A method for establishing and predicting energy-time density atlas model for evaluating rock energy consumption
By constructing a rock energy time-density map model using deep learning denoising models and machine learning methods, the problems of data quality and pattern recognition in rock energy consumption analysis were solved, enabling accurate evaluation and prediction of energy dissipation, and providing a scientific basis for optimizing engineering blasting schemes and energy utilization.
Patent Information
- Application Number
- CN202510107428.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-23
- Publication Date
- 2025-11-18
- Estimated Expiration
- 2045-01-23
AI Technical Summary
Existing technologies lack data quality control and pattern recognition methods in the process of analyzing rock energy density. Complex experimental signals are often accompanied by a lot of noise, and it is difficult to capture nonlinear characteristics and potential patterns, making it difficult to accurately describe the spatiotemporal variation of energy consumption.
A deep learning denoising model is used to process stress-strain signals. A rock energy-time density model is constructed by combining K-means clustering algorithm and kernel density estimation or machine learning methods. A two-dimensional energy-time density map model is constructed by machine learning methods to accurately evaluate rock energy consumption.
It improves the accuracy and reliability of energy density calculation, enhances the ability to capture key features, provides high-quality data support for rock energy dissipation research, is applicable to various rock types and different experimental loading conditions, and provides intuitive visualization and scientific prediction capabilities.
Smart Images
Figure CN119918417B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of rock mechanics, and in particular to an energy-time density graph model establishment and prediction method for evaluating rock energy consumption. BACKGROUND
[0002] In the field of rock mechanics and energy, it is often necessary to extract the required rock material, which will be broken by various external loads. However, the fastest extraction method is still blasting and other dynamic load to break the rock, and then the extracted rock is further refined and processed. It is of great significance to study the energy dissipation characteristics of rock under external dynamic load for understanding its mechanical behavior, fracture mechanism and energy utilization. The traditional energy evaluation method only discusses the dissipated energy and is relatively rough, and it is difficult to accurately describe the spatio-temporal variation law in the energy consumption process. The energy-time density can consider both time and space factors, and can evaluate the energy relationship of rock under various load conditions, and can more comprehensively reflect the energy relationship of rock under various load conditions. The existing technology lacks data quality control and pattern recognition means in the energy-time density analysis process. Complex test signals are often accompanied by a large amount of noise, and the time series of rock dissipation process often has nonlinear characteristics and potential patterns that are difficult to accurately capture by traditional means. With the development of machine learning and deep learning technology, it is possible to realize data preprocessing, feature extraction and pattern prediction through intelligent means. At present, there is a lack of energy-time density range of various rocks. Based on the energy-time density index, machine learning and deep learning and other related technologies are used to establish a graph model as an evaluation method for rock energy consumption, which has important scientific significance and application value for rock energy dissipation research. SUMMARY
[0003] The purpose of the present application is to provide an energy-time density graph model establishment and prediction method for evaluating rock energy consumption, which aims to solve the problem that the traditional energy evaluation method only discusses the dissipated energy and is relatively rough, and it is difficult to accurately describe the spatio-temporal variation law in the energy consumption process.
[0004] In order to achieve the above-mentioned purpose, according to the first aspect of the present application, an energy-time density graph model establishment method for evaluating rock energy consumption is provided, which comprises the following steps:
[0005] Selecting a rock sample, performing a Hopkinson experiment to obtain stress and strain signals;
[0006] Using a deep learning denoising model to process the obtained stress-strain signals to obtain purified stress-strain signals;
[0007] Based on the purified stress-strain signals, calculating dynamic strength data and energy-time density;
[0008] Based on the dynamic strength data and energy-time density of the rock, and by using the K-means clustering algorithm and kernel density estimation or the interpolation model of machine learning, a two-dimensional energy-time density atlas model is constructed.
