A method and system for monitoring the stress of a dredger leg

Abstract

The application provides a kind of dredger leg force monitoring method and system, the method comprises: based on pile-soil interaction theory, construct a finite element model, use hypercube sample sampling to generate samples covering various soil, water depth, depth of penetration and ship motion parameters, input model to obtain pile leg force and seabed displacement angle results, accordingly train neural network back-propagation model.When monitoring, obtain ship horizontal displacement, horizontal angle and upper and lower guide force data, calculate horizontal angle and correct pile leg deformation angle through guide clearance, obtain equivalent horizontal force from guide force, input back-propagation model combined with depth of penetration and water depth, output pile leg force state and seabed displacement angle.The method sets monitoring points at the guide connection, combines finite element and neural network, overcomes the defects that sensors are easily damaged and monitoring is not comprehensive, uses P-Y curve to represent pile-soil interaction, realizes accurate calculation of overall pile leg force and seabed response.

Description

Technical Field

[0001] This application belongs to the field of ship pile monitoring, and in particular relates to a method and system for monitoring the stress on the legs of dredger piles. Background Technology

[0002] Dredger leg stress monitoring technology is used to monitor the stress state of the legs, which are the only fixed support components of the dredger, in real time during dredging operations, aiming to ensure the operational safety and service life of the dredger.

[0003] The existing technology involves directly installing stress or strain sensors on the dredger's leg structure. By drilling holes and wiring in the cylindrical steel leg structure, sensors are installed to directly measure stress or strain data at local locations on the leg. This method allows for direct, localized monitoring of the stress state of the leg by obtaining local stress or strain information at the sensor locations.

[0004] The above-mentioned technical solution has a special cylindrical pile leg structure. The placement of sensors on its body requires drilling and wiring, which will damage the integrity of the pile leg structure and thus affect its load-bearing capacity. The underwater part of the pile leg is in a complex and harsh underwater environment for a long time, and the sensors installed on it are easily damaged, resulting in decreased monitoring data accuracy, shortened equipment lifespan, and difficult and costly maintenance. More importantly, existing technologies have failed to fully consider the complex effects of the ship's rolling and pitching motions on the force transmission path of the pile legs, nor have they conducted quantitative analysis on the interaction between the pile legs and the seabed and its constraint effect on the force on the pile legs. Summary of the Invention

[0005] The purpose of this application is to overcome the defects in the prior art and provide a method and system for monitoring the stress on the legs of a dredger.

[0006] This application provides a method for monitoring the stress on the legs of a dredger, including: A finite element model is constructed based on the pile-soil interaction theory. A hypercube sampling method is used to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, ship motion parameters, and guiding force parameters. The sample input parameters are input into the finite element model to obtain the corresponding pile leg force and seabed top surface displacement and rotation results. A neural network is trained based on the sample input parameters and the results to generate a neural network back-reasoning model. Acquire hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data. Process the acquired hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data into non-directional data. Correct the hull horizontal rotation angle by the guide clearance to obtain the pile leg horizontal rotation angle data. Calculate the equivalent horizontal force based on the upper guide force data and lower guide force data. Acquire the user-input pile leg penetration depth and water depth. The deformation angle of the pile leg, the equivalent horizontal force, the depth of the pile leg into the mud, and the water depth are input into the neural network back-reasoning model to generate the stress state of the pile leg and the displacement and rotation angle of the seabed top surface.

[0007] Optionally, after generating the stress state of the pile leg and the displacement and rotation angle of the seabed top surface, the method further includes: The maximum stress value is determined from the stress state of the pile leg, and the maximum stress value is compared with a first safety threshold set based on the yield strength of the pile leg material; The displacement and rotation angle of the seabed top surface are compared with a second safety threshold set based on the limits of the soil PY curve; An abnormal warning is triggered when the maximum stress value exceeds the first safety threshold, or when the displacement and rotation angle exceed the second safety threshold.

[0008] Optionally, a finite element model is constructed based on the pile-soil interaction theory, including: The PY curve, which characterizes the nonlinear resistance relationship between the soil and the pile leg, is used as the input to the constitutive relationship of the seabed soil and embedded in the finite element model.

[0009] Optionally, a hypercube sampling method is used to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, hull motion parameters, and guiding force parameters, including: Soil type and its corresponding PY curve attributes, water depth, pile leg penetration depth, assumed horizontal displacement and horizontal rotation of the hull, assumed upper guide equivalent horizontal force, and assumed lower guide equivalent horizontal force are used as multiple dimension parameters. The hypercube sampling method is used to systematically combine the multiple dimensional parameters to generate a set of sample input parameters that uniformly covers the multidimensional working space.

