Intelligent Collaborative Sand Dredging System and Method for Water-Rich Sand Layers

By using an intelligent collaborative sand dredging system and a digital twin model, combined with high-pressure nozzles and a rigid chain telescopic mechanism, the problems of low sand dredging efficiency and difficulty in disturbance control in water-rich sand layers have been solved, achieving efficient and safe underground construction.

CN123109233APending Publication Date: 2026-11-17CCCC FIRST HARBOR ENGINEERING CO LTD +3
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202611016709.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-09
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In sand pumping operations in water-rich sand layers, traditional methods are inefficient and difficult to control the disturbance range, which can easily lead to the formation of underground cavities and ground subsidence. Existing monitoring methods cannot achieve real-time early warning and control.

Method used

An intelligent collaborative sand pumping system is adopted, including a water injection unit, a sand pumping unit, a central controller and a power supply. Combining a digital twin model and an auxiliary jet structure, it achieves precise hydraulic disturbance through high-pressure nozzles and a rigid chain telescopic mechanism, dynamically adjusts porosity distribution and operating parameters, and constructs a multi-objective optimization function for real-time optimization.

Benefits of technology

It improves sand dredging efficiency, reduces the risk of ground subsidence, and enables precise control and dynamic adjustment of underground disturbance, ensuring construction safety and efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123109233A_ABST
    Figure CN123109233A_ABST
Patent Text Reader

Abstract

This invention discloses an intelligent collaborative sand pumping system and method for water-rich sand layers, belonging to the field of lakebed sand pumping technology. The sand pumping system consists of a water injection unit, a sand pumping unit, a central control unit, and a power supply unit. Water injection wells are arranged around the sand pumping wells, each equipped with a retractable high-pressure nozzle to achieve precise hydraulic disturbance of the formation. Concentration and mass sensors are used to collect real-time data on sand and water output. The sand pumping method first calibrates the sand layer permeability coefficient function through indoor experiments; then, a seepage-particle coupled digital twin numerical model is built, and adaptive grid partitioning and L-BFGS optimization algorithms are used to invert the formation porosity; a multi-objective function is constructed using the sand pumping rate, sand carrying rate, and formation horizontal disturbance range to optimize the water injection depth, intensity, and angle parameters in parallel, and the field equipment is controlled in real-time via closed-loop control. This invention transforms passive negative pressure sand pumping into active hydraulic sand guiding, balancing sand pumping efficiency and formation disturbance control, reducing the risk of lakebed subsidence and collapse, and solving the model lag problem caused by the nonlinear evolution of formation parameters.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the technical field of lake bottom sand dredging equipment, and relates to an intelligent collaborative sand dredging system and method for water-rich sand layers. Background Technology

[0002] Traditional techniques face significant challenges when dredging sand layers, especially those located beneath clay layers at the lake bottom. Because sand layers exhibit fluid-like properties when infiltrated by high-pressure underground water, traditional methods often operate in a "passive stress" state, relying solely on negative pressure extraction without effectively guiding the movement of particles within the sand layer. This not only leads to low dredging efficiency but also makes it extremely difficult to control the extent of ground disturbance. The most serious problem is that this uncontrollable disturbance easily creates cavities of varying sizes underground. Under the overburden pressure of the clay layer, once these cavities become unstable, they can induce uneven settlement of the ground or lakebed, and even trigger a complete collapse. Furthermore, key parameters such as formation porosity and permeability coefficient exhibit strong nonlinear characteristics and evolve in real time with the dredging process. Traditional monitoring methods and static models struggle to achieve real-time synchronization between the model and field data, causing construction plans to lag behind actual geological changes and failing to provide forward-looking early warnings and control recommendations.

[0003] Prior to this patent, existing patents had attempted to address the monitoring issues in underground engineering by introducing digital twin or numerical simulation technologies. For example, Chinese patent CN112900504A discloses a method and system for identifying leakage in a foundation pit cutoff curtain. This method combines automated monitoring with numerical simulation, using a gradient descent method to invert the permeability coefficients of the soil layer and the curtain. However, its core objective is limited to the "identification" and "early warning" of leakage points; essentially, it is a post-event or in-event monitoring tool, focusing on evaluating the water-stopping effect of existing structures rather than actively generating optimization tools. It cannot physically intervene in the underground flow field, making it difficult to establish a dynamic feedback adjustment mechanism between sand-dredging efficiency and disturbance control.

[0004] In response to the above situation, this paper proposes an intelligent collaborative sand pumping system and method for water-rich sand layers, which makes up for the shortcomings of previous methods in this field. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent collaborative sand pumping system and method for water-rich sand layers.

