CN117092050A - Coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning - Google Patents
Coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning Download PDFInfo
- Publication number
- CN117092050A CN117092050A CN202311176743.XA CN202311176743A CN117092050A CN 117092050 A CN117092050 A CN 117092050A CN 202311176743 A CN202311176743 A CN 202311176743A CN 117092050 A CN117092050 A CN 117092050A
- Authority
- CN
- China
- Prior art keywords
- coal slime
- slime flotation
- data
- mixture sample
- spectrum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 239000003245 coal Substances 0.000 title claims abstract description 233
- 238000005188 flotation Methods 0.000 title claims abstract description 209
- 238000001228 spectrum Methods 0.000 title claims abstract description 93
- 238000001514 detection method Methods 0.000 title claims abstract description 48
- 239000000203 mixture Substances 0.000 claims abstract description 158
- 230000004927 fusion Effects 0.000 claims abstract description 57
- 230000003595 spectral effect Effects 0.000 claims abstract description 48
- 238000000034 method Methods 0.000 claims abstract description 35
- 239000013307 optical fiber Substances 0.000 claims abstract description 35
- 238000002835 absorbance Methods 0.000 claims abstract description 28
- 238000012545 processing Methods 0.000 claims abstract description 8
- 238000000605 extraction Methods 0.000 claims description 23
- 238000011176 pooling Methods 0.000 claims description 16
- 239000010453 quartz Substances 0.000 claims description 11
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N silicon dioxide Inorganic materials O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000003756 stirring Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 5
- 229910052724 xenon Inorganic materials 0.000 claims description 4
- FHNFHKCVQCLJFQ-UHFFFAOYSA-N xenon atom Chemical compound [Xe] FHNFHKCVQCLJFQ-UHFFFAOYSA-N 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000002834 transmittance Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 230000009467 reduction Effects 0.000 claims description 2
- 238000007499 fusion processing Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 10
- 238000013135 deep learning Methods 0.000 abstract description 6
- 238000010183 spectrum analysis Methods 0.000 abstract description 6
- 238000013527 convolutional neural network Methods 0.000 description 15
- 230000006870 function Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 238000005303 weighing Methods 0.000 description 5
- 238000012856 packing Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 239000003795 chemical substances by application Substances 0.000 description 3
- 239000003250 coal slurry Substances 0.000 description 3
- 239000003814 drug Substances 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 239000006260 foam Substances 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 239000002245 particle Substances 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000010521 absorption reaction Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000009614 chemical analysis method Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000002411 thermogravimetry Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000005273 aeration Methods 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 239000002817 coal dust Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000004088 foaming agent Substances 0.000 description 1
- 230000002209 hydrophobic effect Effects 0.000 description 1
- 229910003480 inorganic solid Inorganic materials 0.000 description 1
- 150000002500 ions Chemical class 0.000 description 1
- 230000031700 light absorption Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/28—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q
- G01N1/38—Diluting, dispersing or mixing samples
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/255—Details, e.g. use of specially adapted sources, lighting or optical systems
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/59—Transmissivity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/20—Identification of molecular entities, parts thereof or of chemical compositions
Abstract
The application discloses a coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning, wherein the method comprises the following steps: selecting a coal slime flotation mixture sample; performing absorbance detection on the coal slime flotation mixture sample by using an optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample; constructing a coal slime flotation multi-scale feature fusion model; and carrying out ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain the ash value of the coal slime flotation tailings. By using the application, the high-speed processing of the spectrum data can be realized by utilizing the deep learning and spectrum analysis technology, thereby realizing the rapid detection of the ash value of the coal slime flotation tailings. The method and the system for detecting the coal slime flotation ash based on spectrum multi-mode time sequence learning can be widely applied to the technical field of coal slime flotation ash detection.
Description
Technical Field
The application relates to the technical field of coal slime flotation ash content detection, in particular to a coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning.
Background
The coal slime flotation is an important means for coal production, and the tailing ash information not only reflects the ash data of clean coal, but also reflects the recovery rate information of clean coal. The existing detection method commonly used in the industrial field is to calculate ash content by measuring the weight of residual substances through burning tailings, the detection time is long, or the tailings are floated by touching by hands depending on the observation of a worker according to experience, the method has high randomness, the experience requirement on the worker is high, and the existing technology for detecting the ash content of the coal slime flotation tailings mainly comprises a manual screening method, a microscopic method, a thermogravimetric analysis method, a chemical analysis method and the like. Although the technology is widely applied to the detection of ash values of the coal slime flotation tailings, the technology still has some objective defects that firstly, the manual operation error is large, the manual screening method and the microscopic method need manual operation, the manual screening method and the microscopic method are greatly influenced by human factors in the operation process, the error is large, secondly, the detection speed is slow, each sample needs to be detected one by the manual screening method and the microscopic method, the detection speed is slow, the high-efficiency and quick detection requirements cannot be met, thirdly, the detection precision is low, the precision of the thermogravimetric analysis method and the chemical analysis method is high, but sample treatment, chemical reagents and the like are needed, the operation is complex, the preparation and the treatment process of the samples possibly influence the precision of the test result, fourthly, the time consumption and the cost are high, the detection of the ash values of the coal slime flotation tailings generally needs to be tested on a large number of samples, and the prior art needs to consume a large amount of time and cost, and the detection requirements of high efficiency and low cost cannot be met; therefore, it is necessary to develop a more efficient, convenient, accurate and low-cost technology for detecting ash values of the coal slime flotation tailings so as to meet the demands of the fields of production, environmental protection and the like
Disclosure of Invention
In order to solve the technical problems, the application aims to provide a coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning, which can realize high-speed processing of spectrum data by utilizing deep learning and spectrum analysis technology, thereby realizing rapid detection of coal slime flotation tailing ash content.
