CN116242799B - Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm - Google Patents
Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm Download PDFInfo
- Publication number
- CN116242799B CN116242799B CN202310242896.3A CN202310242896A CN116242799B CN 116242799 B CN116242799 B CN 116242799B CN 202310242896 A CN202310242896 A CN 202310242896A CN 116242799 B CN116242799 B CN 116242799B
- Authority
- CN
- China
- Prior art keywords
- reflecting mirror
- infrared
- base oil
- 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.)
- Active
Links
- 239000002199 base oil Substances 0.000 title claims abstract description 113
- 238000001514 detection method Methods 0.000 title claims abstract description 61
- 230000004927 fusion Effects 0.000 title claims abstract description 24
- 238000013135 deep learning Methods 0.000 title claims abstract description 22
- 230000003287 optical effect Effects 0.000 claims abstract description 45
- 238000000034 method Methods 0.000 claims abstract description 27
- 238000010438 heat treatment Methods 0.000 claims abstract description 26
- 230000007246 mechanism Effects 0.000 claims abstract description 11
- 238000004458 analytical method Methods 0.000 claims abstract description 8
- 238000001228 spectrum Methods 0.000 claims description 68
- 238000002329 infrared spectrum Methods 0.000 claims description 43
- 238000013527 convolutional neural network Methods 0.000 claims description 28
- 235000012431 wafers Nutrition 0.000 claims description 25
- 238000012360 testing method Methods 0.000 claims description 17
- 238000011176 pooling Methods 0.000 claims description 12
- 230000001360 synchronised effect Effects 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 229920006395 saturated elastomer Polymers 0.000 claims description 8
- 238000010183 spectrum analysis Methods 0.000 claims description 8
- 230000003595 spectral effect Effects 0.000 claims description 7
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- PFNQVRZLDWYSCW-UHFFFAOYSA-N (fluoren-9-ylideneamino) n-naphthalen-1-ylcarbamate Chemical compound C12=CC=CC=C2C2=CC=CC=C2C1=NOC(=O)NC1=CC=CC2=CC=CC=C12 PFNQVRZLDWYSCW-UHFFFAOYSA-N 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000002238 attenuated effect Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 229910002092 carbon dioxide Inorganic materials 0.000 claims description 3
- 239000001569 carbon dioxide Substances 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000001105 regulatory effect Effects 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims 1
- 150000001335 aliphatic alkanes Chemical class 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000005102 attenuated total reflection Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 4
- 239000003921 oil Substances 0.000 description 4
- 235000019198 oils Nutrition 0.000 description 4
- 239000010687 lubricating oil Substances 0.000 description 3
- 229920013639 polyalphaolefin Polymers 0.000 description 3
- 239000002904 solvent Substances 0.000 description 3
- 238000010998 test method Methods 0.000 description 3
- 231100000331 toxic Toxicity 0.000 description 3
- 230000002588 toxic effect Effects 0.000 description 3
- 238000005481 NMR spectroscopy Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000004817 gas chromatography Methods 0.000 description 2
- 238000005227 gel permeation chromatography Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 238000002310 reflectometry Methods 0.000 description 2
- 229930195734 saturated hydrocarbon Natural products 0.000 description 2
- 238000002834 transmittance Methods 0.000 description 2
- 101100407030 Arabidopsis thaliana PAO2 gene Proteins 0.000 description 1
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- UFHFLCQGNIYNRP-UHFFFAOYSA-N Hydrogen Chemical compound [H][H] UFHFLCQGNIYNRP-UHFFFAOYSA-N 0.000 description 1
- 101100407037 Oryza sativa subsp. japonica PAO6 gene Proteins 0.000 description 1
- 101150092791 PAO4 gene Proteins 0.000 description 1
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical compound [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 150000001338 aliphatic hydrocarbons Chemical class 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000013078 crystal Substances 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 229910052739 hydrogen Inorganic materials 0.000 description 1
- 239000001257 hydrogen Substances 0.000 description 1
- 238000005984 hydrogenation reaction Methods 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 150000002632 lipids Chemical class 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000002480 mineral oil Substances 0.000 description 1
- 235000010446 mineral oil Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 229920002545 silicone oil Polymers 0.000 description 1
- 229910052717 sulfur Inorganic materials 0.000 description 1
- 239000011593 sulfur Substances 0.000 description 1
- 235000015112 vegetable and seed oil Nutrition 0.000 description 1
- 239000008158 vegetable oil Substances 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
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3577—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing liquids, e.g. polluted water
-
- 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
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
-
- G—PHYSICS
- G02—OPTICS
- G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
- G02B26/00—Optical devices or arrangements for the control of light using movable or deformable optical elements
- G02B26/08—Optical devices or arrangements for the control of light using movable or deformable optical elements for controlling the direction of light
- G02B26/0816—Optical devices or arrangements for the control of light using movable or deformable optical elements for controlling the direction of light by means of one or more reflecting elements
-
- G—PHYSICS
- G02—OPTICS
- G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
- G02B5/00—Optical elements other than lenses
- G02B5/08—Mirrors
-
- G—PHYSICS
- G02—OPTICS
- G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
- G02B7/00—Mountings, adjusting means, or light-tight connections, for optical elements
- G02B7/18—Mountings, adjusting means, or light-tight connections, for optical elements for prisms; for mirrors
- G02B7/182—Mountings, adjusting means, or light-tight connections, for optical elements for prisms; for mirrors for mirrors
-
- 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
- G06N3/08—Learning methods
-
- 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
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N2021/3595—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using FTIR
Abstract
The invention belongs to the technical field of analysis and detection, and particularly relates to a base oil detection device and a detection method based on a deep learning infrared multidimensional fusion algorithm, wherein the detection device comprises a Fourier infrared spectrometer and a sample mechanism arranged in a sample bin of the Fourier infrared spectrometer, and the sample mechanism comprises an L-shaped mounting table, an optical system and a sample frame; the method comprises the steps of dripping base oil on a wafer, placing the wafer at the upper end of a heating block corresponding to a sample hole, receiving infrared beams emitted by a Fourier infrared spectrometer by a detector of the Fourier infrared spectrometer after total reflection of an optical system, and enabling the sample hole and a background hole to be sequentially positioned at the same vertical light converging position in an optical path of the optical system.
Description
Technical Field
The invention belongs to the technical field of analysis and detection, and particularly relates to a base oil detection device and method based on a deep learning infrared multidimensional fusion algorithm.
