Mine blasting analysis method

Abstract

The invention discloses a mine blasting analysis method, and relates to the field of mine blasting. The method comprises the following steps: collecting measurement data of a link to be analyzed in a mine blasting process; the ChatGPT-Matlab framework is utilized to compile an analysis method used by the link to be analyzed; according to the measurement data, an analysis method in the ChatGPT-Matlab framework is operated, and a blasting analysis result of the link to be analyzed is obtained. The technical barrier of green blasting development can be overcome, and the technical use cost is reduced.

Description

Technical Field

[0001] The present application relates to the field of mine blasting, and in particular to a mine blasting analysis method. Background Art

[0002] As the global demand for mineral resources continues to grow, people are paying more and more attention to reducing the impact of mining on the environment. In the transition to green mining, "digitalization" is the most important, and "mechanization and automation", "intelligence", "energy saving and utilization" and "autonomous driving and driverless cars" are ranked second, third and fourth. Although many links in mining have been digitized and intelligentized, the digital intelligence level of the drilling and blasting link in mining is still very low, far from green blasting.

[0003] The key parameters for evaluating green mining are energy consumption, greenhouse gas emissions and chemical use, and each of these three aspects is closely related to mine blasting. Green blasting mainly includes three aspects: ① Improve the energy utilization rate of explosives and reduce the use of explosives; ② Obtain an acceptable rock crushing rate and reduce the production cost and pollution of subsequent grinding and crushing; ③ Control the pollution and negative impact caused by the blasting link. The reasons that restrict the development of green blasting mainly come from two aspects. On the one hand, it comes from the weak green production awareness of mine construction personnel, and on the other hand, it comes from the imperfect application of green blasting technology. Among them, the main factors restricting the application of green blasting technology are the technical barriers and technology use costs caused by the low knowledge level of blasting practitioners. Summary of the invention

[0004] The purpose of this application is to provide a mining blasting analysis method that can overcome the technical barriers to the development of green blasting and reduce the cost of using the technology.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] The present application provides a mine blasting analysis method, comprising: collecting measurement data of a link to be analyzed during a mine blasting process; using a ChatGPT-Matlab framework to write an analysis method used in the link to be analyzed; and running the analysis method in the ChatGPT-Matlab framework based on the measurement data to obtain a blasting analysis result of the link to be analyzed.

[0007] According to the specific embodiments provided in this application, this application has the following technical effects:

[0008] The present application provides a method for analyzing mining blasting. The analysis method used in the link to be analyzed is written using the ChatGPT-Matlab framework. The analysis method in the ChatGPT-Matlab framework is run to obtain the blasting analysis results of the link to be analyzed. After the analysis method is written in the ChatGPT-Matlab framework, the measurement data can be directly used without complicated processing and requiring low learning ability, thus overcoming the technical barriers caused by the low knowledge level of blasting practitioners; and the software use cost of the ChatGPT-Matlab framework is low, which reduces the cost of using technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0010] Figure 1 This is a schematic diagram of a process of a mine blasting analysis method in one embodiment of the present application;

[0011] Figure 2 A schematic diagram of the ChatGPT-Matlab framework application concept provided for another embodiment of the present application;

[0012] Figure 3 A schematic diagram of a method for controlling blasting unit consumption in engineering practice provided by another embodiment of the present application;

[0013] Figure 4 A schematic diagram of an original rock damage analysis diagram provided by another embodiment of the present application;

[0014] Figure 5 A schematic diagram of the analysis results of ImageJ provided in one embodiment of the present application;

[0015] Figure 6 A schematic diagram of the analysis results of the ChatGPT-Matlab framework provided in an embodiment of the present application;

[0016] Figure 7 A schematic diagram of a Wipfrag analysis explosion provided in another embodiment of the present application;

[0017] Figure 8 A schematic diagram of a ChatGPT-Matlab framework analysis explosion provided in another embodiment of the present application;

[0018] Fig. 9 A schematic diagram of a blasting vibration signal collection and processing flow chart provided in another embodiment of the present application;

[0019] Fig.10 A schematic diagram of vibration signal synthesis using a ChatGPT-Matlab framework provided in another embodiment of the present application;

[0020] Fig.11 A schematic diagram of EMD decomposition and Matlab error reporting using the ChatGPT-Matlab framework provided in another embodiment of the present application;

[0021] Fig.12 A schematic diagram of an adjustment error report provided in another embodiment of the present application;

[0022] Fig.13 A schematic diagram of the time amplitude and time-frequency diagram of IMF1-4 provided in another embodiment of the present application;

[0023] Fig.14 A schematic diagram of a ChatGPT-Matlab framework provided for another embodiment of the present application;

[0024] Fig.15 A schematic diagram of HHT analysis and output provided by another embodiment of the present application;

[0025] Fig.16 A schematic diagram of drawing marginal spectra and calculating frequency band energy ratios using a ChatGPT-Matlab framework provided in another embodiment of the present application;

[0026] Fig.17 A schematic diagram of the marginal spectrum and the frequency band energy ratio provided by another embodiment of the present application;

[0027] Fig.18 This is a schematic diagram of the existing membrane sampling method for monitoring blasting fly ash;

[0028] Fig.19 A schematic diagram of a method for photographing dust in an explosion area provided by another embodiment of the present application;

[0029] Fig. 20 A schematic diagram of a fly dust concentration video processing flow provided by another embodiment of the present application;

[0030] Fig.21 A schematic diagram of the results of binarization and grayscale value extraction and processing of a video using the ChatGPT-Matlab framework provided in another embodiment of the present application;

[0031] Fig. 22 A schematic diagram of a ChatGPT-Matlab framework for performing differential analysis on each frame provided in another embodiment of the present application;

[0032] Fig.23 A schematic diagram of a dust concentration curve provided in another embodiment of the present application. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0034] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0035] ChatGPT is an AI-based dialogue system developed by OpenAI. It is based on the GPT (Generative Pre-trained Transformer) model, an advanced language processing technology that can generate coherent and logical text. ChatGPT has a wide range of applications, including but not limited to answering questions, writing articles, generating creative content, writing code, simulating conversations, etc. Due to its natural language understanding ability based on deep learning, ChatGPT can provide corresponding answers or suggestions based on the context, which makes it more natural and helpful in interacting with humans. The emergence of ChatGPT has aroused great interest among ordinary users and many researchers. Research on ChatGPT at home and abroad has grown explosively, mainly focusing on the fields of humanities and social sciences, medicine, and economics.