The first technical scheme adopted by the application is as follows: the coal slime flotation ash content detection method based on spectrum multi-mode time sequence learning comprises the following steps:
selecting a coal slime flotation mixture sample;
performing absorbance detection on the coal slime flotation mixture sample by using an optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample;
constructing a coal slime flotation multi-scale feature fusion model;
and carrying out ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain the ash value of the coal slime flotation tailings.
Further, the step of selecting a sample of the coal slime flotation mixture specifically comprises the following steps:
collecting a certain mass of coal slime mixture with different ash values;
injecting water with a certain mass into the coal slime mixture, and stirring the mixture by a magnetic stirrer to obtain a stirred coal slime mixture;
and sucking a certain mass of the stirred coal slime mixture into a quartz cuvette through a Pasteur pipette, and preparing and obtaining a coal slime flotation mixture sample.
Further, the step of detecting absorbance of the coal slime flotation mixture sample by an optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample specifically comprises the following steps:
based on a transmissivity measuring bracket of the optical fiber spectrometer, isolating a xenon lamp light source in the transmissivity measuring bracket of the optical fiber spectrometer to obtain background light data;
placing the empty quartz cuvette in a transmissivity measuring bracket of the optical fiber spectrometer to obtain reference light data;
placing a quartz cuvette with a coal slime flotation mixture sample in a transmissivity measuring bracket of an optical fiber spectrometer to obtain sample light data;
and calculating by combining the background light data, the reference light data and the sample light data to obtain the spectrum data of the coal slime flotation mixture sample.
Further, the method also comprises the step of circularly acquiring the background light data, the reference light data and the sample light data based on different moments to obtain a plurality of groups of spectrum data of the coal slime flotation mixture samples at different moments.
Further, the calculation expression of the spectrum data of the coal slime flotation mixture sample is specifically shown as follows:
T＝-log 10 I f
in the above formula, T represents the absorbance of a sample of the coal slime flotation mixture, I f Indicating transmittance, I sample Representing sample light data, I off Representing background light data, I sub Representing reference light data.
Further, the coal slime flotation multiscale feature fusion model comprises a data preprocessing module, a feature extraction module, a feature fusion module and a feature classification module, wherein:
the feature extraction module comprises a CNN+LSTM layer and a multi-scale fusion layer, wherein the CNN network layer comprises a one-dimensional convolution layer, a BN+ReLU layer and a maximum pooling layer, and the multi-scale fusion layer comprises a first one-dimensional convolution layer, a second one-dimensional convolution layer and a third one-dimensional convolution layer.
Further, the step of performing ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain the ash value of the coal slime flotation tailings specifically comprises the following steps:
spectral data of the coal slime flotation mixture sample is input into a coal slime flotation multiscale feature fusion model;
performing data format conversion treatment on the spectrum data of the coal slime flotation mixture sample based on the data pretreatment module to obtain pretreated spectrum data of the coal slime flotation mixture sample;
performing characteristic extraction treatment on the pretreated coal slime flotation mixture sample spectral data based on a characteristic extraction module to obtain coal slime flotation mixture sample spectral characteristic data;
carrying out characteristic fusion treatment on the spectral characteristic data of the coal slime flotation mixture sample based on the characteristic fusion module to obtain the spectral characteristic fusion data of the coal slime flotation mixture sample;
and carrying out characteristic classification treatment on the spectral characteristic fusion data of the sample of the coal slime flotation mixture based on the characteristic classification module to obtain the coal slime flotation tailing ash value.
Further, the spectral feature data of the sample of the coal slime flotation mixture includes the spectral feature data of the sample of the first coal slime flotation mixture and the spectral feature data of the sample of the second coal slime flotation mixture, and the step of performing feature extraction treatment on the spectral feature data of the sample of the pretreated coal slime flotation mixture based on the feature extraction module to obtain the spectral feature data of the sample of the coal slime flotation mixture specifically includes:
the pretreated coal slime flotation mixture sample spectrum data are input to a feature extraction module;
carrying out convolution operation treatment on the pretreated coal slime flotation mixture sample spectrum data based on a one-dimensional convolution layer of the CNN network layer to obtain different position characteristics of the coal slime flotation mixture sample spectrum data;
sequentially carrying out normalization treatment and nonlinear transformation treatment on different position characteristics of the coal slime flotation mixture sample spectrum data based on a BN+ReLU layer of a CNN network layer to obtain transformed coal slime flotation mixture sample spectrum data;
carrying out pooling treatment on the transformed coal slime flotation mixture sample spectrum data based on the largest pooling layer of the CNN network layer to obtain the dimension-reduced coal slime flotation mixture sample spectrum data;
capturing and updating the spectrum data of the sample of the coal slime flotation mixture after the dimension reduction based on the LSTM network layer to obtain the spectrum characteristic data of the sample of the first coal slime flotation mixture;
sequentially carrying out pooling treatment, convolution treatment, BN operation and ReLU operation on the pretreated coal slime flotation mixture sample spectrum data based on the first one-dimensional convolution layer of the multi-scale fusion layer to obtain first multi-scale coal slime flotation mixture sample spectrum data;
carrying out convolution treatment, BN operation, reLU operation, convolution treatment, BN operation and ReLU operation on the pretreated coal slime flotation mixture sample spectrum data in sequence based on a second one-dimensional convolution layer of the multi-scale fusion layer to obtain second multi-scale coal slime flotation mixture sample spectrum data;
carrying out convolution treatment, BN operation, reLU operation, convolution treatment, BN operation and ReLU operation on the pretreated coal slime flotation mixture sample spectrum data in sequence based on a third one-dimensional convolution layer of the multi-scale fusion layer to obtain third multi-scale coal slime flotation mixture sample spectrum data;
and performing splicing treatment on the first multi-scale coal slime flotation mixture sample spectral data, the second multi-scale coal slime flotation mixture sample spectral data and the third multi-scale coal slime flotation mixture sample spectral data to obtain second coal slime flotation mixture sample spectral characteristic data.