[0009] Further, the method of selecting rock samples and performing Hopkinson experiments to obtain stress and strain signals specifically comprises:
[0010] A plurality of rock samples are collected and processed into the size of the test piece required for the Hopkinson experiment;
[0011] Hopkinson experiments are performed on the rock samples to be tested, and five experiments are performed on each type of rock to collect the stress and strain signals of the rock under impact load.
[0012] Further, the method of processing the obtained stress-strain signals by using a deep learning denoising model specifically comprises:
[0013] The collected stress and strain signal data are input into the deep learning denoising model, the high-frequency noise in the data is automatically filtered by DAE, and the purified stress-strain signal is obtained.
[0014] Further, the specific acquisition steps of the dynamic strength data include:
[0015] Based on the purified stress-strain signal, and in combination with the one-dimensional stress wave theory of an elastic rod, the calculation method of the dynamic strength data of various rocks is as follows:
[0016]
[0017] In the formula, A, E, and C are the cross-sectional area, elastic modulus, and longitudinal wave velocity of the compression rod, respectively; A s , l s are the cross-sectional area and length of the sample, respectively; ε I (t), ε R (t), and ε T (t) are the incident strain, reflected strain, and transmitted strain of the compression rod;
[0018] Further, the specific acquisition steps of the energy-time density include:
[0019] Based on the purified stress-strain signal, the energy-time density under the same load condition is calculated by using the three-wave method;
[0020] The calculation method of each part of the energy is as follows:
[0021]
[0022]
[0023] W d =W i -W r -W t
[0024] In the formula, W i W r W t W d These are the incident energy, reflected energy, projected energy, and absorbed energy, respectively; σ i , σ r , σ t , respectively, are incident, reflected and transmitted stresses; A0 is the cross-sectional area of the rod; C0 is the elastic wave velocity in the compression member; ρ0 is the elastic rod density;
[0025] Energy dissipation density is defined as the energy dissipated per unit volume.
[0026]
[0027] In the formula: U d For energy density, W d V represents the absorbed energy, and V is the volume.
[0028] The energy-time density of rock is calculated based on the energy dissipation density. The energy-time density reflects the energy dissipated per unit volume of rock per unit time. The calculation formula is as follows:
[0029]
[0030] In the formula, E VT For energy density, U d Let T be the energy density and T be the duration of the reflected wave.
[0031] Furthermore, the method for constructing a two-dimensional energy-time density map model based on the dynamic intensity data and energy-time density of rocks, and utilizing the K-means clustering algorithm and kernel density estimation or machine learning interpolation model, specifically includes:
[0032] (1) Based on the energy-time density and dynamic intensity data of rocks under the same load, with dynamic intensity as the abscissa and energy-time density peak value as the ordinate, draw a scatter plot of basic energy-time density distribution, and generate a two-dimensional energy-time density map model by calculating the category and density probability of each scatter point based on K-means clustering algorithm and kernel density estimation.
[0033] The location of each energy density and dynamic intensity is calculated, and cluster analysis is performed to obtain the density probability of each cluster. Assume the data point set is D = {(x1, y1), (x2, y2), ..., (x...}. n y n)}, where x is the dynamic intensity and y is the peak value of the energy-time density. Based on the K-means clustering algorithm, the objective function of the clustering process can be expressed as:
[0034]
[0035] In the formula: J is the objective function, K is the number of clusters, ck is the center of the k-th cluster, (x i ,y i )-c k The Euclidean distance between the data point and the cluster center;
[0036] Kernel density estimation is used to estimate the probability density of the data points, generating a two-dimensional probability density map:
[0037]
[0038] In the formula, f(x, y) is the probability density function, n is the number of data points, and h is the number of data points. x h y The bandwidth parameter, K, is the kernel function, usually a Gaussian kernel.
[0039] (2) Based on the energy density and dynamic strength data of a single rock under different loads, interpolation and interpolation are performed on the energy density data under different strain rates. The machine learning interpolation model constructs a continuous and smooth energy density spatial distribution map, and a two-dimensional energy density map distribution model is obtained.
