A method for accurately and non-destructively identifying the true concentration of a filling

Abstract

The application discloses a kind of filling body real concentration accurate nondestructive identification method, belong to filling technical field, solve the problem that filling body real concentration after a certain number of days of maintenance in goaf cannot be accurately determined, the method includes: according to the proportioning scheme, tailings, cementing material and water are measured, test block is maintained and handled, test block is scanned by hyperspectral imager, the spectrum curve of test block surface is extracted, modeling data set is collected, modeling data set is divided into training set and test set, concentration prediction model is iteratively trained, filling body surface spectrum curve is loaded, concentration prediction model is executed, and filling body concentration is output;The application uses hyperspectral imaging technology to identify the real concentration of filling body, without damaging the sample, with the advantage of nondestructive testing, after the establishment of filling body concentration identification model, the corresponding concentration can be obtained after scanning the unknown concentration test block, which greatly shortens the experimental detection period and is efficient.

Description

Technical Field

[0001] This invention belongs to the field of filling technology, specifically relating to a method for accurate and non-destructive identification of the true concentration of filling materials. Background Technology

[0002] Cemented tailings backfill is a technology that involves uniformly mixing solid waste, cementing materials, and water to form a backfill slurry, which is then transported to the goaf to control ground pressure. This technology addresses two major hazards: solid waste disposal and goaf collapse, and represents a major development trend in the mining industry.

[0003] The mass concentration of backfill material is a key factor affecting its performance. Higher concentration generally results in higher strength and a denser microstructure. During preparation, the concentration of backfill material can be accurately controlled and measured. However, once the backfill material is transported to the goaf, factors such as groundwater inflow or segregation often lead to differences between the actual and expected concentrations. This deviation is one of the key reasons why backfill material performance deviates from design standards. For cured backfill material, distinguishing or identifying the concentration through visual inspection or existing equipment is extremely difficult. For example, in mining operations, backfill material quality is typically monitored using core sampling, followed by uniaxial compressive strength testing to assess its strength. However, the concentration of the backfill material remains unknown, and it cannot be determined whether the strength difference is due to concentration variations. It is well known that measuring the concentration of fresh backfill slurry is easy, but the true concentration of backfill material cured in the goaf for a certain number of days cannot be accurately determined. Therefore, rapid and non-destructive identification of the actual concentration of the filling material is crucial for accurately controlling filling parameters and achieving intelligent filling. To address the above issues, we propose a method for accurate and non-destructive identification of the true concentration of the filling material. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by providing a method for accurate and non-destructive identification of the true concentration of backfill material, thus solving the problem that the true concentration of backfill material after a certain number of days of curing in goaf areas cannot be accurately measured.

[0005] In existing technologies, it is easy to measure the concentration of fresh backfill slurry, but the true concentration of backfill material after curing for a certain number of days in goaf areas cannot be accurately determined. To address this issue, we propose a precise and non-destructive method for identifying the true concentration of backfill material. The method involves first measuring tailings, cementing material, and water according to a mixing ratio, stirring until homogeneous to obtain a test block, and then taking out the test block after curing for the specified period. A hyperspectral scanning experiment is performed on the test block using a hyperspectral imager to extract the surface spectral curve. Simultaneously, using spectral reflectance as input and test block concentration as output, an algorithm is trained iteratively to predict the concentration, outputting a converged concentration prediction model. Then, backfill material with unknown concentration is obtained, and a hyperspectral scanning experiment is performed on the backfill material using a hyperspectral imager to extract the surface spectral curve. Finally, using the surface spectral curve of the backfill material as input, the concentration prediction model is executed to output the backfill concentration. This invention utilizes hyperspectral imaging technology to identify the true concentration of filling materials without damaging the sample, thus possessing the advantage of non-destructive testing. After the filling material concentration identification model is established, the corresponding concentration can be obtained by scanning a test block with an unknown concentration, greatly shortening the experimental testing cycle and demonstrating high efficiency. Furthermore, this invention has good universality and scalability, and is applicable to all mine filling systems.

[0006] The present invention is implemented as follows: a method for accurate and non-destructive identification of the true concentration of filling material, S10, according to the proportioning scheme, tailings, cementing material and water are measured, stirred until uniform, and a test block is obtained, and the test block is cured.

