Underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling

Abstract

This invention discloses a seepage control method for underground engineering based on multi-objective dewatering scheme optimization and rolling scheduling, belonging to the field of information technology in underground engineering and geotechnical construction. The method proposes a two-stage integrated solution: In the design stage, based on a seepage-stress coupled simulation model, a multi-objective optimization algorithm is used to solve for the Pareto non-dominated solution set of the number, location, and depth of dewatering wells, with the objectives of minimizing structural deformation and engineering costs, and to determine the baseline scheme; In the construction stage, on a fixed dewatering well network, based on real-time monitoring data, a secondary optimization is performed within a rolling time window, with the objectives of maximizing safety margin and minimizing engineering costs, dynamically solving for the target pumping flow rate and start / stop status of each well, and disseminating the results to field equipment for execution via a closed-loop industrial communication protocol. This invention achieves a closed-loop process for the entire dewatering scheme, from design optimization to adaptive control during construction.

Description

Technical Field

[0001] This invention belongs to the field of underground engineering and geotechnical construction informatization, and involves the design, optimization and on-site group control scheduling of dewatering schemes for deep foundation pits. Specifically, it involves an underground engineering seepage control method based on multi-objective dewatering scheme optimization and rolling scheduling. Background Technology

[0002] When constructing deep foundation pits in areas with soft soil, water-rich sand layers, and multiple confined aquifers, it is necessary to control the groundwater level below the excavation face at the target elevation, while also considering the impact of external water level fluctuations on the surrounding environment and structure. The dewatering plan essentially constrains both safety and economic objectives: the safety aspect focuses on pit bottom heave, piping, foundation stability, horizontal displacement of the retaining wall, and settlement of adjacent buildings and structures; the economic aspect focuses on initial construction costs (drilling, well casing and filter media, pump electrical control, pipelines and power distribution, installation and commissioning, etc.) and operational costs (pumping energy consumption, sewage / sludge treatment, maintenance, etc.). Simply relying on experience or single-objective optimization makes it difficult to achieve comprehensive optimization under the coupled conditions of complex hydrogeology and phased excavation.

[0003] Existing methods can be broadly categorized into three types: ① Analytical or empirical methods, which are used to quickly estimate well spacing, well depth, and pumping volume. They are convenient but insufficient for characterizing heterogeneity, anisotropy, and boundary effects; ② Numerical modeling methods, which simulate different well layouts or pumping combinations through seepage-stress coupling and select the best option based on "satisfying constraints." They are precise but computationally burdensome and are mostly used for verification rather than system optimization; ③ Single-objective optimization, such as minimum cost or minimum deformation, can improve the scheme in a certain dimension, but it does not adequately consider the trade-off between "deformation and cost / energy consumption" and engineering-level constraints such as pump curves, frequency bands, power distribution limits, drainage capacity, and phased start-up and shutdown, making it difficult to guarantee a "workable optimal solution."

[0004] In terms of construction phase control, common practices include fixed frequency bands or manual inspections for start-up and shutdown. SCADA or IoT monitoring often remains at the level of information display or alarms, failing to form a closed loop with the design model and scheduling strategy. Thresholds are mostly fixed throughout the entire construction phase, making it difficult to dynamically adjust them according to work processes and operating conditions. In summary, existing solutions suffer from bottlenecks such as single-objective design, offline execution, incomplete constraints, and insufficient robustness against disturbances. Therefore, it is necessary to introduce multi-objective optimization and rolling time-domain control to output Pareto non-dominated solution sets under complete engineering constraints and implement a prediction-optimization-execution-audit closed loop to reduce unit extraction energy consumption and total life-cycle cost while meeting safety constraints. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for seepage control in underground engineering based on multi-objective precipitation scheme optimization and rolling scheduling, which aims to achieve systematic optimization in the precipitation scheme design stage and adaptive closed-loop control in the construction stage.

[0006] In a first aspect, the present invention provides a method for seepage control in underground engineering based on multi-objective precipitation scheme optimization and rolling scheduling, comprising the following steps:

[0007] S1: Offline multi-objective optimization steps during the design phase;

[0008] S11: Obtain a dataset consisting of geometric, hydrogeological, stratigraphic and retaining structure parameters of underground engineering, as well as a constraint set consisting of precipitation depth, structural deformation, well spacing and number of wells, pump characteristics and power, and drainage capacity.

[0009] S12: Based on the dataset, establish a parameterized seepage-stress coupling simulation model, and establish a dual-objective optimization model with the objectives of minimizing the extreme value of structural response and minimizing engineering cost;

[0010] S13: Within the boundary defined by the constraint set, the bi-objective optimization model is solved using a metaheuristic optimization algorithm to obtain a Pareto non-dominated solution set;

[0011] S14: From the Pareto non-dominated solution set, a baseline precipitation scheme is determined according to a preset selection rule. The baseline precipitation scheme defines the number, planar location, and depth of precipitation wells.

[0012] S2: Online rolling scheduling steps during the construction phase;

[0013] S21: Based on the precipitation well network determined by the baseline precipitation scheme, real-time access and processing of monitoring data on water level, flow rate, power consumption, rainfall, and construction conditions;

[0014] S22: Within the preset rolling time window, perform secondary optimization to maximize safety margin and minimize engineering cost, and solve for the start-stop status of each precipitation well and the operating setpoint of the target flow within the current time window;

[0015] S23: Translate the operating setpoint into control commands and send them to the field controller via a communication protocol to schedule the operation of the precipitation well group.

[0016] As an optional implementation of the first aspect of this application, in step S11, the constraint set further includes: the groundwater level drops below the excavation surface by a preset depth; the maximum horizontal displacement of the retaining structure does not exceed a preset threshold; the spacing between dewatering wells is not less than a preset lower limit, and the total number does not exceed a preset upper limit; the pumping volume of a single well meets the flow-head-efficiency curve of the pump and is within the frequency band allowed by the frequency converter; the total pumping power of the well group does not exceed the upper limit of the power distribution capacity, and the total drainage volume does not exceed the upper limit of the drainage or disposal capacity.