[0040] The energy density over time at different strain rates is interpolated using a machine learning interpolation model:
[0041]
[0042] In the formula, denoted as the predicted energy density value, f(x, θ) is the output of the MLP model, W is the weight matrix, b is the bias term, and x is the input feature.
[0043] According to a second aspect of the present invention, a prediction method for an energy-time density spectral model for evaluating rock energy consumption is also provided. The prediction method involves trend prediction based on the two-dimensional energy-time density spectral model to obtain an energy-time density prediction result.
[0044] Furthermore, the method for trend prediction based on the two-dimensional energy-time density map model to obtain the energy-time density prediction result specifically includes:
[0045] Extract the energy-time density with the maximum or minimum feature value, as well as the rising or falling trend of the energy-time density, from the two-dimensional energy-time density map model;
[0046] Trend curves are generated based on the extracted feature values, and based on the generated trend curves, the corresponding energy density results are predicted by inputting new dynamic intensity data, providing a reference for engineering decisions.
[0047] Compared with the prior art, the beneficial effects of the present invention are:
[0048] 1. Improved Accuracy: By introducing a deep learning denoising model to intelligently reduce noise in the experimental data and combining it with deep learning feature extraction techniques, the accuracy and reliability of energy density calculation are effectively improved. This method not only reduces the impact of noise on the results but also enhances the ability to capture key features, providing high-quality data support for subsequent analysis.
[0049] 2. Enhanced visualization: Machine learning-based analysis methods provide intuitive and in-depth visualization of the energy-time density map.
[0050] 3. Wide applicability and dynamic analysis: Applicable to various types of rocks and different experimental loading conditions. Through the prediction model, the energy dissipation under various conditions can be analyzed.
[0051] 4. Prediction and Decision Support: Utilizing the predictive capabilities of deep learning, energy dissipation can be predicted, providing a scientific basis for optimizing engineering blasting schemes, controlling rock fracture, and making energy utilization decisions.
[0052] Based on the implementation methods provided in the above aspects, this application can be further combined to provide more implementation methods. Attached Figure Description
[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0054] Figure 1 A flowchart for calculating the energy density spectrum for evaluating rock energy consumption in this invention has been prepared.
[0055] Figure 2 Figure a shows the original stress-strain data, and Figure b shows the data after DAE smoothing.
[0056] Figure 3 Figure a shows a two-dimensional map model of a typical rock, and Figure b shows a two-dimensional map model of a typical load.
[0057] Figure 4 A flowchart for predicting outcomes and decision analysis;
[0058] Figure 5 Figure a shows a two-dimensional map model of typical rocks in a metal mine, and Figure b shows a two-dimensional map model of typical loads in a metal mine.
[0059] Figure 6 Figure a shows a two-dimensional map model of typical rocks in a limestone mine, and Figure b shows a two-dimensional map model of typical loads in a limestone mine. Detailed Implementation
[0060] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of 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 skilled in the art without creative effort should fall within the scope of protection of the present invention. Unless otherwise specified, the technical means used in the embodiments are conventional means well known to those skilled in the art.
[0061] This invention provides a method for establishing an energy-time density map model to evaluate rock energy consumption. See also... Figure 1 , Figure 1 A flowchart illustrating a method for establishing an energy-time density spectrum model for evaluating rock energy consumption, provided in an embodiment of the present invention. The method for establishing the energy-time density spectrum model includes the following steps:
[0062] Rock samples were selected and SHPB (Hopkinson) experiments were performed to obtain stress and strain signals;
[0063] The acquired stress-strain signal is processed using a deep learning denoising model to obtain a purified stress-strain signal.
[0064] Based on the purified stress-strain signal, dynamic intensity data and energy-time density are calculated;
[0065] Based on the dynamic intensity data and energy-time density of rocks, and using K-means clustering algorithm and KDE (kernel density estimation) probability density algorithm or machine learning interpolation model, a two-dimensional energy-time density map model is constructed.