[0006] The technical problem solved by this invention is achieved through the following technical solution:

[0007] A smart collaborative sand pumping system for water-rich sand layers is characterized by comprising a water injection unit, a sand pumping unit, a central controller, and a power supply. The water injection unit includes an injection well, an injection pipe, a water pump, and a storage tank. The sand pumping unit includes a sand pumping well, a sand collection pipe, a sand pumping vacuum pump, and a sand collection tank. The sand pumping well is installed in the water body, and a concentration detector is installed at the bottom of the sand pumping well. Water injection wells are set around the sand pumping well, and the water injection wells are connected to each other through injection pipes. One of the injection pipes is connected to the storage tank through a pipe connected to the water pump. The sand collection pipe and the sand pumping vacuum pump are connected sequentially through the sand collection pipe, and a mass sensor is installed at the bottom of the sand collection tank. The central controller is connected to the mass sensor and the water pump through data cables, and the power supply is used to power the central controller.

[0008] Furthermore, the distance between the sand extraction well and each water injection well is 10m.

[0009] Furthermore, the water injection well includes a well casing body, a main water injection pipe, and an auxiliary water injection pipe. The main water injection pipe is coaxially mounted inside the well casing body, and auxiliary water injection pipes are mounted inside the well casing body on both sides of the main water injection pipe. A central water injection hole is arranged longitudinally at intervals in the part of the well casing body that extends into the sand layer. Auxiliary jet holes are arranged longitudinally at intervals on both sides of the central water injection hole. The central water injection hole is connected to the main water injection pipe through a connecting pipe. A controlled valve and a double gate valve are installed sequentially from the inside to the outside of the connecting pipe. A high-pressure nozzle is installed in the auxiliary jet hole, and the high-pressure nozzle is connected to the auxiliary water injection pipe through a high-pressure hose.

[0010] Furthermore, it also includes a rigid chain telescopic mechanism. A rigid chain telescopic mechanism is provided on the inner wall of the well casing at the position corresponding to the auxiliary jet hole. The high-pressure hose is connected to the high-pressure nozzle through the rigid chain telescopic mechanism, and the rigid chain telescopic mechanism drives the high-pressure nozzle to perform telescopic movements.

[0011] Furthermore, the rigid chain telescopic mechanism includes a housing, a rigid chain, gears, a chain guide groove, a chain pusher, and an electric motor. The chain guide groove is located within the housing, and the rigid chain is slidably installed within it. The electric motor is connected to the gear, whose teeth mesh with the rigid chain. The chain pusher is rigidly connected to the end of the rigid chain. The end of the high-pressure hose is inserted into the housing and installed on the rigid chain, extending and retracting within the housing along with the rigid chain. The high-pressure nozzle is installed on the end of the rigid chain's sidewall. Through the extension and retraction of the rigid chain, the high-pressure nozzle is driven to pass through the auxiliary jet hole and overcome sand layer resistance to enter the formation, achieving precise point-to-point hydraulic disturbance.

[0012] A smart collaborative sand pumping method for water-rich sand layers, characterized in that the method includes the following steps:

[0013] 1) Obtain samples of the underlying sand layer from field drilling and conduct experiments on sand layers with different porosities. n and head gradient i The permeability test was conducted to obtain the corresponding permeability coefficient. k , fitting k and n , i Functional relationship ;

[0014] 2) The water injection well and the sand extraction well are inserted through the overlying clay layer into the underlying sand layer, and several of the water injection wells are arranged around the sand extraction well;

[0015] 3) Select operating parameters based on engineering experience ,in, H For the depth of the water inlet, p Water injection intensity (water injection volume per unit time) θ To determine the water injection opening angle, a mass sensor below the sand collection tank is continuously monitored to obtain a time-history curve of the total sand and water volume. The rate of sand and water mixture extraction can be obtained through differential calculations. Based on the concentration monitor at the sand extraction wellhead, the sand extraction rate and water extraction rate can be calculated separately. A collaborative sand extraction digital twin model system is set up, constructed, and launched according to the field operating parameters, using the porosity of the underlying sand layer in the geological survey report as the initial value of the permeability coefficient. n 0 is assigned to each element of the seepage grid;

[0016] 4) Using the porosity of each unit as a variable, the porosity of each seepage grid unit is adjusted through an optimization algorithm, based on... Calculate the seepage field and simulate the sand pumping process. And through self-consistent iterative calculation, the head gradient that conforms to the current porosity distribution is obtained in real time. i and their corresponding permeability coefficients k This minimizes the error between the sand discharge rate and water discharge rate at the top port of the sand pumping well obtained by the numerical simulation system and the corresponding rate of the physical sand pumping system. The error calculation formula is as follows: ,in, Q 实 This represents the actual sand extraction rate. w 实 This represents the actual pumping rate. Q 模 To simulate sand extraction rate, w 模 To simulate pumping rate, and The preset weighting coefficients ( By deploying multiple computational scenarios to obtain the porosity sensitivity gradient, the optimal porosity distribution n for the site is obtained through inversion correction. opt =[ n 1,n 2, n 3…, n m ], m This represents the number of seepage grid cells;