[0036] The applications of ChatGPT in different fields include: ChatGPT was used to create multiple-choice questions related to psychology tests. After manual review, it was found that the questions were acceptable and comparable to those created manually. By comparing the problem solutions given by college students with and without ChatGPT, it was found that the use of ChatGPT significantly improved the quality of the solutions and the efficiency of task completion. Participants completed 2 standardized assessments on different general surgery topics before and after using Google Search or ChatGPT. The results showed that ChatGPT was comparable to Google Search in finding answers related to surgery. The potential of ChatGPT for scoring programming tasks was explored. The understanding of cancer genes by ChatGPT was tested, and it was verified that ChatGPT could answer the consulting questions of medical personnel based on the understanding. ChatGPT was applied to common treatment scenarios for knee patients. The results showed that ChatGPT was consistent with the judgments given by professional rehabilitation doctors, demonstrating the application potential of ChatGPT in clinical medicine. The application prospects of question-answering systems in smart agriculture. Let ChatGPT play the role of an investment advisor and conduct a diversified portfolio analysis. The investment results show that ChatGPT's investment portfolio is better than the random investment portfolio, which shows that ChatGPT has the potential to serve as an investment assistant.

[0037] Nowadays, many scholars have mainly studied the natural language feedback ability of ChatGPT, and the application of ChatGPT is mainly concentrated in the natural language discipline, while there are fewer studies on ChatGPT's programming ability and its application in the engineering field. Limited research on ChatGPT's programming performance shows that ChatGPT performs well in solving simple and medium-difficulty programming problems, but the code has low runtime performance when dealing with complex tasks, and its performance in dealing with complex tasks is still far behind that of humans. In addition, in the field of engineering applications, ChatGPT's understanding of cement-based composites was studied through human-computer interaction, and ChatGPT was combined with ChatGPT to identify and analyze ECC cracks, and ChatGPT's design ability for cement-based materials was evaluated. Using ChatGPT's excellent programming ability, examples of two-dimensional seepage simulation, slope stability analysis and image recognition were also completed, and the feasibility and application methods of ChatGPT in geotechnical engineering were explored. Based on ChatGPT's programming ability, a medical image recognition process based on ChatGPT was given through the user's dialogue with ChatGPT, demonstrating the powerful ability of AI in medical image analysis. The application prospects of ChatGPT programming ability in the context of unmanned and intelligent fisheries were discussed, but the discussion was based on assumptions and did not involve the specific implementation process. ChatGPT was asked to reconstruct 40 sets of Java code. During the research process, it was found that ChatGPT had excellent modification performance when the code volume was small, but ChatGPT's response was unpredictable, and different codes would be output under the same prompt. Despite this, ChatGPT can still be used as an excellent auxiliary programming software.

[0038] In the ChatGPT-Matlab framework, ChatGPT-4o is free and open to the public. The permanent use of the Matlab standard version is only $938.4. If the ChatGPT-Matlab framework can be used on a large scale, it can greatly reduce the cost of software use and directly reduce the cost of blasting analysis. At the same time, if the ChatGPT-Matlab framework is mature, it will greatly improve the efficiency of blasting optimization and indirectly reduce the cost of the entire mining process.

[0039] Therefore, this study aims to use the programming capabilities of ChatGPT to solve blasting engineering problems, such as rock damage analysis and vibration signal processing, and to use image change detection technology to try to monitor blasting fly ash concentration.

[0040] In an exemplary embodiment, Figure 1 As shown, a mine blasting analysis method is provided, comprising the following steps 101 to 103. Among them:

[0041] Step 101: Collect measurement data of the link to be analyzed during the mine blasting process.

[0042] Step 102: Use the ChatGPT-Matlab framework to write the analysis method used in the link to be analyzed.

[0043] Step 103: According to the measurement data, the analysis method in the ChatGPT-Matlab framework is run to obtain the blasting analysis result of the link to be analyzed.

[0044] By implementing the above-mentioned steps 101 to 103, after the analysis method is written, the ChatGPT-Matlab framework can directly use the measurement data without complicated processing and requires a lower learning ability, thus overcoming the technical barriers caused by the low knowledge level of blasting practitioners; and the software usage cost of the ChatGPT-Matlab framework is low, which reduces the cost of using the technology.

[0045] In another exemplary embodiment of the present application, the links to be analyzed in the mine blasting process include optimization of explosive consumption, blasting vibration analysis and flying dust monitoring. The specific application of the mine blasting analysis method in these three links to be analyzed are introduced one by one below.

[0046] (1) Optimization of explosive consumption

[0047] In terms of optimizing the amount of explosives used in blasting, domestic and foreign scholars often use Ansys / Ls-dyna to analyze the blasting process. By simulating the blasting process under different explosive unit consumption on the software, the stress distribution and rock damage are analyzed. Combined with the statistical results of rock damage, the explosive unit consumption with the best numerical simulation effect is selected for field tests. The advantage of the numerical simulation method is that it can complete multiple simulation experiments in a short period of time, avoiding the dangers and high costs that may occur in the actual blasting process. In addition, numerical simulation can accurately control blasting parameters such as explosive quantity, hole diameter, hole spacing, etc., so as to carefully study the effects of these parameters on the blasting effect. This high precision and controllability helps to optimize the explosive unit consumption and find the best parameter combination. The numerical simulation method can reduce the number of actual blasting experiments, thereby reducing damage to the environment and waste of explosives. However, the statistical method of rock damage in numerical simulation requires professional knowledge and skills, and has high requirements for researchers. For some small and medium-sized enterprises or companies with weak technical strength, it is difficult to apply.

