Coal quality detection method and device based on near-infrared coupling TRLIBS and medium

Abstract

The invention provides a coal quality detection method, device and equipment based on near-infrared coupling TRLIBS and a medium. The method comprises the steps that the flow of a coal sample with the particle size not larger than 13 mm is detected, and TRLIBS and near-infrared spectrum detection on the coal sample is triggered; the method comprises the following steps: collecting a TRLIBS spectrum of a coal sample by using a spectrograph, collecting a near-infrared diffuse reflection spectrum of the coal sample by using a reflection probe, and carrying out equalization treatment to form TRLIBS and near-infrared diffuse reflection spectrum data; the TRLIBS spectral data are preprocessed, the near infrared diffuse reflection spectral data are processed, and characteristic peaks of hydrogen-containing groups of near infrared spectrums are obtained; carrying out feature level fusion on the preprocessed TRLIBS spectral data and near infrared spectral feature peaks by adopting dynamic weights, carrying out regression analysis on fusion features by utilizing partial least squares regression, and constructing a coal quality parameter prediction model; and detecting an actual coal sample by using the model to obtain a coal quality detection result so as to solve the problems of limited single spectrum accuracy, complex spectrum data fusion algorithm and weak anti-interference capability in the existing coal quality detection technology.

Description

Technical Field

[0001] This invention relates to the field of coal quality analysis technology, and in particular to a coal quality detection method, apparatus, equipment and medium based on near-infrared coupled TRLIBS. Background Technology

[0002] Accurate detection of coal quality parameters is a core aspect of efficient coal utilization and pollution control. Among existing detection technologies, time-resolved laser-induced breakdown spectroscopy (TRLIBS) can rapidly analyze elemental composition, but its quantitative accuracy is limited by matrix effects and spectral fluctuations. Near-infrared spectroscopy (NIRS) can reflect the molecular structure characteristics of organic matter, but its sensitivity to inorganic components such as ash and sulfur is low, and its model adaptability is poor. Single spectral techniques are no longer sufficient to fully meet the needs of rapid detection of multiple parameters in complex coals.

[0003] While existing multimodal fusion methods attempt to integrate the advantages of different spectral technologies, they still have significant limitations: cumbersome preprocessing procedures before fusion, a lack of rationality in feature variable selection, and fixed weight allocation patterns ultimately lead to low model prediction accuracy and weak generalization ability. For example, traditional methods often directly stitch together spectral data from different sources without targeted selection based on variable contribution, resulting in the introduction of a large amount of redundant information; at the same time, they lack effective stability control mechanisms, making it difficult to adapt to the continuous detection needs of industrial environments.

[0004] Therefore, there is an urgent need for a coal quality detection method based on near-infrared coupled TRLIBS to solve the technical problems of limited accuracy of single spectra, complex spectral data fusion algorithms, and weak anti-interference ability in existing coal quality detection technologies. Summary of the Invention

[0005] To overcome the problems existing in related technologies, this disclosure provides a coal quality detection method, device, equipment and medium based on near-infrared coupled TRLIBS, to solve the technical problems of limited single spectrum accuracy, complex spectral data fusion algorithm and weak anti-interference ability in existing coal quality detection technologies.

[0006] This specification provides one or more embodiments of a coal quality detection method based on near-infrared coupled TRLIBS, comprising the following steps: The flow rate of coal samples with a particle size of no more than 13 mm was detected using a laser rangefinder, triggering TRLIBS and near-infrared spectroscopy detection of the coal samples. TRLIBS spectra of the coal sample were acquired using a spectrometer, and near-infrared diffuse reflectance spectra of the coal sample were acquired using a reflectance probe. The spectra were then averaged to generate TRLIBS spectral data and near-infrared diffuse reflectance spectral data. The TRLIBS spectral data is preprocessed to obtain the preprocessed TRLIBS spectrum, and the near-infrared diffuse reflectance spectral data is processed to obtain the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum. The characteristic peaks of hydrogen-containing groups in the preprocessed TRLIBS spectral data and the near-infrared spectral data are fused at the characteristic level using dynamic weights to obtain fused features. Partial least squares regression is then used to perform regression analysis on the fused features to construct a coal quality parameter prediction model. The actual coal sample was tested using the coal quality parameter prediction model to obtain the actual coal quality test results.

[0007] Preferably, the step of using a laser rangefinder to detect the flow rate of coal samples with a particle size of no more than 13 mm, triggering TRLIBS and near-infrared spectroscopy detection of the coal sample, specifically includes the following steps: The coal samples are screened, and those with a particle size of no more than 13 mm are tested. The coal samples are then transported to the testing position by a uniform speed conveyor belt. The thickness of the coal sample at the current detection location is obtained using a laser rangefinder. When the thickness of the coal sample is greater than 30 mm, TRLIBS detection and near-infrared spectroscopy detection of the coal sample are triggered.

