Surrounding rock grade evaluation method and platform based on multi-parameter feature fusion

Abstract

The invention discloses a surrounding rock grade evaluation method and platform based on multi-parameter feature fusion, and belongs to the technical field of soft rock tunneling, and the method comprises the steps: carrying out the information collection of surrounding rock characteristics, and obtaining a surface mechanical feature parameter, an internal structure feature parameter, a dynamic response feature parameter and a support feedback feature parameter; mapping the collected parameters to a four-dimensional feature space to obtain a four-dimensional feature vector X; constructing a scratch mechanical characteristic curve, a sound wave integrity characteristic curve, a drill bit dynamic response curve and a support force feedback curve; performing feature quantification on each curve by adopting machine learning, and constructing a surrounding rock grade evaluation model; on the basis of a surrounding rock grade evaluation model, grade evaluation is conducted on surrounding rock geology, the cutting depth is automatically optimized according to the grading result, drill bit types are switched, dynamic compensation of supporting pressure is achieved, millisecond-grade dynamic grading from plasticity to a brittle failure mode is achieved, construction risks under complex geological conditions are greatly reduced, and operation efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of soft rock tunnel excavation technology, specifically to a method and platform for evaluating the surrounding rock grade based on multi-parameter feature fusion. Background Technology

[0002] Surrounding rock classification is a method in tunnel engineering to classify the stability of engineering rock masses based on indicators such as rock integrity and rock strength. This method provides a basis for the selection of construction methods, support structure design, and material consumption standards by classifying surrounding rock categories.

[0003] In existing technologies, the detection parameters for surrounding rock classification rely solely on displacement gauges or single-point pressure sensors for acquisition. This lacks multi-parameter coupled analysis of rock surface mechanical properties (such as compressive strength and abrasion resistance), internal fracture development (such as longitudinal wave velocity attenuation rate and integrity coefficient), and dynamic responses during operation (such as drill bit speed fluctuations and support force fluctuations). Consequently, it is difficult to capture the continuous evolution process from micro-damage to macro-fracture. Support strategy adjustments depend on human experience, and support platforms need to be dynamically adjusted based on classification results. This results in poor adaptability of support platforms and increases construction risks under complex geological conditions.

[0004] As mining and tunnel engineering extend into deeper fractured zones, the surrounding rock exhibits a complex characteristic of "high rheology, strong heterogeneity, and multiple fractures." There is an urgent need for a dynamic surrounding rock classification and discrimination method and platform that can acquire multi-dimensional parameters in real time, enabling full-scale perception from the mechanical properties of the rock surface to the internal structural features. This would break through the adaptability bottleneck of traditional classification methods under complex geological conditions and provide an integrated solution of "precise perception, intelligent decision-making, and efficient support" for surrounding rock construction. Summary of the Invention

[0005] Therefore, this invention provides a method and platform for evaluating the grade of surrounding rock based on multi-parameter feature fusion, in order to solve the problems existing in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a method for evaluating the grade of surrounding rock based on multi-parameter feature fusion, comprising: Information on the characteristics of the surrounding rock is collected to obtain surface mechanical characteristic parameters, internal structural characteristic parameters, dynamic response characteristic parameters, and support feedback characteristic parameters; Based on surface mechanical characteristic parameters, internal structural characteristic parameters, dynamic response characteristic parameters, and support feedback characteristic parameters, the collected parameters are mapped to a four-dimensional feature space to obtain a four-dimensional feature vector X. Based on the four-dimensional feature vector X, the scratch mechanical feature curve, the acoustic integrity feature curve, the drill bit dynamic response curve, and the support force feedback curve are constructed. Based on the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve and support force feedback curve, machine learning is used to quantify the features of each curve and construct a surrounding rock grade evaluation model. Based on the surrounding rock grade evaluation model, the geological grade of the surrounding rock is evaluated to determine the surrounding rock grade.

[0007] Furthermore, the four-dimensional feature vector X is: ; Where S is the surface mechanical characteristic parameter, W is the internal structural characteristic parameter, D is the dynamic response characteristic parameter, and F is the support feedback characteristic parameter.

