention;

[0036] Figure 8 The response surface and contour plot of the influence of cable tray deflection and wind, wave and current angle on the displacement of positioning piles are shown in the present invention.

[0037] Figure 9The response surface and contour plot of the influence of hull deflection and wind, wave and current angles on the displacement of the positioning pile are shown in the present invention.

[0038] Figure 10 This is a comparison chart of actual and predicted values ​​when inputting the cable tray deflection and wind / wave / current angles for this invention.

[0039] Figure 11 This is a comparison chart of predicted and actual values ​​when inputting the ship's hull deflection and wind, wave and current angles in this invention.

[0040] Figure 12 This is a comparison chart of predicted and actual values ​​when inputting bridge deflection, hull deflection, and wind, wave, and current angles for this invention.

[0041] In the diagram: 1. Trolley section; 2. Positioning pile body; 3. Soil section. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] This embodiment provides a method for predicting the maximum displacement of positioning piles for dredging vessels, such as... Figure 1 As shown, it includes:

[0044] S1. A coupled numerical model of trolley-positioning pile-soil is constructed based on ABAQUS software to calculate the maximum displacement of the positioning pile under various influencing factors, including the bridge deflection angle, the ship deflection angle, and the wind, wave and current angles.

[0045] S2. Based on the response surface methodology, establish a cubic polynomial regression model of the maximum displacement of the positioning pile with respect to the bridge deflection angle, the ship deflection angle, and the wind, wave, and current angle;

[0046] S3. Randomly generate multiple sets of data samples within the range of the input parameters of the cubic polynomial regression model. The data sample sets include: maximum displacement of the positioning pile, bridge deflection angle, ship deflection angle, and wind, wave, and current angle. Divide the data sample sets into training set, validation set, and prediction set.

[0047] S4. Construct a BP neural network model, train and test the BP neural network model using the training set and validation set, and obtain a trained BP neural network model;

[0048] S5. Input the bridge deflection angle, ship deflection angle and wind, wave and current angle from the prediction set into the trained BP neural network and output the predicted value; calculate the relative error and coefficient of determination between the predicted value and the measured value of the maximum displacement of the positioning pile in the prediction set. If the calculated relative error and coefficient of determination meet the designed threshold, the prediction is deemed valid and the predicted value is output as the result; otherwise, the BP neural network is retrained.

[0049] Specifically, firstly, a coupled numerical model of the trolley-positioning pile-soil is constructed based on ABAQUS software to accurately simulate the mechanical response of the positioning pile under the coupled effects of multiple factors such as bridge deflection angle, ship deflection angle, and wind, wave and current angle; secondly, a cubic polynomial regression model of the maximum displacement of the positioning pile with respect to the aforementioned influencing factors is established based on the response surface method. This model reveals the complex mapping relationship between variables through high-order nonlinear fitting.

[0050] Secondly, within the range of model parameter values, multiple sets of data samples containing the maximum displacement of the positioning pile and its influencing parameters are randomly generated and divided into training set, validation set and prediction set; thirdly, a BP neural network model is constructed, and the network weights are optimized using the training set and the hyperparameters are adjusted using the validation set to obtain a prediction model with strong generalization ability.

[0051] Finally, the prediction set is input into the trained BP neural network to generate displacement prediction values. The reliability of the model is verified by calculating the relative error and coefficient of determination between the predicted and measured values. If the threshold is met, the prediction result is output; otherwise, the model is iteratively optimized.

[0052] In a specific embodiment, during the dredging operation, the dredging vessel will inevitably encounter the influence of external environmental loads, and will also have to bear the internal loads generated by its own system operation. The distribution of these loads is as follows: Figure 2 As shown;

[0053] In this embodiment, the Mohr-Coulomb model in ABAQUS software is used as the constitutive simulation of the soil to better analyze the actual impact of pile-soil interaction on the deformation of the positioning pile. To make the simulated stress more uniform, the mesh density is crucial. This is achieved by refining the mesh in the pile head region while maintaining a uniform mesh density in the pile body and the trolley. All element types in the trolley-positioning pile-soil model are C3D8R three-dimensional solid elements. The trolley-positioning pile-soil model is as follows: Figure 3 As shown.

[0054] The Mohr-Coulomb model selected in this embodiment can better reflect the frictional strength characteristics between soil particles, and a smooth potential function is used, which makes the numerical calculation have faster convergence; the trolley material used is S255 steel, and the positioning pile material is S355J2G3.