Background
The lubricating oil is widely applied to the industries of vehicles, machinery and power, and is an important industrial product related to national folk life; more than 95% of the lubricating oil consists of base oil, wherein the base oil is generally divided into five types, and the I type base oil is obtained through extraction, has higher sulfur content and lower viscosity index; the II base oil is mineral oil and is prepared by extracting bottom oil; the III base oil is base oil prepared by a full hydrogenation process, and has higher viscosity index; group IV base oils refer to Polyalphaolefin (PAO) synthetic oils; group V base oils refer to synthetic oils (e.g., lipid oils, silicone oils, etc.) other than group I through IV base oils, vegetable oils, reclaimed base oils, and the like. The five types of base oil have different performances and price which are hundreds of times different, and if the types of the base oil can be rapidly detected and judged, the method has important significance for production management quality detection.
In the prior art, most of the base oil is measured by a Ping Shi viscometer, a gas chromatography, a gel permeation chromatography and a nuclear magnetic resonance spectroscopy (hydrogen spectrum and carbon spectrum), and the detection methods are time-consuming and labor-consuming, have high detection cost, use a large amount of toxic solvents during the test, and cause environmental pollution and other problems.
The base oils of group I, group II, group III and group IV mainly consist of saturated long-chain alkanes, and the spectrum peaks on the infrared spectrum are seriously overlapped, so that the conventional infrared detection device and analysis method cannot distinguish the four base oils.
Meanwhile, the conventional infrared detection device is divided into a transmission mode and a reflection mode: the transmission mode is generally a horizontal light path, see fig. 10, with the following limitations: the sample is placed in a vertical light path, and the flow in the vertical direction of the liquid can influence the signal intensity of the infrared spectrum and the reproducibility of data;
the reflection mode is a horizontal test method generally referred to as an attenuated total reflection method, see fig. 11, and has the following limitations: (1) the infrared light is subject to multiple refraction and attenuated total reflection by a crystal and the like, the energy loss is larger, the detected signal intensity is less than 1/10 of the transmission test signal, and (2) the detection wave band of the attenuated total reflection method is narrower and is generally 4000-675cm -1 While the infrared band collected by the transmission method is 7500 cm to 350cm -1 The method comprises the steps of carrying out a first treatment on the surface of the In addition, the conventional infrared testing device cannot accurately control the temperature, the infrared spectrum is extremely sensitive to the change of external conditions, the testing temperature cannot be accurately controlled, and the application of the infrared spectrum technology is greatly limited.
Disclosure of Invention
The invention aims to solve the problems of high detection cost, pollution and the like in the prior art, and provides a base oil detection device and a base oil detection method based on a deep learning infrared and multidimensional fusion algorithm, so that an infrared spectrum technology can be applied to detection of base oil, the base oil can be rapidly identified, and the base oil detection device and the base oil detection method have the advantages of rapid analysis, no pretreatment and the like, and meanwhile, no toxic solvent is used, so that the detection is pollution-free, and the cost can be reduced to below 10% of the prior art.
The invention adopts the following technical scheme: the base oil detection device based on the deep learning infrared multidimensional fusion algorithm comprises a Fourier infrared spectrometer and a sample mechanism arranged in a sample bin of the Fourier infrared spectrometer, wherein the sample mechanism comprises an L-shaped mounting table 1, an optical system and a sample frame, the sample frame is arranged on a horizontal table of the mounting table 1, and the optical system is arranged on the inner side surface of a vertical table of the mounting table 1;
the sample rack comprises a support vertical rod 31, a support cross rod 32 and a sample base, the lower end of the support vertical rod 31 is vertically arranged on a horizontal platform of the mounting platform 1 through a ball screw nut pair 4, the upper end of the support vertical rod 31 is fixedly connected with the horizontal end of the support cross rod 32 through a horizontally arranged cylindrical adapter 5,
the sample base 3 comprises a square base plate 33 and a heating block 34 arranged at the upper end of the base plate 33, a sample hole 35 and a background hole 36 are respectively formed through the heating block 34 and the corresponding base plate 33, and the base plate 33 is fixedly connected with the horizontal other end of the supporting cross rod 32;
the optical system comprises a first plane mirror 23, a second plane mirror 24, a first concave mirror 21 and a second concave mirror 22, wherein the first plane mirror 23, the second plane mirror 24, the first concave mirror 21 and the second concave mirror 22 are respectively vertically arranged on the inner side surface of the vertical table of the installation table 1 through an L-shaped support 6;
in the test, two identical wafers are taken, base oil is dripped on one wafer and the wafer is placed at the upper end of a heating block 34 corresponding to a sample hole 35, the other wafer is placed at the upper end of the heating block 34 corresponding to a background hole 36, the temperature of the heating block 34 is regulated to be 20-30 ℃, and an infrared beam emitted by a Fourier infrared spectrometer is received by a detector of the Fourier infrared spectrometer after being totally reflected by an optical system; the sample stage is adjusted so that the sample aperture 35 and the background aperture 36 are located in sequence at the same vertical light convergence in the optical path of the optical system, resulting in the infrared spectrum of the base oil.
Further, the horizontal plate of the L-shaped support 6 is fixed on the inner side surface of the vertical table of the mounting table 1 through screws, a pair of mounting grooves are formed in the vertical plate of the L-shaped support 6, and the pair of mounting grooves are spaced and formed along the length direction of the vertical plate;
the first plane mirror 23, the second plane mirror 24, the first concave mirror 21 and the second concave mirror 22 are attached to the vertical plates of the corresponding L-shaped support 6, and a pair of bolts 61 are respectively inserted at the upper and lower ends of the pair of mounting grooves in a matched manner, so as to clamp and fix the corresponding first plane mirror 23, second plane mirror 24, first concave mirror 21 and second concave mirror 22.
Further, the first plane mirror 23, the second concave mirror 22, the first concave mirror 21 and the second plane mirror 24 are correspondingly arranged at four right angles on the inner side surface of the vertical table of the mounting table 1 in sequence;
the second plane reflector 24 is arranged on the optical path of the infrared beam emitted by the Fourier infrared spectrometer and forms an angle of 45 degrees with the optical path; the first concave reflecting mirror 21 and the second plane reflecting mirror 24 are arranged in parallel, and the optical path between them is perpendicular to the optical path of the infrared beam of the light beam; the second concave reflecting mirror 22 and the mirror surface of the first concave reflecting mirror 21 are oppositely arranged to form a 90-degree included angle, and the light path between the second concave reflecting mirror and the mirror surface is parallel to the light path of the infrared light beam; the first plane mirror 23 and the second concave mirror 22 are arranged in parallel, and the optical path between them is perpendicular to the optical path of the infrared beam; the emergent light path of the first plane mirror 23 is parallel to the light path of the infrared light beam; the sample aperture 35 and the background aperture 36 are in turn located at the vertical ray convergence between the first concave mirror 21 and the second planar mirror 24.
Further, the screw of the ball screw nut pair 4 is arranged along the length direction of the horizontal table of the mounting table 1, and one end of the screw is connected with a motor 41; the upper end of the nut seat of the ball screw nut pair 4 is fixedly connected with the lower end of the supporting vertical rod 31.