[0008] Furthermore, the surface mechanical characteristic parameters include the normal load F. n Tangential force F t Scratch depth d, surface hardness H, scratch resistance coefficient R a and energy dissipation rate E d ; The internal structural characteristic parameters include the longitudinal wave velocity V. p Shear wave velocity V s Amplitude attenuation rate α, detection depth h, rock mass integrity coefficient K v Poisson's ratio μ and fracture density index D; The dynamic response characteristic parameters include drill bit rotation speed n, feed rate v, cut volume Q, torque T, and thrust F. p Specific energy index (SE) and rotational speed fluctuation rate ; The support feedback characteristic parameters include the support force F. s Displacement change rate dL / dt, pressure fluctuation coefficient And the deformation modulus E.

[0009] Furthermore, the scratch mechanical characteristic curves include tangential force-scratch depth curves and surface hardness-scratch depth curves; the acoustic integrity characteristic curves include longitudinal wave velocity-detection depth curves and integrity coefficient-detection depth curves; the drill bit dynamic response curves include torque-speed curves and specific energy index-detection depth curves; and the support force feedback curves include support force-displacement curves and pressure fluctuation coefficient-time curves.

[0010] Furthermore, based on the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve, and support force feedback curve, machine learning is used to quantify the features of each curve and construct a surrounding rock grade evaluation model, which specifically includes the following steps: The rate of change (CR), fluctuation frequency (f), and energy spectral density (PSD) of the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve, and support force feedback curve are obtained. Based on the curve change rate CR, fluctuation frequency f, and energy spectral density PSD of each curve, normalization and weighted fusion are performed to obtain the characteristic scalars of the surface mechanics module, the internal structure module, the dynamic response module, and the support feedback module. Based on the characteristic scalars of the surface mechanics module, the internal structure module, the dynamic response module, and the support feedback module, an evaluation function is constructed. The evaluation value MYC is obtained from the evaluation function, and the surrounding rock grade is determined.

[0011] Furthermore, the surface mechanics module feature scalar is obtained by fusing the single-curve scalars of tangential force-scratch depth curve and surface hardness-scratch depth curve; the internal structure module feature scalar is obtained by fusing the single-curve scalars of longitudinal wave velocity-detection depth curve and integrity coefficient-detection depth curve; the dynamic response module feature scalar is obtained by fusing the single-curve scalars of torque-speed curve and specific energy index-detection depth curve; and the support feedback module feature scalar is obtained by fusing the single-curve scalars of support force-displacement curve and pressure fluctuation coefficient-time curve.

[0012] Furthermore, the formula for obtaining the scalar of the single curve is: ; in, It is a single-curve scalar. , and The normalized curve change rate CR, fluctuation frequency f, and energy spectral density PSD are... , and These are the weights of the normalized curve change rate CR, fluctuation frequency f, and energy spectral density PSD, respectively. ; The fusion formula for the module feature scalars is: ; in, Let N be the feature scalar of the i-th module, and N be the number of single curves within the module feature scalar. Let j be the j-th monocursive scalar.

[0013] Furthermore, the evaluation function is: ; Where MYC is the evaluation value, For the i-th feature parameter, Let i be the feature scalar of the i-th module selected in the four-dimensional feature vector. Let i be the spatial weights corresponding to the i feature parameters, and adjust them dynamically according to the environmental parameters. i is the index of different feature parameters, and the value range of i is [1,4].

[0014] A grouting-supported tunneling platform, comprising: The whole rock breaking system includes soft drill bit, hard drill bit and intelligent universal support. The intelligent universal support drives the switching between soft drill bit and hard drill bit. Soft drill bit and hard drill bit are used to tunnel through different surrounding rocks. A multimodal detection system is used to collect multimodal surrounding rock information during the drilling process, process the surrounding rock information, and evaluate the geological grade of the surrounding rock. The dynamic support system includes a retractable hydraulic support, a front-end active support device, and an intelligent control room. The intelligent control room adjusts the support force of the retractable hydraulic support and the front-end active support device to dynamically compensate the support pressure according to the classification results of the surrounding rock grade, so as to meet the needs of subsequent grouting work.