[0010] Optionally, the horizontal rotation angle data is corrected by guide clearance to obtain the pile leg deformation angle, i.e., the pile leg horizontal rotation angle data, including: Based on the known total mechanical clearance between the pile leg and the guide device, and the vertical distance between the upper and lower guide devices, the measurement angle deviation caused by the total mechanical clearance is calculated. The deformation angle of the pile leg is obtained by subtracting the measured angle deviation from the horizontal rotation angle data.

[0011] This application also provides a dredging vessel leg stress monitoring system, including: The model training module is used to construct a finite element model based on the pile-soil interaction theory, and to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, ship motion parameters and guiding force parameters using the hypercubic sample sampling method. The sample input parameters are input into the finite element model to obtain the corresponding pile leg force and seabed top surface displacement and rotation results. Based on the sample input parameters and the results, a neural network is trained to generate a neural network back-reasoning model. The data measurement module is used to acquire hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data. It processes the acquired hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data into non-directional data, corrects the hull horizontal rotation angle data for guide clearance to obtain pile leg horizontal rotation angle data, calculates the equivalent horizontal force based on the upper guide force data and lower guide force data, and acquires the pile leg penetration depth and water depth input by the user. The processing module is used to input the deformation angle of the pile leg, the equivalent horizontal force, the depth of the pile leg into the mud, and the water depth into the neural network back-reasoning model to generate the stress state of the pile leg and the displacement and rotation angle of the seabed top surface.

[0012] Optionally, after generating the stress state of the pile leg and the displacement and rotation angle of the seabed top surface, the method further includes: The maximum stress value is determined from the stress state of the pile leg, and the maximum stress value is compared with a first safety threshold set based on the yield strength of the pile leg material; The displacement and rotation angle of the seabed top surface are compared with a second safety threshold set based on the limits of the soil PY curve; An abnormal warning is triggered when the maximum stress value exceeds the first safety threshold, or when the displacement and rotation angle exceed the second safety threshold.

[0013] Optionally, a finite element model is constructed based on the pile-soil interaction theory, including: The PY curve, which characterizes the nonlinear resistance relationship between the soil and the pile leg, is used as the input to the constitutive relationship of the seabed soil and embedded in the finite element model.

[0014] Optionally, a hypercube sampling method is used to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, hull motion parameters, and guiding force parameters, including: Soil type and its corresponding PY curve attributes, water depth, pile leg penetration depth, assumed horizontal displacement and horizontal rotation of the hull, assumed upper guide equivalent horizontal force, and assumed lower guide equivalent horizontal force are used as multiple dimension parameters. The hypercube sampling method is used to systematically combine the multiple dimensional parameters to generate a set of sample input parameters that uniformly covers the multidimensional working space.

[0015] Optionally, the horizontal rotation angle data is corrected by guide clearance to obtain the pile leg deformation angle, i.e., the pile leg horizontal rotation angle data, including: Based on the known total mechanical clearance between the pile leg and the guide device, and the vertical distance between the upper and lower guide devices, the measurement angle deviation caused by the total mechanical clearance is calculated. The deformation angle of the pile leg is obtained by subtracting the measured angle deviation from the horizontal rotation angle data.