The invention also discloses a detection method of the base oil detection device based on the deep learning infrared multidimensional fusion algorithm, which specifically comprises the following steps:
step S1: respectively placing wafers with the same specification on the heating blocks 34 corresponding to the sample holes 35 and the background holes 36, and dripping base oil on the wafers corresponding to the sample holes 35;
step S2: the infrared light beam emitted by the Fourier infrared spectrometer is received by a detector of the Fourier infrared spectrometer after being totally reflected by the optical system; the position of the base plate 33 is adjusted so that the background hole 36 is positioned at the position where the vertical light rays between the first concave reflecting mirror 21 and the second plane reflecting mirror 24 converge, and a background spectrum is collected; the position of the base plate 33 is adjusted so that the sample hole 35 is positioned at the position where the vertical light rays between the first concave reflecting mirror 21 and the second plane reflecting mirror 24 are converged, and the infrared spectrum of the sample is collected; then dividing the infrared spectrum of the sample by the background spectrum, and deducting interference signals of water vapor and carbon dioxide in the air to obtain the infrared spectrum of the base oil;
step S3: respectively collecting infrared spectrums of base oils of class I, class II, class III and class IV according to the method of S2 to obtain spectrum data points;
step S4: performing variable screening by using interval partial least square to divide the spectrum data point into n sections, respectively calculating a root mean square error value (RMSECV) value of cross verification by adopting a reserved cross verification section, and selecting a section of wavelength which has a low RMSECV value and corresponds to C-H vibration of saturated alkane as an analysis object to establish a spectrum data set S;
wherein: the calculation formula of RMSECV value is as follows:
wherein: n is the number of data set samples, y i 、Respectively an actual measurement value and a predicted value of an ith sample in the data set; step S5: the spectrum data set S is filtered and noisy through continuous wavelet transformation to obtain a spectrum data set M, and a double spectrum is utilized for two purposesThe method comprises the steps of selecting an automatic peak of a double-spectrum two-dimensional synchronous correlation spectrum by dimensional synchronous correlation spectrum analysis, establishing a spectrum matrix data set N, respectively carrying out standardization and normalization treatment on M and N, and then obtaining a data set D after fusion;
wherein: the continuous wavelet transform function is as follows:
where f (t) is the spectral matrix, t is the wavenumber,is a wavelet basis function, alpha is a translation parameter, b is a telescoping factor, C is a wavelet coefficient, and dt represents integration of the wave number t;
step S6: inputting the data set D into a CNN convolutional neural network to carry out CNN model training and prediction, wherein the data set D is used as the input of a model, and the base oil type is used as the output of the model for quick nondestructive detection of the base oil type; the accuracy of the detection result is 100%, and the base oil of the I, II, III and IV types can be accurately identified.
Further, the wafer in the step S1 is a 25 μm thick zinc selenide wafer, and the temperature of the heating block 34 is controlled at 30 ℃.
Further, the two-dimensional synchronous correlation spectrum analysis of the dual spectrum in step S5 refers to: selecting infrared average spectrums of 3 base oils in III base oils or IV base oils from a spectrum data set M as reference spectrums r (v), respectively taking at least 30 base oil infrared spectrums as samples s (v), and carrying out double-spectrum two-dimensional synchronous correlation spectrum analysis, wherein the formula is as follows:
where s (v) is the sample spectrum matrix, r (v) is the reference spectrum matrix, v 1 ，v 2 Representing the two spectra compared respectively.
Further, the CNN model in step S6 includes a one-dimensional convolution layer, an activation layer, a pooling layer, and a full connection layer; the number of convolution kernels of one-dimensional convolution operation is 16, the size is 6*1, and the step length of convolution is 4; and the ReLU completes the activation operation, so that the neurons in the neural network have sparse activation; the pooling operation uses an average pooling model, the pooling window of 4*1 is taken each time, and the step length is set to be 1; inputting the vector into a full connection layer, and outputting a one-dimensional high-order vector; four types of base oil are used, and the number of nodes in the full-connection layer is 2316; the formula of the ReLU function is as follows:
f(x)＝max(0，x)
wherein x is an input value in CNN, and f (x) is an output value;
the CNN convolutional neural network model adopts a learning rate attenuation mechanism, wherein the initial value of the learning rate is set to be 0.001, the attenuation index is set to be 0.1, the learning rate is attenuated gradually over time in the process of network training to dynamically adjust, and the initial value of each layer of weight is subjected to zero-mean Gaussian distribution with standard deviation of 0.1.
Compared with the prior art, the invention has the beneficial effects that:
compared with the prior art, the method can detect the base oil types rapidly, nondestructively and economically in real time by utilizing an infrared spectrum technology, so that the base oil types can be classified, a new detection method is provided for the base oil classification and finished lubricating oil, and compared with the existing detection methods such as Ping Shi viscometer, gas chromatography, gel permeation chromatography and nuclear magnetic resonance spectroscopy, the method does not need toxic solvents, does not need pretreatment, and is rapid in detection, so that the detection cost is reduced to below 10% of the prior art, and the method has a wide application prospect.
Specific: (1) The invention designs a base oil detection device based on a deep learning infrared multidimensional fusion algorithm, when the base oil detection device is used, two identical wafers are taken, base oil is dripped on one wafer and placed at the upper end of a heating block corresponding to a sample hole, one wafer is placed at the upper end of the heating block corresponding to a background hole, the temperature of the heating block is adjusted to be 20-30 ℃, a sample stage is adjusted, the sample hole and the background hole are sequentially positioned at the same vertical light convergence position in an optical path of an optical system, an infrared beam emitted by a Fourier infrared spectrometer is received by a detector of the Fourier infrared spectrometer after being totally reflected by the optical system, a sample infrared spectrum and a background spectrum are respectively obtained, and the sample infrared spectrum is divided by the background spectrum, so that the infrared spectrum of the base oil is obtained, and the base oil detection device has the following advantages:
(1) the optical system comprises a first plane reflecting mirror, a second concave reflecting mirror, and a high-reflectivity reflecting mirror combination of the first concave reflecting mirror and the second plane reflecting mirror, so that the total light transmittance is more than 90%, the obtained infrared spectrum has good quality, and a data base is provided for accurately identifying the types of base oil;
(2) according to the invention, the base plate of the sample base is horizontally and movably arranged in the sample bin of the Fourier infrared spectrometer, and the base oil sample can be directly arranged on the base plate for testing without using special equipment or complex liquid testing equipment; the base plate can finely adjust the position of the base plate through the ball screw nut, so that a detection sample is just processed in a detection light path, and the quality of the obtained infrared spectrum is further ensured;
(3) background interference is deducted in real time in the infrared test process, a high-quality infrared spectrogram is obtained, the problem that the infrared test limits the base oil test in the prior art is solved, the base oil infrared test can be more quickly and economically carried out, and the platform can be further expanded to various in-situ infrared test applications and has a certain guiding significance.