[0008] Preferably, the preprocessing of the TRLIBS spectral data to obtain the preprocessed TRLIBS spectrum specifically includes the following steps: The Raida criterion was used to remove anomalous spectra from TRLIBS spectral data. The sliding window method was used to remove background from the TRLIBS spectral data after removing abnormal spectra, resulting in background-removed spectra. The background-removed spectrum was smoothed and denoised using the Savitzky-Golay method, followed by channel intensity normalization to obtain the normalized spectrum. The normalized spectrum is used to identify characteristic peaks using the variable importance projection method to obtain the preprocessed TRLIBS spectrum.

[0009] Preferably, the step of using the sliding window method to remove background from the TRLIBS spectral data after removing abnormal spectra to obtain background-removed spectra specifically includes the following steps: The TRLIBS spectral data after removing outlier spectra were subjected to equal group segmentation. The minimum value of a single set of TRLIBS spectral data is used as the background spectrum; The background spectrum is obtained by subtracting the TRLIBS spectral data after removing the abnormal spectra from the background spectrum.

[0010] Preferably, the process of processing the near-infrared diffuse reflectance spectral data to obtain the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum specifically includes the following steps: Multivariate scattering correction and first-order derivative processing were performed on near-infrared diffuse reflectance spectral data to obtain first-order derivative spectra; Characteristic peaks of hydrogen-containing groups in near-infrared spectra were extracted using the variable importance projection method.

[0011] This specification provides one or more embodiments of a coal quality detection device based on near-infrared coupled TRLIBS, comprising: The spectral detection trigger module is used to detect the flow rate of coal samples with a particle size of no more than 13 mm using a laser rangefinder, and trigger TRLIBS detection and near-infrared spectral detection of the coal sample. The spectral data acquisition module is used to acquire the TRLIBS spectrum of the coal sample using a spectrometer, acquire the near-infrared diffuse reflectance spectrum of the coal sample using a reflectance probe, and perform mean-averaging processing to form TRLIBS spectral data and near-infrared diffuse reflectance spectral data. The spectral data processing module is used to preprocess the TRLIBS spectral data to obtain the preprocessed TRLIBS spectrum, and to process the near-infrared diffuse reflectance spectral data to obtain the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum. The prediction model construction module is used to perform feature-level fusion of the preprocessed TRLIBS spectral data and the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum using dynamic weights to obtain fusion features, and to perform regression analysis on the fusion features using partial least squares regression to construct a coal quality parameter prediction model. The detection module is used to detect actual coal samples using the coal quality parameter prediction model to obtain actual coal quality detection results.

[0012] Preferably, the spectral detection triggering module includes a screening unit and a detection unit; The screening unit is used to screen coal samples, detect coal samples with a particle size of no more than 13 mm, and transport the coal samples to the detection position by a uniform speed conveyor belt. The detection unit is used to obtain the thickness of the coal sample at the current detection location using a laser rangefinder. When the thickness of the coal sample is greater than 30 mm, it triggers TRLIBS detection and near-infrared spectroscopy detection of the coal sample.

[0013] Preferably, the spectral data processing module includes a TRLIBS spectral preprocessing unit, configured as follows: The Raida criterion was used to remove anomalous spectra from TRLIBS spectral data. The sliding window method was used to remove background from the TRLIBS spectral data after removing abnormal spectra, resulting in background-removed spectra. The background-removed spectrum was smoothed and denoised using the Savitzky-Golay method, followed by channel intensity normalization to obtain the normalized spectrum. The normalized spectrum is used to identify characteristic peaks using the variable importance projection method to obtain the preprocessed TRLIBS spectrum.

[0014] This specification provides one or more embodiments of a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the coal quality detection method based on near-infrared coupling TRLIBS as described above.

[0015] This specification provides one or more embodiments of a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the coal quality detection method based on near-infrared coupled TRLIBS as described above.