[0066] Based on the energy-time density map model for evaluating rock energy consumption provided in this invention, a deep learning denoising model is used to purify the stress-strain signal obtained from the SHPB experiment, improving data quality and the accuracy of subsequent calculations. Using the purified signal, the dynamic strength and energy-time density of the rock are accurately calculated, and an intuitive two-dimensional energy-time density map is constructed by combining algorithms such as K-means clustering and KDE probability density estimation.
[0067] In one embodiment, the method of selecting rock samples and performing SHPB experiments to obtain stress and strain signals specifically includes:
[0068] Collect and process various rock samples into specimens of the required size for SHPB experiments;
[0069] SHPB experiments were performed on the rock samples to be tested, with five experiments conducted for each type of rock. Stress and strain signals of the rocks under impact loading were collected. Figure 2 As shown in Figure a.
[0070] In one embodiment, the method for processing the acquired stress-strain signal using a deep learning denoising model specifically includes:
[0071] The collected stress and strain signal data are input into a deep learning denoising model. A Dependent Engine (DAE) automatically filters high-frequency noise from the data to obtain the purified stress-strain signal, such as... Figure 2 As shown in Figure b.
[0072] In one embodiment, the specific steps for acquiring the dynamic intensity data include:
[0073] Based on the purified stress-strain signals and combined with the one-dimensional stress wave theory of elastic rods, the dynamic strength data of various types of rocks are calculated as follows:
[0074]
[0075] In the formula, A, E, and C represent the cross-sectional area, elastic modulus, and longitudinal wave velocity of the compression member, respectively; A s l s These represent the cross-sectional area and length of the sample, respectively; ε I (t), ε R (t), ε T (t) represents the incident strain, reflected strain, and transmitted strain of the compression bar.
[0076] In one embodiment, the specific steps for obtaining the energy density include:
[0077] Based on the purified stress-strain signal, the energy-time density under the same load conditions is calculated using the three-wave method.
[0078] The energy calculation methods for each part are as follows:
[0079]
[0080] W d =W i -W r -W t
[0081] In the formula, W i W r W t W d These are the incident energy, reflected energy, projected energy, and absorbed energy, respectively; σ i , σ r , σ t , respectively, are incident, reflected and transmitted stresses; A0 is the cross-sectional area of the rod; C0 is the elastic wave velocity in the compression member; ρ0 is the elastic rod density;
[0082] Energy dissipation density is defined as the energy dissipated per unit volume.
[0083]
[0084] In the formula: U d For energy density, W d V represents the absorbed energy, and V is the volume.
[0085] The energy-time density of rock is calculated based on the energy dissipation density. The energy-time density reflects the energy dissipated per unit volume of rock per unit time. The calculation formula is as follows:
[0086]
[0087] In the formula, E VT For energy density, U d Let T be the energy density and T be the duration of the reflected wave.
[0088] In one embodiment, a method for constructing a two-dimensional energy-time density map model based on the dynamic intensity data and energy-time density of rocks, and utilizing K-means clustering algorithm and KDE probability density algorithm or machine learning interpolation model, specifically includes:
[0089] (1) Based on the energy-time density and dynamic intensity data of rocks under the same load, a scatter plot of the basic energy-time density distribution is drawn with dynamic intensity as the abscissa and peak energy-time density as the ordinate. A two-dimensional energy-time density map model is generated by calculating the category and density probability of each scatter point based on the K-means clustering algorithm and the KDE probability density algorithm. Figure 3 As shown in Figure a;
[0090] The location of each energy density and dynamic intensity is calculated, and cluster analysis is performed to obtain the density probability of each cluster. Assume the data point set is D = {(x1, y1), (x2, y2), ..., (x...}. n y n )}, where x is the dynamic intensity and y is the peak value of the energy-time density. Based on the K-means clustering algorithm, the objective function of the clustering process can be expressed as:
[0091]
[0092] In the formula: J is the objective function, K is the number of clusters, ck is the center of the k-th cluster, (x i ,y i )-c k The Euclidean distance between the data point and the cluster center;
[0093] KDE is used to estimate the probability density of the data points, generating a two-dimensional probability density map:
[0094]
[0095] In the formula, f(x, y) is the probability density function, n is the number of data points, and h is the number of data points. x h y The bandwidth parameter, K, is the kernel function, usually a Gaussian kernel.