Further, the step of performing characteristic fusion treatment on the spectral characteristic data of the sample of the coal slime flotation mixture based on the characteristic fusion module to obtain the spectral characteristic fusion data of the sample of the coal slime flotation mixture specifically comprises the following steps:
performing dimension-increasing operation treatment on the spectral characteristic data of the first coal slime flotation mixture sample and the spectral characteristic data of the second coal slime flotation mixture sample to obtain the spectral characteristic data of the first coal slime flotation mixture sample after dimension increase and the spectral characteristic data of the second coal slime flotation mixture sample after dimension increase;
and performing characteristic outer product calculation on the spectrum characteristic data of the first coal slime flotation mixture sample after the dimension increase and the spectrum characteristic data of the second coal slime flotation mixture sample after the dimension increase to obtain spectrum characteristic fusion data of the coal slime flotation mixture sample.
The second technical scheme adopted by the application is as follows: coal slime flotation ash content detecting system based on spectrum multimode time sequence study includes:
the selecting module is used for selecting a coal slime flotation mixture sample;
the absorbance detection module is used for detecting absorbance of the coal slime flotation mixture sample through the optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample;
the construction module is used for constructing a coal slime flotation multiscale feature fusion model;
and the ash detection module is used for carrying out ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain the ash value of the coal slime flotation tailings.
The method and the system have the beneficial effects that: according to the application, the absorbance detection is carried out on the coal slime flotation mixture sample through the optical fiber spectrometer, so that the high-speed processing of the coal slime flotation multi-scale feature fusion model can be realized, the ash detection is carried out on the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model, the deep learning model can effectively learn the features and rules of the spectrum data, and therefore, the coal slime flotation tailing ash value can be accurately predicted, and compared with the traditional detection method, the accuracy is higher, and the rapid detection of the coal slime flotation tailing ash value is realized.
Drawings
FIG. 1 is a flow chart of steps of a coal slime flotation ash detection method based on spectrum multi-mode time sequence learning;
FIG. 2 is a block diagram of a system for detecting ash in coal slime flotation based on spectral multi-modal time series learning;
FIG. 3 is a schematic flow diagram of a prior art clean coal flotation step;
FIG. 4 is a schematic diagram of a magnetic stirrer stirring a beaker sample in accordance with an embodiment of the present application;
FIG. 5 is a schematic representation of different ash value quartz cuvette samples according to an embodiment of the application;
FIG. 6 is a schematic diagram of a coal slime flotation multiscale feature fusion model constructed in accordance with the present application;
FIG. 7 is a schematic diagram of a feature extraction module in a coal slime flotation multiscale feature fusion model constructed in the present application;
fig. 8 is a schematic flow chart of the characteristic fusion treatment of the spectral characteristic data of the coal slime flotation mixture sample according to the present application.
Detailed Description
The application will now be described in further detail with reference to the drawings and to specific examples. The step numbers in the following embodiments are set for convenience of illustration only, and the order between the steps is not limited in any way, and the execution order of the steps in the embodiments may be adaptively adjusted according to the understanding of those skilled in the art.
The technical terms of the present application are explained:
(1) And (3) coal slime flotation: the coal slime is an accessory product generated in the working process of the coal washery, and is a semi-solid object formed by mixing coal dust and water, but the quality of the coal is different, the quality of the coal slime is greatly different, the availability is also greatly different, and the coal slime is various and has wide application;
(2) The steps of the clean coal floatation are as follows: referring to fig. 3, first coal slurry enters a stirring barrel in the form of ore pulp, and is prepared into proper concentration; fully stirring after further adding the medicament, wherein the medicament is a foaming agent and a collecting agent; the stirred coal slurry enters a flotation machine; secondly, the impeller rotates to generate strong stirring and aeration; finally, a large number of bubbles with different sizes are generated in the ore pulp, hydrophobic coal particles are attached to the bubbles due to the absorption agent (collector), the bubbles are brought to the ore pulp surface to be gathered into a mineralized foam layer, the mineralized foam layer is scraped by a foam scraper to be used as clean coal, and when hydrophilic gangue particles do not act with the agent and do not adhere to the bubbles, the hydrophilic gangue particles remain in the ore pulp to become flotation tailings;
(3) Tailing ash: ash, also called ash yield, refers to the mass ratio of coal per unit mass to inorganic solids left after it has been fully combusted. m1 is the mass of the residual quantity, m2 is the mass of the coal sample, and ash is equal to m1/m2;
(4) Optical fiber spectrometer: the optical fiber spectrometer is an instrument for realizing spectrum analysis by utilizing optical fibers to transmit optical signals, compared with the traditional spectrometer, the optical fiber spectrometer adopts optical fibers to isolate a sample from the instrument, so that the direct contact between the sample and the instrument is avoided, the risks of sample pollution and instrument damage are reduced, the optical fiber spectrometer can measure color bands obtained after light rays with different wavelengths are separated by the optical fiber spectrometer, and thus the information of chemical composition, concentration, structure and the like of the sample is determined;
(5) The measurement flow of the optical fiber spectrometer comprises the following steps: the light source generates light, the light is transmitted to the sample through the optical fiber, the sample absorbs part of the light, and the rest of the light is transmitted to the beam splitter through the optical fiber. The optical splitter separates the incident light into a plurality of color bands according to the wavelength, the color bands are transmitted to the detector through the optical fiber, the detector converts the optical signal into an electric signal, and spectral data is output after the processing such as amplification, filtering, digitization and the like;
(6) Absorbance: the absorbance refers to the degree of absorption of light with a specific wavelength by a sample, when the light passes through the sample, molecules or ions in the sample absorb part of the light, the rest of the light passes through the sample and then enters a detector for detection, and the higher the absorbance is, the stronger the absorption of the light by the sample is, the less the rest of the light is, and the smaller the electric signal detected by the detector is;
(7) One-dimensional convolutional neural network: a one-dimensional convolutional neural network is a deep learning model, which is generally used for processing time-series data or one-dimensional signal data, and is composed of a series of convolutional layers, pooling layers and full-connection layers, and is used for extracting features from input data and performing classification or regression tasks, unlike a conventional Convolutional Neural Network (CNN), which performs a convolutional operation on only one dimension of the input data. They are typically used to process time series data such as text, audio, stock prices, etc., the basic structure of a one-dimensional convolutional neural network is similar to that of a conventional CNN, but the convolutional kernels used in the convolutional layers typically slide in only one dimension, and one-dimensional convolutional neural networks can also use different types of pooling layers, such as max-pooling or average-pooling, to reduce the size of the feature map.