[0016] This disclosure provides a coal quality detection method, apparatus, equipment, and medium based on near-infrared coupled TRLIBS. Its advantages lie in utilizing a laser rangefinder to detect the flow rate of coal samples with a particle size not exceeding 13 mm, limiting the coal sample particle size to avoid interference from large particles, ensuring the uniformity of the detected object, triggering TRLIBS and near-infrared spectral detection of the coal sample, and achieving synchronous control of detection upon sample arrival based on the flow rate signal, avoiding errors caused by detection timing deviations, ensuring a continuous detection process, and laying the foundation for the effectiveness of dual-spectral data; using a spectrometer to collect the TRLIBS spectrum of the coal sample, and using a reflection probe to collect the... The near-infrared diffuse reflectance spectrum of the coal sample was obtained and averaged to form TRLIBS spectral data and near-infrared diffuse reflectance spectral data. This allows for targeted acquisition of TRLIBS and near-infrared diffuse reflectance spectra using a spectrometer and reflectance probe, enabling multi-dimensional capture of coal elemental and functional group characteristics. Averaging suppresses random noise, reduces data dispersion, and improves the reliability of the original data, providing high-quality basic material for subsequent processing. The TRLIBS spectral data was preprocessed to obtain preprocessed TRLIBS spectra, and the near-infrared diffuse reflectance spectral data was further processed to obtain near-infrared... Characteristic peaks of hydrogen-containing groups in the spectral data can be used to preprocess TRLIBS spectra to remove interference signals, highlight elemental characteristic lines, and improve the signal-to-noise ratio. First-order derivative processing of the near-infrared spectrum eliminates baseline drift, enhances the identification of characteristic peaks of hydrogen-containing groups, accurately mines core coal quality correlation information, and reduces the interference of redundant data on subsequent modeling. Dynamic weighting is used to perform feature-level fusion of the preprocessed TRLIBS spectral data and the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum to obtain fused features. Partial least squares regression is then used to perform regression analysis on the fused features to construct a coal quality parameter prediction model. The weighted approach achieves adaptive fusion of dual-spectral features, balancing the advantages of both spectra to form a comprehensive and robust fused feature. Partial least squares regression is used to process high-dimensional collinear data, uncovering the mapping relationship between features and coal quality parameters, and constructing a high-precision, high-stability coal quality parameter prediction model. This model is then used to test actual coal samples, obtaining actual coal quality test results without relying on traditional, cumbersome chemical analysis, significantly shortening the testing cycle. The test results are accurate and reliable, meeting the needs of industrial scenarios for coal quality assessment. This provides data support for practical applications such as coal grading and combustion optimization, promoting the industrialization of testing technology. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in one or more embodiments of this specification or in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart illustrating a coal quality detection method based on near-infrared coupled TRLIBS provided for one or more embodiments of this specification; Figure 2 A comparison chart of the received basis low heating value test results with the national standard method provided in one or more embodiments of this specification; Figure 3 A schematic flowchart of a rapid coal quality detection technique using NIRS coupled with TRLIBS provided in one or more embodiments of this specification; Figure 4 Comparison of TRLIBS spectral preprocessing effects provided in one or more embodiments of this specification; Figure 5 A schematic diagram of a coal quality detection device based on near-infrared coupling TRLIBS provided for one or more embodiments of this specification; Figure 6 This is a schematic diagram of the structure of a computer device provided for one or more embodiments of this specification. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this invention.

[0020] The present invention will now be described in detail with reference to specific embodiments and accompanying drawings.

[0021] Method Implementation Examples According to embodiments of the present invention, a coal quality detection method based on near-infrared coupled TRLIBS is provided, such as... Figure 1The diagram shown is a flowchart of the coal quality detection method based on near-infrared coupled TRLIBS provided in this embodiment. In this embodiment, 187 coal samples, including bituminous coal, anthracite, and lignite, were selected. The moisture content on an as-received basis ranged from 8.8% to 39.8%, the ash content on a dry basis ranged from 7.9% to 44.9%, and the lower heating value on an as-received basis ranged from 11.1 to 24.1 MJ / kg. Sample preparation was performed according to GB / T 474-2008. The coal quality detection method based on near-infrared coupled TRLIBS according to this embodiment includes the following steps: S110. Detect the flow rate of coal samples with a particle size not greater than 13 mm using a laser rangefinder, triggering TRLIBS and near-infrared spectroscopy detection of the coal samples.

[0022] S120. The TRLIBS spectrum of the coal sample is acquired using a spectrometer, and the near-infrared diffuse reflectance spectrum of the coal sample is acquired using a reflectance probe. The number of acquisitions by the spectrometer and reflectance probe is set to be within the range of 100 to 200. The acquisition results are averaged to form TRLIBS spectral data and near-infrared diffuse reflectance spectral data. Specifically, after receiving a trigger signal, the laser in the TRLIBS system emits laser light at a frequency of 10 Hz, and the spectrometer acquires the TRLIBS data of the coal sample. Simultaneously, the diffuse reflectance probe acquires the near-infrared spectral data of the coal sample. The acquired TRLIBS and near-infrared spectral data are then processed. The number of acquisitions by the spectrometer and reflectance probe is set to be within the range of 100 to 200, and the acquisition results are averaged to obtain averaged TRLIBS spectral data and near-infrared diffuse reflectance spectral data.

[0023] S130. Preprocess the TRLIBS spectral data, including abnormal spectrum removal, background removal, smoothing and noise reduction, channel intensity normalization and characteristic peak identification, to obtain the preprocessed TRLIBS spectrum. Process the near-infrared diffuse reflectance spectral data to obtain the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum.

[0024] S140. Dynamic weights are used to perform characteristic-level fusion of the preprocessed TRLIBS spectral data and the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum to obtain fusion features. Partial least squares regression is then used to perform regression analysis on the fusion features to construct a coal quality parameter prediction model.