[0096] (2) Based on the energy-time density and dynamic strength data of a single rock under different loads, interpolation and interpolation are performed on the energy-time density data under different strain rates. The machine learning interpolation model constructs a continuous and smooth spatial distribution map of energy-time density, resulting in a two-dimensional energy-time density distribution model, such as... Figure 3 As shown in Figure b;
[0097] The energy density over time at different strain rates is interpolated using a machine learning interpolation model:
[0098]
[0099] In the formula, denoted as the predicted energy density value, f(x, θ) is the output of the MLP model, W is the weight matrix, b is the bias term, and x is the input feature.
[0100] Based on the same inventive concept, this invention also provides a prediction method for an energy time density spectrum model for evaluating rock energy consumption.
[0101] See Figure 4 , Figure 4 An embodiment of the present invention also provides a prediction method for an energy time density spectral model for evaluating rock energy consumption. The prediction method is to perform trend prediction based on the two-dimensional energy time density spectral model to obtain the energy time density prediction result.
[0102] In one embodiment, the method for trend prediction based on the two-dimensional energy-time density map model to obtain the energy-time density prediction result specifically includes:
[0103] Extract the energy-time density with the maximum or minimum feature value, as well as the rising or falling trend of the energy-time density, from the two-dimensional energy-time density map model;
[0104] Trend curves are generated based on the extracted feature values, and based on the generated trend curves, the corresponding energy density results are predicted by inputting new dynamic intensity data, providing a reference for engineering decisions.
[0105] Based on the present invention, a prediction method for evaluating rock energy consumption using an energy-time density map model is provided. This method is applicable to various rock types and different loading conditions. Furthermore, through the two-dimensional energy-time density map model, it can predict energy dissipation, providing a scientific basis for optimizing engineering decisions. In summary, this method improves the accuracy, visualization effect, and applicability of rock energy consumption assessment, and provides strong support for blasting scheme optimization, rock fracture control, and energy utilization.
[0106] Example 1:
[0107] Energy-time-density maps were drawn for typical metal mines, which are rich in a variety of typical rocks.
[0108] Step 1. Data Acquisition and Processing
[0109] a) Collect and process various typical rock specimens to the specimen size required for the SHPB test;
[0110] b) Perform SHPB tests on the rock to be tested. Five tests are conducted for each typical rock type to collect stress and strain signals of the rock under impact load.
[0111] c) Input the collected stress and strain time series data into the deep learning denoising model. The DAE automatically filters the high-frequency noise in the data and retains the main information related to rock stress characteristics, which is convenient for subsequent calculations.
[0112] Step 2. Dynamic intensity calculation
[0113] Based on the one-dimensional stress wave theory of elastic rods and the three assumptions of the SHPB test, the dynamic strength data of various types of rocks are calculated as follows:
[0114]
[0115] In the formula: A, E, and C are the cross-sectional area, elastic modulus, and longitudinal wave velocity of the compression member, respectively; A s l s These represent the cross-sectional area and length of the sample, respectively; ε I (t), ε R (t), ε T (t) represents the incident strain, reflected strain, and transmitted strain of the compression bar.
[0116] Step 3. Calculation of energy density over time
[0117] The three-wave method was used to calculate the energy density of a typical rock under the same load conditions.
[0118] The energy calculation methods for each part are as follows:
[0119]
[0120] W d =W i -W r -W t
[0121] In the formula: W i W r W t W d These are the incident energy, reflected energy, projected energy, and absorbed energy, respectively; σ i , σ r , σ t , respectively, are incident, reflected and transmitted stresses; A0 is the cross-sectional area of the rod; C0 is the elastic wave velocity in the compression member; ρ0 is the elastic rod density;
[0122] Energy dissipation density is defined as the energy dissipated per unit volume.