[0016] The beneficial effects of this application are: This application provides a method for monitoring the stress on the legs of a dredging vessel, comprising: constructing a finite element model based on the pile-soil interaction theory, and using a hypercubic sampling method to generate sample input parameters covering various soil conditions, water depth, leg penetration depth, hull motion parameters, and guiding force parameters; inputting the sample input parameters into the finite element model to obtain the corresponding results of leg stress and seabed top surface displacement and rotation; training a neural network based on the sample input parameters and the results to generate a neural network back-reasoning model; acquiring hull horizontal displacement data, horizontal rotation data, upper guiding force data, and lower guiding force data; correcting the hull horizontal rotation data by guiding clearance to obtain the leg deformation angle, i.e., the horizontal rotation data; calculating the equivalent horizontal force based on the upper and lower guiding force data; and acquiring the leg penetration depth and water depth input by the user; inputting the leg deformation angle, the equivalent horizontal force, the leg penetration depth, and the water depth into the neural network back-reasoning model to generate the stress state of the leg and the displacement and rotation of the seabed top surface. This application achieves monitoring of the stress state of the pile legs by setting monitoring points at the connection between the hull and the pile legs and the guiding device, and combining this with a neural network back-calculation model constructed based on finite element method and hypercubic samples. This addresses the shortcomings of pile leg structural damage, sensor susceptibility to damage, and incomplete monitoring. By considering the characteristic that the pile legs only bear horizontal forces and bending moments, and using the PY curve to characterize the pile-soil interaction, it achieves accurate calculation of the overall stress state of the pile legs and the displacement and rotation angle at the top of the seabed. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the stress monitoring process for the dredger's pile legs in this application; Figure 2 This is an overall schematic diagram of the stress monitoring of the dredger's leg piles in this application; Figure 3 This is a schematic diagram showing the location of the monitoring points in this application; Figure 4 This is a schematic diagram illustrating the logic of hypercubic sample sampling, theoretical analysis, and force inversion in this application; Figure 5 This is a schematic diagram of the dredging vessel leg stress monitoring system in this application. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be provided in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it is to be understood that various forms of implementation of the present disclosure are intended and should not be limited to the embodiments set forth herein. Rather, the embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0019] This application provides a method for monitoring the stress on the legs of a dredger, which is applied to the field of dredger operation safety monitoring technology. It is used to solve the problems that when dredging vessels operate, directly placing sensors on the cylindrical leg body to monitor the stress on the legs will damage the structural integrity of the legs, the sensors are susceptible to corrosion and wear in the underwater environment, and the existing technology does not fully consider the influence of the ship's attitude, including roll and pitch, on the stress on the legs, and it is difficult to quantify the interaction between the legs and the seabed, so it cannot comprehensively and accurately reflect the overall stress state of the legs and warn of potential safety hazards.

[0020] Please refer to Figures 1-3 As shown, this application provides a method for monitoring the stress on the legs of a dredger, comprising: like Figure 3 As shown, upper and lower guide sensors are installed at the connection between the hull and the pile legs below the still water surface. In the cross-sectional view AA of the upper guide device on the pile leg, four pressure sensors are arranged in a cross or ring around the pile leg, namely pressure sensor 1, pressure sensor 2, pressure sensor 3 and pressure sensor 4.

[0021] S101. A finite element model is constructed based on the pile-soil interaction theory, and a hypercube sampling method is used to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, ship motion parameters, and guiding force parameters. The sample input parameters are input into the finite element model to obtain the corresponding pile leg force and seabed top surface displacement and rotation results. Based on the sample input parameters and the results, a neural network is trained to generate a neural network back-reasoning model.

[0022] Based on rigid body mechanics, mechanics of materials, and the theory of pile-soil interaction, a numerical model, namely the finite element model, is constructed to analyze the stress and deformation of the pile leg. The PY curve, which characterizes the nonlinear resistance relationship between the soil and the pile leg, is used as input to the constitutive relationship of the seabed soil and embedded in the finite element model. The PY curve characterizes the force and displacement relationship between the soil and the pile leg, accurately describing the constraint effect of the seabed on the pile leg.

[0023] Please refer to Figure 4 As shown, hypercube sampling includes: In the uplink, monitoring data first enters the data preprocessing stage. In the downlink, a finite element model is first established, and a large amount of sample data is generated using the hypercubic sampling method. This sample data is used to train the neural network model, thereby establishing a mapping relationship from the monitoring data to the stress state of the pile legs. The trained model is finally used for the inversion calculation of pile leg stress; by inputting the preprocessed monitoring data, the final calculation results can be quickly output.

[0024] Then, combining the mechanical parameters of various typical soils and their corresponding PY curve properties, and considering the possible range of actual operation scenarios, the water depth, pile leg penetration depth, and hull horizontal displacement are preset. and horizontal corners The connection point between the pile leg and the upper guide device is subject to horizontal force. The connection between the pile leg and the lower guide device is subject to horizontal force. The system can take multiple values ​​to systematically generate a set of sample input parameters that cover the above-mentioned multidimensional parameters.

[0025] The six independent sampling dimensions are as follows: Seabed soil conditions, including typical soil mechanical parameters and PY curves, water depth, pile leg penetration depth, and ship horizontal displacement. and horizontal corners The connection point between the pile leg and the upper guide device is subject to horizontal force. The connection between the pile leg and the lower guide device is subject to horizontal force. .

[0026] Dredging vessels do not experience significant draft changes during normal operations, and their impact on the stress on the pile legs is negligible. Therefore, when constructing the sample space, draft was not sampled separately to compress the dimensionality of the hypercube sample space.

[0027] The multiple dimensional parameters are systematically combined to ensure that the sample covers various working conditions that may occur in actual operations, while taking into account the uniformity and representativeness of the sample, so as to provide comprehensive basic data for subsequent theoretical analysis, thereby generating the sample input parameter set that uniformly covers the multi-dimensional working condition space.