[0015] Furthermore, the multimodal detection system includes a scratch indentation detection module, an acoustic detection and analysis module, a drill bit intelligent detection module, and a support force dynamic monitoring module. The scratch indentation detection module is used to acquire the surface mechanical characteristic parameters applied by the drill bit to the surrounding rock surface. The acoustic detection and analysis module is used to acquire the internal structural characteristic parameters of the surrounding rock. The drill bit intelligent detection module is used to detect the dynamic response characteristic parameters of the drill bit. The support force dynamic monitoring module is used to monitor the support feedback characteristic parameters of the support frame.

[0016] The present invention has the following advantages: This invention acquires surface mechanical characteristic parameters, internal structural characteristic parameters, dynamic response characteristic parameters, and support feedback characteristic parameters by collecting information on the characteristics of the surrounding rock. The acquired parameters are mapped to a four-dimensional feature space to obtain a four-dimensional feature vector X. Scratch mechanical characteristic curves, acoustic integrity characteristic curves, drill bit dynamic response curves, and support force feedback curves are constructed. Machine learning is used to quantify the features of each curve, constructing a surrounding rock grade evaluation model. Based on the surrounding rock grade evaluation model, the geological grade of the surrounding rock is evaluated, and the cutting depth is automatically optimized, the drill bit type is switched, and dynamic compensation of support pressure is achieved according to the grading results.

[0017] This application constructs a rock grade evaluation model through multimodal detection, achieving millisecond-level dynamic grading from plastic to brittle failure modes. The support platform exhibits strong dynamic adaptability, significantly reducing construction risks under complex geological conditions, improving operational efficiency, and forming a closed-loop control system for the entire process of "detection-grading-support." It provides an innovative solution integrating precise detection, intelligent decision-making, and efficient support for tunnel engineering, mining, and other fields. Attached Figure Description

[0018] To more clearly illustrate the embodiments of the present invention or the technical solutions in 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 merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0019] The structures, proportions, sizes, etc. illustrated in this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed herein, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0020] Figure 1 A flowchart of the method for evaluating the surrounding rock grade during advanced support tunneling operations provided by the present invention; Figure 2 A graph of the scratch mechanical characteristic curve provided by the present invention; Figure 3 A graph of the acoustic wave integrity characteristic curve provided by the present invention; Figure 4 A graph of the dynamic response curve of the drill bit provided by the present invention; Figure 5 A graph of the support force feedback curve provided by the present invention; Figure 6 A schematic diagram of the grouting support tunneling platform; In the diagram: 1. Soft drill bit; 2. Track; 3. Intelligent universal support; 4. Hard drill bit; 5. Composite air supply equipment; 6. Telescopic hydraulic support; 7. Scratch indentation detection module; 8. Fully enclosed hydraulic moving platform; 9. Front-end active support equipment; 10. Intelligent control room; 11. Support force feedback module; 12. Drill bit dynamic response module; 13. Sonic detection and analysis module; 14. Triple vibration buffer system. Detailed Implementation

[0021] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1 A method for evaluating the grade of surrounding rock based on multi-parameter feature fusion, such as Figure 1 As shown, it specifically includes: Step S1: Collect information on the characteristics of the surrounding rock, and obtain surface mechanical characteristic parameters, internal structural characteristic parameters, dynamic response characteristic parameters, and support feedback characteristic parameters; Step S2: Based on surface mechanical characteristic parameters, internal structural characteristic parameters, dynamic response characteristic parameters, and support feedback characteristic parameters, the collected parameters are mapped to a four-dimensional feature space to obtain a four-dimensional feature vector X; Step S3: Based on the four-dimensional feature vector X, construct the scratch mechanical feature curve, the acoustic integrity feature curve, the drill bit dynamic response curve, and the support force feedback curve; Step S4: Based on the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve and support force feedback curve, machine learning is used to quantify the features of each curve and construct a surrounding rock grade evaluation model. Step S5: Based on the surrounding rock grade evaluation model, evaluate the geological grade of the surrounding rock and determine the surrounding rock grade; The four-dimensional feature vector X in step S2 is: ; Where S is the surface mechanical characteristic parameter, W is the internal structural characteristic parameter, D is the dynamic response characteristic parameter, and F is the support feedback characteristic parameter.