(2) The detection method of the base oil detection device provided by the invention realizes that the infrared spectrum technology is applied to the detection of base oil, and can rapidly and accurately identify the type of the base oil;
aiming at the problem that base oils of class I, class II, class III and class IV mainly consist of saturated long-chain alkanes, the spectrum peaks on infrared spectrums are seriously overlapped, and the four types of base oils cannot be distinguished by the existing two-dimensional infrared spectrum technology.
Drawings
FIG. 1 is a schematic structural diagram of a base oil detection device based on a deep learning infrared multidimensional fusion algorithm;
FIG. 2 is a schematic diagram of a relationship system of the present invention;
FIG. 3 is a schematic view of the structure of the sample base of the present invention;
FIG. 4 is a schematic view of the structure of the L-shaped support of the present invention;
FIG. 5 is a schematic view of an adapter according to the present invention;
FIG. 6 is a full spectrum of the infrared spectrum tested in accordance with the present invention;
FIG. 7 is a diagram of the invention for band selection using interval partial least squares;
FIG. 8 is a two-dimensional IR spectrum of a portion of a group I, group II, group III, and group IV base oil in accordance with the invention;
FIG. 9 is a two-dimensional IR spectrum auto-correlation spectrum chart of the present invention;
FIG. 10 is a schematic diagram of an infrared horizontal optical path in the prior art;
fig. 11 is a schematic diagram of the prior art infrared attenuated total reflection.
Wherein: the mounting table 1, the first concave mirror 21, the second concave mirror 22, the first plane mirror 23, the second plane mirror 24, the sample base 3, the support vertical rod 31, the support cross rod 32, the base plate 33, the heating block 34, the sample hole 35, the background hole 36, the ball screw nut pair 4, the motor 41, the adapter 5, the L-shaped support 6, and the pair of bolts 61.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the detailed description is presented by way of example only and is not intended to limit the invention.
Example 1
Referring to fig. 1-5, a base oil detection device based on a deep learning infrared multidimensional fusion algorithm comprises a fourier infrared spectrometer and a sample mechanism arranged in a sample bin of the fourier infrared spectrometer,
the sample mechanism comprises an L-shaped mounting table 1, an optical system and a sample rack, wherein the sample rack is mounted on a horizontal table of the mounting table 1, and the optical system is mounted on the inner side surface of a vertical table of the mounting table 1;
the sample rack comprises a support vertical rod 31, a support cross rod 32 and a sample base 3, the lower end of the support vertical rod 31 is vertically arranged on a horizontal platform of the mounting platform 1 through a ball screw nut pair 4, the upper end of the support vertical rod 31 is fixedly connected with the horizontal end of the support cross rod 32 through a horizontally arranged cylindrical adapter 5,
the sample base 3 comprises a square base plate 33 and a heating block 34 arranged at the upper end of the base plate 33, a sample hole 35 and a background hole 36 are respectively formed through the heating block 34 and the corresponding base plate 33, and the base plate 33 is fixedly connected with the horizontal other end of the supporting cross rod 32;
the optical system comprises a first plane mirror 23, a second plane mirror 24, a first concave mirror 21 and a second concave mirror 22, wherein the first plane mirror 23, the second plane mirror 24, the first concave mirror 21 and the second concave mirror 22 are respectively vertically arranged on the inner side surface of the vertical table of the installation table 1 through an L-shaped support 6;
in the test, two identical wafers are taken, base oil is dripped on one wafer and the wafer is placed at the upper end of a heating block 34 corresponding to a sample hole 35, the other wafer is placed at the upper end of the heating block 34 corresponding to a background hole 36, the temperature of the heating block 34 is regulated to be 20-30 ℃, and an infrared beam emitted by a Fourier infrared spectrometer is received by a detector of the Fourier infrared spectrometer after being totally reflected by an optical system; the sample stage is adjusted so that the sample aperture 35 and the background aperture 36 are located in sequence at the same vertical light convergence in the optical path of the optical system, resulting in the infrared spectrum of the base oil.
The horizontal plate of the L-shaped support 6 is fixed on the inner side surface of the vertical table of the mounting table 1 through screws, a pair of mounting grooves are formed in the vertical plate of the L-shaped support 6, and the mounting grooves are spaced and formed along the length direction of the vertical plate;
the first plane mirror 23, the second plane mirror 24, the first concave mirror 21 and the second concave mirror 22 are attached to the vertical plates of the corresponding L-shaped support 6, and a pair of bolts 61 are respectively inserted at the upper and lower ends of the pair of mounting grooves in a matched manner, so as to clamp and fix the corresponding first plane mirror 23, second plane mirror 24, first concave mirror 21 and second concave mirror 22.
The first plane mirror 23, the second concave mirror 22, the first concave mirror 21 and the second plane mirror 24 are correspondingly arranged at four right angles on the inner side surface of the vertical table of the mounting table 1 in sequence;
the second plane reflector 24 is arranged on the optical path of the infrared beam emitted by the Fourier infrared spectrometer and forms an angle of 45 degrees with the optical path; the first concave reflecting mirror 21 and the second plane reflecting mirror 24 are arranged in parallel, and the optical path between them is perpendicular to the optical path of the infrared beam of the light beam; the second concave reflecting mirror 22 and the mirror surface of the first concave reflecting mirror 21 are oppositely arranged to form a 90-degree included angle, and the light path between the second concave reflecting mirror and the mirror surface is parallel to the light path of the infrared light beam; the first plane mirror 23 and the second concave mirror 22 are arranged in parallel, and the optical path between them is perpendicular to the optical path of the infrared beam; the emergent light path of the first plane mirror 23 is parallel to the light path of the infrared light beam; the sample aperture 35 and the background aperture 36 are in turn located at the point of convergence of the perpendicular light rays between the first concave mirror 21 and the second planar mirror 24, i.e. at point O in fig. 2.
The screw rod of the ball screw nut pair 4 is arranged along the length direction of the horizontal table of the mounting table 1, and one end of the screw rod is connected with a motor 41; the upper end of the nut seat of the ball screw nut pair 4 is fixedly connected with the lower end of the supporting vertical rod 31. As shown in fig. 2, the infrared beam emitted by the fourier infrared spectrometer is received by the detector of the fourier infrared spectrometer after being totally reflected by the optical system, that is, the horizontally leftward focused infrared beam is diverted by the second plane mirror 24, then diverted and converged by the 2 first concave mirrors 21 and 22 for the second time and the third time, and finally diverted by the first plane mirror 23 for the fourth time, and the directions of the four diverts are as follows: vertically upward, horizontally leftward, vertically downward, horizontally leftward, and finally the light beams all enter the detector of the fourier infrared spectrometer. The optical system is a high-reflectivity mirror combination of the first plane mirror 23, the second concave mirror 22, the first concave mirror 21 and the second plane mirror 24, the total light transmittance is more than 90%, and the obtained spectrum quality is good.