[0017] 5) Based on the porosity distribution obtained in step 4), adjust the boundary conditions of the numerical analysis model corresponding to the operating parameters to achieve the following multi-objective optimization function. J Optimal: The multi-objective optimization function J The simulated sand dredging rate (Sand production volume per unit time at the top of the sand-dredging well calculated by the numerical simulation system), and the horizontal disturbance range of the underlying sand layer. R h Weighted composition and simulated sand carrying rate ,and R h The weights are set to the maximum, and the optimal operating parameters are obtained through parallel simulation calculations. ;

[0018] 6) Based on the optimal operating parameters obtained in step 5) Adjust the operating parameters of the physical sand dredging system and continuously monitor on-site data; when the error d When the preset allowable error threshold is exceeded, it is determined that the actual porosity distribution of the sand layer has undergone a non-negligible change. At this point, the process jumps to step 4) to re-invert and correct the site porosity distribution.

[0019] Furthermore, the numerical simulation system includes a solid grid simulating the overlying clay layer, a solid particle aggregate simulating the underlying sand layer, a fluid grid simulating seepage, the well wall of the pumping well, the well wall of the injection well, and boundary conditions.

[0020] Furthermore, step 4) involves dynamic mesh optimization based on the spatial distribution of the porosity, according to the porosity at the current moment in the model. n Change in porosity compared to the previous round of optimization Partitioning: Continuous cells are clustered into a partition, where all grids within that partition share a single porosity and permeability coefficient; if , If a preset grid division threshold is used, the porosity and permeability coefficient of that grid are calculated and inverted separately.

[0021] Moreover, the multi-objective optimization function J The component method is as follows: extract the set of elements in the numerical model whose porosity changes exceed a preset threshold, calculate the geometric envelope of this set on the horizontal projection plane, and define it as the porosity horizontal perturbation range. R The volume of sand produced per unit time at the top of the sand-dredging well, calculated by the numerical simulation system. The sand carrying capacity at the top of the sand-dredging well calculated by the numerical simulation system. Using formulas Calculate the overall score for each working condition, where, , and As the weighting coefficient, and set The maximum priority is given to water injection schemes that focus on horizontal sand transport, guided by the system.

[0022] Furthermore, the inversion correction process described in step 4) uses numerical perturbation to obtain the gradient, and the calculation equation is as follows:

[0023] i) Definition of relative error:

[0024]

[0025] in, Q act This represents the actual sand extraction rate. w act This represents the actual pumping rate. Q sim To simulate sand extraction rate, w sim To simulate pumping rate, and The preset weighting coefficients ( );

[0026] ii) Sensitivity gradient vector calculation:

[0027] Parallel deployment of 1+1+ within each inversion step N One calculation condition, N The number of regions is determined by the porosity of the site in order to obtain the sensitivity gradient vector. :

[0028]

[0029] In the formula, , This is the baseline error calculated based on the current porosity distribution; For the first i The gradient components of each porosity-divided region are determined by applying local porosity perturbations. Calculated;

[0030] iii) Parameter inversion equation:

[0031]

[0032] in, For the search step size, This is a correction matrix generated based on an optimization algorithm to coordinate the correction magnitudes of different partitions.

[0033] Moreover, step 5) specifically includes:

[0034] i) Constructing the water injection parameter space: Selecting the water injection depth of the injection well H Water injection intensity p and the direction of water injection θ As optimization variables, they are discretized and combined within a preset value range to generate variables containing... m A set of working conditions to be calculated ;

[0035] ii) Parallel computing mode: Adjust the boundary conditions of the numerical analysis model corresponding to the operating parameters to perform parallel computing, and calculate the multi-objective optimization function value corresponding to each mode. J i ;

[0036] iii) Parameter locking and output: Extract the global maximum objective function value using a maximum value retrieval algorithm. ;Will Corresponding working conditions Record the optimal water injection parameter combination for the current round. And output it to the actuator.

[0037] Moreover, the correction matrix in the inversion process The generation algorithm is selected from L-BFGS, Adam, RMSProp, or the standard BFGS algorithm.

[0038] Moreover, the correction matrix The generation algorithm is selected from any one of the L-BFGS algorithm, Adam algorithm, RMSProp algorithm, or standard BFGS algorithm, or an improved adaptive optimization algorithm based on the principles of the above algorithms; the algorithm stores the gradients of historical iterations. With parameters n The change in porosity is used to approximate the inverse of the Hessian matrix or to calculate the second-order momentum, so as to dynamically adjust the update weights and search directions of different porosity components.

[0039] The advantages and positive effects of this invention are:

[0040] This invention possesses strong active guidance capabilities. By constructing a collaborative sand-dredging digital twin model and introducing an auxiliary jet structure, it achieves a shift from passive monitoring to actively inducing the "shear effect," effectively guiding sand particles towards the dredging well. Simultaneously, this invention offers high operational efficiency and good safety. The multi-objective optimization function prioritizes the horizontal disturbance range, significantly reducing the risk of ground subsidence caused by underground cavities while improving dredging efficiency. Furthermore, this invention employs adaptive zoning and accelerated optimization algorithms (such as L-BFGS and Adam) to address the slow convergence problem of strongly nonlinear formation parameter inversion, supporting high-frequency dynamic optimization decisions. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the sand dredging system of the present invention;

[0042] Figure 2 This is a schematic cross-sectional view of the sand dredging site of the present invention.

