Red sandstone multi-bench combined mechanical cold excavation construction method

Abstract

The invention relates to the technical field of civil engineering, in particular to a red sandstone multi-bench combined mechanical cold excavation construction method which comprises the following steps: S1, generating a rock mass model by adopting three-dimensional laser scanning and AI geological analysis, and dynamically optimizing mechanical configuration and excavation parameters by combining a BIM numerical simulation rehearsal bench scheme; accurate rehearsal and parameter optimization of a stair flight scheme are achieved through three-dimensional laser scanning, AI geological analysis and BIM numerical simulation, a scientific basis is provided for construction, dynamic adaptation and process correction of excavation operation are achieved with the help of rock mass hardness sensing, intelligent parameter matching and particle size recognition technologies, the construction efficiency and the crushing quality are improved, and the construction cost is reduced. Real-time monitoring, trend prediction and grading response of slope stability are achieved through a multi-source sensing network, a digital twin model and edge calculation, construction safety is guaranteed, and the intelligent level, operation efficiency and safety controllability of red sandstone multi-bench mechanical cold excavation are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of civil engineering, in particular to a red sandstone multi-stage combined mechanical cold excavation construction method. BACKGROUND

[0002] Red sandstone is a sedimentary rock mainly composed of sand particles cemented by cementing materials such as siliceous, calcareous and ferruginous materials, and presents red color due to the presence of iron oxide and other pigment minerals. It is formed in terrestrial or shallow marine sedimentary environment, and has complex lithology and special physical and mechanical properties: the strength of the fresh rock mass is high, but the integrity varies significantly due to the type of cementing material and the degree of weathering, especially after being soaked in water, it is easy to soften, and the joint fissure is developed and the strength is dispersed, which brings many challenges to engineering excavation.

[0003] In the prior art, red sandstone excavation is mostly carried out by blasting or single mechanical excavation: although traditional blasting excavation has high efficiency, it will cause strong disturbance to the rock mass, aggravate the joint expansion and slope instability risk, and the blasting vibration and dust pollution do not meet the environmental protection requirements; single mechanical excavation often has problems such as poor parameter matching, fast equipment wear and tear, and uneven broken particle size due to the inability to adapt to the strength fluctuation characteristics of red sandstone, resulting in low excavation efficiency, and it is difficult to early warn the slope stability through effective monitoring, further increasing the construction safety hazards.

[0004] Therefore, the present application provides a red sandstone multi-stage combined mechanical cold excavation construction method to solve the above technical problems. SUMMARY

[0005] The present application provides a red sandstone multi-stage combined mechanical cold excavation construction method, which realizes precise pre-performance and parameter optimization of the stage scheme through three-dimensional laser scanning, AI geological analysis and BIM numerical simulation, provides scientific basis for construction, realizes dynamic adaptation and process correction of excavation operation by means of rock hardness sensing, intelligent parameter matching and particle size identification technology, improves construction efficiency and broken quality, realizes real-time monitoring, trend prediction and hierarchical response of slope stability through multi-source sensing network, digital twin model and edge computing, ensures construction safety, and significantly improves the intelligent level, operation efficiency and safety controllability of red sandstone multi-stage mechanical cold excavation.

[0006] To achieve the above purpose, the present application provides the following technical scheme:

[0007] The present application provides a red sandstone multi-stage combined mechanical cold excavation construction method, comprising the following steps:

[0008] S1: Adopt three-dimensional laser scanning and AI geological analysis to generate rock mass model, combine with BIM numerical simulation to pre-play bench scheme, dynamically optimize mechanical configuration and excavation parameters;

[0009] S2: According to the generated rock mass model and optimized parameters, combine with rock layer occurrence and stability requirements to divide multi-level bench, determine each layer excavation contour line and control elevation through measurement and lofting;

[0010] S3: Combination of hydraulic breaking hammer, excavator and cutting equipment for layered excavation, dynamic adjustment of operation parameters through rock hardness sensor and intelligent parameter matching system, combined with bucket material particle size identification device for reverse correction of breaking process;

[0011] S4: Mechanically trim the bench working face to the designed slope ratio, use closed transportation system to clean and transport broken materials to the designated stockyard in real time;

[0012] S5: Lay multi-source sensor network at key bench, connect monitoring data to slope digital twin model, predict stability trend through edge computing, automatically trigger construction parameter adjustment instruction when monitoring value approaches threshold, dynamically optimize next level excavation footage, mechanical operation intensity or start prestressed anchoring system.

