Material component detection system and detection method

Abstract

The invention provides a material component detection system and method, and the system comprises a sampling module which is used for obtaining an initial sample from a corresponding process pipeline in response to a sampling instruction of a control system; wherein different sampling instructions correspond to different process pipelines; the sample processing module is used for processing the initial sample to obtain a sample to be detected; the laser detection module is used for carrying out laser detection treatment on the to-be-detected sample to generate a plurality of spectrums; wherein the wave band of each spectrum covers the characteristic peaks of various elements; and the data processing module is used for carrying out feature extraction processing on the spectrum so as to screen out feature peaks meeting a standard spectral line range, and the data processing module is also used for inputting the screened feature peaks as parameters into a preset concentration curve function so as to calculate corresponding component information. According to the material component detection system and detection method provided by the invention, accurate detection of material components can be realized.

Description

Technical Field

[0001] The present invention relates to the field of metallurgy, and in particular to a material composition detection system and a detection method. Background Art

[0002] In the metallurgical industry, accurate detection and real-time monitoring of material composition are the core of process regulation and product quality control. At present, the detection of material composition in the smelting process mainly relies on manual sampling and offline laboratory analysis, which has problems such as insufficient real-time performance, poor reliability, weak guidance and low safety. Manual sampling has a low frequency and is prone to introduce errors. The laboratory analysis cycle is long, and it is difficult to provide timely feedback and adjust process parameters. These limitations make it impossible to achieve real-time detection, fine control and process improvement of material composition, which hinders the intelligent upgrade of the smelting process. Therefore, there is room for improvement. Summary of the invention

[0003] The object of the present invention is to provide a material composition detection system and a detection method, which can realize accurate detection of material composition.

[0004] To solve the above technical problems, the present invention is achieved through the following technical solutions: The present invention provides a material component detection system, comprising: a sampling module, which is used to obtain an initial sample from a corresponding process pipeline in response to a sampling instruction of a control system; wherein different sampling instructions correspond to different process pipelines; A sample processing module, used to process the initial sample to obtain a sample to be tested; A laser detection module, used to perform laser detection processing on the sample to be detected to generate multiple spectra; wherein the wavelength band of each spectrum covers the characteristic peaks of multiple elements; and A data processing module, used to perform spectral denoising and baseline correction processing on each spectrum in turn, and average the processed spectra to generate an average spectrum; The data processing module is also used to calculate the correlation coefficient between each processed spectrum and the average spectrum to screen out the spectrum with a correlation coefficient greater than a preset value, and average the screened spectrum to generate a target spectrum; The data processing module is also used to perform feature extraction processing on the target spectrum to screen out characteristic peaks that meet the standard spectral line range; The data processing module is also used to input the screened characteristic peaks as parameters into a preset concentration curve function to calculate corresponding component information.

[0005] The present invention also provides a material composition detection method, comprising: The sampling module obtains an initial sample from a corresponding process pipeline in response to a sampling instruction of the control system; wherein different sampling instructions correspond to different process pipelines; Processing the initial sample through a sample processing module to obtain a sample to be tested; Performing laser detection processing on the sample to be detected by a laser detection module to generate multiple spectra; wherein the wavelength band of each spectrum covers characteristic peaks of multiple elements; Performing spectral denoising and baseline correction processing on each spectrum in turn through a data processing module, and averaging the processed spectra to generate an average spectrum; The data processing module calculates the correlation coefficient between each processed spectrum and the average spectrum to screen out the spectrum with a correlation coefficient greater than a preset value, and averages the screened spectrum to generate a target spectrum; The target spectrum is subjected to feature extraction processing by the data processing module to screen out feature peaks that meet the standard spectrum range, and the screened feature peaks are input as parameters into a preset concentration curve function to calculate corresponding component information.

