A trench excavating correction system and method

Abstract

The application discloses a slot body excavation deviation rectification system, which comprises an embedded industrial computer as a system information interaction and calculation processing center, a data acquisition unit for acquiring stratum and slot forming state, a dynamic geological database and a digital twin engine connected with the embedded industrial computer for data interaction, and the embedded industrial computer is further electrically connected with a hydraulic servo mechanism, a posture adjusting mechanism and a power modulation module for controlling a double-wheel milling device; the application realizes the axis precision of ±8mm through multi-sensor fusion IMU+ laser+ sound wave and model predictive control MPC, the sound wave scanning predicts the stratum stiffness change 1-3 steps in advance, the control weight is dynamically adjusted, and the deviation convergence speed is improved by 3 times. The application effectively overcomes the problems that the prior art relies on manual experience adjustment or simple feedback control based on PID, and the axis deviation is generally in the deviation of ±30-50mm after the depth exceeds 50m.

Description

Technical Field

[0001] This invention relates to the field of geological treatment technology, and more particularly to the field of continuous wall trench excavation control technology based on the construction of seepage-proof walls to achieve geological waterproofing, specifically to a trench excavation correction system and method. Background Technology

[0002] Cutoff walls, as vertical barrier structures, play a crucial role in water conservancy and hydropower, environmental remediation, and underground engineering, primarily used to control seepage, prevent seepage damage, and enhance project stability. Their construction techniques mainly include grab bucket trenching, impact drilling, sawing, and high-pressure jet grouting. Material selection encompasses concrete, plastic concrete, self-setting mortar, and composite geomembranes to adapt to different geological conditions and seepage control requirements. Currently, cutoff walls are widely used in dam foundation reinforcement, contaminated site isolation, and dike seepage control, especially in deep overburden or complex strata. By forming a continuous, closed barrier, cutoff walls effectively block groundwater seepage paths, reduce uplift pressure, and prevent piping or the spread of chemical pollutants. With technological advancements, the depth, thickness, and precision requirements for cutoff walls are increasing. For example, in pumped storage power stations, the wall depth can reach over 100 meters, and it must be tightly integrated with the bedrock. Furthermore, cutoff walls also serve a structural support function, such as working in conjunction with support systems in deep foundation pit projects, further expanding their engineering value.

[0003] In the construction of cutoff walls, trenching is a core process that directly determines the continuity of the wall and its seepage prevention effect. However, many technical bottlenecks still exist in actual operations, with the problem of trench deviation leading to deviation from the design trajectory being the most prominent. Complex strata, such as gravel layers, strongly weathered bedrock, or alternating soft and hard strata, are prone to causing imbalance in the guidance of trenching equipment. For example, grab buckets or impact drills are prone to "deviation" in hard rock, while in loose layers, uneven lateral resistance may cause trench wall collapse or local enlargement. Deviation problems usually stem from abrupt changes in geological conditions, insufficient accuracy of the equipment guidance system, or improper operating parameters, such as uneven drilling pressure distribution and delayed correction response. Although existing inclination measurement technology can provide real-time feedback on the deviation angle, it is difficult to achieve high-frequency dynamic correction due to limitations in equipment cost and construction efficiency. In actual projects, trench deviation exceeding 2‰ may cause poor overlap between adjacent trench sections, insufficient effective wall thickness, and even the formation of weak seepage zones. For example, in a reservoir, the anti-seepage wall experienced a cumulative deviation of 1.5 meters in the trench openings, causing a localized deviation from the design axis. This necessitated secondary reinforcement grouting as a remedial measure. Furthermore, traditional deviation correction methods rely on manual experience and lack intelligent control mechanisms, which is particularly problematic in deep trench or high-precision projects. There is an urgent need to integrate new technologies such as real-time geological sensing and adaptive deviation correction algorithms to improve trenching quality control. Therefore, this invention is provided to address the problem of deviation between the trenching trajectory and the design axis. Summary of the Invention

[0004] To address the problems in existing anti-seepage wall trenching construction, such as incomplete overlap, insufficient effective wall thickness, and even the formation of weak seepage zones or seepage failure at the overlap, this application provides a trench excavation correction system and method, which can achieve at least one of the following technical effects:

[0005] 1. This invention achieves an axis accuracy of ±8mm by fusing multiple sensors (IMU, laser, acoustic wave, and model predictive control, MPC). Acoustic wave scanning predicts formation stiffness changes 1-3 steps in advance, dynamically adjusting control weights and improving deviation convergence speed by 3 times. This effectively overcomes the problem of existing technologies relying on manual experience or simple PID-based feedback control, which typically results in axis deviations of ±30-50mm at depths exceeding 50m.