[0028] Using hypercube sampling parameters as input, the finite element model is used to perform batch calculations to obtain data on the force on the pile leg, deformation of the pile leg, displacement and rotation of the top surface of the seabed, reaction force at the guide position, and stress distribution of the pile leg under the corresponding working conditions. This data is then used to construct a neural network training dataset.

[0029] Using the working condition parameters of hypercube sampling, i.e. the sample input parameters, as input features, and the pile leg force, deformation, stress, mud surface displacement and rotation angle obtained by finite element calculation as output targets, the neural network model is trained and verified, enabling the model to quickly invert the pile leg force based on monitoring parameters.

[0030] The resulting model works in conjunction with the hypercube sampling algorithm. It is constructed based on rigid body mechanics, mechanics of materials, and pile-soil interaction theory. It fully considers the influence of multiple factors such as ship motion, seabed conditions, user-inputted pile leg penetration depth, and water depth on the pile leg stress. In addition, considering the stress characteristics of the pile leg, it only bears horizontal force and horizontal bending moment and is not coupled with the ship in the vertical direction. The core stress is horizontal shear force and bending moment at the upper and lower guides.

[0031] By monitoring three key parameters—the horizontal displacement of the hull, the horizontal force at the connection point between the pile legs and the upper and lower guide devices—and combining them with a mechanical model, we can not only deduce the overall stress on the pile legs, focusing on the horizontal shear force and the bending moment of the upper and lower guide devices, but also further calculate the displacement and rotation at the top surface of the seabed. This allows us to quantify the interaction between the pile legs and the seabed and the influence of seabed constraints on the stress on the pile legs. The stress-displacement relationship between the soil and the pile legs is represented by the PY curve.

[0032] The construction of the neural network back-calculation model includes three logical steps: the construction of the finite element working condition sample space, the training of the neural network model, and the final calculation of the force on the pile leg.

[0033] The finite element working condition sample space construction adopts the hypercubic sampling method, assuming several typical soil types, combining the mechanical parameters and PY curves of each typical soil type, and simultaneously incorporating the user-input range of pile leg penetration depth, water depth, and draft parameters, assuming several water depths, pile leg penetration depths, hull horizontal displacements, and rotation angles that conform to actual operating scenarios. , And the horizontal force at the connection point between the pile leg and the upper and lower guide devices. and The sample was designed to cover all possible working conditions in actual operations, while also ensuring the uniformity and representativeness of the sample. This provides comprehensive basic data for subsequent theoretical analysis. The draft of the dredger varies little, so it was not considered in order to compress the spatial dimension of the hypercubic sample.

[0034] The neural network model training uses the hypercubic working condition samples constructed above as input conditions. Through batch calculations using the finite element model, the results data such as the force on the pile leg, the deformation of the pile leg, the displacement and rotation of the top surface of the seabed, the reaction force at the guide position, and the stress distribution of the pile leg under the corresponding working conditions are obtained, thereby constructing a neural network training dataset.

[0035] Using the working condition parameters obtained from hypercube sampling as input features, and the pile leg stress, deformation, stress, mud surface displacement and rotation angle obtained from finite element calculation as output targets, the neural network model is trained and validated, enabling the model to quickly invert the pile leg stress based on monitoring parameters.

[0036] The logic for the final calculation of the pile leg stress is as follows: In the model application stage, using the horizontal displacement and rotation data at the lower guide of the pile leg obtained in step 2, the horizontal stress at the upper and lower guides, the user-inputted pile leg penetration depth, and the water depth as input parameters, the mechanical parameters of the soil and the PY curve are calculated in reverse, and the overall stress state of the pile leg is determined. Combining the stress characteristics of the pile leg, the horizontal displacement u and rotation angle a of the pile leg at the top surface of the seabed are calculated, and finally, the accurate calculation of the pile leg stress under multiple working conditions is completed. Through this calculation process, the constraint effect of the seabed on the pile leg, the influence of water depth, draft, pile leg penetration depth, and ship attitude are quantified, and the overall analysis of the pile leg stress is improved.

[0037] S102. Acquire the horizontal displacement data, horizontal rotation data, upper guide force data, and lower guide force data of the hull. Process the acquired horizontal displacement data, horizontal rotation data, upper guide force data, and lower guide force data into non-directional data. Correct the horizontal rotation angle of the hull by adjusting the guide clearance to obtain the deformation angle of the pile leg, i.e., the horizontal rotation angle data. Calculate the equivalent horizontal force based on the upper guide force data and lower guide force data, and obtain the depth of the pile leg into the mud and the water depth.