[0023] Surface mechanical characteristic parameters include normal load F n Tangential force F t Scratch depth d, surface hardness H, scratch resistance coefficient R a and energy dissipation rate E d ; Internal structural characteristic parameters include longitudinal wave velocity V p Shear wave velocity V s Amplitude attenuation rate α, detection depth h, rock mass integrity coefficient K v Poisson's ratio μ and fracture density index D; Dynamic response characteristic parameters include drill bit rotation speed n, feed rate v, cut volume Q, torque T, and thrust F. pSpecific energy index (SE) and rotational speed fluctuation rate ; Support feedback characteristic parameters include support force F s Displacement change rate dL / dt, pressure fluctuation coefficient And the deformation modulus E.

[0024] like Figure 2-5 As shown, the scratch mechanical characteristic curves include the tangential force-scratch depth curve and the surface hardness-scratch depth curve; the acoustic integrity characteristic curves include the longitudinal wave velocity-detection depth curve and the integrity coefficient-detection depth curve; the drill bit dynamic response curves include the torque-speed curve and the specific energy index-detection depth curve; and the support force feedback curves include the support force-displacement curve and the pressure fluctuation coefficient-time curve.

[0025] Based on the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve, and support force feedback curve, machine learning is used to quantify the features of each curve to construct a surrounding rock grade evaluation model, which includes the following steps: Obtain the rate of change (CR), fluctuation frequency (f), and energy spectral density (PSD) of the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve, and support force feedback curve; Based on the curve change rate CR, fluctuation frequency f, and energy spectral density PSD of each curve, normalization and weighted fusion are performed to obtain the characteristic scalars of the surface mechanics module, the internal structure module, the dynamic response module, and the support feedback module. Based on the characteristic scalars of the surface mechanics module, the internal structure module, the dynamic response module, and the support feedback module, an evaluation function is constructed. The evaluation value MYC is obtained from the evaluation function, and the surrounding rock grade is determined.

[0026] Among them, the rate of change of the curve Where x is the horizontal axis of the curve, and f(x) is the vertical axis of the curve; the fluctuation frequency f = number of extreme points / sampling length; energy spectral density ,in, This represents the spectral density.

[0027] The surface mechanics module feature scalar is obtained by fusing the single-curve scalars of tangential force-scratch depth curve and surface hardness-scratch depth curve. The internal structure module feature scalar is obtained by fusing the single-curve scalars of longitudinal wave velocity-detection depth curve and integrity coefficient-detection depth curve. The dynamic response module feature scalar is obtained by fusing the single-curve scalars of torque-speed curve and specific energy index-detection depth curve. The support feedback module feature scalar is obtained by fusing the single-curve scalars of support force-displacement curve and pressure fluctuation coefficient-time curve.

[0028] Furthermore, the formula for obtaining the scalar of a single curve is: ; in, It is a single-curve scalar. , and The normalized curve change rate CR, fluctuation frequency f, and energy spectral density PSD are normalized. Normalization is a common technique in machine learning, which aims to eliminate the influence of dimensions. It will not be elaborated on here. , and These are the weights of the normalized curve change rate CR, fluctuation frequency f, and energy spectral density PSD, respectively. (Adjust according to the curve type; for example, if you are more concerned about PSD in sound wave attenuation, you can set...) ).

[0029] in, , and The calculation formula is: ; ; ; The fusion formula for module feature scalars is: ; in, Let N be the feature scalar of the i-th module, and N be the number of single curves within the module feature scalar. Let j be the j-th monocursive scalar.

[0030] Furthermore, the evaluation function is: ; Where MYC is the evaluation value, For the i-th feature parameter, Let i be the feature scalar of the i-th module selected in the four-dimensional feature vector. Let i be the spatial weights corresponding to the i feature parameters, and adjust them dynamically according to the environmental parameters. i is the index of different feature parameters, and the value range of i is [1,4].

[0031] Specifically, substituting the fusion formula of the module feature scalars, the evaluation function is: ; w i Spatial weights for each parameter (w1=0.3 (surface), w2=0.3 (internal), w3=0.2 (dynamic), w4=0.2 (support)), w i=Standards or expert advice, such as in tunnel engineering, directly set w1=0.3, w2=0.3, w3=0.2, w4=0.2 according to the "Highway Tunnel Design Specification". , , , , , , , These are single-curve scalars representing the tangential force-scratch depth curve, surface hardness-scratch depth curve, longitudinal wave velocity-detection depth curve, integrity coefficient-detection depth curve, torque-speed curve, specific energy index-detection depth curve, support force-displacement curve, and pressure fluctuation coefficient-time curve, respectively.