[0006] 2. This invention can adjust the milling speed from 0.8 to 2.5 m / h and the thrust from 50 to 200 kN in real time according to the formation stiffness, improving efficiency by 40% in hard rock sections. MPC is optimized every 10 m / s, and the correction time accounts for less than 15%. The whole process has an unmanned intervention rate of ≥85%. It effectively solves the problem of low average efficiency caused by frequent shutdowns in exploration when encountering complex formations in existing technologies.

[0007] 3. Based on acoustic early warning, this invention can identify risks such as faults and karst caves 10-30m in advance with a low false alarm rate. Combined with digital twin simulation, it can effectively reduce the accident rate and greatly reduce existing accidents such as hole collapse and stuck drill, thus avoiding potential economic losses.

[0008] 4. This invention dynamically matches the current excavation power by real-time detection of the strata, which can reduce the energy consumption of the excavation equipment on the one hand, and effectively protect the cutting tools from large impacts on the other hand, thus extending the tool life.

[0009] To achieve the above objectives, the technical solution adopted in this application is as follows:

[0010] A trench excavation correction system includes an embedded industrial control computer as the system information interaction and computing processing center, a data acquisition unit connected to the embedded industrial control computer for data interaction and for collecting strata and trenching status, a dynamic geological database and a digital twin engine, and the embedded industrial control computer is also electrically connected to a hydraulic servo mechanism, an attitude adjustment mechanism and a power modulation module for controlling a dual-wheel milling equipment.

[0011] The data acquisition unit includes an IMU attitude detection module and a laser ranging module for detecting the attitude and position of the dual-wheel milling machine, as well as a first acoustic scanning module for acquiring current geological conditions. The IMU attitude detection module and the first acoustic scanning module are both integrated into the head of the dual-wheel milling machine. The laser ranging module includes at least one set of laser transmitters and receivers respectively installed on the ground on both sides of the slot. A pressure-resistant transparent guide tube is provided below the laser transmitter for guiding the laser beam into the slot. A self-cleaning optical window is installed at the bottom of the pressure-resistant transparent guide tube, and a laser ranging target is installed on the head of the dual-wheel milling machine. The attitude adjustment mechanism consists of a pitch mechanism, a yaw mechanism, and an axial mechanism.

[0012] Preferably, the IMU attitude detection module consists of an integrated three-axis gyroscope and a three-axis accelerometer, the laser ranging module includes at least six sets of laser ranging arrays composed of cross-shaped laser rangefinders, and the first acoustic scanning module includes at least six sets arranged in a circular array within the cutter head shield of the dual-wheel milling machine; the adjustment angles of the pitch mechanism and the yaw mechanism are both ±5°, and the adjustment range of the axial mechanism is ±200mm.

[0013] More preferably, it also includes a second acoustic scanning module, which includes a vertical acoustic array mounted on the slot via an auxiliary bracket, with a spacing of 0.6-1 meter between two adjacent vertical acoustic units, and the emitted acoustic frequency is 10kHz-50kHz.

[0014] Based on the correction system provided above, this invention also provides an intelligent correction method for trench excavation, which specifically includes the following steps:

[0015] Step STP100: Ground Penetrating Radar Scanning. The GPR ground penetrating radar emits high-frequency electromagnetic waves and receives reflected signals from the stratigraphic interfaces. The stratigraphic structure is obtained by analyzing the difference in dielectric constant.

[0016] Step STP200, acoustic strata scanning: The second acoustic scanning module emits a 50kHz low-frequency acoustic wave to scan the bottom layer, generates a global stiffness distribution map by receiving reflected signals, and uploads the data to the digital twin engine; The piezoelectric ceramic transducer of the first acoustic scanning module emits a 500kHz high-frequency acoustic pulse, penetrates the mud and enters the near-field stratum, captures the reflected signals of P-waves and S-waves through an array receiver, records the travel time and amplitude, inverts the real-time mechanical parameters and inputs them into the embedded industrial control computer;