[0043] Figure 3 This is a schematic diagram of the water injection well of the present invention;

[0044] Figure 4 This is a cross-sectional view of the water injection well of the present invention;

[0045] Figure 5 This is a schematic diagram of the assembly of the auxiliary water injection pipe and the rigid chain telescopic mechanism of the present invention;

[0046] Figure 6 This is a schematic diagram of the rigid chain telescopic mechanism of the present invention;

[0047] Figure 7 This is a schematic diagram of porosity partitioning in the digital twin model of this invention;

[0048] Figure 8 The relative error of the ninth partition of this invention d Schematic diagram showing the relationship between porosity and porosity;

[0049] Figure 9 The relative error of this invention d A schematic diagram illustrating the relationship between the number of inversions and the number of iterations;

[0050] Figure 10 This is a flowchart of the method of the present invention.

[0051] Explanation of reference numerals in the attached figures:

[0052] 1-Sand dredging vacuum pump, 2-Central controller, 3-Power supply, 4-Water storage tank, 5-Water pump, 6-Sand collection pipe, 7-Ship, 8-Sand dredging well, 9-Water injection pipe, 10-Water injection well, 11-Sand collection bucket, 12-Concentration detector, 13-Clay layer, 14-Sand layer, 15-Auxiliary water injection pipe, 16-Well casing, 17-Main water injection pipe, 18-Central water injection hole, 19-High pressure nozzle, 20-Auxiliary jet hole, 21-Splitting gate valve, 22-Controlled valve, 23-Connecting pipe, 24-Rigid chain telescopic mechanism, 25-Chain pusher, 26-Rigid chain, 27-Chain guide groove, 28-Outer shell, 29-Gear, 30-Electric motor. Detailed Implementation

[0053] The embodiments of the present invention will be further described in detail below with reference to the accompanying drawings:

[0054] An intelligent collaborative sand pumping system for water-rich sand layers is innovative in that it includes a vessel 7, a water injection unit, a sand pumping unit, a central controller 2, and a power supply 3. The water injection unit includes a water injection well 10, a water injection pipe 9, a water pump 5, and a water storage tank 4. The sand pumping unit includes a sand pumping well 8, a sand collection pipe 6, a sand pumping vacuum pump 1, and a sand collection tank 11. A sand pumping well is installed within the water body, and a concentration detector 12 is installed at the bottom of the sand pumping well. Figure 1 As shown, the water pump, water storage tank, central controller, power supply, sand collection pipeline, sand dredging vacuum pump, and sand collection bucket are mounted on the ship; the number of water injection wells in the above-mentioned water injection unit is multiple and can be arranged in any form around the sand dredging unit.

[0055] In this embodiment, water injection wells are set around the sand extraction well, and the water injection wells are connected to each other through water injection pipes. One of the water injection pipes is connected to a water storage tank through a pipe connected to a water pump. A sand collection tank and a sand extraction vacuum pump are connected in sequence through a sand collection pipe on the sand extraction pipe. A mass sensor is set at the bottom of the sand collection tank. The central controller is connected to the mass sensor and the water pump through data lines. The power supply is used to power the central controller.

[0056] The central controller is a parallel computing server, and FLAC is installed. 3D Finite difference element analysis software and PFC 3D Discrete element analysis software; this parallel computing server can run up to 60 three-dimensional water injection and sand pumping calculation examples simultaneously.

[0057] The water injection well includes a well casing 16, a main water injection pipe 17, and an auxiliary water injection pipe 15. The main water injection pipe is coaxially mounted inside the well casing. Auxiliary water injection pipes are mounted inside the well casing on both sides of the main water injection pipe. A longitudinally spaced central water injection hole 18 is provided in the part of the well casing that extends into the sand layer. Auxiliary jet holes 20 are longitudinally spaced on both sides of the central water injection hole. The central water injection hole is connected to the main water injection pipe through a connecting pipe 23. A controlled valve 22 and a double gate valve 21 are installed sequentially from the inside to the outside of the connecting pipe. A high-pressure nozzle 19 is installed in the auxiliary jet hole. The high-pressure nozzle is connected to the auxiliary water injection pipe through a high-pressure hose 16.

[0058] It also includes a rigid chain telescopic mechanism 24, which is provided on the inner wall of the well casing at the position corresponding to the auxiliary jet hole. The high-pressure hose is connected to the high-pressure nozzle through the rigid chain telescopic mechanism, and the rigid chain telescopic mechanism drives the high-pressure nozzle to perform telescopic movements.