[0013] In S1, three-dimensional laser scanning and AI geological analysis are used to generate rock mass model, combined with BIM numerical simulation to pre-play bench scheme, dynamically optimize mechanical configuration and excavation parameters, the specific steps are as follows:

[0014] S1.1: Collect red sandstone outcrop data quickly through three-dimensional laser scanner, combine with AI algorithm to analyze rock mass strength, joint distribution and engineering geological characteristics;

[0015] S1.2: Based on scanning data and geological analysis results, construct three-dimensional visual digital model containing rock mass structure surface, weathering degree and parameters;

[0016] S1.3: Pre-play different bench excavation schemes in digital model, evaluate stability through mechanical simulation calculation, automatically output optimal mechanical combination, bench height and excavation sequence construction parameters.

[0017] The specific steps in S1.3 are as follows:

[0018] S1.3.1: Set multiple bench division schemes in three-dimensional visual digital model, each scheme includes different bench height, slope ratio and excavation sequence;

[0019] S1.3.2: Perform finite element or discrete element mechanical simulation on each scheme, calculate the maximum displacement of the slope, stress concentration area and overall stability safety factor during excavation;

[0020] The slope stability safety factor K is calculated by the following formula:

[0021]

[0022] wherein c i is the cohesion of the ith sliding surface, L i is the length of the ith sliding surface, sigma i is the normal stress, is the internal friction angle, tau i is the shear stress, u i is the pore water pressure;

[0023] S1.3.3: Based on the simulation results, screen the scheme that meets the safety factor K >= 1.3 and has the highest mechanical operation efficiency;

[0024] S1.3.4: Automatically output the optimal mechanical combination, bench height and excavation sequence corresponding to the scheme.

[0025] In S2, according to the generated rock mass model and optimized parameters, combined with the occurrence of rock strata and stability requirements, multi-level benches are divided, and the excavation contour line and control elevation of each layer are determined by measurement and lofting, and the specific steps are as follows:

[0026] S2.1: Based on the three-dimensional visual digital model generated in S1.2 and the optimized parameters output in S1.3, combined with the geological characteristics of the rock strata dip angle and joint development degree, the bench height grading and slope angle design are carried out;

[0027] S2.2: On-site measurement and lofting are carried out using a total station, and the designed graded bench is converted into field markings to clearly define the excavation boundary line and slope top line position of each bench;

[0028] S2.3: The elevation control network for bench excavation is established by scanning with a three-dimensional laser scanner, and construction control elevation reference points are set for each level of bench to guide the mechanical excavation depth control.

[0029] In S3, hydraulic breaking hammer, excavator and cutting equipment are combined for layered excavation, the operation parameters are dynamically adjusted by rock hardness sensor and intelligent parameter matching system, and the breaking process is corrected in reverse by the bucket material particle size identification device, and the specific steps are as follows:

[0030] S3.1: The hydraulic breaking hammer is combined to implement primary rock breaking, the excavator is used to load and transport the broken material, and the cutting equipment is used to complete the contour finishing;

[0031] S3.2: The rock compressive strength is monitored in real time by the rock hardness sensor, and the intelligent control system automatically matches the optimal operation parameters of the breaking hammer impact frequency and the excavator bucket cutting angle;

[0032] S3.3: The vision recognition system carried by the bucket analyzes the size distribution of the crushed material, and feeds back to the intelligent control system to adjust the crushing energy output in real time.

[0033] The specific steps in S3.2 are as follows:

[0034] S3.2.1: The rock hardness sensor collects red sandstone compressive strength data in real time at a sampling frequency of 50-100 Hz, with a monitoring range of 5-30 MPa. The data is transmitted to the intelligent control system after filtering processing;

[0035] S3.2.2: The intelligent control system calls the preset parameter database, which contains the reference operation parameters corresponding to different compressive strength intervals: when the strength is 5-15 MPa, the impact frequency of the crushing hammer is preset to 1200-1500 times / min, and the cutting angle of the excavator bucket is preset to 40°-50°; when the strength is 15-30 MPa, the impact frequency of the crushing hammer is preset to 1500-1800 times / min, and the cutting angle of the excavator bucket is preset to 30°-40°;

[0036] S3.2.3: The system dynamically corrects according to the deviation of the real-time monitoring value and the reference parameter: if the measured strength is higher than the upper limit value of the current interval, automatically increase the crushing hammer frequency by 5%-10%, and at the same time reduce the cutting angle of the excavator bucket by 5°-10°; if the measured strength is lower than the lower limit value of the current interval, automatically reduce the crushing hammer frequency by 5%-10%, and at the same time increase the cutting angle of the excavator bucket by 5°-10°;

[0037] S3.2.4: The corrected parameters are transmitted to the intelligent control system through the wireless transmission module, and the response delay is controlled within 0.3-0.5s.

