A method for obtaining rock mechanical parameters based on acoustic wave detection and artificial intelligence

Abstract

This invention discloses a method for obtaining rock mechanics parameters based on acoustic wave detection and artificial intelligence, comprising the following steps: Step 1, constructing a backpropagation (BP) neural network; Step 2, constructing a rock mass mechanics parameter database, inputting the mechanics parameters from the database into the BP neural network trained in Step 1 to obtain a trained BP neural network; Step 3, obtaining the rock mass to be tested, measuring the uniaxial compressive strength and longitudinal wave velocity of the rock mass, as well as the wave velocity of the rock mass with initial damage, and using these as inputs into the trained BP neural network in Step 2, and outputting the compressive strength σ of the rock mass based on the trained BP neural network. c And the elastic modulus E of the rock mass, the compressive strength σ of the rock mass c The elastic modulus E of the rock mass is used as a reference parameter for tunnel excavation. Using the method in this invention, the input parameters are easily obtained on-site, the computational load is small, and it is unaffected by the type and water content of the rock.

Description

Technical Field

[0001] This invention belongs to the field of geotechnical engineering technology, specifically relating to a method for obtaining rock mechanical parameters based on acoustic wave detection and artificial intelligence. Background Technology

[0002] Rock mass wave velocity is convenient, non-destructive, and rapid to test, and it also shows a good correlation with some rock mass mechanical parameters. Currently, acoustic wave testing technology is widely used in the field of rock mechanics. By conducting indoor mechanical tests on rocks and simultaneously detecting their wave velocity, a quantitative relationship can be established between wave velocity and rock mechanical parameters such as compressive strength σc and elastic modulus E. However, current research on the relationship between rock mass acoustic wave velocity and mechanical parameters mainly relies on empirical formulas and regression fitting analysis. These empirical formulas are only applicable to single conditions, i.e., a single rock type and a single water-bearing state. If the rock type and water-bearing state change, the previously established empirical formulas become inapplicable. For geotechnical engineering in complex geological environments, the strata in the work area are usually a mixture of various rocks with variable water-bearing conditions. In such cases, a large number of empirical formulas need to be established, the calculation process is complex, and corrections are required based on the rock type and water content during actual calculations. Furthermore, water content testing is relatively complex and can usually only be performed in a laboratory, making it difficult to obtain this parameter directly on the construction site. Summary of the Invention

[0003] The purpose of this invention is to provide a method for obtaining rock mechanical parameters based on acoustic wave detection and artificial intelligence. The input parameters are easy to obtain on site, the computational load is small, and it is not affected by the type and water content of the rock.

[0004] This invention employs the following technical solution: a method for obtaining rock mass mechanical parameters based on acoustic wave detection and artificial intelligence, comprising the following steps:

[0005] Step 1: Construct a BP neural network;

[0006] Step 2: Construct a rock mass mechanics parameter database, input the mechanics parameters from the database into the BP neural network training in Step 1, and obtain the trained BP neural network.

[0007] Step 3: Obtain rock samples of the rock mass to be tested, measure the uniaxial compressive strength and longitudinal wave velocity of the rock samples, as well as the wave velocity of the rock mass with initial damage, and use these as inputs to the BP neural network trained in Step 2. Based on the trained BP neural network, output the compressive strength σ of the rock mass. c And the elastic modulus E of the rock mass, the compressive strength σ of the rock mass c The elastic modulus E of the rock mass is used as a reference parameter for tunnel excavation methods.

[0008] Furthermore, the uniaxial compressive strength of the rock sample of the rock mass to be tested was determined by uniaxial compressive strength test.

[0009] Furthermore, the process of constructing the rock mechanics parameter database in step two is as follows:

[0010] Rock samples containing all types of rock masses were collected. For any type of rock sample, at least three water-bearing states and ten damage states were tested to obtain a total of thirty mechanical parameters of rock samples under different working conditions. The mechanical parameters were used as input to the neural network.

[0011] The three moisture states are represented as S-1, S-2 and S-3, where S-1 represents the dry state, i.e., the moisture content is 0%; S-2 represents the semi-saturated state, i.e., the moisture content is 50%; and S-3 represents the fully saturated state, i.e., the moisture content is 100%.

[0012] The ten damage states are represented as D-1 to D-10, respectively.

[0013] The damage state is defined as follows: a continuously increasing load is applied to the rock specimen until cracking occurs, and the load strength that the rock specimen bears when cracking is its ultimate strength; ten different load values ​​are selected sequentially from small to large between the initial load and the ultimate strength, with the magnitude of the load being a percentage of the ultimate strength of the rock specimen as the benchmark; the ten different loads correspond to ten damage states D-1 to D-10 of the rock specimen, where D-1 represents no initial damage, which is the damage state under the initial load value.