[0038] like Figure 2 As shown, this application constructs a complete process for monitoring and processing the stress on the pile legs.

[0039] The system first acquires raw data from upper guide pressure sensors 1, 3, 2, and 4. It then determines whether the data is valid. Only when all sensor data in the X-direction are greater than the noise threshold will the data be retained as if the force in the X-direction of the pile leg guide is greater than or equal to the noise sensor data. Similarly, only when all sensor data in the Y-direction are greater than the noise threshold will the data be retained as if the force in the Y-direction of the pile leg guide is greater than or equal to the noise sensor data. If the data does not meet the threshold requirements, an error is triggered.

[0040] The data that meets the requirements is further processed to obtain the upper guide force. The system acquires data from the lower guide pressure sensor, and its processing flow is the same as that of the upper guide sensor, ultimately obtaining the lower guide force.

[0041] The system receives horizontal displacement data of the hull, processes it through a coordinate transformation module, and outputs the horizontal displacement at the lower guide of the pile leg. The system also utilizes data from the upper guide pressure sensor, which, after processing by a gap adjustment module, outputs the horizontal displacement of the lower guide component. Finally, the upper guide force, lower guide force, horizontal displacement at the lower guide of the pile leg, and the horizontal displacement of the lower guide component are combined to form a monitoring result data packet.

[0042] A horizontal displacement monitoring point is set on the hull of the dredger to collect horizontal displacement data during the operation. The horizontal displacement monitoring point is located near the pile legs of the hull and near the intersection of the longitudinal centerline and the transverse centerline of the hull to ensure that the collected horizontal displacement data can accurately reflect the displacement changes caused by the hull's roll and pitch.

[0043] Horizontal force monitoring points are set at the connection points between the pile legs and the guide device on the hull, and at the connection points between the pile legs and the guide device on the hull. These points are used to collect horizontal force data at the two connection points. The horizontal force monitoring points use non-contact force sensors that are arranged close to the guide device to avoid damaging the connection structure. At the same time, they ensure that the horizontal force data can be accurately collected and eliminate interference from vertical forces.

[0044] During the operation, the horizontal displacement data of the hull was collected in real time and recorded as follows: The horizontal rotation angle data is denoted as The horizontal displacement data at the guide point under the pile leg is denoted as... The horizontal force data at the upper guide point is recorded as follows: The horizontal force data at the lower guide point is denoted as... .

[0045] Considering that pile legs are generally axisymmetric structures, the collected data are all processed into undirected data to compress the dimensionality of the sample space.

[0046] Considering the small gap between the pile legs and the hull guide, the horizontal rotation angle of the hull will differ from the angle of the pile legs after deformation by a value Δθ. The relationship is as follows: in, This represents the horizontal turning angle data of the ship's hull. This indicates the horizontal rotation angle of the pile leg.

[0047] The formula for calculating Δθ is: Where ΔS is the total gap between the pile leg and the guide, and ΔL is the distance between the upper and lower guides.

[0048] Therefore, correcting the horizontal rotation angle data for guide clearance to obtain the leg deformation angle, i.e., the leg horizontal rotation angle data, is achieved by using the above formula relationship from the measured hull horizontal rotation angle. By removing the influence of gaps, the true deformation angle of the pile leg can be obtained. The process involves calculating the equivalent horizontal force based on the upper and lower guide force data. This means simplifying the force data from multiple sensors at the connection point between the pile leg and the upper guide device into a single equivalent horizontal force. The force data from multiple sensors at the connection point between the pile leg and the lower guide device are simplified into an equivalent horizontal force. Obtaining the depth of the pile leg into the mud and the water depth refers to receiving the two key parameters input by the user: the depth of the pile leg into the mud and the water depth.

[0049] Furthermore, by introducing guide clearance correction and considering the gap between the pile leg and the hull guide, the difference compensation for the horizontal rotation angle can be optimized, which can effectively reduce the calculation deviation caused by the structural gap and make the calculation of the pile leg stress and deformation more accurate.

[0050] S103. Input the deformation angle of the pile leg, the equivalent horizontal force, the depth of the pile leg into the mud, and the water depth into the neural network back-reasoning model to generate the stress state of the pile leg and the displacement and rotation angle of the seabed top surface.

[0051] Among them, the deformation angle of the pile leg is Equivalent horizontal force and .

[0052] Using a pre-trained force inference model, combined with the collected data The process involves using four parameters, along with two key parameters input by the user: the depth of the pile leg into the mud and the water depth, to inversely deduce the horizontal displacement u and rotation angle α of the pile leg at the mud surface, as well as the overall stress state of the pile leg.