[0025] Specifically, such as Figure 2The figure shows a comparison between the received basis low calorific value detection results and the national standard method provided in this embodiment. Partial least squares regression is used to perform regression analysis on the fusion characteristics of the characteristic peaks of the preprocessed TRLIBS spectrum and the first derivative near-infrared spectrum to form a coal quality parameter prediction model. First, dynamic weights are used to perform characteristic-level fusion of the characteristic peaks of hydrogen-containing groups in the preprocessed TRLIBS spectrum and the first derivative near-infrared spectrum to obtain fusion characteristics. Specifically, the characteristic peaks in the preprocessed TRLIBS spectrum and the first derivative near-infrared spectrum are analyzed separately, including peak position and intensity. Second, based on these characteristics and the importance of different spectra in the characterization of coal quality parameters, dynamically changing weights are assigned to the characteristic peaks of the two spectra. Through this weight allocation, the characteristic peaks of the two spectra are organically combined to obtain fusion characteristics that more comprehensively reflect coal quality information. Third, partial least squares regression is used to perform regression analysis on the fusion characteristics to form a coal quality parameter prediction model. The obtained fusion characteristics are used as the input variables of the partial least squares regression model, while the known coal quality parameters are used as the output variables of the model. By training with a large amount of sample data, the model learns the intrinsic relationship between fusion features and coal quality parameters, thereby forming a model that can be used to predict coal quality parameters.

[0026] S150. Use the coal quality parameter prediction model to test the actual coal sample and obtain the actual coal quality test results.

[0027] The established coal quality parameter prediction model is used to test actual coal samples to obtain actual coal quality parameter results. The data processing unit outputs the predicted coal quality parameters, such as moisture, ash, volatile matter, sulfur, and calorific value, calculated by the model, intuitively through a host computer interface for easy viewing and recording by operators. The entire process from sample input to result output for a single test can be controlled within 15 seconds; the measurement precision for total moisture is no more than 1.8%, and the measurement accuracy is no more than 2%; the measurement precision for dry basis ash is no more than 1.2%, and the measurement accuracy is no more than 1.5%; the measurement precision for as-borne lower heating value is no more than 0.2 MJ / kg, and the measurement accuracy is no more than 0.315 MJ / kg, etc., which can meet the needs of rapid coal quality analysis in actual production.

[0028] To further demonstrate the practicality of this invention, each coal sample was tested 10 times in the actual testing process. After outlier removal, the effective spectral retention rate reached over 95%; the signal-to-background ratio of characteristic peaks improved by 2-3 times after baseline correction; and the proportion of missing peaks decreased from 15% to 3% after characteristic peak completion. The relative standard deviation (RSD) of near-infrared spectroscopy after multivariate scattering correction improved from 5.2% to 2.1%.

[0029] like Figure 3 The diagram shows a flowchart of the rapid coal quality detection technology using NIRS coupled with TRLIBS provided in this embodiment. Results show that, using this method for spectral data fusion analysis, the correlation coefficient R² for the ash content prediction model is 0.97, and the root mean square error (RMSE) is 0.85%; the R² for the as-received lower heating value prediction model is 0.96, and the RMSE is 0.25 MJ / kg; the R² for the sulfur content prediction model is 0.98, and the RMSE is 0.03%. The detection results for blind samples show that the relative errors for all predicted parameters do not exceed 5%.

[0030] The method provided in this embodiment utilizes a laser rangefinder to detect the flow rate of coal samples with a particle size not exceeding 13 mm. This limitation on particle size avoids interference from large particles, ensuring the uniformity of the detected sample. It triggers TRLIBS and near-infrared spectral detection of the coal sample, achieving synchronous control based on the flow signal to ensure detection upon sample arrival. This avoids errors caused by timing deviations, ensuring a continuous detection process and laying the foundation for the effectiveness of dual-spectral data. The method uses a spectrometer to collect the TRLIBS spectrum of the coal sample and a reflection probe to collect the near-infrared diffuse reflectance spectrum. These are then averaged to form TRLIBS and near-infrared diffuse reflectance spectral data. This allows for targeted acquisition of TRLIBS and near-infrared diffuse reflectance spectra using the spectrometer and reflection probe, enabling multi-dimensional capture of coal elemental and functional group characteristics. Averaging suppresses random noise, reduces data dispersion, and improves the reliability of the original data, providing high-quality foundational material for subsequent processing. The TRLIBS spectral data is preprocessed to obtain preprocessed TRLIBS spectra, and the near-infrared diffuse reflectance spectral data is processed to obtain characteristic peaks of hydrogen-containing groups in the near-infrared spectrum. This enables the detection of TRLIBS and near-infrared diffuse reflectance spectral data. TRLIBS spectral preprocessing removes interference signals, highlights elemental characteristic spectral lines, and improves the signal-to-noise ratio. First-order derivative processing of the near-infrared spectrum eliminates baseline drift, enhances the identification of hydrogen-containing group characteristic peaks, accurately mines core coal quality correlation information, and reduces the interference of redundant data on subsequent modeling. Dynamic weighting is used to perform feature-level fusion of the preprocessed TRLIBS spectral data and the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum to obtain fused features. Partial least squares regression is then used to perform regression analysis on the fused features to construct a coal quality parameter prediction model. Dynamic weighting achieves adaptive fusion of dual-spectral features, balancing the advantages of the two spectra to form a comprehensive and robust fused feature. Partial least squares regression is used to process high-dimensional collinear data, mining the mapping relationship between features and coal quality parameters to construct a high-precision, high-stability coal quality parameter prediction model. This model is then used to test actual coal samples, obtaining actual coal quality test results without relying on traditional cumbersome chemical analysis, significantly shortening the testing cycle. The test results are accurate and reliable, meeting the needs of industrial scenarios for coal quality assessment, providing data support for practical applications such as coal quality grading and combustion optimization, and promoting the industrialization of testing technology.