[0059] The rigid chain telescopic mechanism includes a housing 28, a rigid chain 26, a gear 29, a chain guide groove 27, a chain pusher 25, and an electric motor 30. A chain guide groove is provided inside the housing, and the rigid chain is slidably installed within it. The electric motor is driven by the gear, and the teeth on the gear mesh with the rigid chain. The chain pusher is rigidly connected to the end of the rigid chain. The end of the high-pressure hose passes through the housing and is installed on the rigid chain, extending and retracting within the housing along with the rigid chain. The high-pressure nozzle is installed on the end of the side wall of the rigid chain. Through the telescopic movement of the rigid chain, the high-pressure nozzle is driven to pass through the auxiliary jet hole and overcome the sand layer resistance to enter the formation, achieving precise point-to-point hydraulic disturbance.

[0060] An intelligent collaborative sand-dredging method for water-rich sand layers, the innovation of which is that the method includes the following steps:

[0061] 1) Calibration of permeability coefficient equation: Samples of the underlying sand layer were obtained from field drilling, and tests were conducted on sand layers with different porosities. n and head gradient i The permeability test was conducted to obtain the corresponding permeability coefficient. k , fitting k and n , i Functional relationship ;

[0062] 2) System Setup: The system includes a sand pumping system set up on the project site and a corresponding numerical simulation system established in a computer. In the sand pumping system, water injection wells and sand pumping wells penetrate the overlying clay layer 13 and insert into the underlying sand layer 14. Several water injection wells are arranged around the sand pumping wells. The control unit includes a controlled valve at the water injection port, a water pump at the water injection port, and a double gate at the water injection port. The monitoring unit includes a concentration monitor at the sand pumping wellhead and a mass sensor below the sand collection bucket. The numerical simulation system includes a solid phase grid simulating the overlying clay layer, a solid phase particle aggregate simulating the underlying sand layer, a fluid grid simulating seepage, the well wall of the pumping well, the well wall of the water injection well, and boundary conditions.

[0063] See Figure 2 This is a schematic diagram of the intelligent collaborative sand pumping system for water-rich sand layers according to the present invention. In this embodiment, the site is 40m wide and 40m long, with an upper clay layer thickness of 20m, a lower sand layer thickness of 40m, and a water level 5m above the clay layer. The sand pumping well has a diameter of 1m, and the water injection well has a diameter of 1m. Four water injection wells are arranged in a ring, with a 10m distance between them. The ports of both the water injection and sand pumping wells are buried 20m below the sand layer. At the construction site, a water pump, water injection pipe, water storage tank, and water injection wells are connected sequentially to form a water injection unit, which is connected to a central controller via a data cable. At the construction site, a sand collection tank is connected to the top of the sand pumping well via a sand collection pipe. A mass sensor is installed at the bottom of the sand collection tank to calculate the sand pumping rate. A concentration measuring instrument is installed at the wellhead of the sand pumping well to calculate the water pumping rate. Both sensors are connected to the central controller via data cables.

[0064] Specifically, such as Figure 4 As shown in the geotechnical investigation report, the groundwater level at the site is 5m above the surface. The soil layers involved in this site are divided into two layers: the first layer is an upper cohesive soil layer (initial porosity 0.45, horizontal permeability coefficient 8×10⁻⁶). - 5 cm / s, vertical permeability coefficient 6×10 -5 The porosity (cm / s) was simulated using the finite difference method. The second sand layer (initial porosity 0.35, horizontal permeability coefficient 7×10) was used. -4 cm / s, vertical permeability coefficient 7×10 -4 The speed (cm / s) was simulated using the discrete element method. The parameters selected in the simulation were all based on the geological survey report; the sidewalls of both the water injection unit and the sand pumping unit were simulated using rigid walls with impermeable boundaries.

[0065] 3) System startup: Select operating parameters based on engineering experience. ,in, H For the depth of the water inlet, p Water injection intensity (water injection volume per unit time) θTo determine the water injection opening angle, a mass sensor below the sand collection tank is continuously monitored to obtain a time-history curve of the total sand and water volume. The rate of sand and water mixture extraction can be obtained through differential calculations. Based on the concentration monitor at the sand extraction wellhead, the sand extraction rate and water extraction rate can be calculated separately. A collaborative sand extraction digital twin model system is set up, constructed, and launched according to the field operating parameters, using the porosity of the underlying sand layer in the geological survey report as the initial value of the permeability coefficient. n 0 is assigned to each element of the seepage grid;

[0066] 4) Porosity Distribution Inversion: Using the porosity of each element as a variable, the porosity of each seepage grid element is adjusted through an optimization algorithm, based on... Calculate the seepage field and simulate the sand pumping process. And through self-consistent iterative calculation, the head gradient that conforms to the current porosity distribution is obtained in real time. i and their corresponding permeability coefficients k This minimizes the error between the sand discharge rate and water discharge rate at the top port of the sand pumping well obtained by the numerical simulation system and the corresponding rate of the physical sand pumping system. The error calculation formula is as follows: ,in, Q 实 This represents the actual sand extraction rate. w 实 This represents the actual pumping rate. Q 模 To simulate sand extraction rate, w 模 To simulate pumping rate, and The preset weighting coefficients ( By deploying multiple computational scenarios to obtain the porosity sensitivity gradient, the optimal porosity distribution n for the site is obtained through inversion correction. opt =[ n 1, n 2, n 3…, n m ], m This represents the number of seepage grid cells;

