Online detection method for bauxite component, and related device

Abstract

An online detection method for a bauxite component, and a related device. The method comprises: controlling an LIBS system to collect n pieces of actual spectral data from bauxite to be tested on a preset-speed conveyor belt; performing a preset processing operation on the n pieces of actual spectral data, so as to acquire a set of preprocessed actual spectral data; calculating a screening distance between each piece of preprocessed actual spectral data in the set of preprocessed actual spectral data and a bauxite category spectrum, and adding, to an effective spectral data set, the preprocessed actual spectral data with the screening distance less than or equal to a preset distance, wherein the bauxite category spectrum is obtained by measuring a standard sample with known components on the basis of a static mode; using the average value of the first m pieces of spectral data in the effective spectral data set as actual continuous transmission spectral data; and inputting the actual continuous transmission spectral data into an optimal model, so as to acquire a bauxite component corresponding to each piece of spectral data.

Description

Bauxite composition online detection method and related equipment

[0001] Cross-reference to related applications

[0002] This application claims priority to Chinese Patent Application No. 2024106462671, filed May 23, 2024, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD

[0003] The present disclosure relates to the field of ore detection, and in particular, the present disclosure relates to a bauxite composition online detection method and related equipment. BACKGROUND

[0004] Alumina is an important industrial raw material and resource commodity. Raw ore slurry preparation is the first process of alumina production, and bauxite is the main raw material for raw ore slurry production. The composition of bauxite determines the material ratio of raw ore slurry preparation, and the fluctuation of raw ore slurry composition will affect the entire alumina production line, causing fluctuations in actual dissolution and decomposition rate, increased material consumption and energy consumption, and decreased product quality. Due to factors such as complex composition of bauxite, large variation in chemical composition, dwindling resources, and declining quality, it is difficult to seek solutions from the aspects of ensuring high quality and high stability of bauxite composition.

[0005] Detecting and analyzing the composition of bauxite can help develop an addition strategy for each material composition, guide the batching process, and stabilize the raw ore slurry environment. However, most of the process technology indicators of alumina enterprises currently use manual sampling at regular intervals and laboratory manual analysis methods for sample detection. The sampling time interval is long, the analysis results are lagging, and human errors are likely to occur, which cannot timely and effectively guide ore batching. The main online detection technologies for ore and metal composition at home and abroad include Prompt Gamma Neutron Activation Analysis (PGNAA), Near-Infrared Spectroscopy (NIR), and X-ray Fluorescence Spectroscopy (XRF). However, the main problems are that PGNAA poses a serious radiation risk, NIR cannot directly measure elements and has high requirements for on-site application adaptability, and XRF cannot directly measure elements with small atomic numbers.

[0006] Laser-Induced Breakdown Spectroscopy (LIBS) is a multi-element analysis technology based on atomic emission spectrum, which uses a high-energy laser source to generate high-temperature plasma on the surface of the sample. The elements contained in the sample are vaporized, atomized and excited in the hot plasma, thereby generating atomic and ionic spectra with characteristics of the sample element composition. Compared with traditional analysis methods, LIBS has the characteristics of rapidity, no need for complex sample preparation and remote analysis, and can solve the problem that conventional analysis techniques cannot be used as online detection techniques. It can be used for samples in any physical state, such as solids, liquids and gases including aerosols, and can be used for online analysis.

[0007] In the face of the complexity of ores and the strict time delay requirements of industrial production, the LIBS technology is affected by factors such as matrix effect, self-absorption effect and overlapping peak interference, and has not been applied to online detection of ore composition. There is an urgent need in the art for an online detection method and system for bauxite composition to realize online detection and analysis of bauxite composition, and to realize intelligent control of ore blending, reduce production cost, improve product quality, accelerate automation, informatization and intelligentization of the alumina production process.

[0008] SUMMARY

[0009] A series of simplified concepts are introduced in the summary section, which will be further described in detail in the specific embodiments section. The summary section of the present disclosure does not mean to attempt to define the key features and essential technical features of the claimed technical solutions, nor to attempt to determine the protection scope of the claimed technical solutions.

[0010] In a first aspect, the present disclosure provides an online detection method for bauxite composition, comprising: controlling a LIBS system to collect n actual spectral data on a bauxite to be measured on a preset speed conveyor belt; performing a preprocessing operation on the n actual spectral data to obtain a preprocessed actual spectral data set; calculating a screening distance between each preprocessed actual spectral data in the preprocessed actual spectral data set and an aluminum soil category spectrum, adding the preprocessed actual spectral data with a screening distance less than or equal to a preset distance to an effective spectral data set, the aluminum soil category spectrum being a standard sample with known composition measured based on a static mode; taking the average value of the first m spectral data in the effective spectral data set as actual continuous transmission spectral data; and inputting the actual continuous transmission spectral data into an optimal model to obtain the bauxite composition corresponding to each spectral data.