[0048] In the blasting field practice, after blasting, the unit consumption of explosives is often optimized according to the large block rate, in order to minimize the amount of explosives and maximize the blasting efficiency. Large rock refers to the rock block size that reduces the excavation efficiency or affects the subsequent production during the excavation and loading process after blasting. With the development of computer technology, the acquisition of large block rate has become easier and more accurate. The two computer software Split-Desktop and WipFrag are widely used in blasting block size analysis due to their fool-proof image contour extraction process, and have become commonly used blasting block size analysis software in mines. The blaster uploads the photos of the blast pile, block size, etc. taken on site to the software, sets the scale, and the software automatically captures the rock contour to achieve the purpose of analyzing large blocks. Although the use process of Split-Desktop and WipFrag is simple, the software has poor freedom. It can only use its own contour extraction method to increase or decrease the line density to adjust the contour extraction accuracy, and cannot perform detailed parameter adjustment to achieve the best recognition effect. This directly leads to inaccurate acquisition of large block rate, and it is difficult to find the best unit consumption of explosives in the process of optimizing unit consumption of explosives. Moreover, the method of counting large blocks after blasting has a lag, and cannot predict and optimize blasting parameters in advance. This lag may lead to waste of resources and reduced efficiency.

[0049] The numerical simulation rock damage statistical method and the large block rate statistics have their own advantages and disadvantages, and should be used in combination in the optimization of explosive unit consumption. The numerical simulation rock damage statistical method is suitable for early optimization and parameter adjustment, and can provide high-precision theoretical guidance; the large block rate statistics are suitable for on-site actual effect evaluation and provide intuitive feedback data. The combination of the two can maximize the optimization of explosive unit consumption, improve blasting effects, and reduce costs while ensuring accuracy and practicality.

[0050] Engineering practice blasting unit consumption control method such as Figure 3 As shown in the figure, before blasting, 3D blasting area is modeled with Ansys / Ls-dyna, and rock explosive parameters are set. Then, blasting is simulated, and the rock damage model after blasting is imported into the ChatGPT-Matlab framework for RGB damage analysis to obtain the size of the crushing area, fracture area, and non-damaged area (the analysis part in the figure is red, green, and blue area identification from left to right, and the gray part is the target area). In this process, simulation tests with different parameters are continuously tried to find the blasting parameters that maximize the fracture area and crushing area, and engineering practice is carried out. After the optimal parameters are actually designed to the blasting area on site, the test is carried out according to the on-site charging and blasting process. After the test, rock images are collected after the explosion test, and the rock block size is identified using the ChatGPT-Matlab framework. According to the rock block size, the unit consumption of explosives is continuously fine-tuned. Figure 3 The horizontal axis of the RGB statistical data graph is distance (pixel) and the vertical axis is grayscale value. Figure 3The horizontal axis of the large block analysis chart is the rock diameter (mm), and the coordinate values ​​at both ends are 100mm and 1000mm; the vertical axis is the pass rate (%), from bottom to top, it is 0, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%. The corresponding relationship between rock diameter and pass rate is: 100~20.02%, 200~42.48%, 300~58.60%, 400~72.91%, 500~81.45%, 600~94.12%, 700~100.00%, 800~100.00%, 900~100.00%, 1000~100.00%.

[0051] In view of the above, in an exemplary embodiment, the above step 101 of collecting the measurement data of the link to be analyzed during the mining blasting process can be replaced by the following steps 201 to 203:

[0052] Step 201: Before blasting in the mine, use Ansys / Ls-dyna to model the three-dimensional blasting area.

[0053] Step 202: Set the rock explosive unit consumption.

[0054] Step 203: Based on the rock explosive unit consumption simulation blasting, a rock damage model after mine blasting is obtained as measurement data for optimizing explosive unit consumption.

[0055] In another exemplary embodiment, the process of running the analysis method in the ChatGPT-Matlab framework in step 103 above can be replaced by the following steps 301 to 306:

[0056] Step 301: Import the rock damage model after mine blasting into the ChatGPT-Matlab framework for RGB damage analysis to obtain the crushing area, the crack area and the undamaged area.

[0057] Step 302: Change the rock explosive consumption multiple times, and after each change of the rock explosive consumption, perform "simulated blasting based on the rock explosive consumption to obtain a rock damage model after mine blasting" to determine the rock explosive consumption that maximizes the sum of the crushing zone area and the fracture zone area as the optimal simulated explosive consumption.

[0058] Step 303: Controlling the amount of explosives according to the optimal simulated explosive consumption at the blasting site, and conducting a blasting test.

[0059] Step 304: After the explosion test, rock images are collected and input into the ChatGPT-Matlab framework to identify rock fragmentation.

[0060] Step 305: If the rock fragmentation is greater than the fragmentation threshold, the optimal simulated explosive consumption is adjusted, and the process returns to the step of "controlling the amount of explosives according to the optimal simulated explosive consumption at the blasting site and conducting a blasting test".

[0061] The rock size is usually required to be within 1m.

[0062] Step 306: If the rock fragmentation is less than or equal to the fragmentation threshold, the current optimal simulated explosive consumption is determined as the blasting analysis result of the optimized explosive consumption.

[0063] The following is the ChatGPT response, Matlab analysis and comparison with the conventional method.