Referring to fig. 1, the application provides a coal slime flotation ash detection method based on spectrum multi-mode time sequence learning, which comprises the following steps:
s1, selecting a coal slime flotation mixture sample;
specifically, preparing experimental equipment for selecting coal slime flotation mixture samples, wherein the experimental equipment comprises coal slime mixtures with different ash values, a plurality of quartz cuvettes, a plurality of Pasteur pipettes, a plurality of weighing papers, a plurality of beakers, a plurality of medicine spoons, a weighing instrument and a magnetic stirrer;
referring to fig. 4 and 5, a coal slime mixture with an ash value of 25.6 is prepared, weighing paper is put into a weighing instrument to peel, then the mixture is put into a weighing instrument to weigh 0.2g, then 0.2g of the mixture is poured into a beaker, 50ml of water is added, the beaker is placed into a magnetic stirrer to be fully stirred, after the stirring is completed, 2ml of the mixture is taken into a quartz cuvette by using a Pasteur pipette, so far 0.2g of the sample with the ash value of 25.6 is completely manufactured, and the steps are repeated to manufacture samples with the same sampling gram and different ash values.
S2, detecting absorbance of the coal slime flotation mixture sample through an optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample;
specifically, the components of the optical fiber spectrometer are composed of a Zeeman optical fiber spectrometer S5000-UV-VIS (200-900 nm), a Zeeman ultraviolet quartz optical fiber SMA905, a Zeeman double-pass cuvette bracket TH02 and a Zeeman pulse xenon lamp light source XE0512V 1A;
before sampling, firstly, measuring the background light once by isolating a xenon lamp light source, clicking the measured background light on the right column of software after an object is placed to block the light source, then, measuring the reference light once by placing an empty cuvette, and finally, placing a sample which is sufficiently and uniformly shaken for measurement, wherein the read data is target data: corresponding to the absorbance of the sample. The upper left corner file can be clicked to conduct file export, and file types of txt files, PDF files and EXCEL files can be selected;
in order to collect a large amount of data more rapidly to train a model, the python program can be used for continuous sampling, so that semi-automatic sampling can be realized, the efficiency is high, and the sampling is more accurate;
the operating principle of the optical fiber spectrometer combined with the python program is as follows:
the optical fiber spectrometer flow;
the serial port baud rate of the spectrometer is 115200 (factory default), and because the model of the spectrometer is S2000, the total value of the corresponding pixels is 2088, namely, absorbance data corresponding to 2088 groups of wavelengths can be received (the data form can be a. Txt file, the first column is wavelength, the second column is absorbance corresponding to wavelength, and each. Txt file has 2088 rows of data in total);
according to a serial port communication protocol, the received data of the device is original spectrum data, namely the most original data which is transmitted by a spectrometer and is not calculated, in order to obtain the absorbance of a sample, the first time of receiving the data is required to record background light (namely the original spectrum measured by the spectrometer when a light source is isolated), the second time of receiving the data is required to record reference light (namely the original spectrum measured by the spectrometer when an empty cuvette is placed) for subsequent operation, the third time of receiving the sample light (namely the original spectrum measured by the spectrometer when the sample is placed), and the absorbance of the target data is obtained after operation. Because the prepared sample is a solid-liquid mixture, the sample can be continuously settled along with time migration, the corresponding absorbance is different, and the absorbance data of the corresponding sample at different time is required to be collected so as to enrich the sample data and facilitate the subsequent experiments;
the calculation expression of the spectrum data of the coal slime flotation mixture sample is specifically shown as follows:
T＝-log 10 I f
in the above formula, T represents the absorbance of a sample of the coal slime flotation mixture, I f Indicating transmittance, I sample Representing sample light data, I off Representing background light data, I sub Representing reference light data;
python program usage flow;
clicking a program to run, wherein an operation window prompts 'please input background light', firstly isolating a light source, and then randomly inputting any digital carriage return in the operation window;
at the moment, the operation window prompts that reference light is required to be input, firstly, an empty cuvette is placed in a cuvette base, and then any number is input in the operation window and returned;
at the moment, the window prompts that 'please input sample information', firstly, samples are fully and evenly shaken and then are put into a cuvette base, and then corresponding sample information such as '0.2_25.6' is input into the operation window to be used as a sample label for subsequent experiments;
the program will automatically collect sample absorbance data of 0s,30s,60s,90s,300s in the format of a. Txt file, the file name is "0.2_25.6_0s_0", the first value "0.2" is the gram weight of the coal slurry mixture when it is sampled, the second value "25.6" is the sample ash value, the third value "0s" is the data of the sample collected at time 0, the fourth value "0" is the 1 st repeated sample when the first three values are the same
S3, referring to FIG. 6, constructing a coal slime flotation multiscale feature fusion model;
s4, carrying out ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain a coal slime flotation tailing ash value.