[0032] Based on the calculated evaluation values, the surrounding rock grades are classified as follows: Classification table of surrounding rock grades Surrounding rock grade MYC range Curve feature description Level I ≥85% <![CDATA[The scratch mechanical characteristic curve is smooth, K v > 0.85, σ n < 5%, and the linearity of the support force feedback curve is good]]> Level II 70%-85% <![CDATA[The microfluctuation of the scratch mechanical characteristic curve, 0.7 < K v ≤ 0.85, 5% ≤ σ n < 10%, and the support force feedback curve has slight nonlinearity]]> Level III 55%-70% <![CDATA[The fluctuation of the scratch mechanical characteristic curve is obvious, 0.55 < K v ≤ 0.7, 10% ≤ σ n < 15%, and the plastic deformation characteristics appear in the support force feedback curve]]> Level IV 40%-55% <![CDATA[The multi-peak scratch mechanical characteristic curve, 0.4 < K v ≤ 0.55, σ n ≥ 15%, and the inflection point appears in the support force feedback curve]]> V-Class <40% <![CDATA[The scratch mechanical characteristic curve is irregular, K v ≤ 0.4, σ n fluctuates violently, and the support force feedback curve shows a non-linear decline]]> Among them, based on the tunnel burial depth H and the ground stress σ h The weights are dynamically adjusted based on environmental parameters. ; in, The adjusted weights, For the adjusted weights, λ i The sensitivity coefficient, This is a nonlinear adjustment function. i corresponds to the index of one of the four parameter spaces (surface, interior, dynamic, support).

[0033] This method achieves continuous monitoring and classification of surrounding rock stability through morphological analysis and dynamic comparison of multi-parameter curves. Compared with the traditional point-value evaluation method, it improves the response speed to changes in geological conditions by 50% and the classification accuracy to 92%, showing significant advantages, especially in complex geological areas such as fault fracture zones. To avoid testing errors, other parameters can be used as references, such as the abrupt change in the CR of the scratch mechanical curve, combined with the scratch resistance coefficient R. a The numerical value (increased surface roughness) can more accurately determine whether the fluctuation in the curve is due to actual rock mass breakage or testing errors.

[0034] Example 2 This invention provides a grouting support tunneling platform, such as... Figure 6 As shown, including: The all-rock breaking system includes soft drill bits, hard drill bits, tracks, and intelligent universal supports. The intelligent universal supports drive the soft drill bits and hard drill bits to rotate and switch, and the soft drill bits and hard drill bits are used to excavate different surrounding rocks. Specifically, the whole-lithological rock breaking system includes: (1) Soft drill bit 1: Designed for soft soil layers and weathered rock, it adopts a flexible spiral cutting structure (120mm pitch) and high-toughness titanium alloy material. It achieves low-resistance crushing through a high speed of 200-300r / min, reducing the risk of stuck drill by 60%.

[0035] (2) Hard drill bit 4: It is suitable for hard rock formations (compressive strength > 60MPa), equipped with high-strength alloy teeth (hardness HRC65) and triple vibration buffer system 14 (frequency 5-15Hz), which can withstand 30kN impact load and improve hard rock breaking efficiency by 35%.

[0036] (3) Intelligent universal support 3: hydraulically driven rotary pivot (positioning accuracy ±0.5°), completes soft / hard drill bit switching within 15 seconds, combined with geological data (such as acoustic longitudinal wave velocity (V) p It automatically matches the rock-breaking angle (adjustable from 0-90°) to reduce equipment wear and tear. The hydraulically driven rotary hub used in the intelligent universal support 3 is a commonly used hydraulic rotary mechanism, which will not be described in detail here.

[0037] (4) Track 2: Double track chassis, high strength wear-resistant material combined with precision transmission system, adaptable to muddy and rugged terrain, tension adjustment device ensures smooth movement in complex terrain and provides reliable driving force.

[0038] The multimodal detection system is used to collect multimodal surrounding rock information during the drilling process, process the surrounding rock information, and evaluate the geological grade of the surrounding rock.