[0064] The numerical simulation of rock damage analysis using the ChatGPT-Matlab framework is similar to the analysis of rock damage using ImageJ, which uses the different grayscale values ​​of red, green and blue for identification. The ChatGPT-Matlab framework can directly give the number of pixels of each color, and ImageJ can give a grayscale value map of pixels of different colors. By integrating the broken lines representing different colors in the pixel grayscale value map, the areas of different colors can be obtained. After calculation, the areas of red, green and blue are 11327, 4309 and 28339 pixels respectively. The original rock damage analysis image, ImageJ and ChatGPT-Matlab framework analysis results are shown in the figure below. Figure 4 , Figure 5 and Figure 6 shown. Figure 5 Part (a) shows the ImageJ analysis result with a red area of ​​11327 pixels. Figure 5 Part (b) shows the ImageJ analysis result with a green area of ​​4309 pixels. Figure 5 Part (c) shows the ImageJ analysis result with a blue area of ​​28,339 pixels. Figure 6 Part (a) shows the analysis result of ChatGPT-Matlab framework with a red area of ​​12013 pixels. Figure 6 Part (b) shows the analysis result of ChatGPT-Matlab framework with a green area of ​​3117 pixels. Figure 6Part (c) shows the analysis results of the ChatGPT-Matlab framework with a blue area of ​​26483 pixels. ImageJ's recognition algorithm is more effective in recognizing blue-green areas, while ChatGPT-Matlab framework is more effective in recognizing red areas. However, the results given by the ChatGPT-Matlab framework and ImageJ are similar, which proves that the ChatGPT-Matlab framework and ImageJ are similar in effectiveness in rock damage analysis. ImageJ is a free, small-volume image processing software with rich functions, but the ChatGPT-Matlab framework has a lower learning cost. You can write the code with ChatGPT and import the image without complicated processing.

[0065] The ChatGPT-Matlab framework also has great application potential in the analysis of large blocks after blasting tests. Figure 7 Shows the Wipfrag analysis burst, Figure 8 The ChatGPT-Matlab framework is used to analyze the blast pile. As a large block analysis software commonly used at blasting sites, after several iterations and code optimization, it has demonstrated a strong ability to identify rock contours. The analysis data shows that all rocks are within 700 mm in diameter, 94.12% of the rocks are within 600 mm in diameter, 81.45% of the rocks are within 500 mm in diameter, and 72.91% of the rocks are within 400 mm in diameter. The ChatGPT-Matlab framework analysis results show that all rocks are within 0.2 m in diameter, 11 rocks are between 0.1 m and 0.2 m, and 481 rocks are within 0.1 m, which is very different from the analysis results of Wipfrag. Figure 7 The horizontal and vertical axes of the graphs shown are Figure 3 The same block analysis diagram is shown in .

[0066] Comparing the recognition images together, we found that the reason for the large difference in results may be that the ChatGPT-Matlab framework has poor recognition of the areas with weak light between rocks, and other protrusions on the rock surface may affect the recognition effect. These two items require professional parameter adjustment. Figure 8 As shown in the figure, although the protrusions on the rock surface interfere with the recognition, the ChatGPT-Matlab framework is more sensitive to contour extraction and has broad application prospects. Figure 8 The rightmost horizontal axis represents the rock diameter (meters), and the vertical axis represents the frequency (the number of rocks).

[0067] Using the ChatGPT-Matlab framework for rock damage analysis and post-blasting bulk analysis can further reduce the learning costs of enterprise employees and reduce the enterprise data analysis costs. After further iteration of the code, it can achieve similar effects to professional software, and has great application potential. Moreover, based on accurate rock damage data and bulk analysis data, it can more accurately adjust the unit consumption of explosives, reduce the use of explosives, improve explosives efficiency, and promote the development of green blasting.

[0068] (2) Blasting vibration analysis

[0069] In recent years, with the continuous development of technical means, domestic and foreign scholars have conducted FFT (Fast Fourier Transformation) and HHT (Hilben-Huang) analysis on blasting vibration to observe the time-frequency distribution of blasting vibration energy. Due to the limitation of manpower at the blasting site and the lack of analysis experience, few mines conduct time-frequency analysis on vibration signals after blasting, and these vibration signal time-frequency analysis methods are still limited to cutting-edge research. At present, the level of blasting vibration data analysis on site is still at the level of observing the peak velocity of particle vibration (PPV) in the triaxial vibration signal. The safety range of PPV in the "Safety Regulations for Blasting GB6722-2014" is shown in Table 1.

[0070] Table 1 PPV safety range

[0071]

[0072]

[0073] Blasting vibration monitoring should simultaneously measure the three mutually perpendicular components of particle vibration. The particle vibration velocity in Table 1 is the maximum value of the three-way component, and the vibration frequency is the main vibration frequency. The frequency range is determined based on the waveform measured on site or selected according to the following data: masonry chamber blasting f is less than 20Hz, open-air deep hole blasting f is between 10Hz and 60Hz, open-air shallow hole blasting f is between 40Hz and 100Hz, underground deep hole blasting f is between 30Hz and 100Hz, and underground shallow hole blasting f is between 60Hz and 300Hz.

[0074] At the blasting site, only the vibration signals collected by the sensor are used to observe the peak vibration velocity of the particle, and no vibration signal processing methods such as denoising, signal decomposition, and signal transformation are used for in-depth analysis. Fig. 9 As shown, Fig. 9 The horizontal axis in the lower right corner is time, in ms, and the vertical axis is vibration velocity, in cm / s. Fig. 9The coordinate value at the bottom of the vertical and horizontal coordinates is -54.50, and the two coordinate values ​​between 54.50 and 0.00 are 0.00 and 54.50 from bottom to top. When measuring vibration at the blasting site, first design the vibration measurement position according to the blasting working face. The vibration position is mainly selected near the blasting area and in the buildings around the blasting area. Then the vibration measurement personnel check whether the battery of the blasting vibration equipment is fully charged. On the basis of confirming that the equipment can monitor the blasting vibration normally, dig a pit at the pre-designed measuring point, and inject gypsum and water into the pit to mix, so that the vibration measurement equipment and the ground form a rigid body to ensure the monitoring effect. After the blasting, the blasting vibration equipment is recovered, the equipment is connected to the computer with a data cable, and the blasting vibration data is imported on the ChatGPT-Matlab framework to observe the peak vibration velocity of the three-axis particle. Finally, by looking up the table, observe whether the current blasting meets the regulations. If it exceeds the prescribed allowable range, adjust the maximum charge of the single hole for the next blasting to reduce the blasting vibration.