[0014] Furthermore, the mechanical parameters of the rock specimen are: the uniaxial compressive strength of the rock specimen, the longitudinal wave velocity of the rock specimen, and the wave velocity of the rock specimen with initial damage.

[0015] Furthermore, the ten different load values ​​are selected as follows: the critical point of the plastic stage is used as the initial load value, 95% of the ultimate strength is used as the final load value, and the other load values ​​are selected sequentially within the range of 70% to 95% of the ultimate strength.

[0016] Furthermore, the mechanical parameters of the rock samples were obtained in the following manner:

[0017] The uniaxial compressive strength of the rock specimen is obtained as follows: the compressive strength of the rock specimen is measured by point load test, and the compressive strength of the rock specimen is used as the input uniaxial compressive strength of the rock specimen.

[0018] The wave velocity of a rock sample with initial damage is obtained as follows: the rock sample is pre-compressed and cracked to obtain a rock sample with initial damage, and the wave velocity is measured by the seismic wave method.

[0019] This invention also discloses a method for establishing a rock mass mechanics parameter database in the aforementioned method for obtaining rock mass mechanics parameters based on acoustic detection and artificial intelligence, comprising the following:

[0020] Rock samples containing all types of rock masses were collected. For any type of rock sample, at least three water-bearing states and ten damage states were tested to obtain a total of thirty mechanical parameters under different working conditions. The mechanical parameters were used as input to the neural network.

[0021] The three moisture states are represented as S-1, S-2 and S-3, where S-1 represents the dry state, i.e., the moisture content is 0%; S-2 represents the semi-saturated state, i.e., the moisture content is 50%; and S-3 represents the fully saturated state, i.e., the moisture content is 100%.

[0022] The ten damage states are represented as D-1 to D-10, respectively.

[0023] The damage state is defined as follows: a continuously increasing load is applied to the rock specimen until cracking occurs, and the load strength that the rock specimen bears when cracking is its ultimate strength; ten different load values ​​are selected sequentially from small to large between the initial load and the ultimate strength, with the magnitude of the load being a percentage of the ultimate strength of the rock specimen as the benchmark; the ten different loads correspond to ten damage states D-1 to D-10 of the rock specimen, where D-1 represents no initial damage, which is the damage state under the initial load value.

[0024] The beneficial effects of this invention are: a feasible method for calculating rock mechanical parameters under various lithologies and water-bearing conditions without needing to establish multiple empirical formulas separately based on rock type and water content, and for calculating the initial rock strength σ. c0 Initial P-wave velocity v of rock p0 And the current longitudinal wave velocity v of the rock p The rock strength σ is obtained as an input parameter. c The calculation of the elastic modulus E of rock is reduced, and the input parameters are easier to obtain through triaxial compression tests and wave velocity, making it easy to conduct long-term field tests. Attached Figure Description

[0025] Figure 1 The relationship between rock compressive strength and longitudinal wave velocity;

[0026] Figure 2 The relationship between the rock elastic modulus and the longitudinal wave velocity;

[0027] Figure 3 The relationship between compressive strength and longitudinal wave velocity of rocks under different rock conditions and different water content states;

[0028] Figure 4 This is a data demonstration of rock strength calculated using a neural network.

[0029] Figure 5 This is a data demonstration of how the elastic modulus of rock is calculated using a neural network. Detailed Implementation

[0030] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0031] There is a significant correlation between the wave velocity of rocks and their mechanical parameters. Under constant lithology and water content, the compressive strength and wave velocity of rocks exhibit the following relationship: Figure 1 The exponential relationship is shown. At the same time, a similar relationship exists between the elastic modulus of rock and wave velocity, such as... Figure 2 As shown. Therefore, a fitting formula can be established through data statistics, which is also a commonly used method.

[0032] However, once the lithology or water content changes, the relationship between wave velocity and intensity will change, such as... Figure 3 As shown, the fitting formula is no longer applicable. When the water content of the same type of rock changes, on the one hand, the speed of sound in water (1500 m / s) is much greater than its speed of sound in air (340 m / s); on the other hand, the rock softens when exposed to water, resulting in a change in wave velocity and intensity. For another type of rock, due to differences in rock composition and porosity, the relationship between wave velocity and intensity will also change significantly. Similarly, the relationship between wave velocity and elastic modulus also faces similar issues.