S41, preprocessing data;
specifically, at first, the data preprocessing stage is to perform model training preprocessing on the spectrum data acquired by the experiment. The same sample was input simultaneously with 5 sets of experimental spectral data (0 s,30s,60s,90 and 300 s) from 5 sets of individual 2088x1 (representing the number of sequence length x channels respectively) data to a set of combined 2088x5 (representing the number of sequence length x channels respectively) data formats. The benefits of this are mainly that the network model can better discover the potential characteristics of the sample over time and mention the generalization ability of the model in the subsequent feature extraction stage.
S42, extracting features;
specifically, referring to fig. 7, next is a feature extraction stage, where the preprocessed data is input into a feature extraction network. The feature extraction part is mainly divided into two parts;
the upper half is the cnn+lstm network layer, and data input to this part first goes through the CNN block: firstly, a first one-dimensional convolution layer Conv1D (specific parameters are convolution kernel size: 5, stride: 1, padding mode: valid, output channel number: 32) is used for carrying out convolution operation on different positions in an input sequence in a sliding convolution kernel mode, so that characteristics of different positions are obtained, and the characteristics can represent different properties and structures of the input sequence. Then performing BN operation (Batch Normalization) and a ReLU function; the Batch Normalization (BN) layer is a layer type commonly used in deep neural networks, and can normalize data during training, thereby accelerating training of the network and improving accuracy of the model. ReLU (Rectified Linear Unit) is an activation function commonly used in deep learning that can perform nonlinear transformation on the output of a neural network, thereby enhancing the expressive power and nonlinear feature extraction power of the network. And then passing through a maximum pooling layer (the specific parameters are the pooling core size: 3, the stride: 2, the padding mode: the Same, and the channel number is unchanged). And then the characteristic extraction operation of the CNN block part is completed through a second one-dimensional convolution layer Conv1D (specific parameters are convolution kernel size: 3, stride: 1, padding mode: name and output channel number: 64). Inputting the characteristics of CNN blocks into an LSTM layer, (the specific parameters are that the hidden dimension is 192, the number of layers is 2), wherein the LSTM network is a cyclic neural network capable of effectively processing long-sequence data, and the core is that a gating mechanism is introduced to control the flow of information and the updating of memory state;
the lower half is a multi-scale fusion layer. The method aims to realize sensing of different scales by using convolution kernels of different sizes to find an optimal method for extracting data features, and the module is divided into three parts:
the first part passes through a maximum pooling layer (specific parameters are pooling kernel size: 2, stride: 2, padding mode: valid, channel number is unchanged), then passes through a one-dimensional convolution layer Conv1D (specific parameters are convolution kernel size: 1, stride: 1, padding mode: valid, output channel number: 32), and then passes through BN operation (Batch Normalization) and a ReLU function;
the second part passes through a one-dimensional convolution layer Conv1D (specific parameters are convolution kernel size: 1, stride: 1, packing mode: valid, output channel number: 32), then passes through BN operation (Batch Normalization) and a ReLU function, then passes through a one-dimensional convolution layer Conv1D (specific parameters are convolution kernel size: 3, stride: 1, packing mode: valid, output channel number: 32), and finally passes through BN operation (Batch Normalization) and a ReLU function;
the third part passes through a one-dimensional convolution layer Conv1D (specific parameters are convolution kernel size: 1, stride: 1, packing mode: valid, output channel number: 64), then passes through BN operation (Batch Normalization) and a ReLU function, then passes through a one-dimensional convolution layer Conv1D (specific parameters are convolution kernel size: 5, stride: 1, packing mode: valid, output channel number: 128), and finally passes through BN operation (Batch Normalization) and a ReLU function;
finally, the three extracted features are spliced into a new feature according to the depth direction to be used for subsequent feature fusion.
S43, feature fusion;
specifically, referring to fig. 8, then, a feature fusion part fuses two features obtained in the feature extraction stage of the previous stage, so that the model can obtain more abundant and effective feature information by utilizing the correlation and complementarity between features extracted by different networks, the characterization capability and classification capability of data are improved, firstly, the two obtained features are subjected to dimension-increasing operation, 1 is used for expanding one dimension, and then the obtained features are subjected to outer product calculation. The obtained z-region part contains two kinds of main feature information and also contains potential important information, namely, the feature extraction part is provided with an upper network structure and a lower network structure as shown in fig. 7, and the data respectively pass through the two networks to obtain two feature outputs containing different potential information. The two features are then each dimension-added (one dimension added) for the purpose of the next fusion. Finally, performing outer product calculation on the two features after dimension increase, as shown in fig. 8, to obtain a rectangular feature.
S44, classifying the characteristics.