[0075] In an exemplary embodiment, the ChatGPT-Matlab framework adopts the HHT method to perform blasting vibration analysis, and the above step 102 may be: programming the Hilbert-Huang transform method in the ChatGPT-Matlab framework.

[0076] HHT method is the abbreviation of Hilbert-Huang Transform method. HHT method mainly includes two steps:

[0077] (1) EMD decomposition: decompose the signal x(t) into several IMF components i (t);

[0078] (2) Hilbert transform: Apply the Hilbert transform to each IMF component to obtain its instantaneous frequency and amplitude.

[0079] Huang believes that after vibration occurs, there are some IMF components with non-sinusoidal function characteristics, which together constitute any complex signal including the impact signal, and these IMF components are usually relatively simple. These IMF components can be extracted through EMD decomposition, where the EMD decomposition uses the residual component obtained by the final decomposition as the end mark, and its algorithm expression is:

[0080]

[0081] In the formula, c1(t), c2(t), c3(t),…,c n (t) represents the components of the original signal in different frequency bands, which together constitute the original vibration signal, and their frequencies show a downward trend. The frequency band range of these small frequency bands is not fixed, which proves that EMD decomposition has adaptive characteristics; r n(t) is the final residual obtained by calculation; n is the number of final cycle screening of the formula.

[0082] Perform Hilbert transform on each IMF component obtained by decomposition to obtain the instantaneous frequency corresponding to each component. After integrating all instantaneous spectra, the Hilbert spectrum can be obtained. The calculation steps are as follows: Perform Hilbert transform on all components, the formula is:

[0083]

[0084] This formula represents the convolution operation of the function c(t') and 1 / (πt), where c(t') represents a real-valued function; t' is the real-valued function variable; PV is the Cauchy principal value, and the analytical signal z(t) is constructed based on this: z(t) = c(t) + j·H[c(t)] = a(t)·e j·Φ(t) , where j is the analytical constant; a(t) is the amplitude function; Φ(t) is the phase function, and the formulas are:

[0085]

[0086] In the formula, c(t) is the IMF component of the signal x(t). On this basis, the angular instantaneous frequency ω(t) and the instantaneous frequency f(t) are obtained as follows:

[0087]

[0088] Where Φ(t) is the phase function. After Hilbert transform, the original signal expression is:

[0089]

[0090] In the formula, R e represents the real part of the residual component r; a i (t) and w i (t) are all constants; i is the IMF component number. The spectrum showing the amplitude in time-frequency is called the Hilbert spectrum, and the expression is:

[0091]

[0092] In the formula, ω represents a certain frequency. By integrating the Hilbert spectrum over time t, we can get the Hil-ber t marginal spectrum:

[0093]

[0094] The Hilbert instantaneous energy spectrum is an additional result of the Hilbert spectrum and can be defined as:

[0095]

[0096] Let ChatGPT learn the above process and write HHT analysis code through human-computer interaction. The code mainly includes the following steps:

[0097] ① Read the blasting vibration data, extract the time and vibration data in the three-axis direction, and synthesize a vibration signal, such as Fig.10 shown.

[0098] ②Perform EMD decomposition and draw the amplitude-frequency spectrum of each IMF component. ChatGPT-Matlab framework performs EMD decomposition and Matlab reports errors such as Fig.11 Adjust the error as shown. Fig.12 As shown in the figure. IMF1-4 time amplitude and time frequency diagram Fig.13 shown.

[0099] ③ Perform Hilbert transform on the effective IMF components and draw 3DHilbert spectrum. ChatGPT-Matlab framework is as follows Fig.14 As shown, HHT analysis and output are as follows Fig.15 shown.

[0100] ④ The Hilbert spectrum is integrated over time t to obtain the marginal spectrum, and the energy ratios of different frequency bands are calculated and visualized. The ChatGPT-Matlab framework plots the marginal spectrum and calculates the energy ratios of the frequency bands. Fig.16 As shown, the marginal spectrum and band energy ratio is as follows Fig.17 As shown, Fig.17 Part (a) of the graph shows the marginal spectrum, Fig.17 Part (b) in FIG. 1 shows the frequency band energy ratio. Fig.17 In part (b), the energy ratios of each column from left to right are 37.13%, 11.11%, 10.37%, 7.45%, 6.67%, 8.60%, 7.45%, 4.77%, 1.83%, 1.30%, 1.06%, 0.75%, 0.73%, 0.61%, 0.12%, 0.04%, 0.01%, 0.00%, 0.00%, and 0.00%.

[0101] Based on the above, in another exemplary embodiment, the Hilbert-Huang transform method is written in the ChatGPT-Matlab framework, which may specifically include the following steps 401 to 405:

[0102] Step 401: extracting the time and vibration data of the blasting vibration data in three axes and synthesizing a vibration signal.

[0103] Step 402: Decompose the vibration signal into a plurality of IMF components.

[0104] Step 403: Perform Hilbert transform on each IMF component to obtain instantaneous frequency and amplitude, and draw a 3D Hilbert spectrum.

[0105] Step 404: Integrate the time using the 3D Hilbert spectrum to obtain a marginal spectrum.