[0017] Step STP300: Data fusion and establishment of a digital twin model M0. The attitude data of the dual-wheel milling machine collected by the IMU attitude detection module, the acoustic data B collected by the second acoustic scanning module, and the position data collected by the laser ranging module are synchronized in time and space using a high-precision clock, and coordinates are unified using the ICP algorithm to establish a geological digital twin model M0 for the current construction area. After normal tunneling, the large-scale geological trend is updated every 10 minutes. The first acoustic scanning module monitors the geological conditions within 5m around the milling head in real time, dynamically correcting the digital twin model M0. t+1 ;

[0018] Step STP400: Design a model predictive controller. Using this model predictive controller, establish the dynamic equations of the milling system to predict N. p The pose of the step is calculated, and the control torque τ required by the hydraulic servo mechanism of the dual-wheel milling machine to complete the correction is determined. control The hydraulic servo mechanism is controlled by an embedded industrial computer to correct the deviation until the positional deviation δ reaches a preset value range. The dynamic equation of the milling system is as follows:

[0019]

[0020] Where, q∈R 3 M(q) represents the milling head pose vector, ∈ R 3×3 Represents the inertia matrix; The Coriolis force matrix represents the motion coupling effect; G(q)∈R 3 Represents the gravitational component; Represents the formation friction model; K d ∈R 3 ×3 Represents the elastic reaction force of the formation; δ∈R 3 Represents the pose deviation; Represents the control torque; τ disturbance ∈R 3 Representing the disturbance torque; the design model predictive controller calculates the optimal control sequence U by minimizing the trajectory tracking error and the rate of change of the control quantity:

[0021]

[0022] Where, q ref The axis represents the design reference trajectory of the twin-wheel milling machine; Q represents the state deviation weight; R represents the control quantity deviation weight; N p N represents the prediction step size; c q(k|t) represents the actual control step size; q(k|t) represents the predicted pose; Δu(k|t) represents the rate of change or adjustment of the control quantity.

[0023] Beneficial effects:

[0024] 1. This invention achieves an axis accuracy of ±8mm by fusing multiple sensors (IMU, laser, acoustic wave, and model predictive control, MPC). Acoustic wave scanning predicts formation stiffness changes 1-3 steps in advance, dynamically adjusting control weights and improving deviation convergence speed by 3 times. This effectively overcomes the problem of existing technologies relying on manual experience or simple PID-based feedback control, which typically results in axis deviations of ±30-50mm at depths exceeding 50m.

[0025] 2. This invention can adjust the milling speed from 0.8 to 2.5 m / h and the thrust from 50 to 200 kN in real time according to the formation stiffness, improving efficiency by 40% in hard rock sections. MPC is optimized every 10 m / s, and the correction time accounts for less than 15%. The whole process has an unmanned intervention rate of ≥85%. It effectively solves the problem of low average efficiency caused by frequent shutdowns in exploration when encountering complex formations in existing technologies.

[0026] 3. Based on acoustic early warning, this invention can identify risks such as faults and karst caves 10-30m in advance with a low false alarm rate. Combined with digital twin simulation, it can effectively reduce the accident rate and greatly reduce existing accidents such as hole collapse and stuck drill, thus avoiding potential economic losses.

[0027] 4. This invention dynamically matches the current excavation power by real-time detection of the strata, which can reduce the energy consumption of the excavation equipment on the one hand, and effectively protect the cutting tools from large impacts on the other hand, thus extending the tool life. Attached Figure Description

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

[0029] Figure 1 This is a system architecture block diagram of the present invention.

[0030] Figure 2 This is a flowchart illustrating the correction process of this invention.

[0031] Figure 3 This is a schematic diagram of the construction structure of the present invention.

[0032] In the diagram: 1-Geology; 2-Laser ranging module; 3-Laser target prism; 4-Dual-wheel milling equipment; 5-First acoustic scanning module; 6-Tank; 7-Second acoustic scanning module. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, 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, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0034] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0035] Example 1:

[0036] See Figure 1 The trench excavation correction system includes an embedded industrial control computer as the system information interaction and computing center, a data acquisition unit connected to the embedded industrial control computer for data interaction and for collecting data on the strata and trenching status, a dynamic geological database and a digital twin engine. The embedded industrial control computer is also electrically connected to a hydraulic servo mechanism, an attitude adjustment mechanism and a power modulation module for controlling the dual-wheel milling equipment 4.