[0006] As described above, the present invention provides a material composition detection system and method, which ensure the flexibility and accuracy of the sampling process by automatically controlling the sampling frequency, sampling time and sampling amount; reduce the sample particle size and improve the uniformity by simply processing the sample, thereby reducing the impact on the detection result; use laser online detection technology to achieve real-time and high-accuracy component detection to ensure that the current condition of the material is truly reflected; after the detection is completed, the material is automatically returned to the process pipeline by the return system, without causing material loss and waste.

[0007] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0009] Figure 1 It is a structural schematic diagram of a material composition detection system in one embodiment of the present invention; Figure 2 This is a schematic diagram of a module of a material composition detection system in one embodiment of the present invention; Figure 3 Flow chart of a material composition detection method in one embodiment of the present invention.

[0010] In the figure: 100, sampling module; 200, sample processing module; 210, transfer hopper; 220, transfer belt; 300, laser detection module; 400, data processing module; 500, return material module; 600, control system; 700, process pipeline. DETAILED DESCRIPTION

[0011] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0012] See also Figure 1 The present invention provides a material composition detection system, which can detect the specific components of the material in the process pipeline in the metallurgical industry. The material composition detection system may include a sampling module 100, a sample processing module 200, a laser detection module 300, a data processing module 400 and a return material module 500. Among them, a transfer hopper 210 may be provided between the sampling module 100 and the sample processing module 200, and the transfer hopper 210 may be used to transmit the initial sample. The sample processing module 200, the laser detection module 300 and the data processing module 400 may form a laser component analyzer. A transfer belt 220 may be provided between the sample processing module 200 and the return material module 500, and the transfer belt 220 may be used to transmit the sample to be detected to the return material module 500.

[0013] See also Figure 2 In one embodiment, the sampling module 100 can be used to obtain an initial sample from the corresponding process pipeline 700 in response to a sampling instruction of the control system 600; wherein different sampling instructions correspond to different process pipelines 700. Specifically, the sampling module 100 may include at least one sampler and at least one control valve. Different samplers are connected to the corresponding process pipelines 700, and each sampler is provided with a control valve. The control valve is used to respond to the sampling instruction so that the corresponding sampler obtains an initial sample from the corresponding process pipeline 700. The control system 600 may be a distributed control system (DCS).

[0014] In one embodiment, the sampling module 100 can operate according to the sampling instructions issued by the control system 600. Different sampling instructions correspond to different process pipelines 700, and the control system 600 can select which process pipeline 700 to sample from as needed. The process pipeline 700 is a pipeline system for transporting materials, and the task of the sampling module 100 is to obtain an initial sample from the process pipeline 700. The initial sample refers to an unprocessed material sample directly taken from the process pipeline 700 for subsequent analysis or detection.

[0015] In one embodiment, the sampling module 100 includes at least one sampler. The sampler is a device that actually performs the sampling operation and can control the acquisition of samples by opening and closing. Each sampler is connected to a specific process pipeline 700 to ensure sampling from the specified process pipeline 700. A control valve is provided on each sampler. The function of the control valve is to respond to the sampling instruction and control the opening and closing of the sampler. The opening and closing time of the control valve can be precisely controlled, thereby achieving precise management of the sampling frequency, sampling time and sampling amount.

[0016] In one embodiment, the number of samplers and control valves of the sampling module 100 is custom designed according to the number of process pipelines 700, that is, multiple samplers and control valves can be configured according to actual needs to adapt to different numbers of process pipelines. Through the instructions of the control system 600, the function of sampling from a single process pipeline 700 can be realized. For example, if only sampling from pipeline A is required, the control system 600 will issue instructions to make the sampler and control valve corresponding to pipeline A work, while other samplers and control valves remain closed. The sampling module 100 can also realize simultaneous sampling from multiple process pipelines 700 and mix these samples together. For example, the control system 600 can simultaneously issue different sampling instructions to make the samplers and control valves corresponding to multiple process pipelines 700 (such as pipelines A, B and C) work, so as to simultaneously sample and mix these process pipelines 700.

[0017] See also Figure 1 and Figure 2 In one embodiment, the sample processing module 200 can be used to process the initial sample to obtain the sample to be tested. Specifically, the sample processing module 200 can be used to dry, grind, and compact the initial sample to obtain the corresponding sample to be tested.