[0106] Step 405: Calculate and visualize energy ratios of different frequency bands according to the marginal spectrum.

[0107] In another exemplary embodiment, the above step 103 may be: importing blasting vibration data measured during mine blasting on the ChatGPT-Matlab framework, and outputting the time-frequency distribution of blasting vibration energy as a blasting analysis result of blasting vibration analysis.

[0108] The HHT analysis conducted by the ChatGPT-Matlab framework fills the gap in blasting vibration analysis at engineering sites, reduces the cost of on-site blasting vibration analysis, and improves the efficiency of vibration analysis. Based on the vibration time-frequency energy data, it can optimize the delay time and the single-hole charge amount. Through optimization, it is possible to control blasting vibration, reduce the impact of vibration on the side walls, buildings and engineering equipment, and promote safe and green blasting.

[0109] (3) Fly dust monitoring

[0110] The generation and diffusion of blasting dust mainly consists of three stages, namely the initial stage, the rising diffusion stage, and the settling stage. In the initial stage, the force of the shock wave generated by the explosion of explosives on the dust is much greater than the combined force of the dust's own gravity, the friction between the dust and its attachments, and the adsorption force, causing the dust to quickly rise into the air. The rock layer rises rapidly under the action of the explosion energy. As the shock wave is transmitted to the surface, the surface dust and the rock powder filled in the blast hole will first rise into the air, forming a tall dust column. Then, under the action of the explosion shock wave and the high-temperature and high-pressure blast gas, the rock breaks, and a small amount of dust rushes out of the cracks with the blast gas. Subsequently, under the further action of the shock wave, the rock continues to break, and the blast gas expands rapidly, carrying a large amount of dust and ejecting it. During the collapse of the blast pile, the dust between the rock cracks diffuses along the free surface direction with the displacement of the blast pile.

[0111] In the rising and diffusion stage, the dust will continue to rise under the action of inertia after leaving the surface. At the same time, the explosive gas has just been ejected from the cracks, and the energy has not been fully released, so it will continue to expand. Under the combined effects of its own gravity, air buoyancy, air resistance and the explosive gas, the force magnitude and direction of each dust particle are different, so their movement trajectories are also different. The movement trajectories of a large number of dust particles are superimposed to form a mushroom-shaped dust cloud.

[0112] In the settling stage, the energy of the dust has been exhausted, and its movement is mainly affected by gravity and airflow. The dust is completely diffused and merged with the airflow. As the airflow acts, the distance between the dust particles gradually increases. The length of the diffusion time mainly depends on the strength of the airflow and the gravity of the dust. The stronger the airflow, the longer the diffusion time; heavier dust will settle faster, while lighter dust will diffuse for a long time with the airflow. In addition, the dust cloud formed by the collapse of the explosive pile on the air surface will also diffuse forward with the explosive gas.

[0113] The concentration of blasting dust is usually monitored by the filter membrane sampling method, such as Fig.18 As shown. This method uses dust instruments to monitor blasting dust. First, the wind direction and wind speed must be determined, and then the mining explosion-proof dust sampler is placed downwind. Before starting the machine, a clean filter membrane is placed in the dust collection device. 3 minutes before the blasting, the vacuum pump in the dust sampler will suck the airflow containing blasting dust into the dust sampler, and the dust will adhere to the filter membrane. After the blasting, the dust on the filter membrane can be removed for the detection of dust components.

[0114] This method is widely used in the measurement of blasting fly ash, but it has some limitations. Its sampling range is limited, it is difficult to fully cover the blasting site, it is not real-time enough, it is difficult to reflect instantaneous changes, and environmental interference, complex dust types and high maintenance requirements may affect the accuracy of the results. In addition, there are safety risks for personnel operating at the blasting site, data processing is complex, and equipment and maintenance costs are high, which increases the overall monitoring cost.

[0115] In recent years, with the development of image analysis technology, blasting dust analysis can be more accurate and digital. However, image processing technology involves a series of steps such as image denoising, image enhancement, and grayscale value selection. The learning cost is high and the analyst's coding ability is very high. However, with the development of ChatGPT, users can use ChatGPT's powerful code compilation capabilities to perform a series of tasks. Fig.19 As shown, a drone is used at a fixed position to conduct panoramic monitoring of the explosion area.

[0116] Through dust measurement and image gray value recognition, the relationship between the gray value of the test area and the dust concentration follows this formula:

[0117] C=-0.01157G 2 +2.342G-104.1.

[0118] Where: C is the dust concentration; G is the gray value.

[0119] This application tried this formula, but the recognition effect was not ideal due to the shooting angle, shooting distance and recording equipment. Due to the lack of dust measurement equipment, it is impossible to establish a formula for grayscale value and dust concentration, but according to Yu's research results, it is assumed that grayscale value and dust concentration conform to a linear model and a linear model is used instead. The video processing flow of using blasting video to obtain fly ash concentration is as follows Fig. 20 As shown, Fig. 20 The horizontal axis in the lower left corner represents the frame, and the coordinate values ​​from left to right are 1, 200, 400, 600, 800, 1000, 1200, 1400, and the vertical axis represents the dust concentration, and the coordinate values ​​from bottom to top are 12, 12.5, 13, 13.5, 14, 14.5, 15, 15.5, 16, 16.5. The ChatGPT-Matlab framework code implementation process is as follows:

[0120] First, the background noise of each frame of the video is corrected, binarized, the grayscale value of the dust area is extracted, and the grayscale value is set to a reasonable range. The ChatGPT-Matlab framework performs binarization and grayscale value extraction and processing on the video. Fig.21 As shown, the white points in the Binaryframe in the figure are points within the gray value range.