[0031] In one embodiment, a laser rangefinder is used to detect the flow rate of coal samples with a particle size of no more than 13 mm, triggering TRLIBS and near-infrared spectroscopy detection of the coal samples. This specifically includes the following steps: The coal samples are screened, and those with a particle size of no more than 13 mm are tested. The coal samples are then transported to the testing position by a uniform speed conveyor belt.

[0032] The thickness of the coal sample at the current detection location is obtained using a laser rangefinder. When the thickness of the coal sample is greater than 30 mm, TRLIBS detection and near-infrared spectroscopy detection of the coal sample are triggered.

[0033] Specifically, after being crushed, the coal sample is sieved through a 13mm sieve and evenly spread on a conveyor belt, which then transports it to the detection position at a speed of 0.5-2m / min. When the coal sample reaches the detection position, a laser rangefinder obtains the thickness of the coal sample on the conveyor belt at the current detection position as input to the industrial control computer. When the coal sample thickness exceeds 30mm, the industrial control computer sends a signal to trigger TRLIBS and near-infrared spectroscopy detection of the coal sample.

[0034] The method provided in this embodiment avoids interference from large particle scattering and ensures the uniformity of the coal sample by screening coal samples with a particle size of no more than 13 mm and stably transporting them to the detection position via a uniform conveyor belt. The thickness of the coal sample is monitored by a laser rangefinder, and TRLIBS and near-infrared spectroscopy detection are accurately initiated with a trigger threshold of more than 30 mm. This avoids signal distortion and invalid detection caused by excessively thin coal samples, and ensures that the detection timing matches the coal sample condition, thereby improving detection efficiency and data reliability.

[0035] In one embodiment, TRLIBS spectral data is preprocessed to obtain preprocessed TRLIBS spectra, specifically including the following steps: The Raida criterion was used to remove anomalous spectra from TRLIBS spectral data.

[0036] The sliding window method is used to remove background from the TRLIBS spectral data after removing abnormal spectra to obtain the background-removed spectrum. Specifically, the TRLIBS spectral data after removing abnormal spectra is divided into equal groups, and the minimum value of a single group of TRLIBS spectral data is obtained as the background spectrum. The difference between the TRLIBS spectral data after removing abnormal spectra and the background spectrum is obtained to obtain the background-removed spectrum.

[0037] The Savitzky-Golay method was used to smooth and denoise the background-removed spectrum, followed by channel-by-channel intensity normalization to obtain the normalized spectrum.

[0038] The normalized spectrum was used to identify characteristic peaks using the variable importance projection method to obtain the preprocessed TRLIBS spectrum.

[0039] like Figure 4The image shows a comparison of the TRLIBS spectral preprocessing effects provided in this embodiment. The TRLIBS spectral preprocessing involves several steps. First, the Raida criterion is used to remove abnormal spectra from the TRLIBS spectral data, resulting in spectra with moderate intensity and distinct characteristics. Second, the TRLIBS spectral data after removing abnormal spectra is divided into k equal groups. The intensity values ​​of each wavelength pixel in each group of TRLIBS spectral data are compared, and the minimum value of each group is taken as the background spectrum. The difference between the TRLIBS spectral data after removing abnormal spectra and the background spectrum is calculated to obtain the TRLIBS spectral data after removing the background spectrum. Third, the Savitzky-Golay method is used to smooth and reduce noise in the TRLIBS spectral data after removing the background spectrum. Then, channel intensity normalization is performed, and the four channels in the TRLIBS spectral data after removing the background spectrum are merged to obtain normalized TRLIBS spectral data. Finally, the normalized spectrum is used to identify characteristic peaks through variable importance projection, and the characteristic peaks with high predictive importance for the parameters are calculated, resulting in the preprocessed TRLIBS spectral data.

[0040] The method provided in this embodiment removes abnormal spectra using the Raida criterion, removes background using the sliding window method, smooths and reduces noise using the Savitzky-Golay method, and normalizes the intensity of each channel. Then, it uses the variable importance projection method to screen core feature peaks. This effectively removes interference signals, reduces noise, unifies the data scale, and focuses on key features, significantly improving the signal-to-noise ratio and effectiveness of TRLIBS spectra. This provides high-quality data support for subsequent feature fusion and the construction of coal quality parameter prediction models.