[0018] In one embodiment, the initial sample is an unprocessed material sample taken directly from the process pipeline 700. The sample processing module 200 can perform a variety of processing operations on these initial samples, including drying, grinding and compacting, to obtain samples to be tested that meet the test requirements. Among them, drying refers to removing moisture from the initial sample to make the initial sample dry, which is helpful for subsequent testing, especially when conducting chemical analysis or physical property testing, moisture will affect the accuracy of the test results. Grinding refers to grinding the initial sample into finer particles, which helps to improve the uniformity of the sample and reduce the impact of inconsistent particle size on the test results. Compacting refers to pressing the ground sample into a fixed shape or density, which helps to better control the physical properties of the sample in certain detection methods and ensure the consistency of the detection conditions.

[0019] In one embodiment, the initial sample may contain non-uniform materials, i.e. materials with inconsistent particle size, shape or density. The sample processing module 200 can perform simple processing on these non-uniform materials to reduce the impact of sample particle size and uniformity on the test results. Through processing steps such as drying, grinding and compaction, the uniformity and consistency of the sample can be ensured, thereby improving the accuracy and reliability of the test results.

[0020] In one embodiment, the specific configuration of the sample processing module 200 needs to be determined according to the material characteristics and detection requirements in the actual application scenario. For example, if the material on site contains more moisture, more drying equipment may be required; if the particle size of the material is larger, a more powerful grinding device may be required.

[0021] In one embodiment, the laser detection module 300 can be used to perform laser detection processing on the sample to be detected to generate multiple spectra; wherein the wavelength band of each spectrum covers characteristic peaks of multiple elements.

[0022] In one embodiment, before starting detection, the laser detection module 300 can determine whether to start detection through data from sensors such as distance sensors and weight sensors. For example, the distance sensor can detect whether the distance between the sample to be detected and the detection module is appropriate, and the weight sensor can detect whether the sample to be detected is placed correctly and the weight is within a reasonable range. For example, the sample to be detected can be placed in a fixed position for detection. For another example, the sample to be detected can be detected on a mobile device such as a turntable or a belt.

[0023] In one embodiment, the detection parameters of the laser detection module 300 need to be determined by parameter control experiments to ensure the characteristic and stability of the spectrum. The detection parameters may include laser energy, frequency, spectrometer delay time, and integration time. Laser energy refers to the intensity of the laser emitted by the laser. Frequency refers to the frequency at which the laser emits the laser, that is, the number of laser pulses per unit time. Spectrometer delay time refers to the time delay after laser ablation when the spectrometer starts to collect spectral signals. Integration time refers to the duration of the spectrometer collecting spectral signals.

[0024] In one embodiment, the laser detection module 300 can use laser induced breakdown spectroscopy (LIBS) for detection. Laser induced breakdown spectroscopy refers to a high-energy pulsed laser emitted by a laser, and a laser focusing lens focuses the laser on the surface of the sample to be detected. The high-energy pulse of the laser ablates a portion of the material on the surface of the sample to be detected, generating plasma. The optical signal emitted by the generated plasma is coupled into the optical fiber through the collection optical path and transmitted to the spectral spectrometer and detection device. The spectral spectrometer and detection device acquires and records the spectra emitted by the plasma, which contain characteristic peaks of multiple elements in the sample to be monitored.

[0025] In one embodiment, the data processing module 400 can be used to perform feature extraction processing on the spectrum to screen out characteristic peaks that meet the standard spectral line range. The data processing module is also used to input the screened characteristic peaks as parameters into a preset concentration curve function to calculate the corresponding component information.

[0026] In one embodiment, the data processing module 400 can be used to pre-process the spectrum to generate a target spectrum. Specifically, the data processing module 400 can be used to perform spectrum denoising and baseline correction processing on each spectrum in turn, and average the processed spectra to generate an average spectrum. Subsequently, the data processing module 400 is also used to calculate the correlation coefficient between each processed spectrum and the average spectrum to screen out spectra with correlation coefficients greater than a preset value, and average the screened spectra to generate a target spectrum.