[0017] As an optional implementation of the first aspect of this application, in step S12, the step of establishing a parameterized seepage-stress coupling simulation model based on the dataset includes: implementing it through an automated structure program of finite element software, wherein the automated structure program of finite element software automatically completes the geometric reconstruction, mesh generation, assignment of material properties and boundary conditions, arrangement of calculation procedures and extraction of calculation results of the simulation model according to different combinations of parameters such as the number of precipitation wells, planar location and well depth.

[0018] As an optional implementation of the first aspect of this application, in step S12, the extreme value of the structural response in the bi-objective optimization model is at least one of the maximum horizontal displacement of the retaining structure, the bottom heave of the pit, or the internal force of the support; the engineering cost includes drilling cost, well pipe and filter material cost, water pump and electrical control cost, pipeline and power distribution cost, operating energy consumption cost, and monitoring and environmental protection cost.

[0019] As an optional implementation of the first aspect of this application, in step S14, the step of determining the baseline precipitation scheme from the Pareto non-dominated solution set according to a preset selection rule specifically involves selecting one or more representative solutions based on the congestion distance and a comprehensive score determined by safety and economic preferences.

[0020] As an optional implementation of the first aspect of this application, the online rolling scheduling step during the construction phase, before performing the secondary optimization, further includes: within the rolling time window, using a prediction model to predict the future water level and structural response in the pit based on historical monitoring data and weather forecasts; and calculating and evaluating the current safety margin based on the predicted water level and structural response in the pit, wherein the safety margin serves as the input to the secondary optimization.

[0021] As an optional implementation of the first aspect of this application, in step S22, the period of the rolling time window is 2 to 6 hours; the method also includes a threshold-triggered dynamic control mechanism, which automatically shortens the period of the rolling time window to increase the scheduling frequency when water level deviation exceeds the limit, rainfall increases suddenly, power exceeds the limit, or the deformation rate of the retaining structure is abnormal.

[0022] As an optional implementation of the first aspect of this application, in step S23, the step of translating the operating setpoint into a control command and sending it to the field controller via a communication protocol includes: mapping the target flow rate to the target frequency of the frequency converter or the target opening degree of the valve based on the pump characteristic curve and the field calibration relationship; sending the command containing the target frequency, target opening degree, or start / stop status to the frequency converter, programmable logic controller, or electric valve via Modbus-RTU, Modbus-TCP, OPC-UA, or Ethernet I / O protocol; collecting the feedback information of the execution device and recording a full-link audit log containing input data, optimization targets, issued commands, and execution feedback.

[0023] In a second aspect, embodiments of this application provide an electronic device, which includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the steps of the method described in the first aspect.

[0024] Thirdly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] 1. Systematic and Global Optimum: Through multi-objective optimization, the trade-off space between "safety and economy" was systematically explored, avoiding the limitations of traditional empirical design or single-objective optimization, and finding the comprehensive optimal solution that satisfies all engineering constraints.

[0027] 2. Design and construction integration: Seamlessly connects the global optimization in the design phase with the dynamic scheduling in the construction phase, ensuring the feasibility of the design scheme and enabling the system to adapt to dynamic changes and uncertainties in the construction process.

[0028] 3. Refined and energy-saving: Online rolling scheduling enables refined control of the precipitation well group, avoiding constant pumping and dynamically adjusting according to actual needs, which significantly reduces engineering costs and groundwater extraction.

[0029] 4. Intelligentization and Risk Prevention: Closed-loop control based on real-time data and predictive models improves the system's response speed and robustness to emergencies, enhances engineering safety, and provides data support for post-event review and knowledge transfer through full-process digital recording. Attached Figure Description

[0030] Figure 1 This is a flowchart of an underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling, according to an embodiment of the present invention.

[0031] Figure 2 This is a network diagram showing the layout of dewatering well points in an underground engineering project according to an embodiment of the present invention.

[0032] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0034] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0035] Example 1

[0036] To facilitate understanding, some technical terms used in the following process will be explained.

[0037] Parametric Simulation Model D: Based on the modeling rules and methods specified by different numerical simulation software, a complete geological model is established in the software according to the site information, geological survey report, and groundwater exploration information of the foundation pit project. A finite element analysis model of the foundation pit retaining structure is established based on the current design scheme. Then, the foundation pit project plane is meshed, and an analysis model of the dewatering facilities for the foundation pit retaining structure is established based on the control parameter set F, and dewatering parameters are assigned. The obtained model should be directly importable into the corresponding finite element software for subsequent meshing and simulation calculations, and the calculation results should be extractable using the corresponding methods built into the finite element software.

[0038] Parameter set F: refers to a complete combination of design parameters when each design parameter takes different specified values. The layout design parameters for a single dewatering well include: well location, well depth, pumping flow rate, and whether to activate four control variables. During the design phase, the control variable vector is: The control variable vector during the construction phase is: , where N represents the number of wells.

[0039] Finite element software automated structural program G: This refers to a software system written in a specified language, based on the command tools, programmable API tools, and remote management scripts built into numerical simulation software. It can call the numerical simulation software to perform functions such as modeling, parametric model control, mesh generation, simulation calculation, and result extraction. The software system automatically generates models based on fixed environmental and design parameters. According to the layout rules of dewatering wells, it enables the simulation software to automatically generate simulation models, automatically generate simulation meshes, execute calculations, and extract calculation results (mainly the maximum deformation of the retaining wall).