[0039] Specifically, the multimodal detection system includes a scratch indentation detection module, an acoustic detection and analysis module, a drill bit intelligent detection module, and a support force dynamic monitoring module. The scratch indentation detection module is used to obtain the surface mechanical characteristic parameters applied by the drill bit to the surrounding rock surface, the acoustic detection and analysis module is used to obtain the internal structural characteristic parameters of the surrounding rock, the drill bit intelligent detection module is used to detect the dynamic response characteristic parameters of the drill bit, and the support force dynamic monitoring module is used to monitor the support feedback characteristic parameters of the support frame.

[0040] Specifically, the multimodal detection system includes: (1) Scratch indenter detection module 7: Diamond indenter (cone angle 136°, load accuracy ±0.5%F) s Apply a directional pressure of 5-100N and simultaneously collect parameters: normal load F n Tangential force F t And the scratch depth d, and calculate the surface hardness H=F n / A, (where A is the projected area of ​​the indentation), scratch resistance coefficient R a =F t / F n and energy dissipation rate E d =∫Ft ⋅dd / ∫F n ⋅dd provides basic data for parameter space hierarchies.

[0041] (2) Sound wave detection and analysis module 13: 4 groups of 20-50kHz ultrasonic transducers emit penetrating sound waves, and the collected parameters are: longitudinal wave velocity V p Shear wave velocity V s Calculate the amplitude attenuation rate α and the rock mass integrity coefficient K. v =(V p / V p0 ) 2 (V p0 (V1 = V2) P3 = (V4) P5 = (V6) P7 = (V8) P9 = (V1) P2 = (V2) P3 = (V<sub p 2 -2V s 2 ) / (2(V p 2 -V s 2 The fracture density index D = 1 - K v This enables the quantification of the internal fracture density D and the degree of fragmentation of the surrounding rock (resolution 0.1m).

[0042] (3) Support Force Feedback Module 11: The hydraulic cylinder has a built-in pressure sensor (accuracy ±0.5%) to track the pressure fluctuations of the telescopic support (support force 50-200kN) and the front-end support equipment (maximum pressure 1.5MPa) in real time, and to identify the creep trend of the surrounding rock. Collected parameters: Support force F s Calculate the displacement change rate dL / dt and the pressure fluctuation coefficient. , The time average of the supporting force (e.g., the average value sampled over 1 minute) is used to extract the deformation modulus E=(F s / A) / (ΔL / L0) (A is the support area), ΔL: the overall displacement change of the support monitoring section (not a single point, the average displacement of the key deformation area should be taken, such as the relative displacement between the midpoint of the support top beam and the base), L0: the initial effective length of the support (the design support length during installation, such as the initial length of the hydraulic support after extension is 2.5m).

[0043] (4) Drill bit dynamic response module 12: Hall sensor (accuracy ±1r / min) collects parameters: drill bit rotation speed n, feed speed v, cut amount Q, torque T and thrust F p And calculate the specific energy index SE=(T·n+F) p ·v) / Q, rotational speed fluctuation rate ( The average rotational speed (arithmetic mean) within the sampling period is combined with torque data (to help determine lithology or drill bit load status) to establish a lithology-rotational speed characteristic spectrum (e.g., soft rock rotational speed is stable at 250 r / min ± 10%, hard rock fluctuation > 20%).

[0044] The dynamic support system is used to dynamically compensate for the support pressure based on the classification results of the surrounding rock grade, thus meeting the requirements of subsequent grouting work. The dynamic support system includes a retractable hydraulic support, a front-end active support device, and an intelligent control room, through which the support force of the retractable hydraulic support and the front-end active support device is adjusted.

[0045] Specifically, the dynamic support system includes: (1) Telescopic hydraulic support 6: Stroke 1-3m hydraulic telescopic rod (speed 50mm / s), made of Q345B anti-corrosion steel, with a maximum support force of 200kN, which can be used as a temporary support (anti-collapse) and a netting carrier. The steel mesh can be laid within 10 minutes. The hydraulic telescopic rod is a commonly used hydraulic telescopic structure, which will not be described in detail here.