[0037] The data acquisition unit includes an IMU attitude detection module and a laser ranging module 2 for detecting the attitude and position of the dual-wheel milling machine 4, and a first acoustic scanning module 5 for acquiring the current geological conditions. The IMU attitude detection module and the first acoustic scanning module 5 are both integrated into the head of the dual-wheel milling machine 4. The laser ranging module 2 includes at least one set of laser emitters and receivers respectively installed on the ground on both sides of the slot. A pressure-resistant transparent guide tube is provided below the laser emitter for guiding the laser beam into the slot. A self-cleaning optical window is installed at the bottom of the pressure-resistant transparent guide tube, and a laser ranging target is installed on the head of the dual-wheel milling machine. The attitude adjustment mechanism consists of a pitch mechanism, a yaw mechanism, and an axial mechanism.

[0038] Working principle:

[0039] First, the system structure provided in this embodiment is described. The system includes an embedded industrial control computer as the core of the control layer. This computer serves as the information interaction and computation processing center for the entire system. It establishes and optimizes a digital twin engine that matches the current geological formation by collecting attitude data, position data, and geological data from the data acquisition unit, including data on the attitude of the dual-wheel milling machine 4. It also makes qualitative judgments about whether to perform corrections based on different situations during the actual excavation process, and quantitative judgments about the direction and amount of correction required. Simultaneously, it stores all geological data information and establishes a dynamic geological database. The perception layer, which provides objective, real-time data streams to the control layer, is the data acquisition unit, which acts as the front-end information acquisition unit. This unit includes an IMU attitude detection module and a laser ranging module 2 for detecting the attitude and position of the dual-wheel milling machine 4, and a first acoustic scanning module 5 for acquiring current geological conditions. The first acoustic scanning module 5 mainly acquires the near-layer geological structure adjacent to the milling head of the dual-wheel milling machine 4, playing a role in early prediction to avoid borehole collapse and drill bit jamming. See details... Figure 3 As shown, the first acoustic scanning module 5 can feed back the geological information data of the strata to be excavated to the embedded industrial control computer, which serves as the control layer. After information processing, the embedded industrial control computer sends control commands to the power modulation module in the execution layer, thereby adjusting the output power of the dual-wheel milling machine 4 to adapt to the current geological conditions, avoiding excessive output power leading to increased energy consumption, and also avoiding insufficient output power causing drill jamming. Similarly, the IMU attitude detection module and the laser ranging module 2 can monitor the current posture of the dual-wheel milling machine 4 in real time and compare the actual posture with the preset construction axis through the embedded industrial control computer. If a deviation occurs, control commands are sent to the hydraulic servo mechanism and attitude adjustment mechanism, which serve as the execution layer, to adjust the posture and direction of movement of the dual-wheel milling machine 4, thereby achieving timely correction. It is worth noting that, since different dual-wheel milling machines 4 have different control methods, the hydraulic servo mechanism and the attitude adjustment mechanism will also be different. They can be independent mechanisms or integrated mechanisms. In this embodiment, the hydraulic servo mechanism and the attitude adjustment mechanism are described separately only based on functional distinction, so as to make it easier for those skilled in the art to understand. For the actual structural setup, the two can be set up as a complete mechanism.

[0040] See Figure 3As shown, when trenching construction is required for Geological Site 1, the system can first load mechanical parameters, the design trajectory (or motion axis) of the trenching excavation, and initialize the built-in correction mechanism or algorithm of the embedded industrial control computer. It should be noted that loading the actual mechanical parameters of the trenching excavation is to enable the system to effectively control the current machinery and prevent the execution / control commands issued by the system from exceeding the execution range of the existing machinery. The design trajectory is the trenching trajectory preset by the system, which is the theoretically optimal trajectory for constructing the anti-seepage wall. This ensures that after trenching, the continuous wall poured in the trench 6 can perfectly match the pre-poured node piles to form an integrated concrete structure and achieve the best anti-seepage effect. The purpose of the correction mechanism or algorithm built into the embedded industrial control computer is to initialize the current geological conditions to the theoretical geological conditions before actual construction. However, the initial correction mechanism is based on theoretical or preset geology and does not correspond to the actual geological strata of the trenching site. Therefore, after initialization, the geological conditions are updated based on information collected by the data acquisition unit, including updates to the dynamic geological database and digital twin engine. This ensures that the information collected by the entire system before operation matches the current geological strata, thereby achieving the most accurate control over the actual construction progress, construction status, and correction. Thus, this system achieves accurate and timely correction based on the collected real-time geological information, the mechanical equipment information of the actuators, and the correction mechanism algorithm built into the embedded industrial control computer. Various solutions can be implemented for the correction mechanism and algorithm; the preferred algorithm will be explained in detail in subsequent embodiments.