[0040] Metaheuristic Algorithm H: Based on the algorithm's own rules, it extensively permutes and combines the selectable parameters within their respective value ranges, and quickly searches for the optimal permutation and combination solution from all permutations and combinations by calculating the objective function. Conventional heuristic algorithms can be used, such as global heuristic / evolutionary search (e.g., general breadth-first search, general depth-first search, genetic algorithms, particle swarm optimization, etc.). Improved versions of conventional algorithms, such as NSGA-II, MOEA-D, and MOPSO variants, can also be used. In the current scenario example, the NSGA-II algorithm is recommended.

[0041] Precipitation deployment scheme I: This does not necessarily refer to the optimal value of the objective function. Some tolerance can be allowed based on actual conditions, providing a number of alternative optimal precipitation deployment schemes. For example, if the tolerance is 5, the design schemes corresponding to the objective function are sorted in ascending order of the objective function value, and the top 5 schemes with the smallest values ​​are included in the optimal candidate precipitation deployment schemes.

[0042] Dispatch control module J: Calculates the single-well flow rate setpoint obtained from the optimization model. Start-stop scheme This module transforms commands into field-executable instructions and distributes them to the pump / valve / frequency converter via a hardware and software system. This module includes, but is not limited to: an industrial gateway / edge controller, protocol and communication stacks (such as Modbus-RTU / TCP, OPC-UA, Ethernet I / O, etc.), instruction translation and limiting / ramp processing unit, execution receipt acquisition unit, event and audit log unit, and local / remote human-machine interface. Based on the pump characteristic curve and the field calibration relationship, this module... Mapped to target frequency / opening degree Or equivalent setpoints, generate start / stop With set point The instruction frame is sent to the VFD / PLC / electric valve; and updated according to the rolling time window Δt (preferably 2–6 h, 1–2 h during rainy season or sudden events) or event triggering (water level deviation, rainfall, power limit exceedance, abnormal receipt, etc.). To ensure the executability and safety of the project, the module has built-in upper / lower limit and rate limits, soft start / soft stop, minimum running time, dry run / idle run and interlock protection, and total power. The module implements strategies such as constraint execution, redundancy switching, and maintaining a safe state even in the event of disconnection. When exceeding limits or communication anomalies are detected, it automatically encrypts and schedules or rolls back to a baseline fixed traffic flow, recording end-to-end audit information (input data window, target / constraint, issued instructions, execution receipts, results, and KPIs). The hardware form, communication protocol, and brand of this module are not limited; it can be deployed centrally or in a distributed manner. It supports clock synchronization and offline buffering / breakpoint resumption, and provides manual / automatic one-click switching and access control for on-site emergency and compliant operation.

[0043] Please see Figure 1 This document presents a flowchart of an underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling, as provided in an embodiment of the present invention. The method can be divided into two stages: "offline multi-objective optimization in the design stage" and "online rolling scheduling in the construction stage." In the design stage, a parametric simulation model and a dual / multi-objective optimization model are established to generate a Pareto non-dominated solution set and determine the baseline scheme and feasible region boundary under complete engineering constraints. In the construction stage, without changing the well network and feasible region, rolling time-domain (secondary) optimization is implemented based on real-time monitoring and short-term prediction, periodically generating well group start-up and shutdown. With set point It also configures a threshold trigger (to achieve multi-objective optimization variations or economic constraints) - a dynamic control mechanism for encrypted scheduling, to achieve a stable connection from "scheme" to "control".

[0044] The specific process is as follows:

[0045] S1: Offline multi-objective optimization during the design phase

[0046] The objective of this phase is to produce a Pareto non-dominated solution set and determine a baseline precipitation scheme under complete engineering constraints, providing an executable well network layout and parameter boundaries for the construction phase.

[0047] S11: Obtain a dataset consisting of geometric, hydrogeological, stratigraphic and retaining structure parameters of underground engineering, as well as a constraint set consisting of precipitation depth, structural deformation, well spacing and number of wells, pump characteristics and power, and drainage capacity.

[0048] First, conduct thorough geological surveys to obtain information about the underground engineering project. This includes:

[0049] (a) Site geometry and boundaries: excavation dimensions, excavation depth, and phased excavation schedule;

[0050] (b) Hydrogeology: initial groundwater level (static water level), aquifer thickness, burial depth, recharge or seepage boundary flux, far-field water level and possible recharge sources;

[0051] (c) Stratigraphic and seepage parameters: The permeability coefficient, total unit weight, elastic modulus or elastic modulus versus confining pressure (Ep) of each soil layer is required, as well as Poisson's ratio, internal friction angle, cohesion, consolidation compressibility coefficient, permeability compressibility coefficient, in-situ static earth pressure coefficient, and saturated and unsaturated characteristics of the soil.

[0052] (d) Retaining structure: type (diaphragm wall, SMW, pile wall, etc.), cross-sectional dimensions, thickness, rock penetration / bearing depth, elastic curvature stiffness, bending strength, arrangement of supports / anchors (number of layers, elevation, prestress, cross-sectional parameters), depth of the seepage barrier, and permeability coefficient.

[0053] (e) Dewatering system: grid of candidate dewatering well locations or areas where dewatering wells can be placed, allowable well depth range and well diameter, pump station model library (flow-head-efficiency curve), planned construction period T and day and night pumping periods;

[0054] (f) Key parameters: maximum horizontal displacement S of the retaining wall, allowable settlement of surrounding buildings / pipelines, relative depth of water level a (below the bottom of the pit), well spacing dis, number of wells N;

[0055] Let's denote the above dataset as A. The variables in the dataset will be represented by object and attribute indices in the relevant sections below, such as the lower limit of well spacing: .