[0033] This invention provides a method for obtaining rock mechanical parameters based on acoustic wave detection and artificial intelligence. It employs an artificial intelligence algorithm to aggregate experimental data into a database, which is then used to train a neural network. After the neural network is trained, it can calculate the corresponding rock strength and elastic modulus upon inputting wave velocity and auxiliary control parameters. This method is applicable to rocks with various lithologies and water-bearing conditions. The selected auxiliary control parameter is the uniaxial compressive strength σ of intact rock, i.e., an undamaged rock core. c0 The longitudinal wave velocity v of intact rock p0 The wave velocity (v) was obtained in the laboratory through uniaxial compression testing and wave velocity measurement. Pre-compression cracking of the rock specimen is required to obtain rock with initial damage, and its wave velocity (v) is then measured. p .

[0034] This invention also discloses a method for obtaining rock mass mechanical parameters based on acoustic wave detection and artificial intelligence, comprising the following steps:

[0035] Step 1: Construct a BP neural network;

[0036] Step 2: Construct a rock mass mechanics parameter database, input the mechanics parameters from the database into the BP neural network training in Step 1, and obtain a trained BP neural network; alternatively, a CNN neural network can be used.

[0037] Step 3: Obtain rock samples of the rock mass to be tested, measure the uniaxial compressive strength and longitudinal wave velocity of the rock samples, as well as the wave velocity of the rock mass with initial damage, and use these as inputs to the BP neural network trained in Step 2. Based on the trained BP neural network, output the compressive strength σ of the rock mass. c And the elastic modulus E of the rock mass, the compressive strength σ of the rock mass c The elastic modulus E of the rock mass is used as a reference parameter for tunnel excavation methods. The uniaxial compressive strength of the rock sample of the rock mass to be tested is obtained by uniaxial compressive strength test. The wave velocity of the rock mass to be tested with initial damage is measured by seismic wave method.

[0038] The process of constructing the rock mass mechanics parameter database in step two is as follows:

[0039] Rock samples containing all types of rock masses were collected. For any type of rock sample, at least three water-bearing states and ten damage states were tested to obtain a total of thirty mechanical parameters of the rock samples under different working conditions. The mechanical parameters were used as input to the neural network.

[0040] The three moisture states are represented as S-1, S-2 and S-3, where S-1 represents the dry state, i.e., the moisture content is 0%; S-2 represents the semi-saturated state, i.e., the moisture content is 50%; and S-3 represents the fully saturated state, i.e., the moisture content is 100%.

[0041] The ten damage states are represented as D-1 to D-10 in sequence;

[0042] The damage state is defined as follows: a continuously increasing load is applied to the rock specimen until cracking occurs, and the load strength that the rock specimen bears when cracking is its ultimate strength; ten different load values ​​are selected sequentially from small to large between the initial load and the ultimate strength, with the magnitude of the load being a percentage of the ultimate strength of the rock specimen as the benchmark; the ten different loads correspond to ten damage states D-1 to D-10 of the rock specimen, where D-1 represents no initial damage, which is the damage state under the initial load value.

[0043] The selection of the above ten different load values ​​is as follows: the critical point of the plastic stage is used as the initial load value, 95% of the ultimate strength is used as the final load value, and the other load values ​​are selected sequentially within the range of 70% to 95% of the ultimate strength.

[0044] Specifically, the critical point of the plastic stage is generally about 70% of the ultimate strength. The critical point of the plastic stage is used as the initial load value. The corresponding state under this initial load value is D-1, that is, the state without initial damage. The increment of the next load value adjacent to the initial load value can be set to 2.5%, 5%, or other values. The increment is not fixed. The load value after increasing the increment is within the specified range, that is, it meets the requirements, until 95% of the ultimate strength is used as the final load value.

[0045] The mechanical parameters of the rock samples are: uniaxial compressive strength of the rock samples, longitudinal wave velocity of the rock samples, and wave velocity of the rock samples with initial damage.

[0046] The uniaxial compressive strength of the aforementioned rock specimens was obtained as follows: the compressive strength of the rock specimens was measured through point load tests, and this compressive strength was used as the input uniaxial compressive strength. The wave velocity of the rock specimens with initial damage was obtained as follows: the rock specimens were pre-cracked to obtain rock specimens with initial damage, and the wave velocity was measured using the seismic wave method.

[0047] The BP neural network training input layer has three input variables, namely the aforementioned mechanical parameters. The hidden layer of the BP neural network has 10 nodes; increasing the number of nodes has no significant impact on the accuracy of the prediction data. The output layer has two output variables: the compressive strength σ of the rock mass. c And the elastic modulus E of the rock mass.