[0041] In one embodiment, near-infrared diffuse reflectance spectral data is processed to obtain characteristic peaks of hydrogen-containing groups in the near-infrared spectrum, specifically including the following steps: Multivariate scattering correction and first-order derivative processing were performed on near-infrared diffuse reflectance spectral data to obtain first-order derivative spectra.

[0042] Characteristic peaks of hydrogen-containing groups in near-infrared spectra were extracted using the Variable Importance in Projection (VIP) method.

[0043] First, multivariate scattering correction is performed on the near-infrared spectrum to eliminate scattering interference caused by differences in coal sample particle size and enhance the correlation between the spectrum and composition. Second, first-derivative processing is performed on the near-infrared spectrum to enhance the resolution of functional group characteristic peaks and obtain near-infrared spectral data after first-derivative processing. Finally, the variable importance projection method is used to extract the characteristic peaks of hydrogen-containing functional groups from the near-infrared spectral data after first-derivative processing.

[0044] The method provided in this embodiment eliminates the scattering interference of coal sample particles through multivariate scattering correction, eliminates baseline drift and enhances the resolution of characteristic peaks of hydrogen-containing groups by first-order derivative processing, and then accurately extracts the core characteristic peaks by variable importance projection method. This effectively purifies near-infrared spectral information and focuses on key features related to coal quality, providing high-value data support for subsequent dual-spectral feature fusion and coal quality parameter prediction model construction.

[0045] Device Examples According to embodiments of the present invention, a coal quality detection device based on near-infrared coupled TRLIBS is provided, such as... Figure 5 The diagram shown is a structural schematic of the near-infrared coupled TRLIBS-based coal quality detection device provided in this embodiment. The near-infrared coupled TRLIBS-based coal quality detection device according to this embodiment includes: The spectral detection trigger module 51 is used to detect the flow rate of coal samples with a particle size of no more than 13 mm using a laser rangefinder, and to trigger TRLIBS detection and near-infrared spectral detection of the coal sample.

[0046] The spectral data acquisition module 52 is used to acquire the TRLIBS spectrum of the coal sample using a spectrometer, acquire the near-infrared diffuse reflectance spectrum of the coal sample using a reflectance probe, and perform mean-averaging processing to form TRLIBS spectral data and near-infrared diffuse reflectance spectral data.

[0047] The spectral data processing module 53 is used to preprocess TRLIBS spectral data to obtain preprocessed TRLIBS spectra, and to process near-infrared diffuse reflectance spectral data to obtain characteristic peaks of hydrogen-containing groups in the near-infrared spectrum.

[0048] The prediction model construction module 54 is used to perform feature-level fusion of the characteristic peaks of hydrogen-containing groups in the preprocessed TRLIBS spectral data and near-infrared spectral data using dynamic weights to obtain fusion features, and to perform regression analysis on the fusion features using partial least squares regression to construct a coal quality parameter prediction model.

[0049] The detection module 55 is used to detect actual coal samples using a coal quality parameter prediction model to obtain actual coal quality detection results.

[0050] The apparatus provided in this embodiment includes a spectral detection trigger module 51 that uses a laser rangefinder to detect the flow rate of coal samples with a particle size not exceeding 13 mm. This limits the coal sample particle size to avoid interference from large particles, ensuring the uniformity of the detected object. It triggers TRLIBS and near-infrared spectral detection of the coal sample, achieving synchronous control based on the flow rate signal to ensure detection as soon as the coal sample arrives. This avoids errors caused by timing deviations in detection, ensuring a continuous detection process and laying the foundation for the validity of dual-spectral data. The spectral data acquisition module 52 uses a spectrometer to collect the TRLIBS spectrum of the coal sample and a reflection probe to collect the near-infrared diffuse reflectance spectrum of the coal sample, and then performs averaging. The process involves preprocessing the TRLIBS and near-infrared diffuse reflectance spectra to generate TRLIBS and near-infrared diffuse reflectance spectra. This allows for targeted acquisition of TRLIBS and near-infrared diffuse reflectance spectra using a spectrometer and reflectance probe, enabling multi-dimensional capture of coal elemental and functional group characteristics. Mean square processing suppresses random noise, reduces data dispersion, and improves the reliability of the raw data, providing high-quality foundational material for subsequent processing. The spectral data processing module 53 preprocesses the TRLIBS spectra to obtain preprocessed TRLIBS spectra and processes the near-infrared diffuse reflectance spectra to obtain characteristic peaks of hydrogen-containing groups in the near-infrared spectrum. The system can preprocess TRLIBS spectra to remove interference signals, highlight elemental characteristic spectral lines, and improve the signal-to-noise ratio. It also performs first-order derivative processing on near-infrared spectra to eliminate baseline drift, enhance the identification of hydrogen-containing group characteristic peaks, accurately mine core coal quality correlation information, and reduce the interference of redundant data on subsequent modeling. The prediction model construction module 54 uses dynamic weights to perform feature-level fusion of the preprocessed TRLIBS spectral data and the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum, obtaining fused features. Partial least squares regression is then used to perform regression analysis on the fused features to construct a coal quality parameter prediction model. Dynamic weights are used to achieve dual-spectral... The adaptive fusion of spectral features balances the advantages of two spectra, forming a comprehensive and robust fused feature. Partial least squares regression is used to process high-dimensional collinear data, uncovering the mapping relationship between features and coal quality parameters, and constructing a high-precision, high-stability coal quality parameter prediction model. The detection module 55 uses this model to detect actual coal samples, obtaining actual coal quality test results without relying on traditional, cumbersome chemical analysis, significantly shortening the detection cycle. The test results are accurate and reliable, meeting the needs of industrial scenarios for coal quality assessment, providing data support for practical applications such as coal grading and combustion optimization, and promoting the industrialization of detection technology.