[0056] Furthermore, the set of constraints for construction feasibility is clearly defined, including:

[0057] (a) The groundwater level must be lowered to a depth not less than [amount missing] below the bottom of the pit. rice;

[0058] (b) The maximum horizontal displacement of the retaining wall does not exceed ;

[0059] (c) The spacing between dewatering wells shall not be less than And the total number does not exceed ;

[0060] (d) The pumping capacity of a single well is determined by the optimization algorithm and there is no fixed upper limit; some dewatering wells are allowed to have a pumping capacity of 0 (natural shutdown), so that the optimization algorithm can automatically select the dewatering wells that need to be activated.

[0061] Let the above constraint set be denoted as B. The index representation of the variables in the constraint set is the same as above. The maximum horizontal displacement of the retaining wall is: .

[0062] S12: Based on the dataset, establish a parameterized seepage-stress coupling simulation model, and establish a dual-objective optimization model with the objectives of minimizing the extreme value of structural response and minimizing engineering cost;

[0063] First, based on dataset A, a parameterized seepage-stress coupling simulation model D is established using finite element software (such as Plaxis), discrete element method, or other numerical calculation methods C.

[0064] The seepage-stress coupling simulation model D first establishes a geometric model of the foundation pit project site and retaining structure based on site geometry and boundaries, hydrogeology, strata and seepage parameters, and specific parameters of the retaining structure. It then imports the construction schedule, construction period T, and day / night pumping times for the foundation pit project. The dewatering system model is then parameterized. During the design phase, the installable range of the foundation pit dewatering wells needs to be determined, and a grid of potential well placement areas or locations needs to be established. Allowable well depth, well spacing, and the number of wells are set as control parameters. During the construction phase, only the operating flow rate and start / stop status of each pump need to be controlled.

[0065] The seepage-stress coupling simulation model D can automatically generate a geometric model, divide the mesh, assign material properties and boundary conditions, and simulate the entire excavation and dewatering process based on a set of given dewatering scheme parameters (such as well location, well depth, and pumping rate).

[0066] Optionally, a corresponding finite element software automated structure program G can be developed by combining different parameters. Inputting G enables the parametric modeling, simulation, and final calculation result extraction of the dewatering scheme using automated finite element software, obtaining the maximum horizontal displacement of the foundation pit retaining structure corresponding to the current parameter combination.

[0067] Furthermore, a dual-objective optimization model E is established, comprising a security objective E1 and an economic objective E2.

[0068] (1) The safety target E1 is defined as minimizing the extreme value of the structural response. Usually, the main control deformation index is selected, such as the maximum horizontal displacement of the retaining wall. The formula is expressed as:

[0069] (1)

[0070] in, For the entire construction period, The maximum deformation of the retaining wall of the foundation pit is given at time t during construction.

[0071] (2) The economic objective E2 is defined as minimizing the total project cost. The calculation formula is derived from relevant specifications and textbooks. Some items are only applicable to the design stage and are marked in the description of each indicator:

[0072] (2)

[0073] in, For drilling costs, The unit price is the comprehensive price per unit diameter and depth (unit: yuan / m²). Well diameter (unit: meters) Well depth (unit: meters) Cost of equipment entry and exit for single wells.

[0074] The cost of well casing and filter media, of which Price per unit of well casing (unit: yuan / m). Price of filter media and sealing materials (unit: yuan / m).

[0075] The cost includes the cost of water pumps and electrical control systems. Flow rate to head conversion factor (unit: yuan) ), Design flow rate (unit: ), Design head (unit: ).

[0076] The cost of piping and power distribution for the rainwater system, of which The unit price for water supply and drainage pipelines (unit: yuan / m) The unit price for cable laying (unit: yuan / m) and This refers to the corresponding lengths for drainage pipes and cable laying.

[0077] The cost of energy consumption metering for equipment operation, of which Electricity price (unit: yuan / kWh), The density of water, The system efficiency of the pump.

[0078] The costs include monitoring of dewatering in the foundation pit and environmental treatment. For single-point monitoring costs, To monitor the number of points, This is the unit price for drainage treatment. This represents the drainage volume.

[0079] Other expenses are generally a certain percentage of the costs of the preceding items, i.e. .

[0080] Of the total cost, the first four items are one-time investments, while other operating costs are difficult to predict directly. The first three items can be directly assessed based on current control parameters. Therefore, the entire cost model can be simplified as follows:

[0081] (3)

[0082] By solving the simultaneous equations, a bi-objective optimization function E is obtained, which simultaneously considers the deformation of the foundation pit retaining structure and the economic efficiency of the foundation pit dewatering well layout scheme, as shown in the formula:

[0083] (4)

[0084] in and These are the weighting coefficients for security objectives and the weighting coefficients for economic objectives, respectively. The solution with the highest cost among all the computational examples.

[0085] S13: Within the boundary defined by the constraint set, the bi-objective optimization model is solved using a metaheuristic optimization algorithm to obtain a Pareto non-dominated solution set;

[0086] First, based on the actual constraints, the constraint set B is transformed into mathematical constraints, including:

[0087] Water level control: To ensure that the groundwater level drops below am below the bottom of the pit, of which This indicates the actual water level controlled by precipitation. Indicates the location of the bottom of the pit;

[0088] Deformation control: The upper limit of horizontal displacement of the retaining wall, of which Indicates the actual maximum deformation;

[0089] Well spacing restrictions: To prevent interference between wells, among which Indicates the actual minimum well spacing;

[0090] Well count limit: To prevent excessive well placement, where n represents the actual number of wells;

[0091] Pumping freedom: Automatically selected by optimization algorithms, with no fixed upper limit, among which This indicates the actual pumping flow rate of a single well.

[0092] Furthermore, a suitable metaheuristic algorithm H is selected to solve the optimization objective of the current multi-objective optimization model. Different combinations of parameter F values ​​are generated within the range of F values ​​using the metaheuristic algorithm. By iteratively calling program G through the metaheuristic algorithm, the maximum horizontal deformation of the retaining wall under the current parameter combination is obtained. And calculate equation (3) to evaluate the cost of the current design scheme. The maximum horizontal displacement of the foundation pit retaining structure and the estimated cost obtained by combining all parameters are stored as variables: and Obtained from the calculated assessed cost .