[0048] This invention discloses a method for establishing a rock mass mechanics parameter database in the aforementioned method for obtaining rock mass mechanics parameters based on acoustic wave detection and artificial intelligence, comprising the following:

[0049] Rock samples containing all types of rock masses were collected. For any type of rock sample, at least three water-bearing states and ten damage states were tested to obtain a total of thirty mechanical parameters under different working conditions. The mechanical parameters were used as input to the neural network.

[0050] The three moisture states are represented as S-1, S-2 and S-3, where S-1 represents the dry state, i.e., the moisture content is 0%; S-2 represents the semi-saturated state, i.e., the moisture content is 50%; and S-3 represents the fully saturated state, i.e., the moisture content is 100%.

[0051] The ten damage states are represented as D-1 to D-10, respectively.

[0052] The damage state is defined as follows: a continuously increasing load is applied to the rock specimen until cracking occurs, and the load strength that the rock specimen bears when cracking is its ultimate strength; ten different load values ​​are selected sequentially from small to large between the initial load and the ultimate strength, with the magnitude of the load being a percentage of the ultimate strength of the rock specimen as the benchmark; the ten different loads correspond to ten damage states D-1 to D-10 of the rock specimen, where D-1 represents no initial damage, which is the damage state under the initial load value.

[0053] Rock samples that have undergone pre-compression cracking are collectively referred to as rocks with initial damage. Rock masses are simulated using pre-compression cracked rocks with initial damage.

[0054] A database was established using pre-compression-fractured rock with initial damage to simulate rock masses for predicting rock mass mechanical parameters at construction sites. The mechanical parameters under the aforementioned thirty different working conditions are as follows: D-1 to D-10 under state S-1, yielding ten corresponding working conditions; D-1 to D-10 under state S-2, yielding ten corresponding working conditions; D-1 to D-10 under state S-3, yielding ten corresponding working conditions. Each working condition corresponds to a set of mechanical parameters, and each set of mechanical parameters includes: the uniaxial compressive strength σ of intact rock. c0 The longitudinal wave velocity v of intact rock p0 and wave velocity v of rock mass with initial damage p The rocks in group D-1, having no initial damage, exhibit longitudinal wave velocity and compressive strength consistent with their initial state, i.e., v p =v p0 , σ c =σ c0 .

[0055] As a specific implementation, samples of all rocks within the work area were collected. Data, as shown in Table 1, were obtained through indoor uniaxial compression tests and longitudinal wave velocity measurements, with 30 sets of data for each rock type. All data were used as the training set to train the neural network. The mechanical parameters are as follows:

[0056] Table 1 Examples of rock groupings and related parameters

[0057]

[0058]

[0059]

[0060] Using the method of this invention, since no core drilling was performed, the water content of the rock is at its natural water content. This allows for the determination of the uniaxial compressive strength σ of the intact rock under the current water content conditions. c0 Longitudinal wave velocity v of intact rockp0 In subsequent predictions of the mechanical parameters of the rocks, the influence of water content can be disregarded when collecting rock data on-site.

[0061] When predicting the mechanical parameters of rocks in a field, it is necessary to first identify the rock type. Based on existing experimental data, the uniaxial compressive strength and P-wave velocity of this rock type are input. Then, wave velocity testing is performed on the rock in the field to obtain the wave velocity of the rock mass with initial damage. For the rocks in the field with initial damage, the wave velocity is obtained by testing the wave velocity using the seismic wave method. Once these three data points are acquired, the input layer data, namely the uniaxial compressive strength σ of this rock type, can be imported based on a trained neural network. c0 The longitudinal wave velocity v of this type of rock p0 Wave velocity v of rock mass with initial damage p The corresponding stress value is calculated using a trained neural network, eliminating the need for manual calculation. The longitudinal wave velocity v of this type of rock... p0 The initial longitudinal wave velocity is given.

[0062] like Figure 4 As shown in the figure, the black dots represent training data. This embodiment tested three different types of rocks, with 30 sets of data for each type, totaling 90 sets of data, based on the principles of data selection and grouping. These data include wave velocity, intensity, and elastic modulus under different water-bearing conditions for the three rocks. After the neural network was trained, nine rock datasets were selected to test the accuracy of the method in this invention. In the figure, red dots represent predicted data, and blue dots represent real data obtained through indoor uniaxial compression tests. The data from both are quite similar, indicating that the method in this invention has good computational accuracy.

[0063] The calculation of the elastic modulus is similar, such as... Figure 5 As shown, black dots represent training data, red dots represent predicted data, and blue dots represent real data obtained through indoor uniaxial compression experiments. The data from both sources are quite similar, indicating that the method described in this invention has good computational accuracy.