[0053] This process can reverse-engineer the soil's mechanical parameters and PY curve, and quantify the influence of the seabed on the pile legs, water depth, draft, pile leg penetration depth, and ship attitude, thus improving the overall analysis of the pile leg stress.

[0054] Furthermore, by employing neural network inversion and simplifying the force input, the data from the four force sensors on each of the upper and lower guides are simplified into equivalent horizontal forces. , This significantly reduces the dimensionality of the sample space and the computational cost of the neural network, improving the inversion speed and engineering practicality.

[0055] Furthermore, this application does not damage the main structure of the pile leg, resulting in higher safety and structural integrity. Monitoring points are only arranged at the hull and guide device locations, eliminating the need for drilling holes or wiring on the cylindrical pile leg, thus not damaging the pile leg structure or affecting its load-bearing capacity.

[0056] Furthermore, this application fully conforms to the actual stress characteristics of dredger legs, making the monitoring more targeted. Based on the stress characteristics of dredger cylindrical legs, which only bear horizontal forces and horizontal bending moments and are not coupled with the vertical direction of the hull, it only monitors and inverts horizontal shear forces and upper and lower guide bending moments, eliminating irrelevant stress components, and the monitoring results are more in line with actual working conditions.

[0057] Furthermore, this application considers pile-soil interaction to achieve quantitative analysis of seabed constraints. The PY curve is used to characterize the force-displacement relationship between the soil and the pile leg, which can accurately calculate the horizontal displacement and rotation angle of the pile leg at the top of the seabed and quantify the influence of seabed constraints on the force of the pile leg.

[0058] Furthermore, the use of hypercube sampling provides comprehensive coverage of working conditions and strong adaptability. Through 6-dimensional independent hypercube samples, typical soil conditions, water depth, pile leg penetration depth, horizontal displacement of the hull, and vertical guide force are covered. The samples are uniform and highly representative, and can adapt to different soil types and water depths in dredging operations without the need for frequent model adjustments.

[0059] Furthermore, the equipment is easy to maintain and has a longer service life. All monitoring points are located on the hull, avoiding contact with underwater silt and corrosive environments. The sensors are not easily damaged and are easy to maintain, significantly reducing long-term operating costs.

[0060] After generating the stress state of the pile leg and the displacement and rotation angle of the seabed top surface, the following steps are also included: The maximum stress value is determined from the stress state of the pile leg, and the maximum stress value is compared with a first safety threshold set based on the yield strength of the pile leg material; The displacement and rotation of the seabed top surface are compared with a second safety threshold set based on the limits of the soil PY curve.

[0061] The first safety threshold is the safety threshold of pile leg stress determined based on the pile leg material, representing the ultimate limit of the pile leg structure. The second safety threshold is determined by analyzing the limit values ​​of horizontal displacement u and rotation angle α at the top surface of the seabed using typical soil PY curves obtained through reverse engineering, representing the mechanical properties of the seabed soil.

[0062] By combining the PY curves of typical soils and the mechanical properties of seabed soil, as well as the overall stress state of the pile legs, the safety thresholds for horizontal shear force and bending moment of the pile legs are determined, serving as two sets of criteria for the selection of subsequent alarm thresholds.

[0063] An abnormal warning is triggered when the maximum stress value exceeds the first safety threshold, or when the displacement and rotation angle exceed the second safety threshold.

[0064] Specifically, when the calculated horizontal displacement u and rotation angle α of the pile leg at the top of the seabed exceed the threshold, it indicates that the seabed constraint has failed or the pile leg insertion depth is insufficient, and an early warning is triggered simultaneously; when the calculated maximum stress of the pile leg is close to its material yield strength, it indicates that the pile leg strength is insufficient, and an early warning is triggered simultaneously.

[0065] After the warning signal is triggered, the staff is reminded to take measures such as stopping the machine and adjusting the working posture in time to avoid damage to the pile legs.

[0066] Furthermore, this application has a dual-dimensional early warning mechanism, which provides more comprehensive and reliable safety protection. At the same time, it sets dual safety thresholds based on the limits of seabed soil and the strength limits of pile leg structure, which can not only determine the failure of seabed constraint and insufficient penetration depth, but also determine the excessive stress of pile leg, thus providing more comprehensive and timely early warning.