[0093] Optionally, depending on the different engineering phases, including the design phase and the construction phase, the set of key control parameters F for the design scheme of the dewatering well is determined, and the value range of each parameter in F is defined as the search domain for selecting parameters of the multi-objective optimization model.

[0094] Specifically, a metaheuristic multi-objective optimization algorithm H (such as NSGA-II, MOEA / D, MOPSO, etc.) is employed, using the parameterized simulation model D established in step S12 as the fitness function evaluator. Algorithm H generates and iterates a population within its search space (the range of values ​​for the decision variables). For each individual (representing a precipitation scheme), the simulation model D is called via an automated script to calculate the corresponding maximum horizontal displacement E1 and project cost E2. After multiple generations of evolution, the algorithm converges to a Pareto non-dominated solution set.

[0095] S14: From the Pareto non-dominated solution set, a baseline precipitation scheme is determined according to a preset selection rule. The baseline precipitation scheme defines the number, planar location, and depth of precipitation wells.

[0096] Optionally, communicate with the owner, design firm, and construction company to determine the safety weights. and economic weight Calculate the corresponding values ​​under different parameter combinations Based on the calculated parameter combinations and the corresponding database of calculation results, the design parameter combinations under the minimum objective function are obtained, resulting in the current optimal wellpoint dewatering deployment scheme I.

[0097] Optionally, based on congestion distance and comprehensive weighted score, several representative solutions are selected from the Pareto front, and an optimal baseline precipitation scheme I is jointly determined by the engineering stakeholders. Scheme I fixes the number, planar location, and depth of precipitation wells, serving as the physical basis for the construction phase.

[0098] S2: Online Rolling Scheduling Steps During the Construction Phase

[0099] The goal of this stage is to implement rolling time-domain optimization based on real-time monitoring and short-term forecasting, without altering the existing rainwater well network, and to periodically generate start / stop and flow setpoints for the well group, thereby achieving refined and adaptive control.

[0100] S21: Based on the precipitation well network determined by the baseline precipitation scheme, real-time access and processing of monitoring data on water level, flow rate, power consumption, rainfall, and construction conditions;

[0101] To dynamically adjust the dewatering deployment strategy at each well point during the foundation pit excavation process, a fixed network of dewatering well points is established. The flow rate of each well point and its operational status (whether it is operating normally or temporarily shut down) are analyzed as control parameters. Steps S12 to S14 are repeated in real-time during construction to obtain a control scheme for the single-well flow rate and equipment switching of the dewatering equipment.

[0102] Optionally, real-time monitoring data from the site can be accessed via a water level-flow sensor network, including: water level in each well, outflow rate, power consumption, site rainfall, weather forecast, and current construction stage. The raw data undergoes preprocessing such as time alignment, missing value filling, and anomaly detection.

[0103] S22: Within the preset rolling time window, perform secondary optimization to maximize safety margin and minimize engineering cost, and solve for the start-stop status of each precipitation well and the operating setpoint of the target flow within the current time window;

[0104] Within a rolling time window Δt (e.g., 2-6 hours), using time series forecasting models (such as ARIMA, LSTM) or simplified physical models, combined with the latest monitoring data and weather forecasts, the water level and structural response trends within the pit during the future Δt time window are predicted. Based on the prediction results, the safety margin under the current operating conditions is assessed.

[0105] With "maximizing safety margin + minimizing engineering costs" as the new dual objectives, the various engineering constraints from the design phase are inherited. The decision variable at this stage becomes the target pumping flow rate of each dewatering well within the next time window Δt. and start / stop status Solving this second-order optimization problem yields the optimal well group operation strategy within the current time window.

[0106] S23: Translate the operating setpoint into control commands and send them to the field controller via a communication protocol to schedule the operation of the precipitation well group.

[0107] The target pumping flow rate obtained from the secondary optimization is controlled by a scheduling and control module J. and start / stop status This translates into control commands that can be executed by the field equipment. For example, the target flow rate... The pump's performance curve is mapped to the target frequency of the variable frequency drive (VFD) or the target opening degree of the electric valve.

[0108] These commands are sent to the PLCs or VFDs at each wellhead via industrial communication protocols (such as Modbus-RTU / TCP, OPC-UA) to control the start, stop, and speed of the water pumps. The system also collects the feedback status of the equipment, forming a closed-loop control of "prediction-optimization-issuance-feedback-audit".

[0109] When events such as excessive water level deviation, sudden heavy rainfall, excessive total power, or abnormal equipment communication are detected, the system can trigger an emergency scheduling mechanism, such as shortening the rolling time window Δt to encrypt scheduling, or automatically reverting to a preset conservative operating mode to ensure system safety.

[0110] Example 2

[0111] The following example of a rectangular foundation pit project illustrates the specific application of this invention.

[0112] 1. Project Overview

[0113] In a rectangular foundation pit project (70 m long, 30 m wide, designed excavation depth 18 m), the site cover layer is a 2 m thick layer of miscellaneous fill, underlain by layers of silty clay and fine sand, each with a permeability coefficient of 1.2 × 10⁻⁶. -6 m / s and 8.5 × 10 -5 The groundwater level is measured in m / s; the initial depth of the groundwater is approximately 3 m below the surface. The retaining system consists of an 800 mm thick diaphragm wall supported by three layers of reinforced concrete. The design requires that the water level in the pit be further reduced to at least 1.2 m below the surface at an elevation of -18 m, while simultaneously controlling the maximum horizontal displacement of the retaining wall caused by dewatering to within 15 mm. The original design included 24 dewatering wells, each 30 m deep, with a continuous pumping capacity of 20 m³ / h per well (see...). Figure 2 Based on this setting, the daily pumping volume is estimated to be approximately 9600 m³, which will be used as a baseline for comparison. Subsequent optimization will systematically improve pumping efficiency and cost while meeting control targets.