[0011] In a second aspect, the disclosure provides an online bauxite composition detection device, comprising: a collection unit configured to control a LIBS system to collect n actual spectral data on a bauxite to be detected on a preset speed conveyor belt; a preprocessing unit configured to perform a preprocessing operation on the n actual spectral data to obtain a preprocessed actual spectral data set; a screening unit configured to calculate a screening distance between each preprocessed actual spectral data in the preprocessed actual spectral data set and a bauxite category spectrum, and add the preprocessed actual spectral data with a screening distance less than or equal to a preset distance to an effective spectral data set, wherein the bauxite category spectrum is obtained by measuring a standard sample with known composition in a static mode; an averaging unit configured to take an average of the first m spectral data in the effective spectral data set as actual continuous transmission spectral data; and an acquisition unit configured to input the actual continuous transmission spectral data into an optimal model to obtain a bauxite composition corresponding to each spectral data.

[0012] In a third aspect, the disclosure provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is configured to execute the computer program stored in the memory to implement the steps of the online bauxite composition detection method.

[0013] In a fourth aspect, the disclosure provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the online bauxite composition detection method. BRIEF DESCRIPTION OF DRAWINGS

[0014] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included to provide a description of preferred embodiments, and are not intended to limit the present disclosure. Moreover, like reference numerals designate similar parts throughout the several views. In the drawings:

[0015] FIG. 1 is a flowchart of an online bauxite composition detection method according to some embodiments of the disclosure;

[0016] FIG. 2 is a schematic diagram of an online bauxite composition detection system according to some embodiments of the disclosure;

[0017] FIG. 3 is a schematic diagram of the processing effect of preprocessed LIBS spectra according to some embodiments of the disclosure;

[0018] FIG. 4 is a schematic diagram of a bauxite category spectrum S according to some embodiments of the disclosure;

[0019] FIG. 5 is a schematic diagram of continuous transmission bauxite composition detection according to some embodiments of the disclosure;

[0020] FIG. 6 is a structural schematic diagram of an on-line bauxite composition detection device according to some embodiments of the present disclosure;

[0021] FIG. 7 is a structural schematic diagram of an on-line bauxite composition detection electronic device according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0022] The terms "first", "second", "third", "fourth" and the like in the description and claims of the present disclosure and above-mentioned drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of the terms so construed can be interchanged, under appropriate circumstances, without changing the meaning of the description by the embodiments described herein. Also, the terms "comprise", "comprising", "include", "including", and "includes" as well as any variation thereof, are intended to cover a non-exclusive inclusion, for example, a process, method, system, product or apparatus that comprises a list of steps or units can not necessarily be limited to those steps or units that are clearly listed, but can include other not clearly listed steps or units or steps or units inherent to such process, method, product or apparatus. The technical solutions in the embodiments of the present disclosure will be described clearly and completely below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are only some of the embodiments of the present disclosure, not all the embodiments.

[0023] Referring to FIG. 1, it is a flow schematic diagram of an on-line bauxite composition detection method according to some embodiments of the present disclosure. The on-line bauxite composition detection method according to some embodiments of the present disclosure can include steps S110-S150.

[0024] In step S110, the LIBS system is controlled to collect n actual spectral data on the bauxite to be measured on the preset speed conveyor belt.

[0025] In some embodiments, the LIBS system is configured to collect n actual spectral data of the bauxite to be measured while moving on the conveyor belt at a preset speed. These spectral data are generated by a high-energy laser source to generate high-temperature plasma on the surface of the bauxite sample, and then excite the element emission spectrum in the sample.

[0026] In step S120, the n actual spectral data are preprocessed to obtain a preprocessed actual spectral data set.

[0027] In some embodiments, the n actual spectral data collected are preprocessed. The preprocessing operation can include but is not limited to denoising, baseline correction, normalization and the like, so as to obtain a cleaner and standardized preprocessed actual spectral data set.

[0028] In step S130, a screening distance between each of the pre-processed actual spectrum data and the bauxite category spectrum is calculated, and the pre-processed actual spectrum data with a screening distance less than or equal to a preset distance is added to the effective spectrum data set. The bauxite category spectrum is measured based on a static mode.

[0029] In some embodiments, by calculating a screening distance between each of the pre-processed actual spectrum data and the standard sample spectrum with known composition, the spectrum data with a distance less than or equal to a preset distance is screened and added to the effective spectrum data set. The bauxite category spectrum, i.e. the standard sample spectrum, is measured under static conditions, ensuring the accuracy of the spectrum data.

[0030] In step S140, the average of the first m spectrum data in the effective spectrum data set is taken as the actual continuous transmission spectrum data.

[0031] In some embodiments, the average of the first m spectrum data in the effective spectrum data set is calculated as the actual continuous transmission spectrum data. This average value can represent the spectral characteristics of the bauxite sample to be measured in the continuous transmission process.

[0032] In step S150, the actual continuous transmission spectrum data is input into the optimal model to obtain the bauxite composition corresponding to each spectrum data, wherein the optimal model is trained based on the continuous transmission spectrum data of the standard sample and the bauxite composition corresponding to the continuous transmission spectrum data.

[0033] In some embodiments, the actual continuous transmission spectrum data is input into the optimal model, and the model outputs the bauxite composition corresponding to each spectrum data. The optimal model is trained based on the standard continuous transmission spectrum data, which is obtained by collecting the standard sample spectrum on a preset speed conveyor belt and performing pre-processing and screening distance operations.

[0034] The method of the related art requires manual sampling and laboratory analysis, has a long time interval, and the analysis result lags behind. The present disclosure realizes online real-time detection of bauxite composition through LIBS technology, can obtain data in time, and guide the production process. Through pretreatment and screening operations, the quality and accuracy of the spectral data are ensured, human error and external interference are reduced, and more accurate detection results are provided. Compared with the PGNAA method, the LIBS technology avoids radiation risk and ensures the safety of the operator. The LIBS technology can be used for detection of solid, liquid and gas samples, and does not require complex sample preparation, and is suitable for various industrial environments. Through online detection and intelligent control, the labor cost and laboratory analysis cost are reduced, and the production efficiency and product quality are improved, and the overall production cost is reduced. The present scheme provides a reliable online detection method for the alumina production process, promotes the automation, informatization and intelligent development of the production process, and improves the industry competitiveness.

[0035] In some embodiments, the optimal model is obtained by the following steps:

[0036] The LIBS system is controlled to collect m standard spectral data of a standard sample on the preset speed conveyor belt, and the bauxite composition of the standard sample is obtained.

[0037] The m standard spectral data of the standard sample are subjected to pretreatment operation and screening distance operation, and the average value of the first m spectral data after operation is taken as the continuous transmission spectral data of the standard sample.

[0038] The optimal model is obtained based on the continuous transmission spectral data of the standard sample and the bauxite composition of the standard sample.

[0039] Similar to the data processing of the effective spectral data set, the optimal model is established based on the standard continuous transmission spectral data corresponding to the standard sample, and the standard sample includes one or more. The m standard spectral data collected for each standard sample are subjected to pretreatment operation. The pretreatment operation can include but is not limited to noise reduction, baseline correction, normalization and other processing, so as to obtain cleaner and standardized pretreated standard spectral data set. The screening distance between each pretreated standard spectral data in the above pretreated standard spectral data set and the bauxite category spectrum is calculated, the pretreated standard spectral data with a screening distance less than or equal to a preset distance is added to the effective spectral data set, and the average value of the first m spectral data of the spectral data of the effective spectral data set is taken as the continuous transmission spectral data of the standard sample. The optimal model is obtained based on the continuous transmission spectral data of the standard sample and the bauxite composition of the standard sample.

[0040] Specifically, the standard samples are placed on a set conveyor belt and transmitted at a preset speed, and m standard spectral data are collected using the LIBS system. The collected m standard spectral data are preprocessed to clean, correct, or enhance the signal so as to extract more useful features. Then, a screening distance operation is performed, and abnormal or non-standard data are removed by screening means. The average value of the screened standard spectral data is calculated, and the average value of the data is taken as the continuous transmission spectral data, so as to provide more reliable input for subsequent modeling. A plurality of groups of similar continuous transmission spectral data and the known bauxite composition corresponding to each group of spectral data are collected. The spectral data and the composition data are used as a training set to train the initial model, and finally an optimal model capable of accurately predicting the bauxite composition is obtained. It should be noted that the initial model of the optimal model includes but is not limited to a linear regression model, a partial least squares regression model, and a support vector machine.

[0041] In some embodiments, the online detection method of the bauxite composition according to some embodiments of the present disclosure can further include:

[0042] n1 component different block bauxite samples are prepared and ground and pressed to form the standard sample.

[0043] The LIBS system is controlled to collect n1 standard spectral data of the standard sample in a static state.

[0044] The n1 standard spectral data are preprocessed to obtain a preprocessed standard spectral data set.

[0045] The average value of the spectral data in the preprocessed standard spectral data set is calculated to obtain the bauxite category spectrum.

[0046] In some embodiments, n1 block samples are prepared from bauxite of different sources or components. Then, the block samples are ground and pressed to form uniform standard samples. Grinding and pressing help to reduce the non-uniformity of the samples and improve the accuracy of the detection results. The LIBS system is used to detect the standard samples in a static state to collect n1 standard spectral data. The detection in the static state can ensure the stability and reliability of the spectral data. The n1 standard spectral data are preprocessed, and the preprocessing includes but is not limited to denoising, baseline correction, normalization, etc., to obtain a more standardized preprocessed standard spectral data set. The average value of the spectral data in the preprocessed standard spectral data set is calculated as the bauxite category spectrum. The bauxite category spectrum represents the spectral characteristics of the standard sample and is used for subsequent screening and comparison of actual spectral data.

[0047] In summary, by preparing standard samples and performing pretreatment operations, the high quality and accuracy of bauxite category spectra are ensured, providing reliable references for subsequent actual detection. Through grinding and tabletting, the inhomogeneity of the sample is reduced, and the consistency of the detection results is improved. The standard spectrum data is collected in a static state, avoiding fluctuations and errors that may occur in dynamic detection, ensuring the stability of the spectrum data. The standardized pretreatment operation and bauxite category spectrum calculation provide convenience for subsequent screening and comparison of actual spectrum data, improving the efficiency and accuracy of data analysis.

[0048] In some embodiments, the preparation of n1 compositionally different block bauxite samples and the grinding and tabletting treatment to form the above-mentioned standard samples include:

[0049] The compositionally different block bauxite samples refer to the different content of elements contained in the bauxite; the particle size of the block bauxite sample is less than 20 mm, and n1 is greater than or equal to 30.

[0050] The grinding and tabletting treatment steps include.

[0051] The mortar is used for grinding operation with a maximum duration of less than 2 minutes to form a granular bauxite sample with a particle size of less than or equal to 2 μm.

[0052] The granular bauxite sample is pressed at a pressure of at least 15 MPa for at least 30 seconds to form the block bauxite sample.

[0053] In some embodiments, the compositionally different block bauxite samples are bauxite containing different element contents. By selecting bauxite with different compositions, the standard samples can be ensured to be diverse and more representative.

[0054] The particle size of the block bauxite sample should be less than 20 mm, and at least 30 block samples with different compositions should be prepared to ensure the reliability and comprehensiveness of the data.

[0055] The mortar is used for grinding operation with a maximum duration of less than 2 minutes to form a granular bauxite sample with a particle size of less than or equal to 2 μm. This particle size ensures the fineness and uniformity of the sample, which is helpful for subsequent spectral detection.

[0056] The granular bauxite sample is pressed at a pressure of at least 15 MPa. The pressing time is at least 30 seconds. This high-pressure short-time treatment method can form a dense block bauxite sample, reduce internal voids, and improve the uniformity and stability of the sample.

[0057] In summary, by selecting different components of bauxite, we ensure that the standard sample is representative and can cover various component changes that may occur during production. During sample preparation, we control the sample size to be less than 20mm to ensure that the sample can be effectively processed during grinding. After grinding, the particle size requirement is less than or equal to 2μm to ensure that the sample is fine and uniform, which is beneficial for accurate spectral data acquisition. We use a pressure of at least 15MPa to press the sample to ensure its density and stability, reduce internal voids, and avoid errors caused by sample non-uniformity during testing. The grinding operation time is less than 2 minutes to avoid moisture absorption, and the pressing operation time is at least 30 seconds to ensure that high-quality sample preparation is completed within a limited time and production efficiency is improved.

[0058] In some embodiments, the above-mentioned preprocessing operations include baseline correction operations and noise removal operations.

[0059] The above-mentioned baseline correction operations include one or more of derivative method, iterative polynomial fitting method and penalized least squares method.

[0060] In some embodiments, the purpose of the baseline correction operation is to eliminate baseline drift in the spectral data, thereby improving the accuracy and reliability of the spectral signal. The baseline correction method can include but is not limited to derivative method, iterative polynomial fitting method and penalized least squares method.

[0061] The derivative method eliminates baseline drift by calculating the first or second derivative of the spectral data. This method can effectively remove low-frequency noise and baseline drift, but at the same time may enhance high-frequency noise.

[0062] The iterative polynomial fitting method estimates and eliminates the baseline drift in the spectrum by iteratively fitting a polynomial. This method has high flexibility and can adapt to different shapes of the baseline.

[0063] The penalized least squares method controls the smoothness of the polynomial fitting by introducing a penalty term to achieve baseline correction. This method can well preserve the main features of the spectral signal while eliminating baseline drift.

[0064] The purpose of the noise removal operation is to reduce noise interference in the spectral data and improve the quality of the spectral signal. Common noise removal methods include but are not limited to filtering method and smoothing method.

[0065] The filtering method removes high-frequency noise by using low-pass filters, median filters, etc.

[0066] The smoothing method smooths the spectral data by using sliding window averaging, Savitzky-Golay filtering, etc. to reduce noise.

[0067] The baseline correction and noise removal operation of the embodiments of the present disclosure eliminates the baseline drift and noise interference in the spectral data, and obtains more accurate spectral signals. The selection of multiple baseline correction methods makes the processing process more flexible, and the most suitable method can be selected according to the actual situation to improve the data processing effect. The spectral data after preprocessing has higher quality, reduces the error in subsequent analysis, ensures the reliability of the spectral data, and provides strong support for accurate detection of bauxite composition.

[0068] In some embodiments, the noise removal operation described above includes wavelet transform and / or Savitzky-Golay.

[0069] In some embodiments, both wavelet transform and Savitzky-Golay filtering method can effectively remove the noise in the spectral data and improve the signal quality. Wavelet transform is efficient in processing time and frequency information, and Savitzky-Golay filtering method is excellent in retaining signal peaks. The combination of the two can retain the main characteristics of the spectral signal while removing noise. According to the characteristics of the actual spectral data, wavelet transform and / or Savitzky-Golay filtering method can be selected to flexibly perform noise removal operation and improve the preprocessing effect.

[0070] In some embodiments, the optimal model is a partial least squares regression model, and the optimal model is an R 2 The model is evaluated for detection accuracy and parameter optimization is performed using a grid search method.

[0071] In some embodiments, the spectral data and bauxite composition data are centralized and standardized, and a PLS regression method is used to establish a regression model between the input variable (spectral data) and the output variable (composition data). R 2 As an evaluation index, the detection accuracy of the PLS regression model is measured. The R 2 value is calculated to evaluate the explanatory power and accuracy of the model for bauxite composition detection. The parameter grid of the PLS regression model (such as the number of latent variables) is defined, and the model performance of each parameter combination is evaluated by cross-validation. The parameter combination with the highest R 2 value is selected to obtain the optimal PLS regression model.

[0072] In summary, the embodiments of the present disclosure establish the relationship between the spectral data and the composition data through the PLS regression model, and the R 2The detection accuracy of the model is evaluated to ensure that the model has high-precision component detection capability. The grid search method is used for parameter tuning, which systematically traverses the parameter combinations to ensure that the optimal PLS regression model is obtained, thereby improving the reliability and stability of the model. The PLS regression model can effectively process high-dimensional spectral data, solve the problem of multicollinearity, and is suitable for complex bauxite component detection.

[0073] In some embodiments, the parameters optimized by the above parameter tuning process include the number of principal components, the maximum number of iterations, and the convergence criteria.

[0074] In some embodiments, the number of principal components is an important parameter in partial least squares regression (PLS regression), which determines the number of latent variables extracted from the original spectral data. Selecting an appropriate number of principal components can balance the complexity and interpretability of the model, avoiding overfitting or underfitting.

[0075] The maximum number of iterations determines the upper limit of the number of iterations in the training process of the PLS regression model. Setting a reasonable maximum number of iterations can ensure that the model is fully optimized during the training process, but will not waste computational resources due to excessive iterations.

[0076] The convergence criteria are used to determine whether the model training process has reached an optimal state, usually measured by the change in the loss function. Setting appropriate convergence criteria can avoid the model falling into a local optimal solution, while ensuring that the training process is completed within a reasonable time.

[0077] Through the grid search method, different numbers of principal components are tried, and the number of principal components that can maximize the R 2 value is selected. This ensures that the model can fully extract the features of the spectral data while avoiding overfitting. The maximum number of iterations is set within a reasonable range, and the number of iterations that optimizes the performance of the model is selected through cross-validation, ensuring that the model is fully optimized during the training process. The threshold of the convergence criteria is adjusted, and the most suitable convergence criteria are selected through cross-validation, ensuring that the model can effectively converge during the training process and avoiding falling into a local optimal solution.

[0078] In summary, the method proposed in the embodiments of the present disclosure ensures that the PLS regression model has optimal performance on high-dimensional spectral data through systematic parameter tuning, improving the precision and reliability of bauxite component detection. The parameter range and value in the tuning process can be adjusted according to the specific characteristics of the spectral data, with high flexibility and adaptability. By reasonably setting the maximum number of iterations and the convergence criteria, the model converges to the optimal solution within a reasonable time, improving the training efficiency and model stability.

[0079] In some examples, a LIBS system as shown in FIG. 2 can be employed, which includes a laser, a spectrometer, a detector, and a delay timer. The wavelength of the laser can be 1064 nm, the single pulse energy can be 100 mJ, and the repetition frequency can be 10 Hz; the spectrometer can be a two-channel spectrometer, the wavelength range can be 200 nm-500 nm, and the resolution can be 0.1 nm; the number of pixels of the detector can be 2048; and the delay time of the delay timer can be 300 ns. The detection of the Al, Fe, and Si elements in bauxite is achieved.

[0080] The detection of the Al, Fe, and Si elements in bauxite includes steps S201-S211.

[0081] Step S201: n1 (which can be 30) component different block bauxite samples are prepared, grinding and tabletting are performed, and n1 standard spectral data are collected to obtain a standard spectral data set.

[0082] In step S201, the particle size of the selected bauxite sample is <20 mm; the bauxite sample is ground using a mortar for 2 minutes to prevent moisture absorption, the particle size is controlled to be below 2 μm, and a tablet sample is prepared by pressing at a pressure of 15 MPa for 30 seconds; the LIBS system shown in FIG. 2 is used to collect bauxite spectral data, and during the collection process, each sample is excited 10 times, the average of the 10 spectral data is taken as the spectral data of a single sample, and the data size of the bauxite spectral data set is (4096, 30).

[0083] Step S202: The standard spectral data are preprocessed to reduce baseline drift and continuous background interference in the standard spectral data, and a preprocessed bauxite spectral data set is obtained. Baseline drift is a shift on the overall spectrum caused by factors such as instrument drift and light source fluctuation, and continuous background interference is a background signal of the spectrum caused by factors such as sample scattering and absorption. The first derivative is used to reduce the baseline drift in the spectrum, and by reducing the overall slope and fluctuation of the spectrum, the influence of the baseline drift on the spectrum is reduced. The Savitzky-Golay method is used to reduce the continuous background interference, the data in the sliding window are weighted and filtered, the change information of the signal is effectively retained while the filtering is smoothed, the window width is 5, and the polynomial order is 3. The result of the preprocessing is shown in FIG. 3.

[0084] Step S203: The average of the spectral data obtained in steps S201 and S202 is taken as the bauxite category spectrum S. The spectral data obtained in steps S101 and S102 are 30, and are denoted as {I1, I2, …, I 30}, and the bauxite category spectrum S is calculated by the following formula: FIG. 4.

[0085] S represents the sum of a certain calculation result. I i represents the intensity value of the ith spectral data, a total of 30 data, summed to obtain the value of S.

[0086] Step S204: The bauxite with known composition is transmitted at a uniform speed of 1.5 m / s from the LIBS system shown in Figure 2, the spectral data of the bauxite is collected, and S102 is performed to obtain the preprocessed spectral data set.

[0087] Step S205: Calculate the spectral distance L between each spectral data in the preprocessed spectral data set of S204 and S of S203, remove the spectral data with L>1.0, remove the unqualified spectral data due to sample gap, ensure the validity of the data, and obtain the spectral data set.

[0088] The distance L is calculated as follows:

[0089] L(S', S) represents the distance or difference between two sets of spectral data S' and S. S' represents the preprocessed spectral data set. S represents the original spectral data set.

[0090] In step S206: The first m (which can be 10) spectral data of the spectral data set obtained in S105 are selected and averaged as the 1-time transmission spectral data. The belt transmission time of 1-time transmission is set to 2s, that is, the continuous placement distance of the sample is 3m.

[0091] In step S207: Repeat S204-206 for 90 times to obtain the modeling data set, and divide the modeling data set into training set and validation set according to 7:3. The data size of the modeling data set is (4096, 90).

[0092] In step S208: Use the training set and the validation set to establish the partial least squares regression model (PLSR) for detecting the components of Al, Fe and Si, and use the determination coefficient R 2 to evaluate the detection accuracy.

[0093] Determination coefficient R 2 , which represents the fitting degree of the detection model to the data set. The higher the determination coefficient, the better the fitting of the detection model, and R 2 is calculated as follows:

[0094] Wherein, R 2 is the determination coefficient, which is used to evaluate the fitting effect of the model.

[0095] Y 真实,i represents the true value of the ith sample. Y 预测,ipredi represents the predicted value of the ith sample.

[0096] Step S209: Repeat step S208, and optimize the parameters of PLSR using the grid search method until the R2 of the detection model reaches the optimum, and save the optimal model.

[0097] The parameters of PLSR that need to be optimized include the number of principal components, the maximum number of iterations, and the convergence standard. The number of principal components includes {2, 3, …, 200}, the maximum number of iterations includes {1e3, 1e4, 1e5, 1e6}, and the convergence standard is set to 1e-4.

[0098] In the process of grid search, the training set is divided into a parameter optimization training set and a parameter optimization validation set according to 4:1, and grid search is performed. The optimal parameters of the Al element composition detection model obtained by optimization include the number of principal components 12, the maximum number of iterations 1e 3 , and the convergence standard 1e -4 . The optimal parameters of the Fe element composition detection model obtained by optimization include the number of principal components 13, the maximum number of iterations 1e 3 , and the convergence standard 1e -4 . The optimal parameters of the Si element composition detection model obtained by optimization include the number of principal components 16, the maximum number of iterations 1e 5 , and the convergence standard 1e -4 .

[0099] Step S210: Repeat steps S204 and S205, and take the average of every 10 spectral data as continuous transmission spectral data.

[0100] The belt transmission time of continuous transmission can be set to 4s, that is, the continuous placement distance of the sample is 6m. Since the length of the transmission belt is less than 6m, the blocky bauxite sample is continuously added to the transmission belt during the belt transmission.

[0101] The data size of the continuous transmission spectral data is (4096, n3). Since step S210 is a test for bauxite composition detection in a real production environment, n3 is not a fixed value, but n3≥2 must be ensured in the test. Assuming that the spectral data set obtained in step 105 contains n qualified spectra, n3=n / m.

[0102] In step S211: read the optimal model, input each spectral data of the continuous transmission spectral data into the optimal model, obtain the composition of Al, Fe, and Si in the bauxite for each spectral data, and take the average of the obtained bauxite composition to obtain the composition detection value of Al, Fe, and Si in the continuous transmission bauxite.

[0103] FIG. 5 is a schematic diagram of the present application for 1-time continuous transmission bauxite composition detection. A total of 40 spectral data are collected, and the distance from S is calculated after spectral preprocessing. All 40 data meet the requirements. The average of every 10 spectra is taken to obtain spectra 1-4 in FIG. 5. The composition 1-4 corresponding to the 4 spectra is obtained by model operation. The Al, Fe, and Si element compositions are averaged to obtain the 1-time continuous transmission bauxite composition detection value.

[0104] In summary, the disclosed bauxite composition online detection method and system of the present application is different from the traditional composition detection. Based on the laser-induced breakdown spectroscopy technology and machine learning technology, the bauxite composition information can be detected in real time without sampling and offline analysis. The present application constructs the bauxite category spectrum, filters out the unqualified spectral data due to sample gap, and improves the detection robustness of the blocky sample on the production line. The spectral preprocessing technology is used to eliminate the baseline shift and background noise interference to a certain extent. The 3-level average strategy is adopted to increase the data representativeness of the detection and improve the accuracy of the bauxite composition online detection analysis, thereby promoting the application of the online detection technology.

[0105] As shown in FIG. 6, the present application proposes a bauxite composition online detection device, which comprises:

[0106] The acquisition unit 21 is used to control the LIBS system to collect n actual spectral data on the bauxite to be detected on the preset speed conveyor belt.

[0107] The preprocessing unit 22 is used to perform preprocessing operation on the n actual spectral data to obtain a preprocessed actual spectral data set.

[0108] The screening unit 23 is used to calculate the screening distance between each preprocessed actual spectral data in the preprocessed actual spectral data set and the bauxite category spectrum, and add the preprocessed actual spectral data with a screening distance less than or equal to a preset distance to the effective spectral data set. The bauxite category spectrum is a standard sample with known composition measured based on a static mode.

[0109] The averaging unit 24 is used to take the average value of the first m spectral data in the effective spectral data set as the actual continuous transmission spectral data.

[0110] The acquisition unit 25 is used to input the actual continuous transmission spectral data into an optimal model to obtain the bauxite composition corresponding to each spectral data. The optimal model is trained based on the continuous transmission spectral data of the standard sample and the bauxite composition corresponding to the continuous transmission spectral data.

[0111] As shown in FIG. 7, the electronic device 300 according to the embodiments of the present disclosure includes a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. The processor 320 implements the steps of any method for online detection of bauxite composition described above when executing the computer program 311.

[0112] In summary, the method for online detection of bauxite composition based on LIBS according to some embodiments of the present disclosure includes: controlling a LIBS system to collect n actual spectral data on a bauxite to be detected on a preset speed conveyor belt; performing a preprocessing operation on the n actual spectral data to obtain a set of preprocessed actual spectral data; calculating a screening distance between each preprocessed actual spectral data in the set of preprocessed actual spectral data and an aluminum soil category spectrum, adding the preprocessed actual spectral data with a screening distance less than or equal to a preset distance to a set of valid spectral data, and the aluminum soil category spectrum is a standard sample with known composition measured based on a static mode; taking the average of the first m spectral data in the set of valid spectral data as actual continuous driving spectral data; inputting the actual continuous driving spectral data into an optimal model to obtain the bauxite composition corresponding to each spectral data, and the optimal model is based on standard continuous driving spectral data, and the standard continuous driving spectral data is the average of the first m spectral data obtained based on the LIBS system collecting the standard sample on the preset speed conveyor belt and performing the preprocessing operation and the screening distance operation. The method of the related art requires manual sampling and laboratory analysis, has a long time interval, and the analysis result lags behind. The present disclosure realizes online real-time detection of bauxite composition through LIBS technology, can obtain data in time, and guides the production process. Through preprocessing and screening operations, the quality and accuracy of the spectral data are ensured, human error and external interference are reduced, and more accurate detection results are provided. Compared with the PGNAA method, the LIBS technology avoids radiation risk and ensures the safety of the operator. The LIBS technology can be used for detection of solid, liquid, and gas samples, and does not require complex sample preparation, and is suitable for various industrial environments. Through online detection and intelligent control, the labor cost and laboratory analysis cost are reduced, the production efficiency and product quality are improved, and the overall production cost is reduced. The present scheme provides a reliable online detection method for the production process of alumina, promotes the automation, informatization, and intelligent development of the production process, and improves the competitiveness of the industry.

[0113] Since the electronic device described in the embodiment is the device used in the implementation of the bauxite composition online detection device in the embodiment of the present disclosure, based on the method described in the embodiment of the present disclosure, those skilled in the art can understand the specific implementation of the electronic device in the embodiment and its various forms, so the implementation of the electronic device in the embodiment of the present disclosure is not described in detail, as long as the device used in the implementation of the method in the embodiment of the present disclosure belongs to the scope of the present disclosure.

[0114] In the implementation process, the computer program 311 can implement any implementation in the embodiment corresponding to FIG. 1 when executed by the processor.

[0115] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0116] Those skilled in the art should understand that the embodiments of the present disclosure can be provided as a method, a system, or a computer program product. Therefore, the present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present disclosure can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0117] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices produce a device for implementing the functions specified in one or more flows in the flowcharts and / or one or more blocks in the block diagrams.

[0118] These computer program instructions can also be stored in a computer-readable storage medium that can guide the computer or other programmable data processing devices to work in a specific way, so that the instructions stored in the computer-readable storage medium produce a product including instruction devices, which implement the functions specified in one or more flows in the flowcharts and / or one or more blocks in the block diagrams.

[0119] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart and / or block diagram block or blocks.

[0120] The embodiments of the present disclosure further provide a computer program product, which comprises computer software instructions, and when the computer software instructions are run on a processing device, the processing device executes the process of on-line detection of bauxite composition in the corresponding embodiments.

[0121] The computer program product comprises one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the embodiments of the present disclosure is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer readable storage medium can be any available medium that the computer can store or the data storage device such as server, data center, etc. integrated with one or more available media sets. The available media can be magnetic media (for example, floppy disk, hard disk, magnetic tape), optical media (for example, DVD), or semiconductor media (for example, solid state disk (SSD)) and the like.

[0122] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0123] In several embodiments provided by the present disclosure, it should be understood that the disclosed system, apparatus and method can be implemented in other manners. For example, the apparatus embodiments described above are merely schematic. For example, the division of the units is only a logical function division. For another example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different units, can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electric, mechanical or other form.

[0124] The units described as separated components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0125] In addition, each functional unit in the various embodiments of the present disclosure can be integrated in one processing unit, or each unit can exist physically as separate units, or two or more units can be integrated in one unit. The integrated unit can be implemented in the form of hardware, or in the form of software functional units.

[0126] If the integrated unit is implemented in the form of software functional units and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such an understanding, the technical solutions of the present disclosure essentially, or the part that contributes to the prior art, or all or a part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in the various embodiments of the present disclosure. The foregoing storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and various media that can store program codes.

[0127] The above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than limit them. Although the present disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. A method for online detection of bauxite composition, comprising: controlling a LIBS system to collect n actual spectral data on a bauxite to be detected on a preset speed conveyor belt; performing a preprocessing operation on the n actual spectral data to obtain a preprocessed actual spectral data set; calculating a screening distance between each preprocessed actual spectral data in the preprocessed actual spectral data set and a bauxite category spectrum, and adding the preprocessed actual spectral data with a screening distance less than or equal to a preset distance to a valid spectral data set, wherein the bauxite category spectrum is obtained by measuring a standard sample with known composition in a static manner; taking an average of the first m spectral data in the valid spectral data set as actual continuous driving spectral data; and inputting the actual continuous driving spectral data into an optimal model to obtain a bauxite composition corresponding to each spectral data, wherein the optimal model is trained based on continuous driving spectral data of a standard sample and a bauxite composition corresponding to the continuous driving spectral data.

2. The method for online detection of bauxite composition according to claim 1, wherein the optimal model is obtained by the following steps: controlling the LIBS system to collect m standard spectral data of a standard sample on the preset speed conveyor belt, and obtaining a bauxite composition of the standard sample; performing a preprocessing operation and a screening distance operation on the m standard spectral data of the standard sample, and taking an average of the first m spectral data after the operation as continuous driving spectral data of the standard sample; training based on the continuous driving spectral data of the standard sample and the bauxite composition of the standard sample to obtain the optimal model.

3. The method for online detection of bauxite composition according to claim 1, further comprising: preparing n1 block bauxite samples with different compositions and performing grinding and tabletting treatment to form the standard sample; controlling the LIBS system to collect n1 standard spectral data of the standard sample in a static state; performing a preprocessing operation on the n1 standard spectral data to obtain a preprocessed standard spectral data set; and taking an average of the spectral data in the preprocessed standard spectral data set to obtain the bauxite category spectrum.

4. The method for online detection of bauxite composition according to claim 3, wherein the preparation of n1 block bauxite samples with different compositions and the grinding and tabletting treatment to form the standard sample comprises: the block bauxite samples with different compositions refer to different contents of elements contained in the bauxite; the particle size of the block bauxite sample is less than 20 mm, and n1 is greater than or equal to 30; the grinding and tabletting treatment comprises: performing a grinding operation with a mortar for a longest time of less than 2 minutes to form a granular bauxite sample with a particle size of less than or equal to 2 μm; pressing the granular bauxite sample under a pressure of at least 15 MPa for at least 30 seconds to form the block bauxite sample. the preprocessing operation comprises a baseline correction operation and a noise removal operation; the baseline correction operation comprises one or more of derivative method, iterative polynomial fitting method and penalty least square method.

5. The bauxite composition on-line detection method according to any one of claims 1 to 4, wherein, ​ 6. The bauxite composition on-line detection method according to claim 5, wherein, The noise removal operation includes wavelet transform and / or Savitzky-Golay.

7. The bauxite composition on-line detection method according to claim 1, wherein, The optimal model is a partial least squares regression model, and the optimal model is a model that uses R2 to evaluate detection accuracy and uses a grid search method to optimize parameters.

8. The method of claim 7, wherein, The parameters corresponding to the parameter optimization include the number of principal components, the maximum number of iterations, and the convergence standard.

9. An online bauxite composition detection device, comprising: a collection unit configured to control a LIBS system to collect n actual spectral data on a bauxite to be detected on a preset speed conveyor belt; a preprocessing unit configured to perform a preprocessing operation on the n actual spectral data to obtain a preprocessed actual spectral data set; a screening unit configured to calculate a screening distance between each preprocessed actual spectral data in the preprocessed actual spectral data set and a bauxite category spectrum, and add preprocessed actual spectral data with a screening distance less than or equal to a preset distance to an effective spectral data set, wherein the bauxite category spectrum is obtained based on a static measurement of a standard sample with known composition; an averaging unit configured to average the first m spectral data in the effective spectral data set as actual continuous transmission spectral data; and an acquisition unit configured to input the actual continuous transmission spectral data to an optimal model to obtain a bauxite composition corresponding to each spectral data, wherein the optimal model is trained based on continuous transmission spectral data of a standard sample and a bauxite composition corresponding to the continuous transmission spectral data. a memory and a processor, wherein the processor is configured to implement the steps of the bauxite composition online detection method of any one of claims 1-8 when executing a computer program stored in the memory.

10. An electronic device comprising:

11. A computer readable storage medium comprising a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of the bauxite composition online detection method of any one of claims 1-8. ​

Images