[0038] The specific steps in S3.3 are as follows:

[0039] S3.3.1: The industrial camera installed on the bucket collects the image of the crushed material pile at a frame rate of ≥30fps;

[0040] S3.3.2: Multi-scale morphological segmentation algorithm is used to identify the proportion of oversized particles with a particle size of >100mm;

[0041] S3.3.3: When the proportion of oversized particles exceeds 15%, trigger the third level energy adjustment:

[0042] ① When the proportion of oversized particles is 15%-25%, it is the first level adjustment: increase the impact energy of the crushing hammer by 10%-15%;

[0043] ② When the proportion of oversized particles is 25%-35%, it is the second level adjustment: increase the impact energy by 15%-25% and reduce the excavation footage rate;

[0044] ③ When the percentage of oversized particles is > 35%, it is a third-level adjustment: suspend excavation and start the secondary crushing program;

[0045] S3.3.4: Transmit the adjustment instruction to the crushing hammer PLC controller through the OPC UA protocol.

[0046] In S4, the working face of the ladder section is mechanically trimmed to the design slope, and the closed transportation system is used to clean and transport the crushed material to the designated stockpile in real time. The specific steps are as follows:

[0047] S4.1: Use an intelligent grader equipped with a laser guidance system to accurately trim the excavation surface according to the design slope;

[0048] S4.2: Real-time transfer of crushed material to designated stockpile through belt conveyor system, with supporting dust removal and noise reduction measures.

[0049] In S5, a multi-source sensor network is laid out at key ladder sections, and monitoring data is connected to the digital twin model of the slope. Through edge computing, the stability trend is predicted, and when the monitoring value approaches the threshold, the construction parameter adjustment instruction is automatically triggered. The next level of excavation footage, mechanical operation intensity, or the pre-stressed anchoring system is dynamically optimized. The specific steps are as follows:

[0050] S5.1: Install sensor arrays of displacement meters, stress meters, and osmotic pressure meters at key ladder sections to build a real-time monitoring network and comprehensively collect slope deformation and mechanical response data;

[0051] S5.2: Real-time connection of on-site monitoring data to the digital twin model of the slope, dynamic synchronization update and interactive verification of physical slope and virtual model;

[0052] S5.3: Real-time data analysis based on the edge computing platform, prediction of slope stability evolution trend through machine learning algorithm, and early identification of potential risks;

[0053] S5.4: When the monitoring data reaches the third warning threshold in S5.3.2, the graded response mechanism is automatically triggered:

[0054] ① When the primary warning threshold is reached, the next level of excavation footage is dynamically optimized to 80% of the original plan;

[0055] ② When the intermediate warning threshold is reached, in addition to adjusting the excavation footage, the mechanical operation intensity is reduced to 70%-80% of the original parameters;

[0056] ③ When the advanced warning threshold is reached, excavation work is immediately suspended, the pre-stressed anchoring system is started, and the next level of excavation parameters is adjusted to the safety range.

[0057] The specific steps in S5.3 are as follows:

[0058] S5.3.1: Deploy a lightweight LSTM neural network at the edge computing node, input time series monitoring data including displacement rate, pore water pressure, tilt angle, sampling interval ≤ 5 minutes;

[0059] S5.3.2: Set up a three-level early warning mechanism:

[0060] ① Primary warning: When the predicted displacement in the next 24 hours is > 8mm, mark yellow warning;

[0061] ② Intermediate warning: When the predicted displacement in the next 12 hours is > 10mm, mark orange warning;

[0062] ③ High-level warning: When the predicted displacement in the next 6 hours is > 15mm, mark red warning;

[0063] S5.3.3: When the deviation of the predicted result is > 20% for 3 consecutive times, automatically trigger online fine-tuning of the model;

[0064] S5.3.4: Push the warning result to the on-site monitoring terminal through the MQTT protocol.

[0065] Compared with the prior art, the beneficial effects of the present application are:

[0066] The present application realizes precise pre-performance and parameter optimization of the gradient scheme through three-dimensional laser scanning, AI geological analysis and BIM numerical simulation, provides scientific basis for construction, and realizes dynamic adaptation and process correction of excavation operation through rock hardness sensing, intelligent parameter matching and particle size recognition technology, improves construction efficiency and fragmentation quality, realizes real-time monitoring, trend prediction and hierarchical response of slope stability through multi-source sensing network, digital twin model and edge computing, ensures construction safety, and significantly improves the intelligent level, operation efficiency and safety controllability of red sandstone multi-gradient mechanical cold excavation. BRIEF DESCRIPTION OF DRAWINGS

[0067] Figure 1 The flowchart of the present application, a red sandstone multi-gradient combined mechanical cold excavation construction method.

[0068] Figure 2 The flowchart of rock body modeling and simulation optimization in the present application, a red sandstone multi-gradient combined mechanical cold excavation construction method.

[0069] Figure 3 The flowchart of slope monitoring and early warning in the present application, a red sandstone multi-gradient combined mechanical cold excavation construction method.