[0114] To address the aforementioned problems, the method of this invention sets the number of dewatering wells, planar coordinates, well depth, and pumping flow rate as optimizable decision variables. To ensure controllable water cone interference between wells, and considering site conditions, the model limits the distance between any two wells to no less than 12 m, and the total number of wells to no more than 20. Well depth can be freely adjusted within the range of 24–32 m, and the pumping flow rate… ,when A value of 0 indicates a natural shutdown, used by the algorithm to eliminate redundant wells. Regarding constraints, the model requires a drawdown of at least 1.2 m at the bottom of the pit, limits the horizontal displacement of the retaining wall and surrounding settlement to within design thresholds, and satisfies engineering constraints such as pump curves, frequency band ranges, and upper limits of power distribution. With the dual objectives of "minimizing the maximum displacement of the retaining wall" and "minimizing the total cost of the dewatering project," the cost definition includes at least the one-time investment (CapEx) and may further include operational energy consumption and disposal costs (OpEx); this embodiment primarily uses CapEx for comparison.

[0115] 2. Optimization during the design phase

[0116] (1) Model building

[0117] The precipitation optimization model first defines a set of decision variables: To describe the degrees of freedom of the system: The total number of wells. Let i be the position of the i-th well in the plane coordinate system. For the depth of the well, The design pumping limit for the well at any given time. The start and stop status of the well is represented. Due to site conditions, these variables must fall within the feasible region. The supplementary formula (5) is further used as a constraint under the current engineering conditions to ensure that the water cones between the wells do not overlap adversely.

[0118] (5)

[0119] Add constraint (6) to limit the mechanical arrangement limits:

[0120] (6)

[0121] Add constraint (7) requiring the filter tube to completely penetrate the fine sand aquifer:

[0122] (7)

[0123] Add constraint (8) to control the pumping flow rate of the i-th dewatering well, where From the local Darcy formula The estimate is given by k, where k is the aquifer permeability coefficient and H is the design drawdown. Take the outer diameter of the filter tube. The theoretical influence radius. To give the optimizer the "hole-clearing" degree of freedom, if the algorithm will... When it shrinks to 0, it corresponds to Automatically set to 0 indicates that the well will not be put into operation for the time being.

[0124] (8)

[0125] (2) Objective function

[0126] (a) Safety objective: The time-varying horizontal displacement of the retaining wall throughout the entire excavation process is determined by the seepage-consolidation coupling solution, and its maximum value can be written as:

[0127] (9)

[0128] in, The model is directly output from the Plaxis 3D numerical solution. After each construction stage, the model automatically tracks the top node of the wall and records its extreme value.

[0129] (b) Economic Objective: To ensure that this objective and the construction cost are within the same dimension range, the following approach is used in the bi-objective normalization:

[0130] (10)

[0131] Reduce 3 mm (lower limit of monitoring noise) to Linear mapping to .

[0132] The cost for this project only considers one-time investment expenses. These one-time investments are calculated as a linear sum of drilling costs, well casing and filter media costs, and pump station costs.

[0133] (11)

[0134] in This is the comprehensive unit price for drilling. and These are the diameter and depth of the i-th well, respectively; and The unit prices for well casing and filter media materials are respectively derived from relevant quotas.

[0135] For well diameter; the average value of the quota is taken on site. , , Pump price index coefficient Take as .get:

[0136] (12)

[0137] in and The cost range observed in each iteration is updated in real time to ensure that the normalization scales adaptively as the search converges.

[0138] Based on the above, the unified objective function for multi-objective programming is set as follows:

[0139] (13)

[0140] The project will be jointly agreed upon by the owner and the project team at the bidding planning meeting, taking into account both the project risks and financial constraints. This project will adopt... It leans slightly towards the safe side.

[0141] (3) Optimization solution

[0142] In addition to the Pareto dominance sort, NSGA-II introduces this weighting value only for the automatic recommendation sorting of the decision panel without affecting population selection, so as to take into account the diversity of solution set distribution.

[0143] If the drawdown of the pit bottom water level is insufficient under any working condition Then a penalty is added to the objective function:

[0144] (14)

[0145] In the formula This indicates the current excavation depth of the foundation pit.

[0146] set up If the displacement of the retaining wall exceeds the limit, then apply:

[0147] (15)

[0148] Ensure the algorithm quickly moves away from infeasible regions.

[0149] according to Figure 2 A network of dewatering well points was established, and an automated script based on Python, Plaxis Remote Scripting, was developed. This script inputs the well location, depth, and design flow rate from the "parameter combination table" line by line, instantly generating the corresponding 3D water-soil coupling model according to a unified template. The script can automatically complete all processes, including geometric reconstruction, mesh generation, material property assignment, boundary condition setting, and load case arrangement. The soil parameters for the foundation pit project are shown in Table 1. For each parameter combination, it first copies the standard geometry and replaces the well section keywords within 0.6 seconds, then calls Plaxis's `mesh.generate()` function to refine the local mesh—the minimum side length of elements near the pit wall and well shaft is controlled at 0.25m, the maximum side length in the far field is 3m, and the number of elements in a single model is maintained between 34,000 ± 200. After the grid is completed, the script immediately starts the custom "dewatering-excavation" calculation sequence, with a total calculation time of about 75 seconds. After the calculation is completed, the script captures the horizontal displacement time history of 15 monitoring nodes on the top of the retaining wall within 3 seconds, and stores the extreme values ​​into the database.