Specifically, the feature classification section is last. Performing a single flat operation on the features obtained in the previous stage, flattening the feature matrix into a one-dimensional array, taking the one-dimensional array as an input of a full-connection layer, classifying the features through a two-layer full-connection layer, namely an FC layer, performing Softmax operation, obtaining a data feature output after finishing the module operation, and converting the output of the network into probability distribution by the Softmax function after the feature output passes through the full-connection layer, wherein the probability distribution is given to an output vector, performing exponential operation on each element, and normalizing the result of the exponential operation;
in summary, the specific implementation steps of the application are that a proper amount of coal slime mixture sample can be obtained from a tailing water tank for coal slime flotation in an industrial site, a Pasteur pipette is used for taking 2ml of the coal slime mixture sample into a quartz cuvette, and then an optical fiber spectrometer is used for detecting absorbance of the coal slime mixture sample, and the detected spectral data is led into a trained network model for ash detection, so that compared with the prior art, the method of the application has obvious advantages in precision and speed by utilizing the deep learning and spectral analysis technology, and the traditional chemical analysis technology needs multiple steps of sample collection, processing, analysis and the like, and is time-consuming and labor-consuming, and the accuracy of analysis results is difficult to ensure; the nondestructive detection of the sample can be realized by using the method of the deep learning and spectrum analysis technology, sample treatment is not needed, the detection efficiency and accuracy are greatly improved, and in addition, the method of the deep learning and spectrum analysis technology has stronger operability and adaptability, can be flexibly adjusted and optimized according to actual requirements, and has wider application prospects.
Referring to fig. 2, a coal slime flotation ash detection system based on spectrum multi-modal time series learning includes:
the selecting module is used for selecting a coal slime flotation mixture sample;
the absorbance detection module is used for detecting absorbance of the coal slime flotation mixture sample through the optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample;
the construction module is used for constructing a coal slime flotation multiscale feature fusion model;
and the ash detection module is used for carrying out ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain the ash value of the coal slime flotation tailings.
The content in the method embodiment is applicable to the system embodiment, the functions specifically realized by the system embodiment are the same as those of the method embodiment, and the achieved beneficial effects are the same as those of the method embodiment.
While the preferred embodiment of the present application has been described in detail, the application is not limited to the embodiment, and various equivalent modifications and substitutions can be made by those skilled in the art without departing from the spirit of the application, and these equivalent modifications and substitutions are intended to be included in the scope of the present application as defined in the appended claims.
Claims (10)
1. The coal slime flotation ash content detection method based on spectrum multi-mode time sequence learning is characterized by comprising the following steps of:
selecting a coal slime flotation mixture sample;
performing absorbance detection on the coal slime flotation mixture sample by using an optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample;
constructing a coal slime flotation multi-scale feature fusion model;
and carrying out ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain the ash value of the coal slime flotation tailings.
2. The method for detecting ash content in coal slime flotation based on spectrum multi-mode time sequence learning according to claim 1, wherein the step of selecting the coal slime flotation mixture sample comprises the following steps:
collecting a mixture of coal slime having different ash values;
injecting water into the coal slime mixture, and stirring the mixture by a magnetic stirrer to obtain a stirred coal slime mixture;
and sucking the stirred coal slime mixture into a quartz cuvette through a Pasteur pipette, and preparing a coal slime flotation mixture sample.
3. The method for detecting ash content in coal slime flotation based on spectrum multi-mode time sequence learning according to claim 2, wherein the step of detecting absorbance of the coal slime flotation mixture sample by an optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample specifically comprises the following steps:
based on a transmissivity measuring bracket of the optical fiber spectrometer, isolating a xenon lamp light source in the transmissivity measuring bracket of the optical fiber spectrometer to obtain background light data;
placing the empty quartz cuvette in a transmissivity measuring bracket of the optical fiber spectrometer to obtain reference light data;
placing a quartz cuvette with a coal slime flotation mixture sample in a transmissivity measuring bracket of an optical fiber spectrometer to obtain sample light data;
and calculating by combining the background light data, the reference light data and the sample light data to obtain the spectrum data of the coal slime flotation mixture sample.
4. The method for detecting the ash content in the coal slime flotation based on the spectrum multi-mode time sequence learning according to claim 3, further comprising the step of circularly acquiring background light data, reference light data and sample light data based on different moments to obtain spectrum data of a plurality of groups of coal slime flotation mixture samples at different moments.
5. The method for detecting the ash content in the coal slime flotation based on the spectrum multi-mode time sequence learning according to claim 4, wherein the calculation expression of the spectrum data of the coal slime flotation mixture sample is specifically as follows:
T＝-log 10 I f
in the above formula, T represents the absorbance of a sample of the coal slime flotation mixture, I f Indicating transmittance, I sample Representing sample light data, I off Representing background light data, I sub Representing reference light data.
6. The method for detecting the ash content in the coal slime flotation based on the spectrum multi-mode time sequence learning, which is characterized in that the coal slime flotation multi-scale feature fusion model comprises a data preprocessing module, a feature extraction module, a feature fusion module and a feature classification module, wherein:
the feature extraction module comprises a CNN+LSTM layer and a multi-scale fusion layer, wherein the CNN layer comprises a one-dimensional convolution layer, a BN+ReLU layer and a maximum pooling layer, and the multi-scale fusion layer comprises a first one-dimensional convolution layer, a second one-dimensional convolution layer and a third one-dimensional convolution layer.