[0067] In this example, as Figure 7 As shown, based on porosity n Distribution of changes The site was divided into 12 regions, requiring a total of 13 sets of simulation examples. The first set involved model calculations based on the current porosity, which were completed in step S4. The second to thirteenth sets sequentially applied perturbations to the porosity of the first to twelfth porosity zones. The porosity of the remaining zones remains unchanged, and calculations are performed separately.

[0068] a) Except for the first set of calculations, in each set of calculations The values ​​are -10%, -5%, 5% and 10% respectively, for a total of 52 calculation conditions.

[0069] b) Based on the simulation results above, the central control unit calculates the relative error for each working condition according to the actual sand pumping rate and water pumping rate, as well as the simulated sand pumping rate and water pumping rate. d Simultaneously calculate the porosity sensitivity gradient vector. .

[0070] In this embodiment, i =1-12, This is represented as the gradient of porosity variation across different zones, where , This represents the relative error without applying any perturbation to the porosity of any partition. In the first round of inversion, the linear function fitting of the relative error of partition 9 with porosity is as follows: Figure 6 As shown, the fitting slope is -228, that is... The remaining sets of examples can be processed in the same way.

[0071] c) The correction matrix is ​​calculated using the L-BFGS algorithm. According to the formula The porosity of each zone is adjusted, including the parameters... The value needs to be determined through trial calculations and should ensure computational stability and convergence. Optional, a correction matrix. It can also be obtained using Adam algorithm, RMSProp algorithm or standard BFGS algorithm.

[0072] d) Repeat step ac multiple times to perform soil porosity inversion until the relative error of the current simulation result is found. d Less than the threshold In this example The value is taken as 5%, relative error d The typical change process with the number of porosity inversions is as follows: Figure 7 As shown.

[0073] 5) Optimization of operating parameters: Based on the porosity distribution obtained in step 4), adjust the boundary conditions of the numerical analysis model corresponding to the operating parameters and set the water injection depth. H The water injection depth is 10m-15m, and the water injection intensity is... p The pressure ranges from 0.5MPa to 2.0MPa, with an injection angle of [missing information]. θ The range is 0° to 30°. Define a multi-objective optimization function. , in which settings (Focusing on controlling the range of horizontal disturbances). , This makes the following multi-objective optimization function... J Optimal: The multi-objective optimization functionJ The simulated sand dredging rate (Sand production volume per unit time at the top of the sand-dredging well calculated by the numerical simulation system), and the horizontal disturbance range of the underlying sand layer. R h Weighted composition and simulated sand carrying rate ,and R h The weight is set to the maximum. Optimal operating parameters are obtained through parallel simulation calculations. By utilizing a parallel computing environment, dozens of working conditions were simulated in a short period of time, ultimately identifying the optimal function for multi-objective optimization. J The optimal solution with the largest value On-site construction: Water is injected into the water injection control unit according to the current stage of the precipitation plan, and then the sand pumping unit performs sand pumping operations. During the sand pumping process, the central control unit collects sand pumping rate and water pumping rate data at a certain data acquisition frequency.

[0074] 6) Adjust the operating parameters of the physical sand dredging system: based on the optimal operating parameters obtained in step 5). Adjust the operating parameters of the physical sand dredging system and continuously monitor on-site data; when the error d When the preset allowable error threshold is exceeded, it is determined that the actual porosity distribution of the sand layer has undergone a non-negligible change. At this point, the process jumps to step 4) to re-invert and correct the site porosity distribution.

[0075] The core advantage of this invention's intelligent collaborative sand pumping method for water-rich sand layers lies in the construction of a collaborative sand pumping digital twin model system. This model is not static but can update the permeability coefficient in real time based on the current porosity and pore pressure gradient, thereby accurately capturing the dynamic evolution of the sand. The central controller performs parallel calculations through a multi-objective optimization function, specifically setting the weight of the horizontal disturbance range higher than the sand pumping rate. This means that when searching for the optimal solution, the system prioritizes construction safety and prevents settlement caused by excessive vertical disturbance. Utilizing an adaptive zonal inversion algorithm, the central controller can dynamically correct formation parameters based on measured sand pumping and water pumping data, solving the problem of slow convergence in the inversion of strongly nonlinear formation parameters and ensuring deep synchronization between simulation and field conditions.