[0067] Please refer to Figure 5 As shown, this application also provides a dredging vessel leg stress monitoring system, comprising: The model training module 201 is used to construct a finite element model based on the pile-soil interaction theory, and to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, ship motion parameters and guiding force parameters using the hypercubic sample sampling method. The sample input parameters are input into the finite element model to obtain the corresponding pile leg force and seabed top surface displacement and rotation results. Based on the sample input parameters and the results, a neural network is trained to generate a neural network back-reasoning model. The data measurement module 202 is used to acquire hull horizontal displacement data, upper guide force data, and lower guide force data. It processes the acquired hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data into non-directional data. It corrects the horizontal rotation angle data by the guide clearance to obtain the pile leg deformation angle, i.e., the pile leg horizontal rotation angle data. It calculates the equivalent horizontal force based on the upper guide force data and lower guide force data, and obtains the pile leg penetration depth and water depth. The processing module 203 is used to input the deformation angle of the pile leg, the equivalent horizontal force, the depth of the pile leg into the mud and the water depth into the neural network back-reasoning model to generate the stress state of the pile leg and the displacement and rotation angle of the top surface of the seabed.

[0068] Furthermore, after generating the stress state of the pile leg and the displacement and rotation angle of the seabed top surface, the process also includes: The maximum stress value is determined from the stress state of the pile leg, and the maximum stress value is compared with a first safety threshold set based on the yield strength of the pile leg material; The displacement and rotation angle of the seabed top surface are compared with a second safety threshold set based on the limits of the soil PY curve; An abnormal warning is triggered when the maximum stress value exceeds the first safety threshold, or when the displacement and rotation angle exceed the second safety threshold.

[0069] Furthermore, a finite element model is constructed based on the pile-soil interaction theory, including: The PY curve, which characterizes the nonlinear resistance relationship between the soil and the pile leg, is used as the input to the constitutive relationship of the seabed soil and embedded in the finite element model.

[0070] Furthermore, a hypercube sampling method is employed to generate sample input parameters covering various soil conditions, water depths, pile leg penetration depths, ship motion parameters, and guiding force parameters, including: Soil type and its corresponding PY curve attributes, water depth, pile leg penetration depth, assumed horizontal displacement and horizontal rotation of the hull, assumed upper guide equivalent horizontal force, and assumed lower guide equivalent horizontal force are used as multiple dimension parameters. The hypercube sampling method is used to systematically combine the multiple dimensional parameters to generate a set of sample input parameters that uniformly covers the multidimensional working space.

[0071] Furthermore, the horizontal rotation angle data is corrected by guide clearance to obtain the pile leg deformation angle, i.e., the pile leg horizontal rotation angle data, including: Based on the known total mechanical clearance between the pile leg and the guide device, and the vertical distance between the upper and lower guide devices, the measurement angle deviation caused by the total mechanical clearance is calculated. The deformation angle of the pile leg is obtained by subtracting the measured angle deviation from the horizontal rotation angle data.

[0072] The above embodiments are provided to enable those skilled in the art to understand and apply this application. Those skilled in the art will readily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without inventive effort. Therefore, this application is not limited to the above embodiments, and any improvements and modifications made to this application based on the disclosure thereof should be within the scope of protection of this application.

Claims

1. A method for monitoring the stress on the legs of a dredger, characterized in that, include: A finite element model is constructed based on the pile-soil interaction theory. A hypercube sampling method is used to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, ship motion parameters, and guiding force parameters. The sample input parameters are input into the finite element model to obtain the corresponding pile leg force and seabed top surface displacement and rotation results. A neural network is trained based on the sample input parameters and the results to generate a neural network back-reasoning model. Acquire hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data. Process the acquired hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data into non-directional data. Correct the hull horizontal rotation angle with guide clearance to obtain the pile leg deformation angle, i.e., the pile leg horizontal rotation angle data. Calculate the equivalent horizontal force based on the upper guide force data and lower guide force data. Acquire the user-input pile leg penetration depth and water depth. The deformation angle of the pile leg, the equivalent horizontal force, the depth of the pile leg into the mud, and the water depth are input into the neural network back-reasoning model to generate the stress state of the pile leg and the displacement and rotation angle of the seabed top surface.

2. The method for monitoring the stress on the legs of a dredger according to claim 1, characterized in that, After generating the stress state of the pile leg and the displacement and rotation angle of the seabed top surface, the process also includes: The maximum stress value is determined from the stress state of the pile leg, and the maximum stress value is compared with a first safety threshold set based on the yield strength of the pile leg material; The displacement and rotation angle of the seabed top surface are compared with a second safety threshold set based on the limits of the soil PY curve; An abnormal warning is triggered when the maximum stress value exceeds the first safety threshold, or when the displacement and rotation angle exceed the second safety threshold.

3. The method for monitoring the stress on the legs of a dredger according to claim 1, characterized in that, A finite element model is constructed based on the pile-soil interaction theory, including: The PY curve, which characterizes the nonlinear resistance relationship between the soil and the pile leg, is used as the input to the constitutive relationship of the seabed soil and embedded in the finite element model.