[0070] Figure 4 The flowchart of hierarchical response execution in the present application, a red sandstone multi-gradient combined mechanical cold excavation construction method. DETAILED DESCRIPTION

[0071] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0072] Example:

[0073] like Figures 1-4 As shown, this embodiment provides a multi-stage combined mechanical cold excavation construction method for red sandstone, including the following steps: S1: A rock mass model is generated using three-dimensional laser scanning and AI geological analysis, and the stage plan is pre-simulated using BIM numerical simulation to dynamically optimize the mechanical configuration and excavation parameters; S2: Based on the generated rock mass model and optimized parameters, multi-stage stages are divided according to the rock strata attitude and stability requirements, and the excavation outline and control elevation of each layer are determined by measurement and layout; S3: Hydraulic breakers, excavators, and cutting equipment are used in combination for layered excavation, and the operating parameters are dynamically adjusted by rock mass hardness sensors and intelligent parameter matching systems, combined with a bucket material particle size identification device to reverse correct the crushing process; S4: The working face of the stage is mechanically trimmed to the design slope, and the crushed material is transported to the designated stockpile in real time using a closed transportation system; S5: A multi-source sensor network is deployed in key stages, and the monitoring data is connected to the slope digital twin model. The stability trend is predicted by edge computing, and when the monitoring value approaches the threshold, the construction parameter adjustment command is automatically triggered to dynamically optimize the next level of excavation advance, mechanical operation intensity, or to start the prestressed anchoring system.

[0074] In this embodiment, it should also be noted that in S1, a rock mass model is generated using 3D laser scanning and AI geological analysis. Combined with BIM numerical simulation, the stepped excavation scheme is pre-planned, and the mechanical configuration and excavation parameters are dynamically optimized. The specific steps are as follows: S1.1: Data on red sandstone outcrops is rapidly collected using a 3D laser scanner, and the engineering geological characteristics of rock mass strength and joint distribution are analyzed using AI algorithms. S1.2: Based on the scanning data and geological analysis results, a 3D visualization digital model containing parameters such as rock mass structural surfaces and weathering degree is constructed. S1.3: Different stepped excavation schemes are pre-planned in the digital model. Stability is evaluated through mechanical simulation calculations, and construction parameters for the optimal mechanical combination, stepped height, and excavation sequence are automatically output. The specific steps are as follows: S1.3.1: Multiple stepped division schemes are set in the 3D visualization digital model. Each scheme includes different stepped heights, slope ratios, and excavation sequences. S1.3.2: Finite element or discrete element mechanical simulations are performed on each scheme to calculate the maximum displacement of the slope, stress concentration zone, and overall stability safety factor during excavation. The slope stability safety factor K is calculated using the following formula:

[0075]

[0076] wherein c i is the cohesion of the ith slip surface, L i is the length of the ith slip surface, σ i is the normal stress, is the internal friction angle, τ i is the shear stress, u i is the pore water pressure; S1.3.3: based on the simulation results, screening the scheme that meets the safety factor K≥1.3 and has the highest mechanical operation efficiency; S1.3.4: automatically outputting the optimal mechanical combination, gradient height and excavation sequence corresponding to the scheme.

[0077] Further, it needs to be explained that a BIM software (such as Revit) is used to construct a 1:1 scale three-dimensional visual digital model, and the model contains the following key parameter layers: I, rock mass structural plane layer: labeling the spatial position, ductility and filling properties of joint surface; II, weathering partition layer: different color scales are given according to weathering degree (fully weathered area is red, strongly weathered area is yellow, and moderately weathered area is blue); III, mechanical parameter layer: relating physical and mechanical indexes such as rock mass elastic modulus (5-20 GPa) and Poisson's ratio (0.25-0.35) of each region.

[0078] In the present embodiment, it also needs to be explained that in S2, according to the generated rock mass model and optimized parameters, multi-level gradient sections are divided in combination with the occurrence of rock strata and stability requirements, and the excavation contour line and control elevation of each layer are determined by measurement and lofting, and the specific steps are as follows: S2.1: based on the three-dimensional visual digital model generated in S1.2 and the optimized parameters output in S1.3, in combination with the geological characteristics of rock strata inclination and joint development degree, gradient height grading and slope angle design are carried out; S2.2: field measurement and lofting are carried out by using a total station, the designed graded gradient sections are converted into field marks, and the excavation boundary line and slope top line positions of each gradient section are determined; S2.3: the elevation control network of gradient excavation is established by scanning with a three-dimensional laser scanner, and construction control elevation reference points are set for each gradient section to guide the control of mechanical excavation depth.

[0079] Further, it needs to be explained that the gradient height grading standard is: when the rock strata inclination is ≤30° and the joint spacing is >1m (relatively complete rock mass), the gradient height is taken as 8-10m; when the rock strata inclination is 30°-60° or the joint spacing is 0.5-1m (moderately broken), the gradient height is taken as 5-8m; when the rock strata inclination is >60° or the joint spacing is <0.5m (broken rock mass), the gradient height is ≤5m, and a 2m wide buffer platform is added every 3 gradient sections.