[0076] This invention relates to an intelligent collaborative sand pumping system for water-rich sand layers. The injection well is equipped with main and auxiliary multi-channel water supply pipelines, and a central controller adjusts the injection depth and pressure in real time based on optimal parameters. Furthermore, its rigid chain telescopic mechanism drives a high-pressure nozzle through the well wall, directly overcoming the resistance of the sand layer to penetrate deep into the formation. This point-to-point precise hydraulic disturbance actively induces a "shear effect" in the sand, accurately guiding sand particles towards the pumping well, significantly improving operational efficiency and solving the technical bottleneck of difficult-to-control disturbance range.

[0077] Although embodiments and drawings of the present invention have been disclosed for illustrative purposes, those skilled in the art will understand that various substitutions, variations and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the scope of the present invention is not limited to the contents disclosed in the embodiments and drawings.

Claims

1. An intelligent collaborative sand-dredging system for water-rich sand layers, characterized in that: The system includes a water injection unit, a sand pumping unit, a central controller, and a power supply. The water injection unit includes an injection well, injection pipes, a water pump, and a storage tank. The sand pumping unit includes a sand pumping well, a sand collection pipe, a sand pumping vacuum pump, and a sand collection bucket. A sand pumping well is installed in the water body, and a concentration detector is installed at the bottom of the sand pumping well. Water injection wells are set around the sand pumping well, and the water injection wells are connected to each other through injection pipes. One of the injection pipes is connected to the storage tank through a pipe connected to the water pump. The sand collection pipe is connected to the sand collection bucket and the sand pumping vacuum pump in sequence through the sand pumping pipe. A mass sensor is installed at the bottom of the sand collection bucket. The central controller is connected to the mass sensor and the water pump through data cables. The power supply is used to power the central controller.

2. The intelligent collaborative sand pumping system for water-rich sand layers according to claim 1, characterized in that: The water injection well includes a well casing, a main water injection pipe, and auxiliary water injection pipes. The main water injection pipe is coaxially mounted inside the well casing. Auxiliary water injection pipes are mounted inside the well casing on both sides of the main water injection pipe. A central water injection hole is arranged longitudinally at intervals in the part of the well casing that extends into the sand layer. Auxiliary jet holes are arranged longitudinally at intervals on both sides of the central water injection hole. The central water injection hole is connected to the main water injection pipe through a connecting pipe. A controlled valve and a double gate valve are installed sequentially from the inside to the outside of the connecting pipe. A high-pressure nozzle is installed in the auxiliary jet hole. The high-pressure nozzle is connected to the auxiliary water injection pipe through a high-pressure hose.

3. The intelligent collaborative sand-dredging system for water-rich sand layers according to claim 2, characterized in that: It also includes a rigid chain telescopic mechanism. A rigid chain telescopic mechanism is provided on the inner wall of the well casing at the position corresponding to the auxiliary jet hole. The high-pressure hose is connected to the high-pressure nozzle through the rigid chain telescopic mechanism, and the rigid chain telescopic mechanism drives the high-pressure nozzle to perform telescopic movements.

4. The intelligent collaborative sand-dredging system for water-rich sand layers according to claim 3, characterized in that: The rigid chain telescopic mechanism includes a housing, a rigid chain, gears, a chain guide groove, a chain pusher, and an electric motor. The chain guide groove is located inside the housing, and the rigid chain is slidably installed within it. The electric motor is connected to the gear, whose teeth mesh with the rigid chain. The chain pusher is rigidly connected to the end of the rigid chain. The end of the high-pressure hose passes through the housing and is mounted on the rigid chain, extending and retracting within the housing with the rigid chain. The high-pressure nozzle is mounted on the end of the rigid chain's sidewall. Through the extension and retraction of the rigid chain, the high-pressure nozzle is driven to pass through the auxiliary jet hole and overcome sand layer resistance to enter the formation, achieving precise point-to-point hydraulic disturbance.