4. The method for monitoring the stress on the legs of a dredger according to claim 1, characterized in that, The hypercubic sampling method is used to generate sample input parameters covering various soil conditions, water depths, pile leg penetration depths, ship motion parameters, and guiding force parameters, including: Soil type and its corresponding PY curve attributes, water depth, pile leg penetration depth, assumed horizontal displacement and horizontal rotation of the hull, assumed upper guide equivalent horizontal force, and assumed lower guide equivalent horizontal force are used as multiple dimension parameters. The hypercube sampling method is used to systematically combine the multiple dimensional parameters to generate a set of sample input parameters that uniformly covers the multidimensional working space.

5. The method for monitoring the stress on the legs of a dredger according to claim 1, characterized in that, The horizontal rotation angle data is corrected by adjusting the guide clearance to obtain the pile leg deformation angle, i.e., the pile leg horizontal rotation angle data, including: Based on the known total mechanical clearance between the pile leg and the guide device, and the vertical distance between the upper and lower guide devices, the measurement angle deviation caused by the total mechanical clearance is calculated. The deformation angle of the pile leg is obtained by subtracting the measured angle deviation from the horizontal rotation angle data.

6. A stress monitoring system for dredger legs, characterized in that, The system is used to perform the dredging vessel leg stress monitoring method according to any one of claims 1-5, including: The model training module is used to construct a finite element model based on the pile-soil interaction theory, and to generate sample input parameters covering various soil conditions, water depth, pile leg penetration depth, ship motion parameters and guiding force parameters using the hypercubic sample sampling method. The sample input parameters are input into the finite element model to obtain the corresponding pile leg force and seabed top surface displacement and rotation results. Based on the sample input parameters and the results, a neural network is trained to generate a neural network back-reasoning model. The data measurement module is used to acquire hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data. It processes the acquired hull horizontal displacement data, horizontal rotation angle data, upper guide force data, and lower guide force data into non-directional data, corrects the hull horizontal rotation angle by guide clearance to obtain pile leg deformation horizontal rotation angle data, calculates the equivalent horizontal force based on the upper guide force data and lower guide force data, and acquires the pile leg penetration depth and water depth input by the user. The processing module is used to input the deformation angle of the pile leg, the equivalent horizontal force, the depth of the pile leg into the mud, and the water depth into the neural network back-reasoning model to generate the stress state of the pile leg and the displacement and rotation angle of the seabed top surface.

7. The dredging vessel leg stress monitoring system according to claim 6, characterized in that, After generating the stress state of the pile leg and the displacement and rotation angle of the seabed top surface, the process also includes: The maximum stress value is determined from the stress state of the pile leg, and the maximum stress value is compared with a first safety threshold set based on the yield strength of the pile leg material; The displacement and rotation angle of the seabed top surface are compared with a second safety threshold set based on the limits of the soil PY curve; An abnormal warning is triggered when the maximum stress value exceeds the first safety threshold, or when the displacement and rotation angle exceed the second safety threshold.

8. The dredging vessel leg stress monitoring system according to claim 6, characterized in that, A finite element model is constructed based on the pile-soil interaction theory, including: The PY curve, which characterizes the nonlinear resistance relationship between the soil and the pile leg, is used as the input to the constitutive relationship of the seabed soil and embedded in the finite element model.

9. A dredging vessel leg stress monitoring system according to claim 6, characterized in that, The hypercubic sampling method is used to generate sample input parameters covering various soil conditions, water depths, pile leg penetration depths, ship motion parameters, and guiding force parameters, including: Soil type and its corresponding PY curve attributes, water depth, pile leg penetration depth, assumed horizontal displacement and horizontal rotation of the hull, assumed upper guide equivalent horizontal force, and assumed lower guide equivalent horizontal force are used as multiple dimension parameters. The hypercube sampling method is used to systematically combine the multiple dimensional parameters to generate a set of sample input parameters that uniformly covers the multidimensional working space.

10. A dredging vessel leg stress monitoring system according to claim 6, characterized in that, The horizontal rotation angle data is corrected by adjusting the guide clearance to obtain the pile leg deformation angle, i.e., the pile leg horizontal rotation angle data, including: Based on the known total mechanical clearance between the pile leg and the guide device, and the vertical distance between the upper and lower guide devices, the measurement angle deviation caused by the total mechanical clearance is calculated. The deformation angle of the pile leg is obtained by subtracting the measured angle deviation from the horizontal rotation angle data.

Images