[0123]
[0124] In the formula: U d For energy density, W d V represents the absorbed energy, and V is the volume.
[0125] The energy-time density of rock is calculated based on the energy dissipation density. The energy-time density reflects the energy dissipated per unit volume of rock per unit time. The calculation formula is as follows:
[0126]
[0127] In the formula, E VT For energy density, U d Let T be the energy density and T be the duration of the reflected wave.
[0128] Step 4. Construction of Energy-Time Density Map
[0129] a) Two-dimensional energy-time density map model for typical rocks: Based on the energy-time density and dynamic intensity data of typical rocks under the same load, a scatter plot of the basic energy-time density distribution is drawn with dynamic intensity as the abscissa and peak energy-time density as the ordinate. A two-dimensional energy-time density map model is generated by calculating the category and density probability of each scatter point based on the K-means clustering algorithm and the KDE probability density algorithm. Figure 5 As shown in Figure a.
[0130] b) Two-dimensional energy-time density spatial distribution model under typical loads: Based on the energy-time density and dynamic strength data of a single rock under different loads, interpolation and interpolation are performed on the energy-time density data under different experimental conditions (different strain rates). A continuous and smooth energy-time density spatial distribution map is constructed using a machine learning interpolation model (based on a deep learning interpolation network), resulting in a more comprehensive two-dimensional energy-time density distribution model, such as... Figure 5 As shown in Figure b.
[0131] c) Result prediction and model feature labeling:
[0132] The generated graph model can be further used to generate its trend curve, which can then be used for relevant predictions. At the same time, it can be combined with statistical analysis methods to find its characteristic values, providing a basis for subsequent decision-making.
[0133] Step 5. Graph Analysis and Decision Evaluation
[0134] a) Quantitative analysis: Extracting characteristic values (maximum and minimum energy density, upward and downward trends) from the spectrum;
[0135] b) Result prediction: Based on the generated trend curve, the relevant results can be predicted and analyzed. Inputting dynamic intensity data can predict the energy density results and find the maximum energy density range. This can further guide the use of specific energy parameters for crushing and other operations in production, providing a reference for engineering decisions.
[0136] Example 2:
[0137] Based on the same inventive concept as Example 1, the difference between Example 2 and Example 1 is that the rock types used for calculation are different, and the dynamic characteristics and energy dissipation of different rock types are also different.
[0138] Energy-time-density maps were drawn for typical limestone mines, where the rock types were uniform.
[0139] Step 1. Data Acquisition and Processing
[0140] a) Collect and process various typical rock specimens to the specimen size required for the SHPB test;
[0141] b) Perform SHPB tests on the rock to be tested. Five tests are conducted on each typical rock to collect stress and strain signals of the rock under impact load.
[0142] c) Input the collected stress and strain time series data into the deep learning denoising model. The DAE automatically filters the high-frequency noise in the data and retains the main information related to rock stress characteristics, which is convenient for subsequent calculations.
[0143] Step 2. Dynamic intensity calculation
[0144] Based on the one-dimensional stress wave theory of elastic rods and the three assumptions of the SHPB test, the dynamic strength data of various types of rocks are calculated as follows:
[0145]
[0146] In the formula: A, E, and C are the cross-sectional area, elastic modulus, and longitudinal wave velocity of the compression member, respectively; A s l s These represent the cross-sectional area and length of the sample, respectively; ε I (t), ε R (t), ε T (t) represents the incident strain, reflected strain, and transmitted strain of the compression bar.
[0147] Step 3. Calculation of energy density over time
[0148] The three-wave method was used to calculate the energy density of a typical rock under the same load conditions.
[0149] The energy calculation methods for each part are as follows:
[0150]
[0151] W d =W i -W r -W t
[0152] In the formula: W i W r W t W d These are the incident energy, reflected energy, projected energy, and absorbed energy, respectively; σ i , σ r , σ t , respectively, are incident, reflected and transmitted stresses; A0 is the cross-sectional area of the rod; C0 is the elastic wave velocity in the compression member; ρ0 is the elastic rod density;
[0153] Energy dissipation density is defined as the energy dissipated per unit volume.