5. A smart collaborative sand-dredging method for water-rich sand layers, characterized in that: This method is implemented based on the sand dredging system according to any one of claims 1-4, and the method includes the following steps: 1) Obtain samples of the underlying sand layer from field drilling and conduct experiments on sand layers with different porosities. n and head gradient i The permeability test was conducted to obtain the corresponding permeability coefficient. k , fitting k and n , i Functional relationship ; 2) The water injection well and the sand extraction well are inserted through the overlying clay layer into the underlying sand layer, and several of the water injection wells are arranged around the sand extraction well; 3) Select operating parameters based on engineering experience ,in, H For the depth of the water inlet, p For water injection strength, θ To determine the water injection opening angle, a mass sensor below the sand collection tank is continuously monitored to obtain a time-history curve of the total sand and water volume. The rate of sand and water mixture extraction can be obtained through differential calculations. Based on the concentration monitor at the sand extraction wellhead, the sand extraction rate and water extraction rate can be calculated separately. A collaborative sand extraction digital twin model system is set up, constructed, and launched according to the field operating parameters, using the porosity of the underlying sand layer in the geological survey report as the initial value of the permeability coefficient. n 0 is assigned to each element of the seepage grid; 4) Using the porosity of each unit as a variable, the porosity of each seepage grid unit is adjusted through an optimization algorithm, based on... Calculate the seepage field and simulate the sand pumping process. And through self-consistent iterative calculation, the head gradient that conforms to the current porosity distribution is obtained in real time. i and their corresponding permeability coefficients k This minimizes the error between the sand discharge rate and water discharge rate at the top port of the sand pumping well obtained by the numerical simulation system and the corresponding rate of the physical sand pumping system. The error calculation formula is as follows: ,in, Q 实 This represents the actual sand extraction rate. w 实 This represents the actual pumping rate. Q 模 To simulate sand extraction rate, w 模 To simulate pumping rate, and The preset weighting coefficients, By deploying multiple computational scenarios to obtain the porosity sensitivity gradient, the optimal porosity distribution n for the site is obtained through inversion correction. opt =[ n 1, n 2, n 3 …, n m ], m This represents the number of seepage grid cells; 5) Based on the porosity distribution obtained in step 4), adjust the boundary conditions of the numerical analysis model corresponding to the operating parameters to achieve the following multi-objective optimization function. J Optimal: The multi-objective optimization function J The simulated sand dredging rate Horizontal disturbance range of the underlying sand layer R h Weighted composition and simulated sand carrying rate ,and R h The weights are set to the maximum, and the optimal operating parameters are obtained through parallel simulation calculations. ; 6) Based on the optimal operating parameters obtained in step 5) Adjust the operating parameters of the physical sand dredging system and continuously monitor on-site data; when the error d When the preset allowable error threshold is exceeded, it is determined that the actual porosity distribution of the sand layer has undergone a non-negligible change. At this point, the process jumps to step 4) to re-invert and correct the site porosity distribution.

6. The intelligent collaborative sand pumping method for water-rich sand layers according to claim 5, characterized in that: The numerical simulation system includes a solid grid simulating the overlying clay layer, a solid particle aggregate simulating the underlying sand layer, a fluid grid simulating seepage, the well wall of the pumping well, the well wall of the injection well, and boundary conditions.

7. The intelligent collaborative sand pumping method for water-rich sand layers according to claim 5, characterized in that: Step 4) involves dynamic mesh optimization based on the spatial distribution of porosity, according to the porosity at the current moment in the model. n Change in porosity compared to the previous round of optimization Partitioning: Continuous cells are clustered into a partition, where all grids within that partition share a single porosity and permeability coefficient; if , If a preset grid division threshold is used, the porosity and permeability coefficient of that grid are calculated and inverted separately.

8. The intelligent collaborative sand pumping method for water-rich sand layers according to claim 5, characterized in that: The multi-objective optimization function J The component method is as follows: extract the set of elements in the numerical model whose porosity changes exceed a preset threshold, calculate the geometric envelope of this set on the horizontal projection plane, and define it as the porosity horizontal perturbation range. R ; The volume of sand produced per unit time at the top of the sand-dredging well calculated by the numerical simulation system The sand carrying capacity at the top of the sand-dredging well calculated by the numerical simulation system. ; Using formula Calculate the overall score for each working condition, where, , and As the weighting coefficient, and set The maximum priority is given to water injection schemes that focus on horizontal sand transport, guided by the system.

9. The intelligent collaborative sand pumping method for water-rich sand layers according to claim 5, characterized in that: Step 4) The inversion correction process uses numerical perturbation to obtain the gradient, and the calculation equation is as follows: i) Definition of relative error: in, Q act This represents the actual sand extraction rate. w act This represents the actual pumping rate. Q sim To simulate sand extraction rate, w sim To simulate pumping rate, and The preset weighting coefficients ( ); ii) Sensitivity gradient vector calculation: Parallel deployment of 1+1+ within each inversion step N One calculation condition, N The number of regions is determined by the porosity of the site in order to obtain the sensitivity gradient vector. : In the formula, , This is the baseline error calculated based on the current porosity distribution; For the first i The gradient components of each porosity-divided region are determined by applying local porosity perturbations. Calculated; iii) Parameter inversion equation: in, For the search step size, This is a correction matrix generated based on an optimization algorithm to coordinate the correction magnitudes of different partitions.

10. The intelligent collaborative sand pumping method for water-rich sand layers according to claim 5, characterized in that: Step 5) specifically includes: i) Constructing the water injection parameter space: Selecting the water injection depth of the injection well H Water injection intensity p and the direction of water injection θ As optimization variables, they are discretized and combined within a preset value range to generate variables containing... m A set of working conditions to be calculated ; ii) Parallel computing mode: Adjust the boundary conditions of the numerical analysis model corresponding to the operating parameters to perform parallel computing, and calculate the multi-objective optimization function value corresponding to each mode. J i ; iii) Parameter locking and output: Extract the global maximum objective function value using a maximum value retrieval algorithm. ;Will Corresponding working conditions Record the optimal water injection parameter combination for the current round. And output it to the actuator.

Citation Information

Patent Citations

  • Foundation pit waterproof curtain water leakage identification method and system

    CN112900504A