[0080] The formula for designing the slope angle is: slope angle α = β - γ, wherein β is the natural repose angle of the rock mass (35°-45° for red sandstone), and γ is the correction coefficient (5°-10° for joint development, and 3°-5° for rock strata inclined in the same direction).

[0081] In the present embodiment, it is also necessary to point out that in S3, the hydraulic breaking hammer, excavator and cutting equipment are combined to carry out layer-by-layer excavation, the operation parameters are dynamically adjusted by the rock hardness sensor and the intelligent parameter matching system, and the broken material particle size recognition device is reversely corrected. The specific steps are as follows: S3.1: the hydraulic breaking hammer is combined to implement primary breaking of the rock mass, the excavator is used to load and transport the broken material, and the cutting equipment is used to complete contour finishing; S3.2: the rock hardness sensor is used to monitor the compressive strength of the rock in real time, and the intelligent control system automatically matches the optimal operation parameters of the breaking hammer impact frequency and the excavator bucket cutting angle; the specific steps are as follows: S3.2.1: the rock hardness sensor collects the compressive strength data of red sandstone in real time at a sampling frequency of 50-100 Hz, the monitoring range covers 5-30 MPa, and the data is transmitted to the intelligent control system after filtering processing; S3.2.2: the intelligent control system calls the preset parameter database, which contains the reference operation parameters corresponding to different compressive strength intervals: when the strength is 5-15 MPa, the breaking hammer impact frequency is preset to 1200-1500 times / min, and the excavator bucket cutting angle is preset to 40°-50°; when the strength is 15-30 MPa, the breaking hammer impact frequency is preset to 1500-1800 times / min, and the excavator bucket cutting angle is preset to 30°-40°; S3.2.3: the system dynamically corrects according to the deviation of the real-time monitoring value and the reference parameter: if the measured strength is higher than the upper limit value of the current interval, the breaking hammer frequency is automatically increased by 5%-10%, and the excavator bucket cutting angle is simultaneously reduced by 5°-10°; if the measured strength is lower than the lower limit value of the current interval, the breaking hammer frequency is automatically reduced by 5%-10%, and the excavator bucket cutting angle is simultaneously increased by 5°-10°; S3.2.4: the corrected parameters are synchronously transmitted to the intelligent control system through the wireless transmission module, and the response delay is controlled within 0.3-0.5 s. S3.3: the visual recognition system carried by the bucket is used to analyze the particle size distribution of the broken material, and the feedback is transmitted to the intelligent control system to adjust the broken energy output in real time. The specific steps are as follows: S3.3.1: the industrial camera installed on the bucket collects the broken material pile image at a frame rate of ≥30 fps; S3.3.2: the multiscale morphological segmentation algorithm is used to identify the proportion of super-diameter particles with a particle size of >100 mm; S3.3.3: when the proportion of super-diameter particles exceeds 15%, the three-level energy adjustment is triggered: ① when the proportion of super-diameter particles is 15%-25%, the first-level adjustment is carried out: the breaking hammer impact energy is increased by 10%-15%; ② when the proportion of super-diameter particles is 25%-35%, the second-level adjustment is carried out: the impact energy is increased by 15%-25% and the excavation footage rate is reduced; ③ when the proportion of super-diameter particles is >35%, the third-level adjustment is carried out: the excavation is suspended and the secondary breaking program is started; S3.3.4: the adjustment instruction is transmitted to the breaking hammer PLC controller through the OPC UA protocol.

[0082] Further, it needs to be explained that the preset parameter database in S3.2.2 is constructed by "field test + machine learning optimization": 1. Early test: more than 100 groups of control tests are conducted in typical red sandstone areas (5-30 MPa strength interval) to record the relationship between the frequency of the breaking hammer, the angle of the bucket, and the breaking efficiency and the equipment loss rate under different strengths; 2. Database optimization: the test data is trained using a BP neural network, and the "highest efficiency and loss rate <5%" is taken as the target to output the reference parameters for each strength interval; 3. Dynamic calibration: every 5000m 3 of construction, the actual data (such as the particle size qualification rate in S3.3 feedback) is returned to the database to automatically correct the reference parameters (correction amplitude ≤10%). The industrial camera is equipped with a dustproof glass cover and a light supplement lamp, and image acquisition automatically performs white balance correction and defogging processing to ensure clear imaging under low illumination or dust conditions. The steps of the multi-scale morphological segmentation algorithm in S3.3.2 are as follows: 1. Image preprocessing: the images collected by the industrial camera are grayed, denoised (Gaussian filtering), and contrast enhanced; 2. Multi-scale segmentation: three different size (3x3, 5x5, 7x7) structure elements are used to perform morphological gradient operation on the image to extract the edges of particles of different sizes; 3. Particle size calculation: the pixel calibration (1 pixel = 0.5 mm) is used to convert the circumscribed circle diameter of the particles in the image into the actual particle size, and the proportion of the number of particles >100 mm is counted; 4. Accuracy verification: manual sieving method (100 mm sieve) is used to calibrate the recognition results every hour to ensure that the error is ≤3%.