[0154]
[0155] In the formula: U d For energy density, W d V represents the absorbed energy, and V is the volume.
[0156] The energy-time density of rock is calculated based on its energy dissipation density. Energy-time density reflects the energy dissipated per unit volume of rock per unit time. The calculation formula is as follows:
[0157]
[0158] In the formula, E VT For energy density, U d Let T be the energy density and T be the duration of the reflected wave.
[0159] Step 4. Construction of Energy-Time Density Map
[0160] a) Two-dimensional energy-time density map model for typical rocks: Based on the energy-time density and dynamic intensity data of typical rocks under the same load, a scatter plot of the basic energy-time density distribution is drawn with dynamic intensity as the abscissa and peak energy-time density as the ordinate. A two-dimensional energy-time density map model is generated by calculating the category and density probability of each scatter point based on the K-means clustering algorithm and the KDE probability density algorithm. Figure 6 As shown in Figure a.
[0161] b) Two-dimensional energy-time density spatial distribution model under typical loads: Based on the energy-time density and dynamic strength data of a single rock under different loads, interpolation and interpolation are performed on the energy-time density data under different experimental conditions (different strain rates). A continuous and smooth energy-time density spatial distribution map is constructed using a machine learning interpolation model (based on a deep learning interpolation network), resulting in a more comprehensive two-dimensional energy-time density distribution model, such as... Figure 6 As shown in Figure b.
[0162] c) Result prediction and model feature labeling:
[0163] The generated graph model can be further used to generate its trend curve, which can then be used for relevant predictions. At the same time, it can be combined with statistical analysis methods to find its characteristic values, providing a basis for subsequent decision-making.
[0164] Step 5. Graph Analysis and Decision Evaluation
[0165] a) Quantitative analysis: Extract characteristic values (maximum and minimum energy density, upward and downward trends) from the spectrum.
[0166] b) Result prediction: Based on the generated trend curve, the relevant results can be predicted and analyzed. Inputting dynamic intensity data can predict the energy density results and find the maximum energy density range. This can further guide the use of specific energy parameters for crushing and other operations in production, providing a reference for engineering decisions.
[0167] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.
Claims
1. A method for establishing an energy-time density spectrum model for evaluating rock energy consumption, characterized in that, The method for establishing the energy-time density map spectral model includes the following steps: Rock samples were selected for Hopkinson experiments to obtain stress and strain signals. The acquired stress-strain signal is processed using a deep learning denoising model to obtain a purified stress-strain signal. Based on the purified stress-strain signal, dynamic intensity data and energy-time density are calculated; Based on the dynamic intensity data and energy-time density of rocks, and using the K-means clustering algorithm and kernel density estimation or machine learning interpolation model, a two-dimensional energy-time density map model is constructed. The method for constructing a two-dimensional energy-time density map model based on dynamic intensity data and energy-time density of rocks, and using K-means clustering algorithm and kernel density estimation or machine learning interpolation model, specifically includes: (1) Based on the energy density and dynamic intensity data of rocks under the same load, with dynamic intensity as the abscissa and energy density peak value as the ordinate, draw a scatter plot of basic energy density distribution. Based on K-means clustering algorithm and kernel density estimation, calculate the category and density probability of each scatter point to generate a two-dimensional energy density map model. The location of each energy density and dynamic intensity is calculated, and cluster analysis is performed to obtain the density probability of each cluster. Assume the data point set is... D ={( x 1, y 1), x 2, y 2), ..., ( x n , y n )},in x For dynamic intensity, y To determine the peak value of the energy density, based on the K-means clustering algorithm, the objective function of the clustering process can be expressed as: ； In the formula: J Let be the objective function. K For the number of clusters, c k For the first k The center of each cluster, The Euclidean distance between the data point and the cluster center; Kernel density estimation is used to estimate the probability density of the data points, generating a two-dimensional probability density map: ； In the formula, f ( x , y ) is the probability density function. n The number of data points, h x , h y Bandwidth parameters K The kernel function is usually a Gaussian kernel; (2) Based on the energy density and dynamic strength data of a single rock under different loads, interpolation and interpolation are performed on the energy density data under different strain rates. The machine learning interpolation model constructs a continuous and smooth energy density spatial distribution map, and a two-dimensional energy density map distribution model is obtained. The energy density over time at different strain rates is interpolated using a machine learning interpolation model: ； In the formula, The predicted energy density value, f ( x , θ ) represents the output of the MLP model. W This is the weight matrix. b For bias terms, x For input features.