[0027] In one embodiment, the data processing module 400 may first perform spectral denoising on each spectrum to generate a corresponding first intermediate spectrum. Specifically, the data processing module 400 may use Fourier transform, Savitzky-Golay filtering, Kalman filtering or wavelet transform, etc. to perform denoising on each spectrum. Fourier transform refers to converting spectral data from the time domain (or space domain) to the frequency domain, removing high-frequency noise and then converting it back to the time domain to achieve denoising. Savitzky-Golay filtering refers to smoothing data by polynomial weighted fitting to reduce noise. Kalman filtering refers to a recursive filter that estimates the state of the system through prediction and update steps, and is suitable for denoising of dynamic systems. Wavelet transform refers to decomposing spectral data into components of different frequencies, and achieving denoising by removing high-frequency noise components.

[0028] In one embodiment, the data processing module 400 using Savitzky-Golay filtering is used as an example for explanation. Savitzky-Golay filtering smoothes data by fitting a polynomial to data points in a sliding window. During the fitting process, the weight of each data point is determined according to its position in the window, and is usually weighted using a polynomial coefficient. The filter uses a sliding window of a fixed length, and the window size can be selected according to the characteristics of the data and the denoising requirements. The window slides on the data points, moving one data point at a time, and fitting a polynomial to the data points in the window. By selecting the appropriate polynomial order and window size, the residual between the fitted polynomial and the original data can be minimized, thereby achieving denoising and smoothing to generate the corresponding first intermediate spectrum. Among them, the difference between the fitted polynomial and the original data is called the residual.

[0029] In one embodiment, in the above fitting process, the calculation formula of Savitzky-Golay filtering can be expressed as: . Wherein, Y is a window fitting value matrix, here it is the spectral intensity, unit au. X represents the variable matrix, here it is the wavelength, unit nm. A represents the polynomial fitting coefficient matrix. E represents the residual matrix. m represents the half window size. d-1 represents the maximum number of fittings. N is the window size, where the window size refers to the window size of the processing within a spectrum. The total length of the spectrum is divided by the length of a single processing to obtain the number of windows, and the length of a single processing is the window size. The value of N conforms to N=2m+1. E is the residual obtained from the fitting. Each window corresponds to a residual. The smallest residual is selected as the minimum residual of the spectrum.

[0030] In one embodiment, the data processing module 400 can perform baseline correction processing on each first intermediate spectrum to generate a corresponding second intermediate spectrum. Baseline correction is a commonly used spectral data preprocessing method for removing baseline drift or background interference in the spectrum. Baseline drift refers to the nonlinear deviation of the spectral background signal, which may cause deviations in the detection results. Through baseline correction, the quality of spectral data can be improved, thereby improving the accuracy and reliability of subsequent analysis.

[0031] In one embodiment, since the first intermediate spectrum may contain baseline drift or background interference, it is necessary to perform baseline correction processing on it to remove the background signal or baseline drift in the first intermediate spectrum, so that the processed second intermediate spectrum is purer. Specifically, some baseline points can be selected in the first intermediate spectrum first. The baseline points are points in the spectrum with no or few characteristic peaks, usually located on the spectral background signal. The selected baseline points should be distributed as evenly as possible in various regions of the spectrum. Subsequently, a polynomial fitting method can be used to fit the baseline of the spectrum through the selected baseline points. Subsequently, the baseline value of each point can be calculated based on the polynomial fitting results, where the baseline value is the predicted value of the polynomial at the corresponding wavelength position. Finally, the actual signal value in the first intermediate spectrum can be subtracted from the corresponding baseline value to generate a second intermediate spectrum after baseline correction.

[0032] In one embodiment, in the above-mentioned baseline correction process, the polynomial fitting formula can be expressed as: Where Y is the spectral intensity on the ordinate, in units of au. X is the wavelength on the abscissa, in units of nm. n is the polynomial order, which depends on the characteristics of the data and the degree of fitting required. A0~A n are the fitting coefficients.