[0150] Table 1. Soil and ground parameters for foundation pit engineering

[0151]

[0152] The NSGA-II multi-objective evolutionary algorithm was developed and launched. The algorithm generates an initial generation of 40 individuals within the feasible region using a randomized Latin hypercube, followed by recombination with a crossover probability of 0.9 and a mutation probability of 0.1. Real-valued genes are simulated binary crossovers, while discrete genes (well number N and well index) are exchanged at a single point. After each evolutionary round, the algorithm performs a non-dominated sort of the population based on two objectives: maximum wall displacement and capital cost, and uses crowding distance to maintain solution set diversity. If an individual violates the drawdown or well spacing constraints, the script immediately adds 10 to its objective function. 6 A penalty term of orders of magnitude automatically eliminates it during the selection phase. The NSGA-II algorithm underwent 80 iterations and called 3,200 simulations, with a total duration of approximately 1 hour and 55 minutes. In the later stages of evolution, the variances of both objectives converged to 10. -3 The magnitude indicates that the population has stabilized near the Pareto front.

[0153] The algorithm writes 3,200 calculation results into the database in the format of "parameter combination - safety index - cost", and generates unique keys using an auto-incrementing index. For ease of subsequent querying, each record also includes a hash symbol string (e.g., "[5,7, 12, 16, 19, 21, 25, 28, 33, 37, 40, 42, 45, 48]") and an 80-dimensional evolutionary generation label. Two clustered indexes are used on the database side: one sorted by cost in ascending order, and the other by wall top displacement in ascending order. Combined with JSON fields storing complete displacement time history data, this ensures that retrieving the lowest cost scheme that meets safety requirements from any threshold takes no more than 30ms. Statistical analysis shows that the cost sample range falls between 280,000 and 640,000 yuan, while the wall top displacement range is 6.8mm to 13.6mm; these extreme values ​​are dynamically updated in the subsequent normalization formula. , and , .

[0154] After the data is successfully imported, the weighted aggregation process is automatically triggered. The system fixes the safety weight α at 0.55 according to the minutes of the meeting between the owner and the designer, and then calculates... The overall score was used to rank the options. The results showed that the lowest-scoring (i.e., the optimal) option was a combination of 14 wells, a well depth of 30 m, and a single-well design flow rate of 25 m³ / h. The capital investment for this option was calculated to be 291,000 yuan, including 101,000 yuan for drilling costs, 78,000 yuan for well casing and filter media costs, and 112,000 yuan for pumps and electrical control costs. Meanwhile, the numerical model reported a maximum horizontal displacement of only 9.2 mm for the retaining wall, with a safety margin of 38.7%, fully meeting the control targets.

[0155] (4) Scheme determination

[0156] The final complete design scheme is as follows: In terms of layout, the long side of the foundation pit is 70 m and the short side is 30 m. The well locations are evenly distributed along a 5 m buffer zone outside the continuous wall, with a total of 14 deep wells, numbered W1–W14 counterclockwise from the northwest corner. Their planar coordinates (with the northwest corner of the foundation pit as the origin, the long side as x, and the short side as y) are as follows:

[0157] W1 (0 m, 0 m), W2 (12 m, 0 m), W3 (24 m, 0 m), W4 (36 m, 0 m), W5 (48 m, 0 m), W6 (60 m, 0 m), W7 (70 m, 5 m), W8 (70 m, 17 m), W9 (70 m, 29 m), W10 (60 m, 34 m), W11 (48 m, 34 m), W12 (36 m, 34 m), W13 (24 m, 34 m), W14 (12 m, 34 m).

[0158] All wells are drilled to a uniform depth of 30 m, with the filter section ranging from -24 m to -30 m. The entire section uses φ400 mm stainless steel bridge-type filter pipes and is filled with 4–8 mm quartz sand to fill the annular gaps. Each well is equipped with a 37 kW variable frequency deep well pump, designed to pump up to 25 m³ / h. The pump control frequency range is 30–50 Hz, with continuous speed adjustment. An integrated electrical control box and a flow meter are installed at the wellhead. All pumps can be controlled via an IoT system.

[0159] 3. Construction phase scheduling

[0160] (1) Hardware architecture

[0161] After the well network is designed and fixed, the flow rate of each well will be set during the construction period. Start-stop status As a control variable during the scheduling period, the system executes a closed loop of 'prediction-optimization-issuance-receipt-audit' according to a rolling time window Δt (preferably 2–6 h, 1–2 h during rainy seasons or sudden events). Specifically, it periodically collects water level / flow / electricity and rainfall data, performs time alignment and anomaly detection; obtains a safety margin based on short-term forecasts, constructs a two-level optimization model of 'maximizing safety margin + minimizing energy consumption', and generates the current period's... Send data to VFD / valve via Modbus-RTU / TCP or OPC-UA and collect receipts; when water level deviation / rainfall / power exceedance or receipt abnormality occurs, encrypt the scheduling or revert to the baseline fixed flow (N+1 redundancy), and add control hardware as shown in Table 2.

[0162] Table 2 Dynamic Control Scheduling Hardware and Software Architecture

[0163]

[0164] (2) Operation strategy

[0165] Rolling cycle: Under normal operating conditions, the rolling time window Δt is set to 4 hours.

[0166] Secondary optimization: Every 4 hours, the system collects monitoring data from the past 4 hours and predicts water level changes for the next 4 hours. With the hard constraint of "maintaining the water level within the range of [-19.4m, -19.2m]" (safety margin) and the objective of "minimizing total energy consumption", the optimal combination of operating frequencies for 14 wells is solved.

[0167] Dynamic Trigger: The system is set to dynamically start and stop based on water level. When the monitored water level in the pit is above -19.2m, all wells operate at a higher frequency (e.g., 45Hz). When the water level drops below -19.4m, the system automatically shuts down the two wells with the highest energy consumption per unit, and the remaining wells operate at a lower frequency (e.g., 35Hz). If the water level rises above -19.3m, the shut-down wells resume low-frequency operation in a "stop first, start later" sequence.