[0041] Example 2:

[0042] This embodiment optimizes the system structure in several aspects based on Embodiment 1. Specifically, the IMU attitude detection module consists of an integrated three-axis gyroscope and a three-axis accelerometer, achieving an accuracy of 0.001°. Its working principle involves calculating pose by measuring angular velocity and linear acceleration. Its core technology is based on attitude angle integration and position integration. The attitude angle integral equation of the gyroscope is:

[0043] θ(t)=θ0+∫0 t ω(t)dt

[0044] Where ω is the triaxial angular velocity and θ0 is the initial attitude angle;

[0045] The position integral and the integral equation of the accelerometer are as follows:

[0046] p(t)=p0+∫∫0 t a(t)dt 2

[0047] Where a is the triaxial acceleration after deducting the gravitational component;

[0048] The laser ranging module includes at least six sets of laser ranging arrays composed of cross-shaped laser rangefinders arranged on both sides of the slot opening, with a wavelength of 532nm and a measurement accuracy of ±1mm. It also includes a laser target prism 3 mounted on the milling head. The first acoustic scanning module includes at least six sets of ultrasonic arrays arranged in a circular array within the cutter head shield of the dual-wheel milling machine, emitting at frequencies of 500kHz-2MHz. The acoustic transmitter and receiver arrays are directly embedded within the cutter head shield of the dual-wheel milling machine, arranged in a ring, spaced 60° apart, and close to the working surface of the slot wall. Since the excavation direction is always downward, the six sets of milling head probes rotate with the milling head. Therefore, the signal arrival times of different probes differ. To obtain objective and accurate information, it is necessary to fuse the data from multiple probes. This embodiment uses a time-difference positioning method, specifically: utilizing the signal arrival time difference Δt of the six sets of milling head probes. i The azimuth angle θ and distance r of the stratigraphic anomaly are calculated, thereby enabling the fusion of multi-probe data, wherein:

[0049]

[0050] Where c represents the speed of light, ε r This represents the current dielectric constant of the mud.

[0051] The advantages of this design are: 1. The acoustic transmitter and receiver array are directly embedded in the cutter head shield of the dual-wheel milling machine, arranged in a ring (6 groups, spaced 60° apart), close to the working surface of the trench wall; 2. Real-time synchronous detection of 360° ground changes around the milling head; In terms of protective design, a tungsten carbide protective cover with a thickness of 5-8mm is used, which can effectively resist the impact of gravel and avoid damage from external forces during excavation. To further highlight the technical advantages of this structural design, equivalent model tests were conducted, and the differences between using only the first acoustic scanning module or the second acoustic scanning module 7 and the dual-position synergistic effect are shown in Table 1 below.

[0052] Stratigraphic modeling resolution <![CDATA[0.5m 3 ]]> <![CDATA[0.1m 3 ]]> <![CDATA[0.05m 3 ]]> Abnormal detection rate <![CDATA[60%(>2m 3 )]]> <![CDATA[85%(>0.5m 3 )]]> <![CDATA[98%(>0.2m 3 )]]> Position control accuracy (100m) ±25mm ±15mm ±8mm Adjustment time required when encountering faults >30 minutes 10-15 minutes <5 minutes

[0053] Table 1 Comparison of technical effects using a single acoustic scanning module and a dual-module approach.

[0054] Furthermore, the surface is coated with a mud-repellent material with a contact angle >120° to reduce mud adhesion. It is worth noting that the mud-repellent material is a crucial improvement for ensuring the overall system's correction accuracy. Due to the specific working environment of this invention, mud adhesion significantly affects the accuracy of the acquired signal. The mud-repellent material used in this embodiment varies depending on the application scenario, as shown in Table 2 below:

[0055]

[0056] Table 2. Composition of sludge-dredging materials for different application scenarios

[0057] The pitch and yaw mechanisms are both adjustable at ±5°, and the axial mechanism is adjustable within ±200mm.