[0055] During the simulation, the influence of various working conditions on the displacement of the positioning pile was calculated using the coupled numerical model of the trolley-positioning pile-soil, such as... Figure 4 , 5 As shown in Figures 6 and 7, the operating conditions include: bridge deflection angle, hull deflection angle, and wind, wave, and current angle.

[0056] Specifically, the implementation steps for establishing a cubic polynomial regression model of the maximum displacement of the positioning pile with respect to the bridge girder deflection angle, the ship hull deflection angle, and the wind, wave, and current angle based on the response surface method include:

[0057] S21. The bridge deflection angle, hull deflection angle and wind, wave and current angle are used as input variables, and the range of values ​​for the input variables is limited.

[0058] S22. Generate an input set of combinations of input variables through experimental design;

[0059] S23. Calculate the maximum displacement of the positioning pile corresponding to the input set using the coupled numerical model of the trolley-positioning pile-soil.

[0060] S24. Combining the input set and the maximum displacement of the positioning pile, and using a cubic polynomial model, construct the mapping relationship between the bridge deflection angle, the ship deflection angle, and the wind, wave and current angle and the maximum displacement of the positioning pile, i.e., a cubic polynomial regression model.

[0061] In this embodiment, the response surface and contour plot of the influence of the bridge deflection angle, hull deflection angle, and wind, wave, and current angle on the displacement of the positioning pile are as follows: Figure 7 , 8 As shown in Figures 9 and 9.

[0062] Specifically, the expression for constructing the cubic polynomial regression model is as follows:

[0063]

[0064] In the formula, Y is the maximum displacement of the positioning pile; A is the bridge deflection angle; B is the hull deflection angle; and C is the wind, wave and current angle.

[0065] In a specific embodiment, multiple sets of data samples are randomly generated within the range of the input parameters of the cubic polynomial regression model. The data sample sets include: maximum displacement of the positioning piles, bridge deflection angle, ship deflection angle, and wind, wave, and current angles. The data sample sets are then divided into a training set, a validation set, and a prediction set. 70% of the sample data is used as the training set, 15% as the validation set, and 15% as the prediction set. The training set is used to train the BP neural network model, the validation set is used to prevent overfitting of the BP neural network model, and the prediction set is used to verify the prediction results of the BP neural network model.

[0066] In a specific embodiment, a BP neural network model is constructed and trained using a training set. The training process is implemented using MATLAB, and the specific implementation process is shown in Tables 1, 2, and 3.

[0067] Table 1. Values ​​of relevant training parameters for BP.

[0068]

[0069] Table 2 Data Input

[0070]

[0071]

[0072] Table 3 Prediction Implementation Process

[0073]

[0074]

[0075] Specifically, through relative error and coefficient of determination R 2 The stability of the sample and the predictive ability of the BP neural network model are evaluated; the relative error is the deviation between the predicted value output by the BP neural network model and the measured value of the maximum displacement of the predicted piles in the set.

[0076] The coefficient of determination is the goodness of linear fit between the predicted value output by the BP neural network model and the measured value of the maximum displacement of the predicted piles in the prediction set. It is calculated based on the goodness of linear fit formula and is used to quantify the correlation strength between the predicted value and the measured value.

[0077] The formula for the goodness of fit of the linear fit is:

[0078]

[0079] In the formula, T test To predict the measured value of the maximum displacement of the centralized positioning piles; T sim The predicted value output by the BP neural network model; This is the average value of the measured values ​​used to predict the maximum displacement of the centralized positioning piles;

[0080] In this embodiment, because the generated random sample data is relatively discrete and fluctuates significantly, the range of the relative error is slightly larger, and the coefficient of determination R... 2 A value greater than 0.70 indicates that the sample has good stability and the neural network has good predictive ability. A relative error range of 0-15% is considered reasonable.

[0081] In this embodiment, the comparison between the predicted and measured values ​​of the BP neural network model is as follows: Figure 10 , 11 As shown in Figure 12; Figure 10 and 11 When the bridge deflection angle and wind, wave and current direction angle are used as inputs, the predicted maximum displacement of the positioning piles shows a significant consistency with the measured values, and the predicted results are generally close to the true values.