[0007] S20, after the curing period, the test block is taken out and a hyperspectral scanning experiment is performed on the test block using a hyperspectral imager. The spectral data is calibrated and preprocessed for reflectance, and the region of interest is further divided to extract the surface spectral curve of the test block.

[0008] S30: Collect the modeling dataset, divide the modeling dataset into training set and test set, use spectral reflectance as input and sample block concentration as output to train the algorithm, iteratively train the concentration prediction model, and output a converged concentration prediction model.

[0009] S40, Obtain infill bodies of unknown concentration, perform hyperspectral scanning experiments on the infill bodies using a hyperspectral imager, perform reflectance calibration and preprocessing on the spectral data, further divide the region of interest, and extract the spectral curves of the infill body surface;

[0010] S50, load the spectral curve of the filling body surface, use the spectral curve of the filling body surface as input, execute the concentration prediction model, and output the filling body concentration.

[0011] The variables in the mixing scheme include concentration, sand-cement ratio, water-cement ratio, and binder ratio.

[0012] The hyperspectral imager is not less than 10 -1λ, the number of channels is no less than 100.

[0013] In step S20, the curing period of the test block shall not be less than 28 days, the number of scans of a single test block shall not be less than 5 times during the curing period, and each test block shall scan 2-4 faces. Each test block surface shall be divided into no less than 100 regions of interest. The reflectance of each region of interest shall be obtained from the average reflectance value of no less than 20 pixels. The spectral curve of each filling body shall be the average reflectance of all regions of interest.

[0014] In step S30, the test set accounts for no less than 20% of the total dataset, and the selected concentration prediction model has an accuracy of no less than 85% on the test set.

[0015] In step S40, the filling material of unknown concentration includes laboratory-prepared and field-collected core samples. If the filling material is wetted with water, it is air-dried for 2-6 hours before being subjected to hyperspectral scanning.

[0016] Compared with the prior art, the embodiments of this application have the following main advantages:

[0017] This invention utilizes hyperspectral imaging technology to identify the true concentration of filling materials without damaging the sample, thus possessing the advantage of non-destructive testing. After the filling material concentration identification model is established, the corresponding concentration can be obtained by scanning a test block with an unknown concentration, greatly shortening the experimental testing cycle and demonstrating high efficiency. Furthermore, this invention has good universality and scalability, and is applicable to all mine filling systems. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the implementation process of the method for accurate and non-destructive identification of the true concentration of filling materials provided by the present invention.

[0019] Figure 2 The diagram shows the results of hyperparameter optimization in the concentration prediction model. Detailed Implementation

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0021] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0022] In existing technologies, it is easy to measure the concentration of fresh backfill slurry, but the true concentration of backfill material after curing for a certain number of days in goaf areas cannot be accurately determined. To address this issue, we propose a precise and non-destructive method for identifying the true concentration of backfill material. The method involves first measuring tailings, cementing material, and water according to a mixing ratio, stirring until homogeneous to obtain a test block, and then taking out the test block after curing for the specified period. A hyperspectral scanning experiment is performed on the test block using a hyperspectral imager to extract the surface spectral curve. Simultaneously, using spectral reflectance as input and test block concentration as output, an algorithm is trained iteratively to predict the concentration, outputting a converged concentration prediction model. Then, backfill material with unknown concentration is obtained, and a hyperspectral scanning experiment is performed on the backfill material using a hyperspectral imager to extract the surface spectral curve. Finally, using the surface spectral curve of the backfill material as input, the concentration prediction model is executed to output the backfill concentration. This invention utilizes hyperspectral imaging technology to identify the true concentration of filling materials without damaging the sample, thus possessing the advantage of non-destructive testing. After the filling material concentration identification model is established, the corresponding concentration can be obtained by scanning a test block with an unknown concentration, greatly shortening the experimental testing cycle and demonstrating high efficiency. Furthermore, this invention has good universality and scalability, and is applicable to all mine filling systems.

[0023] This invention provides a method for accurate and non-destructive identification of the true concentration of filling materials. Figure 1 A schematic diagram illustrating the implementation process of a method for accurate and non-destructive identification of the true concentration of filling materials is shown. The method specifically includes:

[0024] S10, measure tailings, cementing material and water according to the proportioning scheme, stir until uniform, obtain test blocks, and cure the test blocks.