[0121] Next, perform differentiation and draw the dust concentration curve. The ChatGPT-Matlab framework performs differentiation on each frame. Fig. 22 As shown in the figure, the white point in the Differenceframe is the point where this frame changes from the previous frame. Fig.23 In the dust concentration curve shown, it is obvious that there may be a reasonable gray value interval or noise before detonation, which interferes with the dust concentration identification. During this blasting process, the dust concentration peaked at the 372nd frame and then fell back quickly. Fig.23 The horizontal axis represents the frame, and the vertical axis represents the dust concentration.

[0122] Therefore, in an exemplary embodiment, when the link to be analyzed is flying dust monitoring, the above step 101 collects the measurement data of the link to be analyzed during the mine blasting process, specifically including: using a drone to perform panoramic monitoring of the blasting area, and obtaining a blasting video of the entire blasting process of the mine as the measurement data for flying dust monitoring.

[0123] In another exemplary embodiment, when the link to be analyzed is fly ash monitoring, the above step 103 of running the analysis method in the ChatGPT-Matlab framework can be replaced by the following steps 501 to 505:

[0124] Step 501: After denoising each frame of the blasting video, convert it into a grayscale image.

[0125] Step 502: Binarize the grayscale image to obtain a binary image of each frame in the blasting video.

[0126] Step 503: extracting the grayscale value of the dust area in each frame of the binary image, and calculating the dust concentration of each frame of the binary image based on a linear model between the grayscale value and the dust concentration.

[0127] Step 504: Differentiate each binary image frame from the previous binary image frame to obtain a differential frame image.

[0128] Step 505: Draw a dust concentration curve according to the dust concentration of each frame of the binary image.

[0129] In another exemplary embodiment, the linear model between the gray value and the dust concentration is:

[0130] C=0.1G.

[0131] Where C is the dust concentration and G is the gray value.

[0132] The analysis of fly dust through the above human-computer interaction process shows that the ChatGPT-Matlab framework can effectively monitor blasting fly dust through grayscale values, filling the gap that mine fly dust is difficult to monitor and providing a low-cost and easy-to-use fly dust monitoring method.

[0133] In another exemplary embodiment of the present application, an improved prompter is designed using the ChatGPT-Matlab framework, which logically prompts ChatGPT to derive programming code in Matlab, and the improved prompter repeatedly checks the results until an acceptable result is obtained. Figure 2 As shown, after the above step 102, the method may further include steps 201 to 205.

[0134] Step 201: Run the programmed analysis method according to the measurement data.

[0135] Step 202: If the written code of the analysis method cannot be run, the code vulnerability is repaired and the process returns to the step of "running the written analysis method according to the measurement data".

[0136] Step 203: If the written code of the analysis method is executable, determine whether the running result is within a preset deviation range, and obtain a determination result.

[0137] Step 204: When the judgment result is yes, adjust the parameters in the analysis method, and return to the step of "running the written analysis method according to the measurement data" until the accuracy of the running result is equal to the preset accuracy.

[0138] Step 205: When the judgment result is no, perform error analysis, adjust the analysis method or adjust the code of the analysis method, and return to the step of "running the written analysis method according to the measurement data".

[0139] Figure 2 It shows the application idea of ​​ChatGPT-Matlab framework. Figure 2 "Designing problems for generative AI based on analysis needs" means selecting a method from a variety of analysis methods for a link to be analyzed and writing code. For example, in the blasting vibration analysis link, the HHT method is selected to analyze the vibration signal, so ChatGPT is asked to write an HHT analysis code. Figure 2 The "acceptable results" in the analysis is explained as follows: For example, in the analysis of numerical simulation results, imageJ is used as a benchmark to compare the number of red, green and blue pigment points analyzed by imageJ and ChatGPT codes. If the number and proportion of pixels of the three colors are similar, it is acceptable. Otherwise, it is unacceptable, indicating that there is a problem with the code. Figure 2 The "adjustment of analysis parameters" in the code is explained as follows: for example, in the case of color pigment point recognition, the color recognition sensitivity in the code is adjusted by 0-100, in the blasting vibration analysis code, the error is reduced by setting different numbers of IMF components, and in the dust video analysis link, the parameters are adjusted by changing the recognition area and delineating the dust grayscale. Figure 2 The "error analysis" in the article is explained as follows: This error analysis refers to two situations: one is Matlab error analysis, and the other is unacceptable result adjustment. For the first case, the personnel can see the problem based on the Matlab error. For example, if the code cannot run normally, the code bug needs to be fixed according to the specific error reason. If the code runs normally but the result is unacceptable, it is up to the researchers to determine whether it is necessary to change an analysis method or adjust the code parameters and sequence.

[0140] This application discusses the development dilemma of green blasting technology and proposes a new application idea based on the ChatGPT-Matlab framework, aiming to solve the technical barriers and high cost problems currently faced by the green blasting field. Through research, the following main conclusions were drawn:

[0141] 1. Optimization of explosive consumption: Using the ChatGPT-Matlab framework, we can optimize the explosive consumption by combining numerical simulation with large-scale field data. This framework can accurately analyze the rock damage after blasting, provide the best amount of explosives, reduce the amount of explosives used, and improve blasting efficiency.

[0142] 2. Blasting vibration analysis: This framework can conduct in-depth analysis of blasting vibration signals and apply the Hilbert-Huang transform (HHT) method to decompose and identify key components in the vibration signal. This method can effectively control blasting vibration, reduce the impact on the surrounding environment and buildings, and improve safety.

[0143] 3. Fly dust monitoring: Through image processing technology, the ChatGPT-Matlab framework can accurately monitor and analyze the fly dust generated during the blasting process. This method overcomes the limitations of traditional fly dust monitoring methods and provides a low-cost, high-precision fly dust concentration monitoring solution, which helps reduce environmental pollution.