[0083] In this embodiment, it also needs to be explained that in S4, the working surface of the gradient section is mechanically trimmed to the designed slope rate, and a closed transportation system is used to real-time clean and transport the broken materials to the designated stockyard, and the specific steps are as follows: S4.1: an intelligent grader equipped with a laser guidance system is used to accurately trim the excavation surface according to the designed slope rate; S4.2: the broken materials are real-time transported to the designated stockyard through a belt conveying system, and dust removal and noise reduction measures are matched.

[0084] Further, it needs to be explained that the belt conveying system is equipped with a fully enclosed guide groove and a negative pressure dust removal device, and the dust emission concentration during conveying is ≤10 mg / m 3 , and the noise control is below 75 dB(A).

[0085] In this embodiment, it also needs to be explained that in S5, a multi-source sensor network is arranged in the key section, the monitoring data is accessed into the slope digital twin model, the stability trend is predicted through edge computing, and the construction parameter adjustment instruction is automatically triggered when the monitoring value approaches the threshold value, the next level excavation footage, mechanical operation intensity or the prestressed anchoring system is dynamically optimized, and the specific steps are as follows: S5.1: installing a sensor array of displacement meters, stress meters and osmotic pressure meters in the key section, constructing a real-time monitoring network, and comprehensively collecting slope deformation and mechanical response data; S5.2: real-time access of the on-site monitoring data into the slope digital twin model, dynamic synchronous updating and interactive verification of the physical slope and the virtual model; S5.3: real-time data analysis based on the edge computing platform, prediction of the slope stability evolution trend through the machine learning algorithm, and early identification of potential risks; the specific steps are as follows: S5.3.1: deploying a lightweight LSTM neural network on the edge computing node, inputting time series monitoring data including displacement rate, pore water pressure and inclination angle, and sampling interval ≤5 minutes; S5.3.2: setting a three-level early warning mechanism: ① primary warning: when the predicted displacement in the next 24 hours is >8mm, mark yellow warning; ② intermediate warning: when the predicted displacement in the next 12 hours is >10mm, mark orange warning; ③ advanced warning: when the predicted displacement in the next 6 hours is >15mm, mark red warning; S5.3.3: when the deviation of the continuous 3 times of prediction results is >20%, automatically trigger the online fine tuning of the model; S5.3.4: pushing the warning results to the on-site monitoring terminal through the MQTT protocol. S5.4: when the monitoring data reaches the three-level warning threshold in S5.3.2, automatically trigger the graded response mechanism: ① when the primary warning threshold is reached, dynamically optimize the next level excavation footage to 80% of the original plan; ② when the intermediate warning threshold is reached, in addition to adjusting the excavation footage, reduce the mechanical operation intensity to 70%-80% of the original parameters; ③ when the advanced warning threshold is reached, immediately suspend the excavation operation, start the prestressed anchoring system and adjust the next level excavation parameters to the safety range.

[0086] Further, it needs to be explained that the training data source and model structure of the lightweight LSTM neural network: 1. Training data: containing 3 types of samples (totaling 100,000+ groups): historical engineering data: monitoring data (displacement, stress, osmotic pressure) and instability cases similar to the red sandstone slope; simulation data: 2000 kinds of slope evolution data under different working conditions generated through the mechanical simulation of S1.3; real-time data on site: monitoring data (1 group every 5 minutes) in the first 3 months of construction. 2. Network structure: input layer (3 features: displacement rate, pore water pressure, inclination angle)→2 layers of LSTM hidden layer (64 neurons each)→output layer (predicted displacement value in the next 24h / 12h / 6h); 3. Loss function: mean square error is adopted, and the training is stopped when the prediction error is ≤8%.

[0087] In the description of the specification, reference to "one embodiment", "an example", "a specific example" or the like means that a particular feature, structure, material or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. The appearances of the phrases "in one embodiment", "an example", "a specific example" or the like in various places in the specification are not necessarily referring to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0088] The preferred embodiments of the application disclosed above are only to help explain the application. The preferred embodiments do not describe all the details of the application and limit the application to the specific embodiments described. Obviously, many modifications and variations can be made in light of the contents of the specification. The specification selects and specifically describes these embodiments in order to better explain the principles and practical application of the application, so that those skilled in the art can well understand and utilize the application. The application is limited only by the claims and their full scope and equivalents.