[0046] (2) Front active support equipment 9: Five-axis adjustable support arm (angle accuracy ±1°), built-in pressure closed-loop control system, automatically increases the support pressure to 1.2MPa according to the surrounding rock classification results (such as Class III surrounding rock), and compensates for the deformation of the surrounding rock (accuracy ±2mm). The front active support equipment is a commonly used support equipment in surrounding rock excavation, and will not be described in detail here.

[0047] (3) The intelligent control room features a 10:12-inch high-definition touch screen that links to multi-source data and integrates an edge computing industrial control computer. It outputs the "nine-step method" construction parameters in real time (such as cutting depth 0.8m / cycle, drill bit speed 180r / min), supports remote (within a 500m range) and local dual-mode control, and has a command response delay of <50ms. The "nine-step method" construction parameters are the construction parameters set according to the evaluated surrounding rock grade during surrounding rock construction. Referencing the highway tunnel design specifications, they will not be elaborated here.

[0048] The grouting support tunneling platform also includes an environmental protection and auxiliary system, which is used to purify the air quality in the work area and help meet the needs of high-altitude operations.

[0049] Specifically, environmental protection and support systems include: (1) Composite air protection equipment 5: Three-stage filtration system (primary filter + activated carbon + HEPA + axial flow fan (air volume 2000m³ / h), real-time purification of pollutants such as PM2.5 (filtration efficiency 99.97%) and CO (adsorption rate 98%), and the air quality in the work area meets the GBZ2.1 standard.

[0050] (2) Fully enclosed hydraulic mobile platform 8: 3m×2m lifting working surface (load capacity 500kg), equipped with anti-fall guardrail and laser rangefinding anti-collision system, supports ±2m lateral movement and 0-5m lifting, meeting the needs of high-altitude operations such as anchor bolt installation and equipment maintenance.

[0051] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A method for evaluating the grade of surrounding rock based on multi-parameter feature fusion, characterized in that, include: Information on the characteristics of the surrounding rock is collected to obtain surface mechanical characteristic parameters, internal structural characteristic parameters, dynamic response characteristic parameters, and support feedback characteristic parameters; Based on surface mechanical characteristic parameters, internal structural characteristic parameters, dynamic response characteristic parameters, and support feedback characteristic parameters, the collected parameters are mapped to a four-dimensional feature space to obtain a four-dimensional feature vector X. Based on the four-dimensional feature vector X, the scratch mechanical feature curve, the acoustic integrity feature curve, the drill bit dynamic response curve, and the support force feedback curve are constructed. Based on the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve and support force feedback curve, machine learning is used to quantify the features of each curve and construct a surrounding rock grade evaluation model. Based on the surrounding rock grade evaluation model, the geological grade of the surrounding rock is evaluated to determine the surrounding rock grade.

2. The method for evaluating the grade of surrounding rock based on multi-parameter feature fusion as described in claim 1, characterized in that, The four-dimensional feature vector X is: ； Where S is the surface mechanical characteristic parameter, W is the internal structural characteristic parameter, D is the dynamic response characteristic parameter, and F is the support feedback characteristic parameter.

3. The surrounding rock grade evaluation method based on multi-parameter feature fusion as described in claim 1, characterized in that, The surface mechanical characteristic parameters include the normal load F. n Tangential force F t Scratch depth d, surface hardness H, scratch resistance coefficient R a and energy dissipation rate E d ; The internal structural characteristic parameters include the longitudinal wave velocity V. p Shear wave velocity V s Amplitude attenuation rate α, detection depth h, rock mass integrity coefficient K v Poisson's ratio μ and fracture density index D; The dynamic response characteristic parameters include drill bit rotation speed n, feed rate v, cut volume Q, torque T, and thrust F. p Specific energy index (SE) and rotational speed fluctuation rate ; The support feedback characteristic parameters include the support force F. s Displacement change rate dL / dt, pressure fluctuation coefficient And the deformation modulus E.

4. The method for evaluating the grade of surrounding rock based on multi-parameter feature fusion as described in claim 1, characterized in that, The scratch mechanical characteristic curves include tangential force-scratch depth curves and surface hardness-scratch depth curves; the acoustic integrity characteristic curves include longitudinal wave velocity-detection depth curves and integrity coefficient-detection depth curves; the drill bit dynamic response curves include torque-speed curves and specific energy index-detection depth curves; and the support force feedback curves include support force-displacement curves and pressure fluctuation coefficient-time curves.