[0168] Emergency Response: When the rain gauge detects heavy rain (e.g., >10mm / h), the system automatically shortens the rolling time window to 1 hour, raises the safety lower limit of the target water level, increases the pumping intensity, and responds to runoff replenishment in advance.

[0169] The above methods not only yielded an economical and safe precipitation scheme during the design phase, but also enabled refined and intelligent control of the precipitation process during the construction phase, further reducing project costs and improving the system's robustness against external disturbances.

[0170] Optionally, embodiments of this application also provide an electronic device, including a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the various processes of the above-described embodiment of an underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling, and can achieve the same technical effect. To avoid repetition, they will not be described again here.

[0171] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described embodiment of an underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling, and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0172] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0173] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0174] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0175] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A multi-objective precipitation scheme based optimization and rolling scheduling of underground engineering seepage control method, characterized in that, Comprise: S1: design stage offline multi-objective optimization step; S11: obtain a data set composed of the parameters of the geometry, hydrogeology, stratum and enclosure structure of the underground engineering, and a constraint set composed of the precipitation depth, structure deformation, well spacing and well number, pump characteristics and power, drainage capacity; S12: based on the data set, a parameterized seepage-stress coupling simulation model is established, and a double-objective optimization model with the minimum of structure response extreme value and the minimum of engineering cost as the target is established; S13: within the boundary defined by the constraint set, a meta-heuristic optimization algorithm is used to solve the double-objective optimization model to obtain a Pareto non-inferior solution set; S14: from the Pareto non-inferior solution set, a baseline precipitation scheme is determined according to a pre-set selection rule, and the baseline precipitation scheme defines the number, plane position and well depth of the precipitation well; S2: construction stage online rolling scheduling step; S21: based on the precipitation well network determined by the baseline precipitation scheme, real-time monitoring data of water level, flow, power, rainfall and construction conditions are accessed and processed; S22: within a pre-set rolling time window, a two-level optimization is performed to maximize the safety margin and minimize the engineering cost, and the start-stop state and target flow operating set point of each precipitation well in the current time window are solved and generated; S23: the operating set point is translated into a control instruction and issued to the field controller through a communication protocol to schedule the operation of the precipitation well group.

2. The underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to claim 1, characterized in that, In step S11, the constraint set further comprises: The groundwater level is lowered to below the excavation surface by not less than a pre-set depth; The maximum horizontal displacement of the enclosure structure does not exceed a pre-set threshold; The distance between precipitation wells is not less than a pre-set lower limit, and the total number does not exceed a pre-set upper limit; The single-well pumping capacity meets the flow-head-efficiency curve of the water pump and is within the frequency range allowed by the frequency converter; The total pumping power of the well group does not exceed the upper limit of the power distribution capacity, and the total drainage capacity does not exceed the upper limit of the drainage or disposal capacity.

3. The underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to claim 1, characterized in that, In step S12, the step of establishing a parameterized seepage-stress coupling simulation model based on the data set comprises: Through a finite element software automatic structure program, the finite element software automatic structure program automatically completes the geometric reconstruction, mesh division, material property and boundary condition assignment, calculation procedure arrangement and calculation result extraction of the simulation model according to different combinations of the number, plane position and well depth of the precipitation well.

4. The underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to claim 1, characterized in that, In step S12, the structure response extreme value in the double-objective optimization model is at least one of the maximum horizontal displacement of the enclosure structure, the pit bottom heave or the support internal force; the engineering cost includes drilling cost, well pipe and filter material cost, water pump and electric control cost, pipeline and power distribution cost, operation energy consumption cost, and monitoring and environmental protection treatment cost.

5. The underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to claim 1, characterized in that, In step S14, the step of determining the baseline precipitation scheme from the Pareto non-inferior solution set according to the pre-set selection rule is specifically selecting one or more representative solutions according to the crowding distance and the comprehensive score determined by the safety and economic preference.

6. The underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to claim 1, characterized in that, In the construction stage online rolling scheduling step, before performing the two-level optimization, it further comprises: Within the rolling time window, a prediction model is used to predict the future pit water level and structure response according to the historical monitoring data and weather forecast; Based on the predicted water level in the pit and the structure response, a current safety margin is calculated and evaluated as an input for the secondary optimization.

7. The underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to claim 1 or 6, characterized in that, In step S22, the period of the rolling time window is 2 to 6 hours; the method further comprises a dynamic regulation mechanism triggered by threshold, when water level deviation is out of limit, rainfall increases suddenly, power is out of limit or the deformation rate of the enclosure structure is abnormal, the period of the rolling time window is shortened automatically to increase the scheduling frequency.

8. The underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to claim 1, characterized in that, In step S23, the operation set point is translated into control instructions, which are issued to the field controller through a communication protocol, including: According to the water pump characteristic curve and the field calibration relationship, the target flow is mapped to the target frequency of the frequency converter or the target opening of the valve; Through Modbus-RTU, Modbus-TCP, OPC-UA or Ethernet I / O protocol, the instructions containing target frequency, target opening or start-stop state are issued to the frequency converter, programmable logic controller or electric valve; The receipt information of the execution device is collected, and the whole-link audit log containing input data, optimization target, issued instruction and execution receipt is recorded.

9. An electronic device, comprising: The processor, memory and program or instructions stored on the memory and executable on the processor are included, and the program or instructions are executed by the processor to realize the steps of the underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to any one of claims 1-8.

10. A readable storage medium, characterized by, The program or instructions are stored on the readable storage medium, and the program or instructions are executed by the processor to realize the steps of the underground engineering seepage control method based on multi-objective precipitation scheme optimization and rolling scheduling according to any one of claims 1-8.

Images