Claims

1. A red sandstone multi-stage combined mechanical cold excavation construction method, characterized in that, The method comprises the following steps: S1: generating a rock mass model by using three-dimensional laser scanning and AI geological analysis, combining BIM numerical simulation to pre-play benching scheme, dynamically optimizing mechanical configuration and excavation parameters; S2: according to the generated rock mass model and optimized parameters, combining the occurrence and stability requirements of rock strata, dividing multi-stage benches, determining the excavation contour line and control elevation of each layer by measurement and lofting; S3: combined use of hydraulic breaking hammer, excavator and cutting equipment for layered excavation, dynamically adjusting operation parameters through rock hardness sensor and intelligent parameter matching system, and reverse correcting the breaking process combined with the bucket material particle size identification device; S4: mechanically finishing the bench working face to the designed slope ratio, and using a closed transportation system to clean and transport the broken material to the designated stockpile in real time; S5: laying a multi-source sensing network at key benches, connecting monitoring data to the digital twin model of the slope, predicting the stability trend through edge computing, and automatically triggering construction parameter adjustment instructions when the monitoring value approaches the threshold, dynamically optimizing the next stage excavation footage, mechanical operation intensity or starting the prestressed anchoring system.

2. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 1, characterized in that, In S1, the three-dimensional laser scanning and AI geological analysis are used to generate a rock mass model, combined with BIM numerical simulation to pre-play benching scheme, dynamically optimizing mechanical configuration and excavation parameters, the specific steps are as follows: S1.1: quickly collect red sandstone outcrop data by three-dimensional laser scanner, and analyze the engineering geological properties of rock mass strength and joint distribution combined with AI algorithm; S1.2: based on the scanning data and geological analysis results, a three-dimensional visual digital model containing rock mass structure surface and weathering degree parameters is constructed; S1.3: pre-play different benching excavation schemes in the digital model, evaluate the stability through mechanical simulation calculation, and automatically output the construction parameters of the optimal mechanical combination, bench height and excavation sequence.

3. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 2, characterized in that, The specific steps in S1.3 are as follows: S1.3.1: set multiple benching division schemes in the three-dimensional visual digital model, each scheme including different bench height, slope ratio and excavation sequence; S1.3.2: perform finite element or discrete element mechanical simulation on each scheme, calculate the maximum displacement of the slope, stress concentration area and overall stability safety factor during excavation; wherein the slope stability safety factor K is calculated by the following formula: where c i is the cohesion of the ith slip surface, L i is the length of the ith slip surface, σ i is the normal stress, is the internal friction angle, τ i is the shear stress, u i is the pore water pressure; S1.3.3: based on the simulation results, select the scheme that meets the safety factor K≥1.3 and has the highest mechanical operation efficiency; S1.3.4: automatically output the optimal mechanical combination, bench height and excavation sequence corresponding to the scheme.

4. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 1, characterized in that, In S2, according to the generated rock mass model and optimized parameters, combined with the occurrence and stability requirements of rock strata, multi-stage benches are divided, and the excavation contour line and control elevation of each layer are determined by measurement and lofting, the specific steps are as follows: S2.1: based on the three-dimensional visual digital model generated in S1.2 and the optimization parameters output in S1.3, combined with the geological characteristics of rock strata inclination and joint development degree, design the bench height classification and slope angle; S2.2: use a total station to measure and loft on site, convert the designed classified benches into field markings, and clearly define the excavation boundary and slope top line positions of each bench. S2.3: Establish the elevation control network of the bench excavation by the three-dimensional laser scanner, and set up the construction control elevation reference points for each bench to guide the mechanical excavation depth control.

5. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 1, characterized in that, The S3 combines the use of hydraulic breaking hammer, excavator and cutting equipment to excavate layer by layer, dynamically adjusts the operation parameters through the rock hardness sensor and intelligent parameter matching system, and reversely corrects the breaking process combined with the shovel material particle size identification device. The specific steps are as follows: S3.1: The combination of hydraulic breaking hammer is used to implement primary breaking of rock mass, excavator is used to load and transport broken materials, and cutting equipment is used to complete contour finishing; S3.2: The rock compressive strength is monitored in real time by the rock hardness sensor, and the intelligent control system automatically matches the optimal operation parameters of the breaking hammer impact frequency and the excavator bucket cutting angle; S3.3: The visual recognition system carried by the shovel is used to analyze the particle size distribution of the broken material, and the feedback is transmitted to the intelligent control system to adjust the breaking energy output in real time.

6. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 5, characterized in that, The specific steps in S3.2 are as follows: S3.2.1: The rock hardness sensor collects the red sandstone compressive strength data in real time at a sampling frequency of 50-100 Hz, and the monitoring range covers 5-30 MPa. The data is transmitted to the intelligent control system after filtering processing; S3.2.2: The intelligent control system calls the preset parameter database, which contains the reference operation parameters corresponding to different compressive strength intervals: when the strength is 5-15 MPa, the breaking hammer impact frequency is preset to 1200-1500 times / min, and the excavator bucket cutting angle is preset to 40°-50°; when the strength is 15-30 MPa, the breaking hammer impact frequency is preset to 1500-1800 times / min, and the excavator bucket cutting angle is preset to 30°-40°; S3.2.3: The system dynamically corrects according to the deviation of the real-time monitoring value and the reference parameter: if the measured strength is higher than the upper limit value of the current interval, the breaking hammer frequency is automatically increased by 5%-10%, and the excavator bucket cutting angle is simultaneously reduced by 5°-10°; if the measured strength is lower than the lower limit value of the current interval, the breaking hammer frequency is automatically reduced by 5%-10%, and the excavator bucket cutting angle is simultaneously increased by 5°-10°; S3.2.4: The corrected parameters are transmitted to the intelligent control system through the wireless transmission module, and the response delay is controlled within 0.3-0.5s.

7. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 5, characterized in that, The specific steps in S3.3 are as follows: S3.3.1: The industrial camera mounted on the shovel collects the broken material pile image at a frame rate of ≥30fps; S3.3.2: The multi-scale morphological segmentation algorithm is used to identify the proportion of super-diameter particles with a particle size of >100mm; S3.3.3: When the proportion of super-diameter particles exceeds 15%, trigger the three-level energy adjustment: ① When the proportion of super-diameter particles is 15%-25%, it is a first-level adjustment: increase the breaking hammer impact energy by 10%-15%; ② When the proportion of super-diameter particles is 25%-35%, it is a second-level adjustment: increase the impact energy by 15%-25% and reduce the excavation footage rate; ③ When the proportion of super-diameter particles is >35%, it is a third-level adjustment: suspend excavation and start the secondary breaking program; S3.3.4: The adjustment instruction is transmitted to the breaking hammer PLC controller through the OPC UA protocol.

8. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 1, characterized in that, The S4 is characterized in that the bench face is mechanically trimmed to the design slope, and a closed transportation system is used to real-time clean and transport the broken materials to the designated stockyard, and the specific steps are as follows: S4.1: The intelligent grader is equipped with a laser guidance system, and the excavation surface is accurately trimmed according to the design slope; S4.2: The broken materials are real-time transported to the designated stockyard through the belt conveying system, and the dust removal and noise reduction measures are matched.

9. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 1, characterized in that, The S5 is characterized in that a multi-source sensing network is arranged at the key bench, the monitoring data is connected to the digital twin model of the slope, the stability trend is predicted through edge computing, the construction parameter adjustment instruction is automatically triggered when the monitoring value approaches the threshold value, the next level excavation footage, mechanical operation intensity or the prestressed anchoring system is dynamically optimized, and the specific steps are as follows: S5.1: The sensor array of displacement meter, stress meter and osmotic pressure meter is installed at the key bench, the real-time monitoring network is constructed, and the slope deformation and mechanical response data are comprehensively collected; S5.2: The real-time monitoring data is connected to the digital twin model of the slope, and the physical slope and the virtual model are dynamically synchronized and interactively verified; S5.3: Real-time data analysis is carried out based on the edge computing platform, the slope stability evolution trend is predicted through the machine learning algorithm, and the potential risk is identified in advance; S5.4: When the monitoring data reaches the third warning threshold in S5.3.2, the hierarchical response mechanism is automatically triggered: ① When the primary warning threshold is reached, the next level excavation footage is dynamically optimized to 80% of the original plan; ② When the intermediate warning threshold is reached, in addition to adjusting the excavation footage, the mechanical operation intensity is reduced to 70%-80% of the original parameter; ③ When the high-level warning threshold is reached, the excavation operation is immediately suspended, the prestressed anchoring system is started, and the next level excavation parameter is adjusted to the safety range.

10. The multi-bench combined mechanical cold excavation construction method of red sandstone according to claim 9, characterized in that, The specific steps in S5.3 are as follows: S5.3.1: A light-weight LSTM neural network is deployed on the edge computing node, and the time series monitoring data includes displacement rate, pore water pressure and inclination angle, and the sampling interval is ≤5 minutes; S5.3.2: A three-level warning mechanism is set: ① Primary warning: when the predicted displacement in the next 24 hours is >8mm, a yellow warning is marked; ② Intermediate warning: when the predicted displacement in the next 12 hours is >10mm, an orange warning is marked; ③ High-level warning: when the predicted displacement in the next 6 hours is >15mm, a red warning is marked; S5.3.3: When the deviation of the continuous three prediction results is >20%, the online fine-tuning of the model is automatically triggered; S5.3.4: The warning result is pushed to the on-site monitoring terminal through the MQTT protocol.

Images