5. The surrounding rock grade evaluation method based on multi-parameter feature fusion as described in claim 1, characterized in that, Based on the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve, and support force feedback curve, machine learning is used to quantify the features of each curve to construct a surrounding rock grade evaluation model, which specifically includes the following steps: Obtain the rate of change (CR), fluctuation frequency (f), and energy spectral density (PSD) of the scratch mechanical characteristic curve, acoustic integrity characteristic curve, drill bit dynamic response curve, and support force feedback curve; Based on the curve change rate CR, fluctuation frequency f, and energy spectral density PSD of each curve, normalization and weighted fusion are performed to obtain the characteristic scalars of the surface mechanics module, the internal structure module, the dynamic response module, and the support feedback module. Based on the characteristic scalars of the surface mechanics module, the internal structure module, the dynamic response module, and the support feedback module, an evaluation function is constructed. The evaluation value MYC is obtained from the evaluation function, and the surrounding rock grade is determined.

6. The method for evaluating the grade of surrounding rock based on multi-parameter feature fusion as described in claim 5, characterized in that, The surface mechanics module feature scalar is obtained by fusing the single-curve scalars of tangential force-scratch depth curve and surface hardness-scratch depth curve. The internal structure module feature scalar is obtained by fusing the single-curve scalars of longitudinal wave velocity-detection depth curve and integrity coefficient-detection depth curve. The dynamic response module feature scalar is obtained by fusing the single-curve scalars of torque-speed curve and specific energy index-detection depth curve. The support feedback module feature scalar is obtained by fusing the single-curve scalars of support force-displacement curve and pressure fluctuation coefficient-time curve.

7. The method for evaluating the grade of surrounding rock based on multi-parameter feature fusion as described in claim 6, characterized in that, The formula for obtaining the scalar of the single curve is: ； in, It is a single-curve scalar. , and The normalized curve change rate CR, fluctuation frequency f, and energy spectral density PSD are... , and These are the weights of the normalized curve change rate CR, fluctuation frequency f, and energy spectral density PSD, respectively. ; The fusion formula for the module feature scalars is: ； in, Let N be the feature scalar of the i-th module, and N be the number of single curves within the module feature scalar. Let j be the j-th monocursive scalar.

8. The method for evaluating the grade of surrounding rock based on multi-parameter feature fusion as described in claim 7, characterized in that, The evaluation function is: ； Where MYC is the evaluation value, For the i-th feature parameter, Let i be the feature scalar of the i-th module selected in the four-dimensional feature vector. Let i be the spatial weights corresponding to the i feature parameters, and adjust them dynamically according to the environmental parameters. i is the index of different feature parameters, and the value range of i is [1,4].

9. A grouting-supported tunneling platform for implementing the surrounding rock grade evaluation method as described in any one of claims 1-8, characterized in that, include: The whole rock breaking system includes soft drill bit, hard drill bit and intelligent universal support. The intelligent universal support drives the switching between soft drill bit and hard drill bit, which are used to excavate different surrounding rocks. A multimodal detection system is used to collect multimodal surrounding rock information during the drilling process, process the surrounding rock information, and evaluate the geological grade of the surrounding rock. The dynamic support system includes a retractable hydraulic support, a front-end active support device, and an intelligent control room. The intelligent control room adjusts the support force of the retractable hydraulic support and the front-end active support device to dynamically compensate the support pressure according to the classification results of the surrounding rock grade, so as to meet the needs of subsequent grouting work.

10. The grouting support tunneling platform as described in claim 9, characterized in that, The multimodal detection system includes a scratch indentation detection module, an acoustic detection and analysis module, a drill bit intelligent detection module, and a support force dynamic monitoring module. The scratch indentation detection module is used to acquire the surface mechanical characteristic parameters applied by the drill bit to the surrounding rock surface. The acoustic detection and analysis module is used to acquire the internal structural characteristic parameters of the surrounding rock. The drill bit intelligent detection module is used to detect the dynamic response characteristic parameters of the drill bit. The support force dynamic monitoring module is used to monitor the support feedback characteristic parameters of the support frame.

Images