[0058] Further preferably, this embodiment also includes a second acoustic scanning module 7. The second acoustic scanning module 7 includes a vertical acoustic array mounted on the slot opening via an auxiliary support. Adjacent vertical acoustic units are spaced 0.6-1 meters apart, and the emitted acoustic frequency is 10kHz-50kHz, used for large-scale stratigraphic modeling. It is worth noting that the function of the second acoustic scanning module 7 is different from that of the first acoustic scanning module; their focuses and the technical problems they solve are not the same. Specifically, the second acoustic scanning module 7 uses low frequencies and addresses the understanding of a completely new geological structure. It is used to update the dynamic geological database and digital twin engine before construction, ensuring that the information held by the system remains basically consistent with the current geological conditions during construction. Its drawback is that the accuracy of the specific geological conditions at a certain depth needs improvement. Therefore, the first acoustic scanning module can work in conjunction with the second acoustic scanning module... 7. The technologies complement each other. Because the first acoustic scanning module is installed on the milling head, it descends gradually with the milling head as trenching progresses, always remaining in close contact with the current geological stratum. This allows for precise detection of the specific geology of the stratum to be excavated. For example, it can accurately detect whether the stratum to be excavated at a certain depth is hard rock, gravel, soft soil, or a loose composite geological structure prone to borehole collapse. This allows for timely adjustment of the power of the dual-wheel milling equipment, pre-adjusting the system's operating conditions to suit the current geological structure and reducing the probability of borehole collapse, stuck drill, and other accidents. In summary, the first acoustic scanning module uses high-frequency pulses with a main frequency of 500kHz and a pulse width of 1μs for high-resolution detection at close range (<5m); the second acoustic scanning module 7 uses low-frequency pulses with a main frequency of 50kHz and a pulse width of 10μs for long-distance geological imaging. The combined design and installation at different locations enable precise monitoring of the construction status before and during construction, regardless of the location, stratum, or geological conditions. This solves the problems of borehole collapse and stuck drill bits caused by inaccurate geological assessments at different depths in existing technologies. To further enhance the anti-interference capability of the first acoustic scanning module in a mud environment, this embodiment also provides an adaptive frequency selection method. By installing a density sensor on the milling head, the mud density is collected in real time and fed back to the system control center. The system optimizes the frequency according to the following mechanism, thereby improving the actual anti-interference effect, as shown in Table 3 below:

[0059] 1 <1.2 500kHz Low attenuation, high resolution 2 1.2-1.4 250kHz Balancing penetration and resolution 3 >1.4 100kHz Avoid complete attenuation of high-frequency signals

[0060] Table 3 Frequency Selection Table for Different Mud Densities

[0061] Example 3:

[0062] This embodiment provides an intelligent deviation correction method for trench excavation based on the deviation correction system provided in any of the above embodiments, specifically including the following steps:

[0063] Step STP100: Ground Penetrating Radar Scanning. The GPR ground penetrating radar emits high-frequency electromagnetic waves and receives reflected signals from the stratigraphic interfaces. The stratigraphic structure is obtained by analyzing the difference in dielectric constant.

[0064] Step STP200, acoustic strata scanning: The second acoustic scanning module emits a 50kHz low-frequency acoustic wave to scan the bottom layer, generates a global stiffness distribution map by receiving reflected signals, and uploads the data to the digital twin engine; The piezoelectric ceramic transducer of the first acoustic scanning module emits a 500kHz high-frequency acoustic pulse, penetrates the mud and enters the near-field stratum, captures the reflected signals of P-waves and S-waves through an array receiver, records the travel time and amplitude, inverts the real-time mechanical parameters and inputs them into the embedded industrial control computer;

[0065] Step STP300: Data fusion and establishment of a digital twin model M0. The attitude data of the dual-wheel milling machine collected by the IMU attitude detection module, the acoustic data B collected by the second acoustic scanning module, and the position data collected by the laser ranging module are synchronized in time and space using a high-precision clock, and coordinates are unified using the ICP algorithm to establish a geological digital twin model M0 for the current construction area. After normal tunneling, the large-scale geological trend is updated every 10 minutes. The first acoustic scanning module monitors the geological conditions within 5m around the milling head in real time, dynamically correcting the digital twin model M0. t+1 ;

[0066] Step STP400: Design a model predictive controller. Using this model predictive controller, establish the dynamic equations of the milling system to predict N. p The pose of the step is calculated, and the control torque τ required by the hydraulic servo mechanism of the dual-wheel milling machine to complete the correction is determined. control The hydraulic servo mechanism is controlled by an embedded industrial computer to correct the deviation until the positional deviation δ reaches a preset value range. The dynamic equation of the milling system is as follows:

[0067]