[0033] In one embodiment, the data processing module 400 may perform an averaging process on all the second intermediate spectra to generate an average spectrum. The purpose of the averaging process is to reduce the influence of random noise and improve the signal-to-noise ratio of the spectral signal by summarizing multiple second intermediate spectra. Random noise is usually irregular in repeated measurements, and averaging can make these noises cancel each other out, thereby obtaining a smoother and more stable spectral signal.

[0034] In one embodiment, the data processing module 400 can average all received second intermediate spectra according to the corresponding wavelength positions, that is, for each wavelength position, average the signal values ​​of the position in multiple second intermediate spectra to generate the signal average value at the wavelength position. After averaging, different wavelength positions are associated with corresponding signal average values ​​to generate an average spectrum. Each signal value in the average spectrum is the average value of the signal values ​​of multiple second intermediate spectra at the corresponding wavelength position.

[0035] In one embodiment, the data processing module 400 can calculate the correlation coefficient between each second intermediate spectrum and the average spectrum, and determine whether the correlation coefficient is greater than a preset value: if greater, retain the corresponding second intermediate spectrum; otherwise, delete the corresponding second intermediate spectrum. The purpose of abnormal spectrum removal is to exclude those abnormal spectra caused by measurement errors, instrument failures or other reasons from all second intermediate spectra. These abnormal spectra may interfere with the results of data analysis and reduce the quality of the overall data. Spectral correlation refers to the degree of similarity between two segments of spectral data. Mathematically, the correlation coefficient is usually used to quantify this degree of similarity. The value range of the correlation coefficient is between -1 and 1, 1 represents a complete positive correlation, -1 represents a complete negative correlation, and 0 represents no correlation.

[0036] In one embodiment, the calculation formula for calculating the correlation coefficient between the second intermediate spectrum and the average spectrum by the data processing module 400 can be expressed as: . Among them, r can be expressed as the correlation coefficient. y and x are the spectral intensity on the ordinate and the wavelength on the abscissa. n is the number of points on the abscissa / vertical axis in the spectrum. is the mean of all x, is the mean of all y's. is the i-th point in the horizontal axis. is the i-th point in the ordinate. The data processing module 400 determines whether to retain or delete the spectrum by calculating the correlation coefficient between each spectrum and the average spectrum. The preset value can be adjusted according to specific experimental conditions and data analysis requirements, and the specific size is not limited. For example, the preset value can be within the range of 0.7 to 0.8.

[0037] In one embodiment, the data processing module 400 may average all the retained second intermediate spectra to generate a target spectrum. The process of generating the target spectrum may be the same as the process of generating the average spectrum, which will not be described in detail here.

[0038] In one embodiment, the data processing module 400 can perform feature extraction processing on the target spectrum to screen out characteristic peaks that meet the standard spectral line range. Specifically, the data processing module 400 can first perform feature extraction processing on the target spectrum to generate an output projection matrix. Among them, the eigenvector in the output projection matrix is ​​the characteristic peak of the corresponding element. Subsequently, the data processing module 400 can compare the eigenvector in the output projection matrix with the standard spectral line range: if the eigenvector meets the standard spectral line range, it is determined to be a characteristic peak; otherwise, it is determined to be a non-characteristic peak.

[0039] In one embodiment, the data processing module 400 can calculate the sample matrix of the target spectrum. The purpose of the sample matrix is ​​to convert the data of the target spectrum into a matrix form that is convenient for multivariate analysis. The target spectrum contains multiple wavelength positions (horizontal coordinates) and corresponding spectral intensities (vertical coordinates). For each horizontal coordinate (wavelength position) in the target spectrum, their average value can be calculated. For each vertical coordinate (spectral intensity) in the target spectrum, their average value can also be calculated. For all horizontal coordinate points and vertical coordinate points, the corresponding average values ​​are subtracted respectively to generate new data points. Finally, the new data points after subtracting the average value are organized into a matrix, namely the sample matrix. The rows of the sample matrix represent the wavelength positions, the columns represent the spectral intensities, and each element in the matrix represents the spectral intensity after subtracting the average value. Among them, the calculation formula of the sample matrix can be expressed as: .

[0040] In one embodiment, the data processing module 400 can calculate the corresponding characteristic covariance matrix based on the sample matrix. The characteristic covariance matrix (Covariance Matrix) is a symmetric matrix used to describe the covariance between each variable in a multidimensional data set. Each element in the characteristic covariance matrix represents the covariance between two variables, and the elements on the diagonal represent the variance of each variable. The data processing module 400 calculates the characteristic covariance matrix based on the sample matrix, and generates a matrix describing the covariance relationship between each wavelength position through matrix multiplication and normalization operations. Among them, the calculation formula of the characteristic covariance matrix C can be expressed as: .

[0041] In one embodiment, the data processing module 400 can calculate the eigenvalues ​​and eigenvectors of the eigencovariance matrix. Specifically, the eigenvalue is a scalar that describes the degree of scaling of the eigenvector under the transformation of the eigencovariance matrix. The larger the eigenvalue, the greater the data variation in the direction of the eigenvector. An eigenvector is a vector that represents the main direction of change of the data in that direction, and each eigenvector corresponds to an eigenvalue. The eigenvalues ​​and eigenvectors can be obtained by solving the following eigenvalue equation: Cv=λv, where λ is the eigenvalue and v is the eigenvector. By solving the characteristic polynomial det(C-λI)=0, where I is the unit matrix and det represents the determinant, the eigenvalue λi can be obtained. For each eigenvalue λi, the corresponding eigenvector can be obtained by solving the linear equation (C-λiI)vi=0.

[0042] In one embodiment, the data processing module 400 can sort the eigenvalues ​​from large to small, select multiple eigenvalues ​​with the highest order, and use the corresponding eigenvectors as column vectors to form an eigenvector matrix. The selection is based on retaining the variation information of most data while reducing the dimension of the data. After the eigenvalues ​​are sorted from large to small, it can be clear which directions have the largest data variation. The eigenvector corresponding to the largest eigenvalue represents the main direction of change of the data, the eigenvector corresponding to the second largest eigenvalue represents the second main direction of change, and so on.

[0043] In one embodiment, the data processing module 400 can calculate the output projection matrix based on the eigenvector matrix and the sample matrix. The eigenvector in the output projection matrix is ​​the characteristic peak of the corresponding element. The output projection matrix can be a new matrix calculated by multiplying the sample matrix and the eigenvector matrix. The output projection matrix F (m × k) can be expressed as: F (m × k) = DataAdjust (m × n) × E (n × k). Among them, DataAdjust (m × n) is the sample matrix. E (n × k) is the eigenvector matrix. m is the number of spectra. n is the number of features.

[0044] In one embodiment, the data processing module 400 can compare the eigenvectors in the output projection matrix with the standard spectral line range: if the eigenvector meets the standard spectral line range, it is determined to be a characteristic peak; otherwise, it is determined to be a non-characteristic peak. Specifically, the standard spectral line range refers to the wavelength or energy range of the characteristic peak known in a specific database. Each eigenvector in the output projection matrix can be compared with the standard spectral line range to determine whether it meets the standard spectral line range. If the eigenvector meets the standard spectral line range, the wavelength position corresponding to the eigenvector is determined to be a characteristic peak. If the eigenvector does not meet the standard spectral line range, the wavelength position corresponding to the eigenvector is determined to be a non-characteristic peak.

[0045] In one embodiment, the data processing module 400 is also used to input the screened characteristic peaks as parameters into a preset concentration curve function to calculate the corresponding component information. Specifically, the concentration curve function is established through the early modeling stage and is used to describe the relationship between the concentration of a specific component and its spectral characteristics. These relationships are usually linear or nonlinear and are obtained by fitting experimental data. In the modeling stage, it is necessary to collect spectral data for samples with different concentration gradients, perform preprocessing, and finally establish a concentration curve function. The data processing module 400 can substitute the extracted characteristic peaks as parameters into the preset concentration curve function to calculate the corresponding component information. Concentration curve function It can be expressed as: Among them, I C and I RRespectively represent the spectral line intensity (characteristic peak) of the measured element and the internal standard element. C is the reference content of the measured element obtained by the reference method. K and B represent the slope and intercept respectively.

[0046] In one embodiment, the data processing module 400 is further used to send the calculated component information to the control system 600 for subsequent process control.

[0047] In one embodiment, the return material module 500 can be used to respond to the return material instruction of the control system 600 to return the sample to be tested to the corresponding process pipeline 700. Specifically, the return material module 500 is a module responsible for returning the sample to be tested to the process pipeline 700 in the production process. The main function of the return material module 500 is to return the sample to the production process according to the return material instruction of the control system 600 after the sample test is completed.

[0048] In one embodiment, the material return method of the material return module 500 may include pneumatic sample delivery, gravity self-flow sample delivery, spiral reamer sample delivery, etc. Among them, pneumatic sample delivery refers to using the airflow generated by compressed air or a fan to blow the sample back into the process pipeline 700. Gravity self-flow sample delivery refers to using the gravity of the material itself to naturally flow and send the sample back into the process pipeline 700 through a slope or a pipeline. Spiral reamer sample delivery refers to using a spiral reamer (such as a screw conveyor) to push the sample back into the process pipeline 700. By setting the material return module 500, material loss and waste will not be caused.

[0049] It can be seen that in the above scheme, the sampling frequency, sampling time and sampling volume are automatically controlled to ensure the flexibility and accuracy of the sampling process; the sample particle size is reduced and the uniformity is improved by simple sample processing, thereby reducing the impact on the test results; the laser online detection technology is used to achieve real-time and high-accuracy component detection to ensure that the current condition of the material is truly reflected; after the test is completed, the material is automatically returned to the process pipeline by the return system, without causing material loss and waste.

[0050] See also Figure 3 The present invention also provides a material composition detection method, which can be applied to the above detection system. The detection method may include the following steps: Step S10: The sampling module obtains an initial sample from the corresponding process pipeline in response to the sampling instruction of the control system; wherein different sampling instructions correspond to different process pipelines; Step S20: Processing the initial sample through the sample processing module to obtain a sample to be tested; Step S30, performing laser detection processing on the sample to be detected by the laser detection module to generate multiple spectra; wherein the wavelength band of each spectrum covers characteristic peaks of multiple elements; Step S40, performing spectral denoising and baseline correction processing on each spectrum in turn through the data processing module, and averaging the processed spectra to generate an average spectrum; Step S50, calculating the correlation coefficient between each processed spectrum and the average spectrum through the data processing module to screen out the spectrum with a correlation coefficient greater than a preset value, and averaging the screened spectrum to generate a target spectrum; Step S60: Perform feature extraction processing on the target spectrum through the data processing module to screen out characteristic peaks that meet the standard spectral line range, and input the screened characteristic peaks as parameters into a preset concentration curve function to calculate corresponding component information.

[0051] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0052] The embodiments of the present invention disclosed above are only used to help illustrate the present invention. The embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A material composition detection system, characterized in that: include: A sampling module, used to obtain an initial sample from a corresponding process pipeline in response to a sampling instruction of a control system; wherein different sampling instructions correspond to different process pipelines; A sample processing module, used to process the initial sample to obtain a sample to be tested; A laser detection module, used to perform laser detection processing on the sample to be detected to generate multiple spectra; wherein the wavelength band of each spectrum covers the characteristic peaks of multiple elements; and A data processing module, used to perform spectral denoising and baseline correction processing on each spectrum in turn, and average the processed spectra to generate an average spectrum; The data processing module is also used to calculate the correlation coefficient between each processed spectrum and the average spectrum to screen out the spectrum with a correlation coefficient greater than a preset value, and average the screened spectrum to generate a target spectrum; The data processing module is also used to perform feature extraction processing on the target spectrum to screen out characteristic peaks that meet the standard spectral line range; The data processing module is also used to input the screened characteristic peaks as parameters into a preset concentration curve function to calculate corresponding component information.