Example 2
Referring to FIGS. 6 to 9, a detection method of base oil based on deep learning infrared multidimensional fusion algorithm IS provided, wherein the detection device in the above embodiment 1 IS adopted for detection, before sample measurement IS started, a Siemens Feicolet IS50 type Fourier infrared spectrometer IS started, an MCT detector IS preheated for 1 hour, and the wave number range IS 650-4000cm -1 At intervals of 2cm -1 The number of scans was 32.
The method specifically comprises the following steps:
step S1: respectively placing wafers with the same specification on the heating blocks 34 corresponding to the sample holes 35 and the background holes 36, and dripping base oil on the wafers corresponding to the sample holes 35; the temperature of the heating block 34 is controlled at 30 ℃;
step S2: the infrared light beam emitted by the Fourier infrared spectrometer is received by a detector of the Fourier infrared spectrometer after being totally reflected by the optical system; the position of the base plate 33 is adjusted so that the background hole 36 is positioned at the position where the vertical light rays between the first concave reflecting mirror 21 and the second plane reflecting mirror 24 converge, and a background spectrum is collected; the position of the base plate 33 is adjusted so that the sample hole 35 is positioned at the position where the vertical light rays between the first concave reflecting mirror 21 and the second plane reflecting mirror 24 are converged, and the infrared spectrum of the sample is collected;
then dividing the infrared spectrum of the sample by the background spectrum, and deducting interference signals of water vapor and carbon dioxide in the air to obtain the infrared spectrum of the base oil;
each sample is tested by one infrared spectrum, and the background spectrum of the zinc selenide sample cell is collected before each test, so that the background and baseline drift of air and instruments is eliminated, and the instrument test error is reduced; and the quality and accuracy of the background spectrum affects the quality of the sample spectrum.
Step S3: respectively collecting infrared spectrums of various types of base oils including I type base oil, II type base oil, III type base oil and IV type base oil (marks 25, 150, 250, 100BS, 150SBS, 1GN, 150SN, 100N, 150N, 250N, 500N, 350SN, 900SN, 1200N, PAO2, PAO4, PAO6, PAO10, PAO40 and PAO 100) according to the method of S2 to obtain spectrum data points; as shown in FIG. 6, the infrared spectra of the base oil samples are very similar, and almost no obvious difference is seen in FIG. 8, since the base oil is mainly composed of long-chain saturated aliphatic hydrocarbon, the infrared spectra of the four base oils are highly similar, and the infrared 2920cm of the base oil samples are infrared -1 ,2850cm -1 The peak of (C) corresponds to the stretching vibration of saturated hydrocarbon in the base oil, 1460cm -1 ,1376cm -1 Corresponds to the flexural vibration of saturated hydrocarbon, 722cm -1 The absorption peak of (2) corresponds to the in-plane swing vibration of long-chain saturated alkane, so the infrared spectrum acquisition range of the base oil is 525-4000cm -1 The sampling interval is 0.5cm -1 Thus 6949 spectral data points total.
Step S4: performing variable screening by using Interval Partial Least Squares (iPLS), dividing 6949 spectrum data points into 50 sections, respectively calculating a root mean square error value (RMSECV) of cross verification by adopting each section of cross verification, selecting a section of wavelength with low RMSECV value and corresponding to saturated alkane C-H vibration as an analysis object, and establishing a spectrum data set S;
the result is shown in FIG. 7, 1454-1521cm -1 And 2801-2995cm -1 Has low RMSECV value, and due to C-H vibration of saturated alkane, the two sections are 33-35 sections with corresponding infrared spectrum of 2801-2995cm -1 As an analysis object, a spectral data set S is established.
Wherein the calculation formula of the RMSECV value is as follows:
wherein: n is the number of data set samples, y i 、Respectively an actual measurement value and a predicted value of an ith sample in the data set;
step S5: the spectral data set S is noisy by continuous wavelet transform to obtain a spectral data set M,
wherein the continuous wavelet transform function is as follows:
where f (t) is the spectral matrix, t is the wavenumber,is a wavelet basis function, alpha is a translation parameter, b is a telescoping factor, C is a wavelet coefficient, and dt represents integration of the wave number t;
in the embodiment, daubechies db3 wavelet basis function is selected, and the decomposition scale is 2;
the two-dimensional synchronous correlation spectrum analysis of the double spectrum refers to: selecting infrared average spectrums of class III base oils 100N, 150N and 250N from a spectrum data set M as reference spectrums r (v), respectively taking 204 base oil infrared spectrums as samples s (v), carrying out double-spectrum two-dimensional synchronous correlation spectrum analysis,
the partial results are shown in FIG. 8, which shows the partial two-dimensional infrared spectrograms of the base oils of the invention of the group I, the group II, the group III and the group IV, and obvious differences can be seen from the two-dimensional infrared spectrograms;
the formula is as follows:
where s (v) is the sample spectrum matrix, r (v) is the reference spectrum matrix, v 1 ，v 2 Respectively represent to proceedThe two spectra are compared.
Then selecting the automatic peak of the two-dimensional synchronous correlation spectrum of the double spectrum,
as shown in FIG. 9, the two-dimensional infrared spectrum autocorrelation spectrum of the present invention has a simple pattern and a significant difference compared with the original infrared spectrum and the two-dimensional infrared spectrum. And establishing a spectrum matrix data set N according to the data set, respectively carrying out standardization and normalization processing on the M and the N, and then obtaining a data set D after fusion.
Step S6: and inputting the data set D into a CNN convolutional neural network to perform CNN model training and prediction, wherein the data set D is used as the input of a model, and the base oil type is used as the output of the model for rapid nondestructive detection of the base oil type.
The CNN model comprises a one-dimensional convolution layer, an activation layer, a pooling layer and a full-connection layer; the number of convolution kernels of one-dimensional convolution operation is 16, the size is 6*1, and the step length of convolution is 4; and the ReLU completes the activation operation, so that the neurons in the neural network have sparse activation; the pooling operation uses an average pooling model, the pooling window of 4*1 is taken each time, and the step length is set to be 1; inputting the vector into a full connection layer, and outputting a one-dimensional high-order vector; four types of base oil are used, and the number of nodes in the full-connection layer is 2316; the formula of the ReLU function is as follows:
f(x)＝max(0，0)
wherein x is an input value in CNN, and f (x) is an output value;
the CNN convolutional neural network model adopts a learning rate attenuation mechanism, wherein the initial value of the learning rate is set to be 0.001, the attenuation index is set to be 0.1, the learning rate is attenuated gradually over time in the process of network training to dynamically adjust, and the initial value of each layer of weight is subjected to zero-mean Gaussian distribution with standard deviation of 0.1.