[0068] Where, q∈R 3 M(q) represents the milling head pose vector, ∈ R 3×3 The inertia matrix represents the mass and moment of inertia of the milling head. It is a known quantity and can be verified through an unloaded acceleration test. The Coriolis force matrix represents the motion coupling effect and is calculated online based on a kinematic model. It is used to calculate the cross product of the milling head angular velocity and linear velocity in real time; G(q)∈R 3 This represents the gravitational component, with gravitational acceleration g = 9.81 m / s². 2 The attitude sensor provides feedback on the tilt angle θ and the gravity component G. x =mgsinθ, G y =mg cosθ; Representative formation friction model, including mud and combined formation resistance; K d ∈R 3×3 Representing the elastic reaction force of the formation, the formation stiffness is inverted by acoustic scanning; δ∈R 3 The positional deviation is represented by K, which is obtained in real time through laser ranging. In this embodiment, K is updated every 0.5m. d Hard rock structure K d =MN / m, sand layer K d =20MN / m; Represents the control torque; τ disturbance ∈R 3 The disturbance torque, which is influenced by relatively minor factors and is mainly caused by geological abrupt changes and equipment vibration, is estimated using an extended state observer (ESO). The design model predictive controller calculates the optimal control sequence U by minimizing the trajectory tracking error and the rate of change of the control variable.

[0069]

[0070] Where, q ref The axis representing the design reference trajectory of the twin-wheel milling machine is a known quantity, generated through importing from design drawings or B-spline interpolation; Q represents the state deviation weight, and R represents the control quantity deviation weight. Both of these weight parameters are known and set based on empirical trial and error; N p N represents the prediction step size. c q(k|t) represents the actual control step size, which is set according to the system's dynamic response time. Different system hardware and response speeds can have different step sizes. q(k|t) represents the predicted pose, which is obtained by ground-based prediction based on the dynamic model. The recursive process is as follows:

[0071]

[0072] In each control cycle of MPC, the future N is predicted using the current state q(t) and the control sequence U. p The step trajectory; Δu(k|t) represents the rate of change of the control variable or the adjustment, which can be obtained online using an interior point solver, such as OSQP.

[0073] For ease of understanding of this embodiment, the working logic for the correction is shown in Table 4 below.

[0074]

[0075] Table 4. Correction Scenarios and Functional Component Status Table of this Embodiment

[0076] According to Table 4 above, the actual construction deployment can be divided into the initialization phase, the system adaptation and learning phase, the full-speed construction phase, and the anomaly handling phase.

[0077] The initialization phase includes initializing the parameters or states of the system's mechanical equipment, calibrating the zero points and proportional coefficients of each sensor in the data acquisition unit, such as laser ranging, IMU, and acoustic scanning.

[0078] The system adaptation and learning phase refers to the initial stage of construction where the system control actuators excavate and form trenches. On one hand, it observes whether the system has adaptability and compatibility with the current geological conditions and construction. On the other hand, it collects data for identification and updating the digital twin engine, laying the foundation for subsequent high-power tunneling. In this embodiment, this phase advances at a speed of 0.5 m / h, adjusting the state deviation weight Q to achieve a lateral deviation convergence speed t. s <3s, the system presets an error range of ±2%. When the actual tunneling depth reaches 10m, the full-speed construction stage can be carried out. Of course, for better verification and adaptation, the tunneling depth of this stage can also be defined as 15m, which can be set manually.

[0079] During the full-speed construction phase, MPC closed-loop control is initiated, and optimization is performed every 10ms; the formation stiffness K is saved every 0.5m. d And the coefficient of friction, while building components and saving construction logs.

[0080] The anomaly handling phase is characterized by sudden events that occur throughout any of the above phases. It may never happen, but it could occur at any time. Depending on the actual construction situation, when the pose deviation exceeds 2σ (assuming σ is a historical statistical value), the following three-level response is triggered:

[0081] First level: Increase the weights of the Q matrix;

[0082] Level 2: Reduce speed to 50% and recalibrate;

[0083] Level 3: Emergency stop and initiation of manual intervention.