Claims

1. A method for obtaining rock mass mechanical parameters based on acoustic wave detection and artificial intelligence, characterized in that, Includes the following steps: Step 1: Construct a BP neural network; Step 2: Construct a rock mass mechanics parameter database, and input the mechanics parameters in the database into the BP neural network in Step 1 for training to obtain a trained BP neural network; Step 3: Obtain rock samples of the rock mass to be tested, measure the uniaxial compressive strength and longitudinal wave velocity of the rock samples, as well as the wave velocity of the rock mass with initial damage, and use these as inputs to the BP neural network trained in Step 2. Based on the trained BP neural network, output the compressive strength σ of the rock mass. c And the elastic modulus E of the rock mass, the compressive strength σ of the rock mass c The elastic modulus E of the rock mass is used as a reference parameter for tunnel excavation methods; The process of constructing the rock mass mechanics parameter database in step two is as follows: Rock samples containing all types of rock masses were collected. For any type of rock sample, at least three water-bearing states and ten damage states were tested to obtain a total of thirty mechanical parameters of the rock samples under different working conditions. The mechanical parameters were used as input to the neural network. The three moisture states are represented as S-1, S-2, and S-3, respectively, where S-1 represents the dry state (0% moisture content), S-2 represents the semi-saturated state (50% moisture content), and S-3 represents the fully saturated state (100% moisture content). The ten damage states are represented as D-1 to D-10 in sequence; The damage state is defined as follows: a continuously increasing load is applied to the rock specimen until cracking occurs, and the load strength that the rock specimen withstands when cracking is its ultimate strength; ten different load values ​​are selected sequentially from small to large between the initial load and the ultimate strength, with the load magnitude based on the ultimate strength of the rock specimen. The selection of the ten different load values ​​is as follows: the critical point of the plastic stage is used as the initial load value, 95% of the ultimate strength is used as the final load value, and the other load values ​​are selected sequentially within the range of 70% to 95% of the ultimate strength; the ten different loads correspond to ten damage states D-1 to D-10 of the rock specimen, where D-1 represents no initial damage, which is the damage state under the initial load value.

2. The method for obtaining rock mass mechanical parameters based on acoustic wave detection and artificial intelligence as described in claim 1, characterized in that, The uniaxial compressive strength of the rock sample of the rock mass to be tested was measured by uniaxial compressive strength test.

3. The method for obtaining rock mass mechanical parameters based on acoustic wave detection and artificial intelligence as described in claim 2, characterized in that, The mechanical parameters of the rock sample are: the uniaxial compressive strength of the rock sample, the longitudinal wave velocity of the rock sample, and the wave velocity of the rock sample with initial damage.

4. The method for obtaining rock mass mechanical parameters based on acoustic wave detection and artificial intelligence as described in claim 3, characterized in that, The mechanical parameters of the rock sample were obtained in the following manner: The uniaxial compressive strength of the rock sample is obtained in the following way: the compressive strength of the rock sample is measured by point load test, and the compressive strength of the rock sample is used as the input uniaxial compressive strength of the rock sample. The wave velocity of the rock sample with initial damage was obtained by pre-compressing and cracking the rock sample to obtain a rock sample with initial damage, and then testing the wave velocity using the seismic wave method.

5. The method for establishing a rock mass mechanics parameter database in the rock mass mechanics parameter acquisition method based on acoustic detection and artificial intelligence as described in any one of claims 1-4, characterized in that, Includes the following steps: Rock samples containing all types of rock masses were collected. For any type of rock sample, at least three water-bearing states and ten damage states were tested to obtain a total of thirty mechanical parameters under different working conditions. These mechanical parameters were used as inputs to a neural network. The three moisture states are represented as S-1, S-2, and S-3, respectively, where S-1 represents the dry state (0% moisture content), S-2 represents the semi-saturated state (50% moisture content), and S-3 represents the fully saturated state (100% moisture content). The ten damage states are represented as D-1 to D-10 in sequence; The damage state is defined as follows: a continuously increasing load is applied to the rock specimen until cracking occurs, and the load strength that the rock specimen bears when cracking is its ultimate strength; ten different load values ​​are selected sequentially from small to large between the initial load and the ultimate strength, with the load magnitude based on the ultimate strength of the rock specimen, the critical point of the plastic stage as the initial load value, and 95% of the ultimate strength as the final load value, and the other load values ​​are selected sequentially within the range of 70% to 95% of the ultimate strength; the ten different loads correspond to ten damage states D-1 to D-10 of the rock specimen, where D-1 represents no initial damage, which is the damage state under the initial load value.

Images