2. The method for establishing an energy-time density map model for evaluating rock energy consumption according to claim 1, characterized in that, The method for selecting rock samples and performing Hopkinson experiments to obtain stress and strain signals specifically includes: Collect and process various rock samples to the specimen size required for the Hopkinson's experiment; Hopkinson's experiments were performed on the rock samples to be tested, with five experiments conducted for each type of rock to collect stress and strain signals under impact loading.
3. The method for establishing an energy-time density spectrum model for evaluating rock energy consumption according to claim 1, characterized in that, The methods for processing the acquired stress-strain signals using deep learning denoising models specifically include: The collected stress and strain signal data are input into a deep learning denoising model. The high-frequency noise in the data is automatically filtered out by the DAE to obtain the purified stress-strain signal.
4. The method for establishing an energy-time density map model for evaluating rock energy consumption according to claim 1, characterized in that, The specific steps for acquiring the dynamic intensity data include: Based on the purified stress-strain signal and combined with the one-dimensional stress wave theory of elastic rods, the calculation method for the dynamic strength data of various types of rocks is as follows: ； In the formula, A , E , C These are the cross-sectional area, elastic modulus, and longitudinal wave velocity of the compression member, respectively. A s , l s These are the cross-sectional area and length of the sample, respectively. , , The incident strain, reflected strain, and transmitted strain of the compression bar are given.
5. The method for establishing an energy-time density spectrum model for evaluating rock energy consumption according to claim 1, characterized in that, The specific steps for obtaining the energy density include: Based on the purified stress-strain signal, the energy-time density under the same load conditions is calculated using the three-wave method. The energy calculation methods for each part are as follows: ； ； ； ； In the formula, W i , W r , W t , W d These are the incident energy, reflected energy, projected energy, and absorbed energy, respectively; σ i , σ r , σ t , respectively, are incident, reflected and transmitted stresses; A0 is the cross-sectional area of the rod; C0 is the elastic wave velocity in the compression member; ρ 0 represents the density of the elastic rod; Energy dissipation density is defined as the energy dissipated per unit volume. ； In the formula: U d Energy density, W d To absorb energy, V For volume; The energy-time density of rock is calculated based on the energy dissipation density. The energy-time density reflects the energy dissipated per unit volume of rock per unit time. The calculation formula is as follows: ； In the formula, E VT For energy density, U d Energy density, T This represents the duration of the reflected wave.
6. A prediction method for an energy-time density map model for evaluating rock energy consumption, wherein the model is established based on the establishment method described in any one of claims 1-5, characterized in that, The prediction method is to perform trend prediction based on the two-dimensional energy-time density map model to obtain the energy-time density prediction result.
7. The prediction method for the energy-time density map model for evaluating rock energy consumption according to claim 6, characterized in that, The method for trend prediction based on the two-dimensional energy-time density map model to obtain energy-time density prediction results specifically includes: Extract the energy-time density with the maximum or minimum feature value, as well as the rising or falling trend of the energy-time density, from the two-dimensional energy-time density map model; Trend curves are generated based on the extracted feature values, and based on the generated trend curves, the corresponding energy density results are predicted by inputting new dynamic intensity data, providing a reference for engineering decisions.
Citation Information
Patent Citations
Fault automatic detection and repair method for self-healing intelligent power line
CN118739184A