[0084] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A trench excavation correction system, characterized in that: The system includes an embedded industrial control computer that serves as the information interaction and computing center of the system, a data acquisition unit connected to the embedded industrial control computer for data interaction and for collecting data on the formation and trenching status, a dynamic geological database and a digital twin engine. The embedded industrial control computer is also electrically connected to a hydraulic servo mechanism, an attitude adjustment mechanism and a power modulation module for controlling the twin-wheel milling equipment. The data acquisition unit includes an IMU attitude detection module and a laser ranging module for detecting the attitude and position of the dual-wheel milling machine, as well as a first acoustic scanning module for acquiring current geological conditions. The IMU attitude detection module and the first acoustic scanning module are both integrated into the head of the dual-wheel milling machine. The laser ranging module includes at least one set of laser transmitters and receivers respectively installed on the ground on both sides of the slot opening. A pressure-resistant transparent guide tube is provided below the laser transmitter for guiding the laser beam into the slot. A self-cleaning optical window is installed at the bottom of the pressure-resistant transparent guide tube, and a laser ranging target is installed on the head of the dual-wheel milling machine. The attitude adjustment mechanism consists of a pitch mechanism, a yaw mechanism, and an axial mechanism.

2. The trench excavation correction system according to claim 1, characterized in that: The IMU attitude detection module consists of an integrated three-axis gyroscope and a three-axis accelerometer. The laser ranging module includes at least six laser ranging arrays composed of cross-shaped laser rangefinders. The first acoustic scanning module includes at least six arrays arranged in a circular pattern within the cutter head shield of the dual-wheel milling machine. The adjustment angles of the pitch and yaw mechanisms are both ±5°, and the adjustment range of the axial mechanism is ±200mm.

3. The trench excavation correction system according to claim 2, characterized in that: It also includes a second acoustic scanning module, which includes a vertical acoustic array mounted on the slot via an auxiliary bracket. The interval between two adjacent vertical acoustic units is 0.6-1 meter, and the emitted acoustic frequency is 10kHz-50kHz.

4. A method for intelligent correction of deviations during trench excavation, characterized in that: The correction system described in claim 3 is used to implement this, specifically including the following steps: Step STP100: Ground Penetrating Radar Scanning. The GPR ground penetrating radar emits high-frequency electromagnetic waves and receives reflected signals from the stratigraphic interfaces. The stratigraphic structure is obtained by analyzing the difference in dielectric constant. Step STP200, acoustic strata scanning: The second acoustic scanning module emits a 50kHz low-frequency acoustic wave to scan the bottom layer, generates a global stiffness distribution map by receiving reflected signals, and uploads the data to the digital twin engine; The piezoelectric ceramic transducer of the first acoustic scanning module emits a 500kHz high-frequency acoustic pulse, penetrates the mud and enters the near-field stratum, captures the reflected signals of P-waves and S-waves through an array receiver, records the travel time and amplitude, inverts the real-time mechanical parameters and inputs them into the embedded industrial control computer; Step STP300: Data fusion and establishment of a digital twin model M0. The attitude data of the dual-wheel milling machine collected by the IMU attitude detection module, the acoustic data B collected by the second acoustic scanning module, and the position data collected by the laser ranging module are synchronized in time and space using a high-precision clock, and coordinates are unified using the ICP algorithm to establish a geological digital twin model M0 for the current construction area. After normal tunneling, the large-scale geological trend is updated every 10 minutes. The first acoustic scanning module monitors the geological conditions within 5m around the milling head in real time, dynamically correcting the digital twin model M0. t+1 ; Step STP400: Design a model predictive controller. Using this model predictive controller, establish the dynamic equations of the milling system to predict N. p The pose of the step is calculated, and the control torque τ required by the hydraulic servo mechanism of the dual-wheel milling machine to complete the correction is determined. control The hydraulic servo mechanism is controlled by an embedded industrial computer to correct the deviation until the positional deviation δ reaches a preset value range. The dynamic equation of the milling system is as follows: Where, q∈R 3 M(q) represents the milling head pose vector, ∈ R 3×3 Represents the inertia matrix; The Coriolis force matrix represents the motion coupling effect; G(q)∈R 3 Represents the gravitational component; Represents the formation friction model; K d ∈R 3×3 Represents the elastic reaction force of the formation; δ∈R 3 Represents the pose deviation; Represents the control torque; τ disturbance ∈R 3 Representing the disturbance torque; the design model predictive controller calculates the optimal control sequence U by minimizing the trajectory tracking error and the rate of change of the control quantity: Where, q ref The axis represents the design reference trajectory of the twin-wheel milling machine; Q represents the state deviation weight; R represents the control quantity deviation weight; N p N represents the prediction step size; c q(k|t) represents the actual control step size; q(k|t) represents the predicted pose; Δu(k|t) represents the rate of change or adjustment of the control quantity.

Images