[0025] It should be noted that the variables in the aforementioned proportioning scheme include concentration, sand-cement ratio, water-cement ratio, and binder ratio. The concentration range and gradient are adjusted according to the mining production. When preparing the test block proportions, the current industry standard "Standard for Basic Performance Test Methods of Building Mortar" JGJ / T 70 is generally referenced.

[0026] S20, after the curing period, the test block is taken out and a hyperspectral scanning experiment is performed on the test block using a hyperspectral imager. The spectral data is calibrated and preprocessed for reflectance, and the region of interest is further divided to extract the surface spectral curve of the test block.

[0027] In this embodiment, the hyperspectral imager is not less than 10 -1 λ, the number of channels is not less than 100, the curing period of the test block is not less than 28 days, the number of scans of a single test block is not less than 5 during the curing period, and each test block scans 2-4 surfaces. Each test block surface is divided into not less than 100 regions of interest. The reflectance of each region of interest is obtained from the average reflectance value of not less than 20 pixels. The spectral curve of each filling body is the average reflectance of all regions of interest.

[0028] It should be noted that, using a domestic metal mine as a case study, the cement-to-tailings ratio was set at 1:4, with mass concentration gradients of 61%, 64%, 67%, 70%, and 73%. According to this mix design, the weighed tailings, cement, and water were poured into a mixing tank and stirred until uniformly mixed. The filling slurry was then poured into a 7.07×7.07×7.07cm mold. After 24 hours, the mold was removed, and the test blocks were placed in a standard curing chamber for curing. Spectroscopic scanning experiments were conducted at curing ages of 3, 7, 11, 14, 17, 21, 25, and 28 days.

[0029] The hyperspectral imager has a resolution of 7 nm and includes 256 bands. The entire instrument is housed in a sealed black box to prevent external light from affecting the experiment. Four halogen lamps are installed inside the black box as the light source. Specview software is used to control the spectrometer, including adjusting the camera's focus and exposure. Before testing, calibration is performed using a 100% reflectance white board, assuming zero absorption in all bands. To ensure the image is centered, the infill specimen is placed directly below the hyperspectral camera, and each specimen is scanned sequentially according to its number.

[0030] The scanned files were imported into Specview software for pixel-by-pixel reflectance calibration. Simultaneously, spectral smoothing was performed to reduce the impact of noise and spurious spectra. ENVI software was used to extract the spectral values ​​from the calibrated spectrum. Taking the entire surface of the infill as the research object, 25 pixel blocks (5×5) were selected as regions of interest. Each region of interest was treated as a whole, and the average spectrum was calculated as the spectral value for that region. To reduce error, approximately 200 regions of interest were extracted from each infill surface, and the average of these 200 spectra was used as the highlight curve for that block. To remove specific noise or redundant information and extract a cleaner spectral signal, local maxima were found in the data. These maxima were connected to form an upper envelope, and the reflectance value for each band was obtained by dividing the original data by the upper envelope.

[0031] S30: Collect the modeling dataset, divide the modeling dataset into training set and test set, use spectral reflectance as input and sample block concentration as output to train the algorithm, iteratively train the concentration prediction model, select the algorithm with the best accuracy and robustness as the concentration prediction model, and output the converged concentration prediction model.

[0032] It should be noted that the test set accounts for no less than 20% of the total dataset, and the selected concentration prediction model has an accuracy of no less than 85% on the test set.

[0033] S40, Obtain infill bodies of unknown concentration, perform hyperspectral scanning experiments on the infill bodies using a hyperspectral imager, perform reflectance calibration and preprocessing on the spectral data, further divide the region of interest, and extract the spectral curves of the infill body surface;

[0034] In this embodiment of the invention, the filling material of unknown concentration includes laboratory-prepared and field-collected core samples. If the filling material is wetted with water, it is air-dried for 2-6 hours before being subjected to hyperspectral scanning.

[0035] S50, load the spectral curve of the filling body surface, use the spectral curve of the filling body surface as input, execute the concentration prediction model, and output the filling body concentration.