Establishing CNN prediction model on the correction set, inputting spectrum in the prediction set into CNN convolutional neural network model to obtain discrimination result of sample base oil type of the prediction set, comparing with actual type, and comparing with the result shown in Table 1,
table 1 results of correction and prediction models under different modeling methods, variable selection and fusion methods
Table 1 shows the comparison results of the original spectrum full-band PLS-DA model and the CNN model, namely 2801-2995cm -1 Modeling results of wave band PLS-DA and CNN and 2801-2995cm -1 The modeling result of the wave band multidimensional spectrum CNN is compared with the modeling result to prove that the training set accuracy train_acc and the test set accuracy test_acc deep learning (CNN) discrimination model are better than the traditional PLS-DA effect no matter after full wave band screening or after wave band screening and before data fusion. In the deep learning model, the accuracy of the training set of the full spectrum model, the band selection model and the band selection 2T2D model can reach 1, and the accuracy of the prediction set is close to 1. However, the accuracy of the training set and the prediction set reaches 1 by combining the data fusion-band selection-2T 2D model with the deep learning, so that the base oil type can be accurately identified, namely the base oil detection device and the test method based on the deep learning infrared multidimensional fusion algorithm have the highest base oil detection accuracy, the robustness is enhanced, and the base oil type detection device and the test method can be well used for accurately and rapidly identifying the base oil type.
It will be readily appreciated by those skilled in the art that the foregoing description is merely a preferred embodiment of the invention and is not intended to limit the invention, but any modifications, equivalents, improvements or alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (7)
1. A base oil detection method based on a deep learning infrared multidimensional fusion algorithm,
the base oil detection method is based on a base oil detection device based on a deep learning infrared multidimensional fusion algorithm,
the base oil detection device comprises a Fourier infrared spectrometer and a sample mechanism arranged in a sample bin of the Fourier infrared spectrometer,
the sample mechanism comprises an L-shaped mounting table (1), an optical system and a sample rack, wherein the sample rack is mounted on a horizontal table of the mounting table (1), and the optical system is mounted on the inner side surface of a vertical table of the mounting table (1);
the sample rack comprises a support vertical rod (31), a support cross rod (32) and a sample base (3), the lower end of the support vertical rod (31) is vertically arranged on a horizontal platform of the mounting platform (1) through a ball screw nut pair (4), the upper end of the support vertical rod (31) is fixedly connected with the horizontal end of the support cross rod (32) through a horizontally arranged cylindrical adapter (5),
the sample base (3) comprises a square base plate (33) and a heating block (34) arranged at the upper end of the base plate (33), a sample hole (35) and a background hole (36) are respectively formed through the heating block (34) and the corresponding base plate (33), and the base plate (33) is fixedly connected with the horizontal other end of the supporting cross rod (32);
the optical system comprises a first plane reflecting mirror (23), a second plane reflecting mirror (24), a first concave reflecting mirror (21) and a second concave reflecting mirror (22), wherein the first plane reflecting mirror (23), the second plane reflecting mirror (24), the first concave reflecting mirror (21) and the second concave reflecting mirror (22) are vertically arranged on the inner side surface of a vertical table of the mounting table (1) through L-shaped supporting seats (6) respectively;
in the test, two identical wafers are taken, base oil is dripped on one wafer and the wafer is placed at the upper end of a heating block (34) corresponding to a sample hole (35), the other wafer is placed at the upper end of the heating block (34) corresponding to a background hole (36), the temperature of the heating block (34) is regulated to be 20-30 ℃, and infrared light beams emitted by a Fourier infrared spectrometer are received by a detector of the Fourier infrared spectrometer after being totally reflected by an optical system; the sample stage is adjusted so that a sample hole (35) and a background hole (36) are sequentially positioned at the same vertical light converging position in the light path of the optical system to obtain the infrared spectrum of the base oil, and the base oil detection method specifically comprises the following steps:
step S1: respectively placing wafers with the same specification on heating blocks (34) corresponding to the sample holes (35) and the background holes (36), and dripping base oil on the wafers corresponding to the sample holes (35);
step S2: the infrared light beam emitted by the Fourier infrared spectrometer is received by a detector of the Fourier infrared spectrometer after being totally reflected by the optical system; the position of the base plate (33) is adjusted, so that a background hole (36) is positioned at the position where the vertical light rays between the first concave reflecting mirror (21) and the second plane reflecting mirror (24) are converged, and a background spectrum is collected; then, the position of the base plate (33) is adjusted to enable the sample hole (35) to be positioned at the position where the vertical light rays between the first concave reflecting mirror (21) and the second plane reflecting mirror (24) are converged, and the infrared spectrum of the sample is collected;
then dividing the infrared spectrum of the sample by the background spectrum, and deducting interference signals of water vapor and carbon dioxide in the air to obtain the infrared spectrum of the base oil;
step S3: respectively collecting infrared spectrums of base oils of class I, class II, class III and class IV according to the method of S2 to obtain spectrum data points;
step S4: performing variable screening by using iPLS interval partial least square to divide spectrum data points into n sections, respectively calculating RMS error values of RMSECV cross verification by adopting each section of cross verification, selecting a section of wavelength which has low RMSECV value and corresponds to saturated alkane C-H vibration as an analysis object, and establishing a spectrum data set S;
wherein: the calculation formula of RMSECV value is as follows:
wherein: n is the number of data set samples, y i 、Respectively an actual measurement value and a predicted value of an ith sample in the data set;
step S5: noise filtering is carried out on the spectrum data set S through continuous wavelet transformation to obtain a spectrum data set M, automatic peaks of a double-spectrum two-dimensional synchronous correlation spectrum are selected through double-spectrum two-dimensional synchronous correlation spectrum analysis, a spectrum matrix data set N is established, standardization and normalization processing are carried out on the M and the N respectively, and then a data set D is obtained after fusion;
wherein: the continuous wavelet transform function is as follows:
where f (t) is the spectral matrix, t is the wavenumber,is a wavelet basis function, alpha is a translation parameter, b is a telescoping factor, C is a wavelet coefficient, and dt represents integration of the wave number t;
step S6: inputting the data set D into a CNN convolutional neural network to carry out CNN model training and prediction, wherein the data set D is used as the input of a model, and the base oil type is used as the output of the model for quick nondestructive detection of the base oil type; the accuracy of the detection result is 100%, and the base oil of the I, II, III and IV types can be accurately identified.
2. The base oil detection method based on the deep learning infrared multidimensional fusion algorithm according to claim 1, wherein the method is characterized by comprising the following steps of: the wafer in the step S1 is a zinc selenide wafer with the thickness of 25 micrometers, and the temperature of the heating block (34) is controlled to be 30 ℃.