[0051] In one embodiment, the spectral detection triggering module 51 includes a screening unit and a detection unit. The screening unit is used to screen the coal sample, detect coal samples with a particle size of no more than 13 mm, and transport the coal sample to the detection position via a uniform speed conveyor belt. The detection unit is used to obtain the thickness of the coal sample at the current detection position using a laser rangefinder. When the coal sample thickness is greater than 30 mm, it triggers TRLIBS detection and near-infrared spectral detection of the coal sample.

[0052] The device provided in this embodiment screens coal samples with a particle size of no more than 13 mm and stably transports them to the detection position via a uniform conveyor belt, avoiding interference from large particle scattering and ensuring the uniformity of the coal sample state. It uses a laser rangefinder to monitor the thickness of the coal sample and accurately starts TRLIBS and near-infrared spectroscopy detection with a trigger threshold of more than 30 mm. This avoids signal distortion and invalid detection caused by excessively thin coal samples and ensures that the detection timing matches the coal sample state, thereby improving detection efficiency and data reliability.

[0053] In one embodiment, the spectral data processing module includes a TRLIBS spectral preprocessing unit, configured as follows: The Raida criterion was used to remove anomalous spectra from TRLIBS spectral data.

[0054] The sliding window method was used to remove background from the TRLIBS spectral data after removing abnormal spectra, resulting in background-removed spectra.

[0055] The Savitzky-Golay method was used to smooth and denoise the background-removed spectrum, followed by channel-by-channel intensity normalization to obtain the normalized spectrum.

[0056] The normalized spectrum was used to identify characteristic peaks using the variable importance projection method to obtain the preprocessed TRLIBS spectrum.

[0057] The device provided in this embodiment removes abnormal spectra using the Raida criterion, removes background using the sliding window method, smooths and reduces noise using the Savitzky-Golay method, and normalizes the intensity of each channel. Then, it uses the variable importance projection method to screen core feature peaks. This effectively removes interference signals, reduces noise, unifies data scale, and focuses on key features, significantly improving the signal-to-noise ratio and effectiveness of TRLIBS spectra. This provides high-quality data support for subsequent feature fusion and the construction of coal quality parameter prediction models.

[0058] The embodiments of the present invention are device embodiments corresponding to the above method embodiments. The specific operations of each module processing step can be understood with reference to the description of the method embodiments, and will not be repeated here.

[0059] like Figure 6 As shown, the present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements the coal quality detection method based on near-infrared coupling TRLIBS in the above embodiments, or when the computer program is executed by a processor, it implements the coal quality detection method based on near-infrared coupling TRLIBS in the above embodiments.

[0060] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0061] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for apparatus or system embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The apparatus and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and the contents not described in detail in the specification of the present invention are known to those skilled in the art.

Claims

1. A coal quality detection method based on near-infrared coupled TRLIBS, characterized in that, Includes the following steps: The flow rate of coal samples with a particle size of no more than 13 mm was detected using a laser rangefinder, triggering TRLIBS and near-infrared spectroscopy detection of the coal samples. TRLIBS spectra of the coal sample were acquired using a spectrometer, and near-infrared diffuse reflectance spectra of the coal sample were acquired using a reflectance probe. The spectra were then averaged to generate TRLIBS spectral data and near-infrared diffuse reflectance spectral data. The TRLIBS spectral data is preprocessed to obtain the preprocessed TRLIBS spectrum, and the near-infrared diffuse reflectance spectral data is processed to obtain the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum. The characteristic peaks of hydrogen-containing groups in the preprocessed TRLIBS spectral data and the near-infrared spectral data are fused at the characteristic level using dynamic weights to obtain fused features. Partial least squares regression is then used to perform regression analysis on the fused features to construct a coal quality parameter prediction model. The actual coal sample was tested using the coal quality parameter prediction model to obtain the actual coal quality test results.