[0036] This invention utilizes hyperspectral imaging technology to identify the true concentration of filling materials without damaging the sample, thus possessing the advantage of non-destructive testing. After the filling material concentration identification model is established, the corresponding concentration can be obtained by scanning a test block with an unknown concentration, greatly shortening the experimental testing cycle and demonstrating high efficiency. Furthermore, this invention has good universality and scalability, and is applicable to all mine filling systems.

[0037] It should be noted that during the training of the concentration prediction model, the modeling dataset was divided into a 70% training set and a 30% test set. Support Vector Machine (SVM) algorithm was used for modeling. In SVM, the adjustment of hyperparameters is crucial to model performance. The concentration prediction model hyperparameters include C and Gamma. In this embodiment, the value range of C is (0.1, 1.0, 10, 100, 200, 500), and the value range of Gamma is (0.001, 0.01, 0.1, 1, 10). The hyperparameters were optimized using a grid optimization algorithm combined with 5-fold cross-validation. Figure 2 The graph shows the hyperparameter optimization results for the concentration prediction model. The highest accuracy is achieved when hyperparameter C = 200 and Gamma = 10. The accuracy on the training set and the test set is 99.31% and 98.23%, respectively.

[0038] In summary, this invention provides a method for accurate and non-destructive identification of the true concentration of filling materials. The embodiments of this invention utilize hyperspectral imaging technology to identify the true concentration of filling materials without damaging the sample, thus possessing the advantages of non-destructive testing. After the filling material concentration identification model is established, the corresponding concentration can be obtained by scanning a test block with an unknown concentration, greatly shortening the experimental testing cycle and demonstrating high efficiency. Furthermore, this invention has good universality and scalability, and is applicable to all mine filling systems.

[0039] It should be noted that, for the sake of simplicity, the foregoing embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to the present invention. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0040] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on these embodiments, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can still combine, add, delete, or otherwise adjust the features of the various embodiments of the present invention according to the circumstances without conflict or creative effort, thereby obtaining different technical solutions that do not fundamentally depart from the concept of the present invention. These technical solutions also fall within the scope of protection of the present invention.

Claims

1. A method for accurate and non-destructive identification of the true concentration of filling materials, characterized in that, include: S10, measure tailings, cementing material and water according to the proportioning scheme, stir until uniform, obtain test blocks, and cure the test blocks. S20, after the curing period, the test blocks are taken out and a hyperspectral scanning experiment is performed on the test blocks using a hyperspectral imager. The reflectance of the spectral data is calibrated and preprocessed, and the regions of interest are further divided. The surface spectral curves of the test blocks are extracted. The curing period of the test blocks is not less than 28 days. During the curing period, the number of scans of a single test block is not less than 5 times, and 2-4 surfaces of each test block are scanned. The surface of each test block is divided into not less than 100 regions of interest. The reflectance of each region of interest is obtained from the average reflectance value of not less than 20 pixels. The spectral curve of each filling body is the average reflectance of all regions of interest. S30. Collect the modeling dataset, divide the modeling dataset into a training set and a test set, train the algorithm with spectral reflectance as input and sample concentration as output, iteratively train the concentration prediction model, and output a converged concentration prediction model. The test set accounts for no less than 20% of the total dataset, and the accuracy of the selected concentration prediction model on the test set is no less than 85%. S40, Obtain infill material of unknown concentration, perform hyperspectral scanning experiment on infill material using hyperspectral imager, perform reflectance calibration and preprocessing on spectral data, further divide region of interest, extract spectral curve of infill material surface, infill material of unknown concentration includes laboratory prepared and core sample blocks from the sampling site. If infill material is wetted by water, perform hyperspectral scanning after air drying for 2-6 hours. S50, load the spectral curve of the filling body surface, use the spectral curve of the filling body surface as input, execute the concentration prediction model, and output the filling body concentration.

2. The method for accurate and non-destructive identification of the true concentration of filling material as described in claim 1, characterized in that: The variables in the mixing scheme include concentration, sand-cement ratio, water-cement ratio, and binder ratio.

3. The method for accurate and non-destructive identification of the true concentration of filling materials as described in claim 2, characterized in that: The hyperspectral imager is no less than λ, the number of channels is no less than 100.

Images