3. The base oil detection method based on the deep learning infrared multidimensional fusion algorithm according to claim 1, wherein the method is characterized by comprising the following steps of: the two-dimensional synchronous correlation spectrum analysis of the double spectrum in the step S5 refers to: and selecting infrared average spectrums of 3 base oils in the III base oils or IV base oils from the spectrum data set M as reference spectrums r (v), respectively taking not less than 30 base oil infrared spectrums as samples s (v), and carrying out double-spectrum two-dimensional synchronous correlation spectrum analysis.
4. The base oil detection method based on the deep learning infrared multidimensional fusion algorithm according to claim 1, wherein the method is characterized by comprising the following steps of: the CNN model in the step S6 comprises a one-dimensional convolution layer, an activation layer, a pooling layer and a full connection layer; the number of convolution kernels of one-dimensional convolution operation is 16, the size is 6*1, and the step length of convolution is 4; and the ReLU completes the activation operation, so that the neurons in the neural network have sparse activation; the pooling operation uses an average pooling model, the pooling window of 4*1 is taken each time, and the step length is set to be 1; inputting the vector into a full connection layer, and outputting a one-dimensional high-order vector; four types of base oil are used, and the number of nodes in the full-connection layer is 2316; the formula of the ReLU function is as follows:
f(x)＝max(0，x)
wherein x is an input value in CNN, and f (x) is an output value;
the CNN convolutional neural network model adopts a learning rate attenuation mechanism, the initial value of the learning rate is set to be 0.001, the attenuation index is 0.1, the learning rate is attenuated gradually along with time in the process of network training to dynamically adjust, and the initial value of each layer of weight is subjected to zero-mean Gaussian distribution with standard deviation of 0.1.
5. The base oil detection method based on the deep learning infrared multidimensional fusion algorithm according to claim 1, wherein the method is characterized by comprising the following steps of: the horizontal plate of the L-shaped support (6) is fixed on the inner side surface of the vertical table of the mounting table (1) through screws, a pair of mounting grooves are formed in the vertical plate of the L-shaped support (6), and the mounting grooves are spaced and formed along the length direction of the vertical plate;
the first plane reflector (23), the second plane reflector (24), the first concave reflector (21) or the second concave reflector (22) are attached to the vertical plate of the corresponding L-shaped support (6), and a pair of bolts (61) are respectively inserted at the upper end and the lower end of the pair of mounting grooves in a matched mode and used for clamping and fixing the corresponding first plane reflector (23), second plane reflector (24), first concave reflector (21) or second concave reflector (22) from the upper end and the lower end.
6. The method for detecting base oil based on the deep learning infrared multidimensional fusion algorithm according to claim 5, wherein the method is characterized by comprising the following steps: the first plane reflecting mirror (23), the second concave reflecting mirror (22), the first concave reflecting mirror (21) and the second plane reflecting mirror (24) are sequentially and correspondingly arranged at four right angles on the inner side surface of the vertical table of the mounting table (1);
the second plane reflector (24) is arranged on the optical path of the infrared light beam emitted by the Fourier infrared spectrometer and forms an angle of 45 degrees with the optical path; the first concave reflecting mirror (21) and the second plane reflecting mirror (24) are oppositely arranged in parallel, and the light path between the first concave reflecting mirror and the second plane reflecting mirror is perpendicular to the light path of the infrared light beam; the second concave reflecting mirror (22) and the mirror surface of the first concave reflecting mirror (21) are oppositely arranged to form a 90-degree included angle, and the light path between the second concave reflecting mirror and the mirror surface is parallel to the light path of the infrared light beam; the first plane reflecting mirror (23) and the second concave reflecting mirror (22) are oppositely arranged in parallel, and the light path between the first plane reflecting mirror and the second plane reflecting mirror is perpendicular to the light path of the infrared light beam; the emergent light path of the first plane reflector (23) is parallel to the light path of the infrared light beam;
the sample hole (35) and the background hole (36) are sequentially positioned at the vertical light converging position between the first concave reflecting mirror (21) and the second plane reflecting mirror (24).
7. The base oil detection method based on the deep learning infrared multidimensional fusion algorithm according to claim 1, wherein the method is characterized by comprising the following steps of: the screw rod of the ball screw nut pair (4) is arranged along the length direction of the horizontal table of the mounting table (1), and one end of the screw rod is connected with a motor (41); the upper end of the nut seat of the ball screw nut pair (4) is fixedly connected with the lower end of the supporting vertical rod (31).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310242896.3A CN116242799B (en) | 2023-03-14 | 2023-03-14 | Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310242896.3A CN116242799B (en) | 2023-03-14 | 2023-03-14 | Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116242799A CN116242799A (en) | 2023-06-09 |
| CN116242799B true CN116242799B (en) | 2023-08-18 |
Family
ID=86624030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310242896.3A Active CN116242799B (en) | 2023-03-14 | 2023-03-14 | Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116242799B (en) |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH05232018A (en) * | 1992-02-24 | 1993-09-07 | Nec Corp | Infrared absorption spectrum measuring method for solid |
| CN101694455A (en) * | 2009-09-29 | 2010-04-14 | 武汉理工大学 | Horizontal sample frame of Fourier infrared spectrograph |
| CN101782511A (en) * | 2010-03-31 | 2010-07-21 | 中国人民解放军总后勤部油料研究所 | Method for rapidly identifying type and grade of lubricating oil |
| CN101900672A (en) * | 2009-05-27 | 2010-12-01 | 中国石油化工股份有限公司 | Method for quickly identifying class and viscosity grade of lubricating oil |
| CN102841071A (en) * | 2011-06-23 | 2012-12-26 | 中国石油化工股份有限公司 | Method for identifying types of crude oil by using two-dimensional correlation infrared asynchronization spectrum |
| WO2016023779A1 (en) * | 2014-08-11 | 2016-02-18 | Topnir Systems Sas | Methods of determining properties of oil |
| CN107179292A (en) * | 2016-03-10 | 2017-09-19 | 中国农业机械化科学研究院 | Different near infrared spectrum variable preferred result fusion methods and application |
| CN107748146A (en) * | 2017-10-20 | 2018-03-02 | 华东理工大学 | A kind of crude oil attribute method for quick predicting based near infrared spectrum detection |