2. The material composition detection system according to claim 1, characterized in that: The sampling module includes at least one sampler and at least one control valve; different samplers are connected to corresponding process pipelines, and each sampler is provided with a control valve; the control valve is used to respond to the sampling instruction so that the corresponding sampler obtains an initial sample from the corresponding process pipeline.

3. The material composition detection system according to claim 1, characterized in that: The sample processing module is used to dry, grind and compact the initial sample to obtain the corresponding sample to be tested.

4. The material composition detection system according to claim 1, characterized in that: The data processing module realizes the generation of the average spectrum through the following steps: Performing spectral denoising processing on each of the spectra to generate a corresponding first intermediate spectrum; Performing baseline correction processing on each of the first intermediate spectra to generate a corresponding second intermediate spectrum; averaging the signal values ​​at the same wavelength position in all the second intermediate spectra to generate a signal average value at the wavelength position; The different wavelength positions are associated with the corresponding signal averages to generate an average spectrum.

5. The material composition detection system according to claim 4, characterized in that: The data processing module realizes the generation of the target spectrum through the following steps: Calculating the correlation coefficient between each of the second intermediate spectrum and the average spectrum, and determining whether the correlation coefficient is greater than a preset value: if so, retaining the corresponding second intermediate spectrum; otherwise, deleting the corresponding second intermediate spectrum; The retained second intermediate spectra are averaged to generate a target spectrum.

6. The material composition detection system according to claim 1, characterized in that: The data processing module implements the screening of characteristic peaks through the following steps: Performing feature extraction processing on the target spectrum to generate an output projection matrix; wherein the feature vector in the output projection matrix is ​​a feature peak of the corresponding element; The characteristic vector in the output projection matrix is ​​compared with the standard spectral line range: if the characteristic vector conforms to the standard spectral line range, it is determined to be a characteristic peak; otherwise, it is determined to be a non-characteristic peak.

7. The material composition detection system according to claim 6, characterized in that: The data processing module generates an output projection matrix by the following steps: Calculating a sample matrix of the target spectrum, and calculating a characteristic covariance matrix according to the sample matrix; Calculating the eigenvalues ​​and eigenvectors of the eigencovariance matrix; The eigenvalues ​​are sorted in descending order, a plurality of eigenvalues ​​with the highest order are selected, and the eigenvectors corresponding to the eigenvalues ​​are used as column vectors to form an eigenvector matrix; An output projection matrix is ​​calculated according to the eigenvector matrix and the sample matrix; wherein the eigenvectors in the output projection matrix are characteristic peaks of corresponding elements.

8. The material composition detection system according to claim 1, characterized in that: The data processing module is also used to send the calculated component information to the control system.

9. The material composition detection system according to claim 1, characterized in that: It also includes a material return module, which is used to respond to the material return instruction of the control system to return the sample to be detected to the corresponding process pipeline.

10. A material composition detection method, characterized in that: include: The sampling module obtains an initial sample from a corresponding process pipeline in response to a sampling instruction of the control system; wherein different sampling instructions correspond to different process pipelines; Processing the initial sample through a sample processing module to obtain a sample to be tested; Performing laser detection processing on the sample to be detected by a laser detection module to generate multiple spectra; wherein the wavelength band of each spectrum covers characteristic peaks of multiple elements; Performing spectral denoising and baseline correction processing on each spectrum in turn through a data processing module, and averaging the processed spectra to generate an average spectrum; The data processing module calculates the correlation coefficient between each processed spectrum and the average spectrum to screen out the spectrum with a correlation coefficient greater than a preset value, and averages the screened spectrum to generate a target spectrum; The target spectrum is subjected to feature extraction processing by the data processing module to screen out feature peaks that meet the standard spectrum range, and the screened feature peaks are input as parameters into a preset concentration curve function to calculate corresponding component information.

Images