[0144] In summary, the application of ChatGPT-Matlab framework in green blasting technology has shown significant advantages, which can effectively reduce enterprise costs, improve blasting efficiency and safety, and promote the development of green mining. In the future, with the continuous advancement and optimization of technology, the application prospects of this framework in blasting engineering are broad and have important research and practical value.

[0145] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0146] The present application uses specific examples to illustrate the principles and implementation methods of the present application. The above examples are only used to help understand the method and core ideas of the present application. At the same time, for those skilled in the art, according to the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present application.

Claims

1. A mine blasting analysis method, characterized in that: include: Collect measurement data for analysis during mining blasting; Use the ChatGPT-Matlab framework to write the analysis method used in the link to be analyzed; According to the measurement data, the analysis method in the ChatGPT-Matlab framework is run to obtain the blasting analysis results of the link to be analyzed.

2. The mine blasting analysis method according to claim 1, characterized in that: The analysis method used in the analysis link is written using the ChatGPT-Matlab framework, which also includes: running the prepared analysis method according to the measurement data; If the written code of the analysis method cannot be run, the code vulnerability is repaired and the process returns to step "running the written analysis method according to the measurement data"; If the written code of the analysis method is executable, then determining whether the running result is within a preset deviation range, and obtaining a determination result; When the judgment result is yes, adjust the parameters in the analysis method, and return to the step of "running the written analysis method according to the measurement data" until the accuracy of the running result is equal to the preset accuracy; When the judgment result is no, an error analysis is performed, the analysis method is adjusted or the code written for the analysis method is adjusted, and the process returns to the step of "running the written analysis method according to the measurement data".

3. The mine blasting analysis method according to claim 1, characterized in that: When the link to be analyzed is the optimization of explosive unit consumption, the measurement data of the link to be analyzed during the mine blasting process is collected, including: Before blasting in the mine, use Ansys / Ls-dyna to model the 3D blasting area; Set the unit consumption of rock explosives; Based on the rock explosive unit consumption simulation blasting, a rock damage model after mine blasting is obtained as measurement data for optimizing explosive unit consumption.

4. The mine blasting analysis method according to claim 3, characterized in that: When the link to be analyzed is the optimization of explosive unit consumption, the analysis method in the ChatGPT-Matlab framework is run according to the measurement data to obtain the blasting analysis results of the link to be analyzed, which specifically include: The rock damage model after mine blasting was imported into the ChatGPT-Matlab framework for RGB damage analysis to obtain the crushed area, crack area, and undamaged area. Changing the rock explosive unit consumption multiple times, and after each change of the rock explosive unit consumption, executing "simulating blasting based on the rock explosive unit consumption to obtain a rock damage model after mine blasting", determining the rock explosive unit consumption that maximizes the sum of the crushing area and the fracture area as the optimal simulated explosive unit consumption; At the blasting site, the amount of explosives is controlled according to the optimal simulated explosive unit consumption, and a blasting test is performed; After the blast test, rock images were collected and input into the ChatGPT-Matlab framework to identify rock fragmentation. If the rock fragmentation is greater than the fragmentation threshold, the optimal simulated explosive consumption is adjusted, and the process returns to step "controlling the amount of explosives according to the optimal simulated explosive consumption at the blasting site and conducting a blasting test"; If the rock fragmentation is less than or equal to the fragmentation threshold, the current optimal simulated explosive consumption is determined as the blasting analysis result of the optimized explosive consumption.

5. The mine blasting analysis method according to claim 1, characterized in that: When the link to be analyzed is blasting vibration analysis, the ChatGPT-Matlab framework is used to write the analysis method used in the link to be analyzed, including: Write the Hilbert-Huang transform method in the ChatGPT-Matlab framework.

6. The mine blasting analysis method according to claim 5, characterized in that: Write the Hilbert-Huang transform method in the ChatGPT-Matlab framework, including: Extract the time and vibration data of blasting vibration data in three axes and synthesize a vibration signal; Decomposing the vibration signal into a plurality of IMF components; Perform Hilbert transform on each IMF component to obtain instantaneous frequency and amplitude, and plot the 3D Hilbert spectrum; The marginal spectrum is obtained by integrating the 3D Hilbert spectrum over time; From the marginal spectra, the energy ratios of different frequency bands are calculated and visualized.

7. The mine blasting analysis method according to claim 5, characterized in that: When the link to be analyzed is blasting vibration analysis, the analysis method in the ChatGPT-Matlab framework is run according to the measurement data to obtain the blasting analysis results of the link to be analyzed, which specifically include: The blasting vibration data measured during the mine blasting process were imported on the ChatGPT-Matlab framework, and the time-frequency distribution of the blasting vibration energy was output as the blasting analysis result of the blasting vibration analysis.

8. The mine blasting analysis method according to claim 1, characterized in that: When the link to be analyzed is fly dust monitoring, the measurement data of the link to be analyzed during the mine blasting process is collected, including: Use drones to conduct panoramic monitoring of the blasting area and obtain blasting videos of the entire blasting process in the mine as measurement data for fly dust monitoring.

9. The mine blasting analysis method according to claim 8, characterized in that: When the link to be analyzed is fly ash monitoring, the analysis method in the ChatGPT-Matlab framework is run according to the measurement data to obtain the blasting analysis results of the link to be analyzed, specifically including: After denoising each frame of the explosion video, convert it into a grayscale image; Binarize the grayscale image to obtain a binary image of each frame in the blasting video; Extract the gray value of the dust area in each frame of the binary image, and calculate the dust concentration of each frame of the binary image based on the linear model between the gray value and the dust concentration; Difference each binary image frame from the previous binary image frame to obtain a differential frame image; According to the dust concentration of each frame of binary image, a dust concentration curve is drawn.

10. The mine blasting analysis method according to claim 9, characterized in that: The linear model between gray value and dust concentration is: C = 0.1G; Where C is the dust concentration and G is the gray value.

Images