| CN107966420A (en) * | 2016-10-19 | 2018-04-27 | 中国石油化工股份有限公司 | A kind of method by near infrared spectrum prediction oil property |
| CN109142265A (en) * | 2018-08-24 | 2019-01-04 | 武汉轻工大学 | A kind of edible oil type method for quick identification of near infrared spectrum Wavelet Fusion |
| CN109724938A (en) * | 2017-10-30 | 2019-05-07 | 中国石油化工股份有限公司 | By the method for near infrared spectrum prediction lube base oil nature |
| CN110398472A (en) * | 2019-06-28 | 2019-11-01 | 天津大学 | A kind of vegetable oil identification classification method based on two-dimensional correlation spectra quantization |
| CN110749575A (en) * | 2019-10-17 | 2020-02-04 | 夏永刚 | Traditional Chinese medicine polysaccharide two-dimensional infrared spectrum identification prediction model and construction method and application thereof |
| CN110907388A (en) * | 2019-11-22 | 2020-03-24 | 光钙(上海)高科技有限公司 | Oil spill type identification method based on infrared spectroscopic analysis |
| CN111257305A (en) * | 2020-03-31 | 2020-06-09 | 北京农业智能装备技术研究中心 | Two-dimensional correlation LIBS spectral measurement method, system and device |
| WO2022160662A1 (en) * | 2021-02-01 | 2022-08-04 | 广东省农业科学院蔬菜研究所 | Method for measuring content of sugar in pumpkins by means of near infrared spectrum instrument |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008032694A (en) * | 2006-07-04 | 2008-02-14 | Dkk Toa Corp | Method of identifying kind of oil, and kind of oil identifying unit |
| US9222881B2 (en) * | 2012-02-24 | 2015-12-29 | Massachusetts Institute Of Technology | Vibrational spectroscopy for quantitative measurement of analytes |
-
2023
- 2023-03-14 CN CN202310242896.3A patent/CN116242799B/en active Active
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH05232018A (en) * | 1992-02-24 | 1993-09-07 | Nec Corp | Infrared absorption spectrum measuring method for solid |
| CN101900672A (en) * | 2009-05-27 | 2010-12-01 | 中国石油化工股份有限公司 | Method for quickly identifying class and viscosity grade of lubricating oil |
| CN101694455A (en) * | 2009-09-29 | 2010-04-14 | 武汉理工大学 | Horizontal sample frame of Fourier infrared spectrograph |
| CN101782511A (en) * | 2010-03-31 | 2010-07-21 | 中国人民解放军总后勤部油料研究所 | Method for rapidly identifying type and grade of lubricating oil |
| CN102841071A (en) * | 2011-06-23 | 2012-12-26 | 中国石油化工股份有限公司 | Method for identifying types of crude oil by using two-dimensional correlation infrared asynchronization spectrum |
| WO2016023779A1 (en) * | 2014-08-11 | 2016-02-18 | Topnir Systems Sas | Methods of determining properties of oil |
| CN107179292A (en) * | 2016-03-10 | 2017-09-19 | 中国农业机械化科学研究院 | Different near infrared spectrum variable preferred result fusion methods and application |
| CN107966420A (en) * | 2016-10-19 | 2018-04-27 | 中国石油化工股份有限公司 | A kind of method by near infrared spectrum prediction oil property |
| CN107748146A (en) * | 2017-10-20 | 2018-03-02 | 华东理工大学 | A kind of crude oil attribute method for quick predicting based near infrared spectrum detection |
| CN109724938A (en) * | 2017-10-30 | 2019-05-07 | 中国石油化工股份有限公司 | By the method for near infrared spectrum prediction lube base oil nature |
| CN109142265A (en) * | 2018-08-24 | 2019-01-04 | 武汉轻工大学 | A kind of edible oil type method for quick identification of near infrared spectrum Wavelet Fusion |
| CN110398472A (en) * | 2019-06-28 | 2019-11-01 | 天津大学 | A kind of vegetable oil identification classification method based on two-dimensional correlation spectra quantization |
| CN110749575A (en) * | 2019-10-17 | 2020-02-04 | 夏永刚 | Traditional Chinese medicine polysaccharide two-dimensional infrared spectrum identification prediction model and construction method and application thereof |
| CN110907388A (en) * | 2019-11-22 | 2020-03-24 | 光钙(上海)高科技有限公司 | Oil spill type identification method based on infrared spectroscopic analysis |
| CN111257305A (en) * | 2020-03-31 | 2020-06-09 | 北京农业智能装备技术研究中心 | Two-dimensional correlation LIBS spectral measurement method, system and device |
| WO2022160662A1 (en) * | 2021-02-01 | 2022-08-04 | 广东省农业科学院蔬菜研究所 | Method for measuring content of sugar in pumpkins by means of near infrared spectrum instrument |
Non-Patent Citations (1)
| Title |
|---|
| 矿物油型和聚α-烯烃型基础油的鉴别方法;王宸;《检验检疫学刊》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116242799A (en) | 2023-06-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109211803B (en) | Device for rapidly identifying micro plastic based on microscopic multispectral technology | |
| EP0841548A2 (en) | Systems and methods for processing and analyzing terahertz waveforms | |
| CN106525759B (en) | A method of honey types are identified based on decaying total reflection Terahertz dielectric spectra | |
| CN101750024B (en) | Near infrared spectrum detection method for insulation coating layer thickness of silicon steel | |
| CN103134765A (en) | Chinese medicine sample authenticity preliminary screening method based on terahertz time-domain spectrum | |
| CN107703097B (en) | Method for constructing model for rapidly predicting crude oil property by using near-infrared spectrometer | |
| CN107101948B (en) | Sea surface oil spill detection method based on polarization absorption characteristic | |
| CN103674854A (en) | Apple grading method based on synergic interval partial least square discriminant analysis | |
| Ren et al. | White dwarf-main sequence binaries from LAMOST: the DR1 catalogue | |
| CN109444092B (en) | Test system and method for identifying oil product types by combining terahertz and laser | |
| CN107631993A (en) | A kind of infrared spectrum recognition methods based on the residual differential spectrum analysis of steam | |
| CN105784672A (en) | Drug detector standardization method based on dual-tree complex wavelet algorithm | |
| CN111380809A (en) | Method for testing oil film type based on polarization characteristic | |
| CN111351640B (en) | Optical lens quality detection system and method based on double-optical-comb interferometry | |
| KR102141228B1 (en) | Method and apparatus for measuring physical quantity of a thin layer using terahertz spectroscopy | |
| CN116242799B (en) | Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm | |
| CN110907388A (en) | Oil spill type identification method based on infrared spectroscopic analysis | |
| CN111781584A (en) | Based on target radiation O2Passive distance measuring method of absorption correlation K distribution method | |
| CN116718613A (en) | Terahertz spectrum-based epoxy glass fiber composite material internal defect detection method | |
| CN106525764A (en) | Material surface finish detection method based on terahertz time-domain spectroscopy | |
| CN108489928B (en) | Method for detecting textile fiber components by short-wave infrared extinction spectrum | |
| CN106970042B (en) | Method for detecting impurity and moisture content of carrageenin | |
| CN112198136A (en) | Nondestructive detection method for turbine oil acid value based on mid-infrared spectrum | |
| CN110836860B (en) | Surface enhanced infrared substrate based on metamaterial and molecular detection method thereof | |
| CN114112985A (en) | Near-infrared spectrometer and near-infrared online detection method |
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 |