[0082] like Figure 12 When the input parameters are the bridge deflection angle, the ship deflection angle, and the wind, wave, and current angle, the calculated maximum displacement of the positioning pile is consistent with the trend of the measured value. By inputting the three factors, the predicted value is close to the actual value, indicating that the prediction method has high accuracy and practicality for the prediction results of two factors and three input factors, reflecting the high calculation efficiency and ease of use of the prediction method.

[0083] The present invention has the following beneficial effects:

[0084] This invention combines the Mohr-Coulomb pile-soil model, response surface methodology, and BP neural network to construct a predictive model for the maximum displacement of positioning piles in dredging projects, filling a technological gap in parametric prediction in this field. It establishes a quantitative relationship between bridge deflection angle, hull deflection angle, wind, wave, and current angle, and displacement through a cubic polynomial regression model, replacing the traditional random data input method. A dual verification mechanism (relative error and coefficient of determination) ensures that the prediction results meet engineering accuracy requirements. The rapid and reliable prediction results of this invention can provide a basis for adjusting the operational attitude of dredging vessels, guiding relevant decision-makers to make more informed decisions and ensuring the smooth progress of engineering projects.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the maximum displacement of positioning piles for dredging vessels, characterized in that, include: S1. A coupled numerical model of trolley-positioning pile-soil is constructed based on ABAQUS software to calculate the maximum displacement of the positioning pile under various influencing factors, including the bridge deflection angle, the ship deflection angle, and the wind, wave and current angles. S2. Based on the response surface methodology, establish a cubic polynomial regression model of the maximum displacement of the positioning pile with respect to the bridge deflection angle, the ship deflection angle, and the wind, wave, and current angle; S3. Randomly generate multiple sets of data samples within the range of the input parameters of the cubic polynomial regression model. The data sample sets include: maximum displacement of the positioning pile, bridge deflection angle, ship deflection angle, and wind, wave, and current angle. Divide the data sample sets into training set, validation set, and prediction set. S4. Construct a BP neural network model, train and test the BP neural network model using the training set and validation set, and obtain a trained BP neural network model; S5. Input the bridge deflection angle, ship deflection angle and wind, wave and current angle from the prediction set into the trained BP neural network and output the predicted value; calculate the relative error and coefficient of determination between the predicted value and the measured value of the maximum displacement of the positioning pile in the prediction set. If the calculated relative error and coefficient of determination meet the designed threshold, the prediction is deemed valid and the predicted value is output as the result; otherwise, the BP neural network is retrained.

2. The method for predicting the maximum displacement of positioning piles for dredging vessels according to claim 1, characterized in that, In the coupled numerical model of the trolley-positioning pile-soil, the Mohr-Coulomb model is used to simulate the constitutive structure of the soil; by refining the mesh in the pile head area and using a mesh division method with uniform mesh density in the pile body and the trolley, uniform stress simulation is achieved.

3. The method for predicting the maximum displacement of positioning piles for dredging vessels according to claim 1, characterized in that, The implementation steps for establishing a cubic polynomial regression model of the maximum displacement of the positioning pile with respect to the bridge deflection angle, the ship deflection angle, and the wind, wave, and current angle based on the response surface method include: S21. The bridge deflection angle, hull deflection angle and wind, wave and current angle are used as input variables, and the range of values ​​for the input variables is limited. S22. Generate an input set of combinations of input variables through experimental design; S23. Calculate the maximum displacement of the positioning pile corresponding to the input set using the coupled numerical model of the trolley-positioning pile-soil. S24. Combining the input set and the maximum displacement of the positioning pile, and using a cubic polynomial model, construct the mapping relationship between the bridge deflection angle, the ship deflection angle, and the wind, wave and current angle and the maximum displacement of the positioning pile, i.e., a cubic polynomial regression model.

4. The method for predicting the maximum displacement of positioning piles for dredging vessels according to claim 1 or 3, characterized in that, The expression for constructing the cubic polynomial regression model is as follows: In the formula, Y is the maximum displacement of the positioning pile; A is the bridge deflection angle; B is the hull deflection angle; and C is the wind, wave and current angle.

5. The method for predicting the maximum displacement of positioning piles for dredging vessels according to claim 1, characterized in that, The relative error is the deviation between the predicted value output by the BP neural network model and the measured value of the predicted maximum displacement of the concentrated positioning piles. The coefficient of determination is the goodness of linear fit between the predicted value output by the BP neural network model and the measured value of the maximum displacement of the positioning piles in the prediction set. It is calculated based on the goodness of linear fit formula and is used to quantify the correlation strength between the predicted value and the measured value.

Images