2. The coal quality detection method based on near-infrared coupled TRLIBS as described in claim 1, characterized in that, The process of using a laser rangefinder to detect the flow rate of coal samples with a particle size of no more than 13 mm, triggering TRLIBS and near-infrared spectroscopy detection of the coal samples, specifically includes the following steps: The coal samples are screened, and those with a particle size of no more than 13 mm are tested. The coal samples are then transported to the testing position by a uniform speed conveyor belt. The thickness of the coal sample at the current detection location is obtained using a laser rangefinder. When the thickness of the coal sample is greater than 30 mm, TRLIBS detection and near-infrared spectroscopy detection of the coal sample are triggered.

3. The coal quality detection method based on near-infrared coupled TRLIBS as described in claim 1, characterized in that, The preprocessing of the TRLIBS spectral data to obtain the preprocessed TRLIBS spectrum specifically includes the following steps: The Raida criterion was used to remove anomalous spectra from TRLIBS spectral data. The sliding window method was used to remove background from the TRLIBS spectral data after removing abnormal spectra, resulting in background-removed spectra. The background-removed spectrum was smoothed and denoised using the Savitzky-Golay method, followed by channel intensity normalization to obtain the normalized spectrum. The normalized spectrum is used to identify characteristic peaks using the variable importance projection method to obtain the preprocessed TRLIBS spectrum.

4. The coal quality detection method based on near-infrared coupled TRLIBS as described in claim 3, characterized in that, The process of removing background from TRLIBS spectral data after removing outlier spectra using the sliding window method to obtain background-removed spectra includes the following steps: The TRLIBS spectral data after removing outlier spectra were subjected to equal group segmentation. The minimum value of a single set of TRLIBS spectral data is used as the background spectrum; The background spectrum is obtained by subtracting the TRLIBS spectral data after removing the abnormal spectra from the background spectrum.

5. The coal quality detection method based on near-infrared coupled TRLIBS as described in claim 1, characterized in that, The process of processing the near-infrared diffuse reflectance spectral data to obtain the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum specifically includes the following steps: Multivariate scattering correction and first-order derivative processing were performed on near-infrared diffuse reflectance spectral data to obtain first-order derivative spectra; Characteristic peaks of hydrogen-containing groups in near-infrared spectra were extracted using the variable importance projection method.

6. A coal quality detection device based on near-infrared coupled TRLIBS, characterized in that, include: The spectral detection trigger module is used to detect the flow rate of coal samples with a particle size of no more than 13 mm using a laser rangefinder, and trigger TRLIBS detection and near-infrared spectral detection of the coal sample. The spectral data acquisition module is used to acquire the TRLIBS spectrum of the coal sample using a spectrometer, acquire the near-infrared diffuse reflectance spectrum of the coal sample using a reflectance probe, and perform mean-averaging processing to form TRLIBS spectral data and near-infrared diffuse reflectance spectral data. The spectral data processing module is used to preprocess the TRLIBS spectral data to obtain the preprocessed TRLIBS spectrum, and to process the near-infrared diffuse reflectance spectral data to obtain the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum. The prediction model construction module is used to perform feature-level fusion of the preprocessed TRLIBS spectral data and the characteristic peaks of hydrogen-containing groups in the near-infrared spectrum using dynamic weights to obtain fusion features, and to perform regression analysis on the fusion features using partial least squares regression to construct a coal quality parameter prediction model. The detection module is used to detect actual coal samples using the coal quality parameter prediction model to obtain actual coal quality detection results.

7. The coal quality detection device based on near-infrared coupled TRLIBS as described in claim 6, characterized in that, The spectral detection triggering module includes a screening unit and a detection unit; The screening unit is used to screen coal samples, detect coal samples with a particle size of no more than 13 mm, and transport the coal samples to the detection position by a uniform speed conveyor belt. The detection unit is used to obtain the thickness of the coal sample at the current detection location using a laser rangefinder. When the thickness of the coal sample is greater than 30 mm, it triggers TRLIBS detection and near-infrared spectroscopy detection of the coal sample.

8. The coal quality detection device based on near-infrared coupled TRLIBS as described in claim 6, characterized in that, The spectral data processing module includes a TRLIBS spectral preprocessing unit, configured as follows: The Raida criterion was used to remove anomalous spectra from TRLIBS spectral data. The sliding window method was used to remove background from the TRLIBS spectral data after removing abnormal spectra, resulting in background-removed spectra. The background-removed spectrum was smoothed and denoised using the Savitzky-Golay method, followed by channel intensity normalization to obtain the normalized spectrum. The normalized spectrum is used to identify characteristic peaks using the variable importance projection method to obtain the preprocessed TRLIBS spectrum.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the coal quality detection method based on near-infrared coupling TRLIBS as described in any one of claims 1 to 5.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the coal quality detection method based on near-infrared coupling TRLIBS as described in any one of claims 1 to 5.

Images