7. The method for detecting ash content in coal slime flotation based on spectrum multi-mode time sequence learning according to claim 6, wherein the step of detecting ash content in spectrum data of a coal slime flotation mixture sample based on a coal slime flotation multi-scale feature fusion model to obtain ash content value of coal slime flotation tailings specifically comprises the following steps:
spectral data of the coal slime flotation mixture sample is input into a coal slime flotation multiscale feature fusion model;
performing data format conversion treatment on the spectrum data of the coal slime flotation mixture sample based on the data pretreatment module to obtain pretreated spectrum data of the coal slime flotation mixture sample;
performing characteristic extraction treatment on the pretreated coal slime flotation mixture sample spectral data based on a characteristic extraction module to obtain coal slime flotation mixture sample spectral characteristic data;
carrying out characteristic fusion treatment on the spectral characteristic data of the coal slime flotation mixture sample based on the characteristic fusion module to obtain the spectral characteristic fusion data of the coal slime flotation mixture sample;
and carrying out characteristic classification treatment on the spectral characteristic fusion data of the sample of the coal slime flotation mixture based on the characteristic classification module to obtain the coal slime flotation tailing ash value.
8. The method for detecting ash content in a coal slime flotation mixture sample based on spectrum multi-mode time sequence learning according to claim 7, wherein the spectrum characteristic data of the coal slime flotation mixture sample includes spectrum characteristic data of a first coal slime flotation mixture sample and spectrum characteristic data of a second coal slime flotation mixture sample, and the step of performing feature extraction processing on the pretreated coal slime flotation mixture sample spectrum data based on a feature extraction module to obtain the spectrum characteristic data of the coal slime flotation mixture sample specifically comprises the following steps:
the pretreated coal slime flotation mixture sample spectrum data are input to a feature extraction module;
carrying out convolution operation treatment on the pretreated coal slime flotation mixture sample spectrum data based on a one-dimensional convolution layer of the CNN network layer to obtain different position characteristics of the coal slime flotation mixture sample spectrum data;
sequentially carrying out normalization treatment and nonlinear transformation treatment on different position characteristics of the coal slime flotation mixture sample spectrum data based on a BN+ReLU layer of a CNN network layer to obtain transformed coal slime flotation mixture sample spectrum data;
carrying out pooling treatment on the transformed coal slime flotation mixture sample spectrum data based on the largest pooling layer of the CNN network layer to obtain the dimension-reduced coal slime flotation mixture sample spectrum data;
capturing and updating the spectrum data of the sample of the coal slime flotation mixture after the dimension reduction based on the LSTM network layer to obtain the spectrum characteristic data of the sample of the first coal slime flotation mixture;
sequentially carrying out pooling treatment, convolution treatment, BN operation and ReLU operation on the pretreated coal slime flotation mixture sample spectrum data based on the first one-dimensional convolution layer of the multi-scale fusion layer to obtain first multi-scale coal slime flotation mixture sample spectrum data;
carrying out convolution treatment, BN operation, reLU operation, convolution treatment, BN operation and ReLU operation on the pretreated coal slime flotation mixture sample spectrum data in sequence based on a second one-dimensional convolution layer of the multi-scale fusion layer to obtain second multi-scale coal slime flotation mixture sample spectrum data;
carrying out convolution treatment, BN operation, reLU operation, convolution treatment, BN operation and ReLU operation on the pretreated coal slime flotation mixture sample spectrum data in sequence based on a third one-dimensional convolution layer of the multi-scale fusion layer to obtain third multi-scale coal slime flotation mixture sample spectrum data;
and performing splicing treatment on the first multi-scale coal slime flotation mixture sample spectral data, the second multi-scale coal slime flotation mixture sample spectral data and the third multi-scale coal slime flotation mixture sample spectral data to obtain second coal slime flotation mixture sample spectral characteristic data.
9. The method for detecting the ash content in the coal slime flotation based on the spectrum multi-mode time sequence learning according to claim 8, wherein the step of obtaining the spectrum characteristic fusion data of the sample of the coal slime flotation mixture based on the characteristic fusion processing of the spectrum characteristic data of the sample of the coal slime flotation mixture by the characteristic fusion module specifically comprises the following steps:
performing dimension-increasing operation treatment on the spectral characteristic data of the first coal slime flotation mixture sample and the spectral characteristic data of the second coal slime flotation mixture sample to obtain the spectral characteristic data of the first coal slime flotation mixture sample after dimension increase and the spectral characteristic data of the second coal slime flotation mixture sample after dimension increase;
and performing characteristic outer product calculation on the spectrum characteristic data of the first coal slime flotation mixture sample after the dimension increase and the spectrum characteristic data of the second coal slime flotation mixture sample after the dimension increase to obtain spectrum characteristic fusion data of the coal slime flotation mixture sample.
10. Coal slime flotation ash content detecting system based on spectrum multimode time sequence study, which is characterized by comprising the following modules:
the selecting module is used for selecting a coal slime flotation mixture sample;
the absorbance detection module is used for detecting absorbance of the coal slime flotation mixture sample through the optical fiber spectrometer to obtain spectrum data of the coal slime flotation mixture sample;
the construction module is used for constructing a coal slime flotation multiscale feature fusion model;
and the ash detection module is used for carrying out ash detection on the spectral data of the coal slime flotation mixture sample based on the coal slime flotation multi-scale feature fusion model to obtain the ash value of the coal slime flotation tailings.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311176743.XA CN117092050B (en) | 2023-09-13 | 2023-09-13 | Coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311176743.XA CN117092050B (en) | 2023-09-13 | 2023-09-13 | Coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117092050A true CN117092050A (en) | 2023-11-21 |
| CN117092050B CN117092050B (en) | 2024-02-27 |
Family
ID=88770053
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311176743.XA Active CN117092050B (en) | 2023-09-13 | 2023-09-13 | Coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117092050B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5065416A (en) * | 1987-12-04 | 1991-11-12 | Process Technology, Inc. | On-line slurry ash monitoring system |
| US5258825A (en) * | 1991-11-13 | 1993-11-02 | Perten Instruments North America, Inc. | Optical compositional analyzer apparatus and method for detection of ash in wheat and milled wheat products |
| CN105486661A (en) * | 2016-01-28 | 2016-04-13 | 山西汾西矿业（集团）有限责任公司 | Near-infrared spectrum coal ash content rapid detection method based on Gaussian process |
| CN108489912A (en) * | 2018-05-11 | 2018-09-04 | 东北大学 | A kind of coal constituent analysis method based on coal spectroscopic data |
| CN111753912A (en) * | 2020-06-28 | 2020-10-09 | 中国矿业大学 | Coal slime flotation clean coal ash content prediction method based on deep learning |
| US20220051074A1 (en) * | 2018-10-10 | 2022-02-17 | Zhejiang University | Quantitative spectral data analysis and processing method based on deep learning |
| CN114636687A (en) * | 2022-03-14 | 2022-06-17 | 国能神皖能源有限责任公司 | Small sample coal quality characteristic analysis system and method based on deep migration learning |
| CN115861235A (en) * | 2022-12-05 | 2023-03-28 | 中国矿业大学 | Flotation tailing ash content prediction method based on multi-feature data fusion |
| US20230252761A1 (en) * | 2021-01-26 | 2023-08-10 | Wuhan University | Method for classifying hyperspectral images on basis of adaptive multi-scale feature extraction model |
-
2023
- 2023-09-13 CN CN202311176743.XA patent/CN117092050B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5065416A (en) * | 1987-12-04 | 1991-11-12 | Process Technology, Inc. | On-line slurry ash monitoring system |
| US5258825A (en) * | 1991-11-13 | 1993-11-02 | Perten Instruments North America, Inc. | Optical compositional analyzer apparatus and method for detection of ash in wheat and milled wheat products |
| CN105486661A (en) * | 2016-01-28 | 2016-04-13 | 山西汾西矿业（集团）有限责任公司 | Near-infrared spectrum coal ash content rapid detection method based on Gaussian process |
| CN108489912A (en) * | 2018-05-11 | 2018-09-04 | 东北大学 | A kind of coal constituent analysis method based on coal spectroscopic data |
| US20220051074A1 (en) * | 2018-10-10 | 2022-02-17 | Zhejiang University | Quantitative spectral data analysis and processing method based on deep learning |
| CN111753912A (en) * | 2020-06-28 | 2020-10-09 | 中国矿业大学 | Coal slime flotation clean coal ash content prediction method based on deep learning |
| US20230252761A1 (en) * | 2021-01-26 | 2023-08-10 | Wuhan University | Method for classifying hyperspectral images on basis of adaptive multi-scale feature extraction model |
| CN114636687A (en) * | 2022-03-14 | 2022-06-17 | 国能神皖能源有限责任公司 | Small sample coal quality characteristic analysis system and method based on deep migration learning |
| CN115861235A (en) * | 2022-12-05 | 2023-03-28 | 中国矿业大学 | Flotation tailing ash content prediction method based on multi-feature data fusion |
Non-Patent Citations (2)
| Title |
|---|
| 郭西进 等: "基于深度卷积网络的煤泥浮选尾矿灰分检测方法研究", 煤炭技术, vol. 39, no. 02, pages 144 - 146 * |
| 韩宇 等: "基于CNN−BP 的浮选尾煤灰分智能检测方法", 工矿自动化, vol. 49, no. 3, pages 100 - 106 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117092050B (en) | 2024-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shruti et al. | Blanks and bias in microplastic research: Implications for future quality assurance | |
| CN101587109B (en) | Determination method and intruments of hexavalent chromium | |
| CN110033133A (en) | Silt source tracing method in a kind of river and lake system | |
| CN112730326A (en) | Intelligent identification device and method for rock slices | |
| Pollina et al. | PlanktoScope: affordable modular quantitative imaging platform for citizen oceanography | |
| CN117092050B (en) | Coal slime flotation ash content detection method and system based on spectrum multi-mode time sequence learning | |
| CN1245218A (en) | Solid-phase one-by-one base nucleic acid analysis method and its instrument | |
| Hamdhani et al. | Performance of a handheld chlorophyll-a fluorometer: Potential use for rapid algae monitoring | |
| CN201517993U (en) | Micro-fluidic chip detecting device based on photoacoustic technique | |
| CN1458529A (en) | Colorimetric component analysis method and its device based on image technology | |
| Cai et al. | Online detection of concentrate grade in the antimony flotation process based on in situ Raman spectroscopy combined with a CNN-GRU hybrid model | |
| CN115861235A (en) | Flotation tailing ash content prediction method based on multi-feature data fusion | |
| CN113295682B (en) | Phenolic compound analysis method based on polyphenol oxidase activity nanoenzyme | |
| CN1546987A (en) | Minitype biochemical spectrum analyzer | |
| CN109696393A (en) | A kind of method that determination of the environment tests polystyrene microsphere content under simulated conditions | |
| CN112461780B (en) | Evaluation and calculation method for original total organic carbon content of hydrocarbon source rock | |
| CN110760574B (en) | Device and method for measuring base | |
| Xu et al. | Efficient and accurate microplastics identification and segmentation in urban waters using convolutional neural networks | |
| CN2575678Y (en) | Globe satellite positioning sea chart and fish school detector | |
| Dyachenko et al. | Method development for microplastic analysis in wastewater | |
| CN200993634Y (en) | Sediment instrument | |
| CN106680267A (en) | Quantitative analysis system for urine iodine | |
| CN113406038A (en) | Optical detection method and device for pH value of water | |
| CN105067575A (en) | Method for detecting mercury in water based on quantum dot fluorescence probe, and apparatus thereof | |
| CN106124464B (en) | A kind of fibre-optical sensing device for detection of heavy metal ion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |