阅读说明：本技术 基于近红外建模的成品油添加剂检测分析方法及装置 (Near-infrared modeling-based product oil additive detection and analysis method and device ) 是由 丁焰 杜彪 尹航 吉喆 鲁冰 朱云鹏 裴修尧 周玉山 于 2021-07-29 设计创作，主要内容包括：本公开提供一种基于近红外建模的成品油添加剂检测分析方法及装置,能够根据成品油添加剂样品的近红外检测值以及各个成分的湿化学值进行结合关联,进而构建对应的数据库,形成近红外分析模型,这样,就可以利用近红外分析模型对待检测的成品油添加剂的各项成分含量数据进行检测分析,使得检测得到的各成分含量的准确度得到有效提高。(The invention provides a method and a device for detecting and analyzing a finished oil additive based on near-infrared modeling, which can be used for carrying out combined association according to a near-infrared detection value of a finished oil additive sample and a wet chemical value of each component so as to construct a corresponding database and form a near-infrared analysis model, so that each component content data of the finished oil additive to be detected can be detected and analyzed by using the near-infrared analysis model, and the accuracy of each detected component content is effectively improved.)

1. A product oil additive detection and analysis method based on near infrared modeling is characterized by comprising the following steps:

s1, obtaining n finished oil additive samples;

s2, acquiring the near infrared spectrum of each product oil additive sample by using near infrared equipment;

s3, detecting the components of each product oil additive sample by using a wet chemical method to obtain a wet chemical value corresponding to the components of each product oil additive sample;

s4, selecting a region with the largest correlation of information quantity from the near infrared spectrum of each product oil additive sample as a near infrared spectrum region by utilizing a pre-constructed region selection neural network model;

s5, performing first derivative treatment on the near infrared spectrum region of each product oil additive sample, and performing multiple linear regression treatment to obtain a near infrared detection value;

s6, correlating the near infrared detection value and the wet chemical value of each product oil additive sample by utilizing partial least squares operation processing to obtain correlated data of each product oil additive sample;

s7, calculating the root mean square error between the near infrared detection value associated with each product oil additive sample and the corresponding wet chemical value, and deleting the associated data of the product oil additive samples with the root mean square error larger than the error threshold;

s8, constructing a database by using the correlated data of the residual product oil additive samples, and taking the database as a near infrared analysis initial model;

s9, performing cross inspection on the near-infrared analysis initial model by using n finished oil additive samples to obtain a near-infrared analysis model;

s10, acquiring a corresponding near infrared spectrum to be analyzed by using infrared equipment for the finished oil additive to be detected, and processing the finished oil additive to be detected by using a humidification method to obtain a wet chemical value to be analyzed corresponding to the component of the finished oil additive to be detected;

and S11, inputting the near infrared spectrum to be analyzed of the finished oil additive to be detected and the corresponding wet chemical value to be analyzed into the near infrared analysis model, and processing by using the near infrared model to determine the content data of each component of the finished oil additive to be detected.

2. The method of claim 1, wherein the constructing of the region-selective neural network model comprises:

SA1, constructing a deep neural network with one input layer, a plurality of hidden layers and one output layer, wherein the number of the hidden layers corresponds to the category number of the product oil additives;

SA2, acquiring a predetermined number of finished oil additives, acquiring near infrared spectrums of the finished oil additives by using near infrared equipment, marking target areas in spectral areas with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrums of the predetermined number of marked finished oil additives as training samples;

SA3, classifying the training samples according to the categories of the finished oil additives;

SA4, extracting training samples of a target class from the training samples of multiple classes;

SA5, inputting the training sample of the target class to an input layer of the deep neural network, preprocessing the training sample of the target class by using the input layer, and sending the preprocessed training sample to a hidden layer corresponding to the target class for processing;

SA6, the hidden layer corresponding to the target class sends the processing result to the output layer, and the output layer arranges the processing result into a corresponding spectrum region result for output;

SA7, judging whether the output spectrum region result is the same as the corresponding target region mark, if so, selecting the next training sample of the target class for training, otherwise, calculating a loss function between the output spectrum region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer corresponding to the target class in the deep neural network according to the loss value until the output spectrum region result is the same as the corresponding target region mark;

SA8, continuously repeating the steps SA5 to SA7 until all training samples of the target class are trained to obtain a trained neural network;

SA9, extracting training samples of a new target class from the training samples of the remaining classes, repeating the processes from the step SA5 to the step SA8 to train the trained neural network until all training samples of all classes are completely trained, and taking the finally obtained neural network as a region selection neural network model.

3. The method of claim 1, wherein the constructing of the region-selective neural network model comprises:

SB1, obtaining a predetermined number of finished oil additives, obtaining the near infrared spectrum of each finished oil additive by using near infrared equipment, marking the target region in the spectral region with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrum of the predetermined number of marked finished oil additives as a training sample;

SB2, classifying the training samples according to the categories of the finished oil additives to obtain training samples of M categories;

SB3, constructing M deep neural networks with an input layer, a hidden layer and an output layer;

SB4, obtaining a training sample of a target class and a corresponding deep neural network, and training the selected deep neural network by using the training sample of the target class to obtain a neural network initial model capable of identifying the spectral region with the maximum correlation of the near infrared spectrum of the corresponding class, wherein M deep neural networks are trained to obtain M neural network initial models corresponding to class labels;

SB5, integrating the M neural network initial models, adding a category judgment layer before the M neural network initial models, and adding an output result sorting layer after the M neural network initial models to obtain the region selection neural network model.

4. The method according to claim 3, wherein step SB4 specifically comprises:

SB41, inputting the training samples of the target class to the input layer of the corresponding deep neural network, preprocessing the training samples of the target class by using the input layer, and sending the preprocessed training samples to the hidden layer for processing;

SB42, the hidden layer sends the processing result to the output layer, and the output layer arranges the processing result into the corresponding spectrum region result for output;

SB43, judging whether the output spectrum region result is the same as the corresponding target region mark, if so, selecting the next training sample of the target class for training, otherwise, calculating a loss function between the output spectrum region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer in the deep neural network according to the loss value until the output spectrum region result is the same as the corresponding target region mark;

and SB44, repeating the steps from SB41 to SB43 until all the training samples of the target class are trained to obtain the initial model of the neural network.

5. The method according to any one of claims 1 to 4, wherein step S9 specifically comprises:

s9-1, sorting N finished oil additive samples N₁，N₂……N_nSelecting N from N finished oil additive samples₁The finished oil additive sample is used as a test sample, and the rest finished oil additive sample is used as a modeling sample;

s9-2, utilizing the modeling sample to construct a near infrared analysis initial model again on the constructed near infrared analysis initial model according to the scheme of the steps S2 to S8;

s9-3, acquiring near infrared spectra corresponding to various components of a test sample by using infrared equipment, processing the test sample by using a humidification method to obtain wet chemical values corresponding to the various components, inputting the wet chemical values into a reconstructed near infrared analysis initial model for processing, outputting content data of the various components of the test sample, and comparing the output content data of the various components of the test sample with real data of the test sample to obtain the accuracy of the test;

s9-4, sequentially selecting the next product oil additive sample as a test sample according to the sequence, taking the rest product oil additive samples as modeling samples, and repeating the schemes of the steps S9-2 and S9-3 until the n product oil additive samples are respectively taken as test samples to obtain the accuracy rate p of corresponding test₁，p₂……p_n；

S9-5, calculating the accuracy rate p of the test₁，p₂……p_nAnd judging whether P is larger than or equal to P, if so, taking the finally obtained near-infrared analysis initial model as a near-infrared model, otherwise, obtaining a new finished oil additive sample again, repeating the processes from S9-1 to S9-4 by using the new finished oil additive sample until the obtained P is larger than or equal to P, and taking the finally obtained near-infrared analysis initial model as the near-infrared model, wherein P is an accuracy threshold value.

6. The method according to any one of claims 1-5, wherein after step S9, the method further comprises:

s9', obtaining a product oil additive samples as a correction set, obtaining b product oil additive samples as a verification set, building a near-infrared analysis verification model on the built near-infrared analysis model again by using the product oil additive samples in the correction set according to the scheme from the step S2 to the step S8, sequentially inputting the product oil additive samples in the verification set into the near-infrared analysis verification model to output verification results, comparing the verification results with the real results of the product oil additive samples in the verification set, and determining the verification accuracy;

and S9 ', judging whether the verification accuracy exceeds a verification threshold value, if so, taking the obtained near-infrared analysis verification model as a final near-infrared analysis model, otherwise, obtaining a new correction set and a new verification set, repeating the scheme of the step S9' until the obtained verification accuracy exceeds the verification threshold value, and taking the obtained near-infrared analysis verification model as the final near-infrared analysis model.

7. The method according to claim 1, wherein the multiple linear regression processing procedure in step S5 specifically includes:

multiplying at least one principal component spectrum in the near-infrared spectrum region after the first derivative processing by the corresponding principal component score coefficient to obtain corresponding principal component spectrum data, and constructing at least one principal component matrix according to the at least one principal component spectrum data;

and constructing at least one main linear function according to the at least one principal component matrix, and performing linear regression processing on the at least one main linear function to obtain a corresponding near-infrared detection value.

8. The method according to claim 7, wherein the operation processing procedure of the partial least squares method in step S6 specifically includes:

constructing a corresponding spectrum matrix according to the near-infrared detection value, constructing a main wet chemical matrix according to a main wet chemical value in the wet chemical values, correlating the spectrum matrix with the main wet chemical matrix, and calculating a prediction linear relation function of the two matrixes;

step S8, specifically, it is: and (4) constructing a database by using the predicted linear relation functions obtained by the residual product oil additive samples, and taking the database as a near infrared analysis initial model.

9. A product oil additive detection and analysis device based on near infrared modeling is characterized by comprising:

the acquisition module is used for acquiring n finished oil additive samples;

the near-infrared acquisition module is used for acquiring the near-infrared spectrum of each product oil additive sample by using near-infrared equipment;

the wet chemical processing module is used for detecting the components of each product oil additive sample by using a wet chemical method to obtain a wet chemical value corresponding to the components of each product oil additive sample;

the selecting module is used for selecting a region with the maximum correlation of information quantity from the near infrared spectrum of each product oil additive sample as a near infrared spectrum region by utilizing a pre-constructed region selection neural network model;

the mathematical processing module is used for performing first derivative processing on the near infrared spectrum region of each product oil additive sample and then performing multiple linear regression processing to obtain a near infrared detection value;

the correlation processing module is used for correlating the near infrared detection value and the wet chemical value of each product oil additive sample by utilizing partial least square operation processing to obtain correlated data of each product oil additive sample;

the deleting module is used for calculating the root mean square error between the correlated near-infrared detection value and the corresponding wet-chemical value of each product oil additive sample, and deleting the correlated data of the product oil additive samples with the root mean square error larger than the error threshold;

the construction module is used for constructing a database by utilizing the correlated data of the residual product oil additive samples, and taking the database as a near infrared analysis initial model;

the cross inspection module is used for carrying out cross inspection on the near-infrared analysis initial model by using n finished oil additive samples to obtain a near-infrared analysis model;

the pretreatment module to be detected is used for acquiring a corresponding near infrared spectrum to be analyzed of the finished oil additive to be detected by using infrared equipment, and processing the finished oil additive to be detected by using a humidification method to obtain a wet chemical value to be analyzed corresponding to the component of the finished oil additive to be detected;

and the detection analysis module is used for inputting the near infrared spectrum to be analyzed corresponding to each component of the finished oil additive to be detected and the wet chemical value to be analyzed corresponding to each component into the near infrared analysis model, and processing and determining the content data of each component of the finished oil additive to be detected by using the near infrared model.

10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of any one of claims 1 to 8 when executing the program.

Technical Field

The disclosure relates to the technical field of oil product detection, in particular to a method and a device for detecting and analyzing a product oil additive based on near-infrared modeling.

Background

The near infrared spectrum analysis method is widely applied to oil property analysis, and compared with the traditional laboratory test method, the method has the advantages of high analysis speed, high precision, low consumption and the like.

However, in the near infrared spectrum analysis in the prior art, only the near infrared spectrum of the finished oil additive is generally analyzed, and the content of each component of the finished oil additive or the corresponding data index is determined according to the near infrared spectrum. The single near infrared spectrum analysis may cause inaccurate detection of various indexes of the finished oil additive, and the detection precision is not high.

Disclosure of Invention

In view of the above, the present disclosure is directed to a method and an apparatus for detecting and analyzing a product oil additive based on near infrared modeling, which can solve or partially solve the above technical problems.

In view of the above, the present disclosure provides, in a first aspect, a method for detecting and analyzing a product oil additive based on near-infrared modeling, including:

s1, obtaining n finished oil additive samples;

s2, acquiring the near infrared spectrum of each product oil additive sample by using near infrared equipment;

s3, detecting the components of each product oil additive sample by using a wet chemical method to obtain a wet chemical value corresponding to the components of each product oil additive sample;

s4, selecting a region with the largest correlation of information quantity from the near infrared spectrum of each product oil additive sample as a near infrared spectrum region by utilizing a pre-constructed region selection neural network model;

s5, performing first derivative treatment on the near infrared spectrum region of each product oil additive sample, and performing multiple linear regression treatment to obtain a near infrared detection value;

s6, correlating the near infrared detection value and the wet chemical value of each product oil additive sample by utilizing partial least squares operation processing to obtain correlated data of each product oil additive sample;

s7, calculating the root mean square error between the near infrared detection value associated with each product oil additive sample and the corresponding wet chemical value, and deleting the associated data of the product oil additive samples with the root mean square error larger than the error threshold;

s8, constructing a database by using the correlated data of the residual product oil additive samples, and taking the database as a near infrared analysis initial model;

s9, performing cross inspection on the near-infrared analysis initial model by using n finished oil additive samples to obtain a near-infrared analysis model;

s10, acquiring a corresponding near infrared spectrum to be analyzed by using infrared equipment for the finished oil additive to be detected, and processing the finished oil additive to be detected by using a humidification method to obtain a wet chemical value to be analyzed corresponding to the component of the finished oil additive to be detected;

and S11, inputting the near infrared spectrum to be analyzed of the finished oil additive to be detected and the corresponding wet chemical value to be analyzed into the near infrared analysis model, and processing by using the near infrared model to determine the content data of each component of the finished oil additive to be detected.

Further, the construction process of the area selection neural network model comprises the following steps:

SA1, constructing a deep neural network with one input layer, a plurality of hidden layers and one output layer, wherein the number of the hidden layers corresponds to the category number of the product oil additives;

SA2, acquiring a predetermined number of finished oil additives, acquiring near infrared spectrums of the finished oil additives by using near infrared equipment, marking target areas in spectral areas with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrums of the predetermined number of marked finished oil additives as training samples;

SA3, classifying the training samples according to the categories of the finished oil additives;

SA4, extracting training samples of a target class from the training samples of multiple classes;

SA5, inputting the training sample of the target class to an input layer of the deep neural network, preprocessing the training sample of the target class by using the input layer, and sending the preprocessed training sample to a hidden layer corresponding to the target class for processing;

SA6, the hidden layer corresponding to the target class sends the processing result to the output layer, and the output layer arranges the processing result into a corresponding spectrum region result for output;

SA7, judging whether the output spectrum region result is the same as the corresponding target region mark, if so, selecting the next training sample of the target class for training, otherwise, calculating a loss function between the output spectrum region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer corresponding to the target class in the deep neural network according to the loss value until the output spectrum region result is the same as the corresponding target region mark.

SA8, continuously repeating the steps SA5 to SA7 until all training samples of the target class are trained to obtain a trained neural network;

SA9, extracting training samples of a new target class from the training samples of the remaining classes, repeating the processes from the step SA5 to the step SA8 to train the trained neural network until all training samples of all classes are completely trained, and taking the finally obtained neural network as a region selection neural network model.

Further, the construction process of the area selection neural network model comprises the following steps:

SB1, obtaining a predetermined number of finished oil additives, obtaining the near infrared spectrum of each finished oil additive by using near infrared equipment, marking the target region in the spectral region with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrum of the predetermined number of marked finished oil additives as a training sample;

SB2, classifying the training samples according to the categories of the finished oil additives to obtain training samples of M categories;

SB3, constructing M deep neural networks with an input layer, a hidden layer and an output layer;

SB4, obtaining a training sample of a target class and a corresponding deep neural network, and training the selected deep neural network by using the training sample of the target class to obtain a neural network initial model capable of identifying the spectral region with the maximum correlation of the near infrared spectrum of the corresponding class, wherein M deep neural networks are trained to obtain M neural network initial models corresponding to class labels;

SB5, integrating the M neural network initial models, adding a category judgment layer before the M neural network initial models, and adding an output result sorting layer after the M neural network initial models to obtain the region selection neural network model.

Further, step SB4 specifically includes:

SB41, inputting the training samples of the target class to the input layer of the corresponding deep neural network, preprocessing the training samples of the target class by using the input layer, and sending the preprocessed training samples to the hidden layer for processing;

SB42, the hidden layer sends the processing result to the output layer, and the output layer arranges the processing result into the corresponding spectrum region result for output;

SB43, judging whether the output spectrum region result is the same as the corresponding target region mark, if so, selecting the next training sample of the target class for training, otherwise, calculating a loss function between the output spectrum region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer in the deep neural network according to the loss value until the output spectrum region result is the same as the corresponding target region mark.

And SB44, repeating the steps from SB41 to SB43 until all the training samples of the target class are trained to obtain the initial model of the neural network.

Further, step S9 specifically includes:

s9-1, sorting N finished oil additive samples N₁，N₂……N_nSelecting N from N finished oil additive samples₁The finished oil additive sample is used as a test sample, and the rest finished oil additive sample is used as a modeling sample;

s9-2, utilizing the modeling sample to construct a near infrared analysis initial model again on the constructed near infrared analysis initial model according to the scheme of the steps S2 to S8;

s9-3, acquiring near infrared spectra corresponding to various components of a test sample by using infrared equipment, processing the test sample by using a humidification method to obtain wet chemical values corresponding to the various components, inputting the wet chemical values into a reconstructed near infrared analysis initial model for processing, outputting content data of the various components of the test sample, and comparing the output content data of the various components of the test sample with real data of the test sample to obtain the accuracy of the test;

s9-4, sequentially selecting the next product oil additive sample as a test sample according to the sequence, taking the rest product oil additive samples as modeling samples, and repeating the schemes of the steps S9-2 and S9-3 until the n product oil additive samples are respectively taken as test samples to obtain the accuracy rate p of corresponding test₁，p₂……p_n；

S9-5, calculating and checkingAccuracy p of₁，p₂……p_nAnd judging whether P is larger than or equal to P, if so, taking the finally obtained near-infrared analysis initial model as a near-infrared model, otherwise, obtaining a new finished oil additive sample again, repeating the processes from S9-1 to S9-4 by using the new finished oil additive sample until the obtained P is larger than or equal to P, and taking the finally obtained near-infrared analysis initial model as the near-infrared model, wherein P is an accuracy threshold value.

Further, after step S9, the method further includes:

s9', obtaining a product oil additive samples as a correction set, obtaining b product oil additive samples as a verification set, building a near-infrared analysis verification model on the built near-infrared analysis model again by using the product oil additive samples in the correction set according to the scheme from the step S2 to the step S8, sequentially inputting the product oil additive samples in the verification set into the near-infrared analysis verification model to output verification results, comparing the verification results with the real results of the product oil additive samples in the verification set, and determining the verification accuracy;

and S9 ', judging whether the verification accuracy exceeds a verification threshold value, if so, taking the obtained near-infrared analysis verification model as a final near-infrared analysis model, otherwise, obtaining a new correction set and a new verification set, repeating the scheme of the step S9' until the obtained verification accuracy exceeds the verification threshold value, and taking the obtained near-infrared analysis verification model as the final near-infrared analysis model.

Further, the multiple linear regression processing procedure in step S5 specifically includes:

multiplying at least one principal component spectrum in the near-infrared spectrum region after the first derivative processing by the corresponding principal component score coefficient to obtain corresponding principal component spectrum data, and constructing at least one principal component matrix according to the at least one principal component spectrum data;

and constructing at least one main linear function according to the at least one principal component matrix, and performing linear regression processing on the at least one main linear function to obtain a corresponding near-infrared detection value.

Further, the operation processing procedure of the partial least squares method in step S6 specifically includes:

constructing a corresponding spectrum matrix according to the near-infrared detection value, constructing a main wet chemical matrix according to a main wet chemical value in the wet chemical values, correlating the spectrum matrix with the main wet chemical matrix, and calculating a prediction linear relation function of the two matrixes;

step S8, specifically, it is: and (4) constructing a database by using the predicted linear relation functions obtained by the residual product oil additive samples, and taking the database as a near infrared analysis initial model.

The second aspect of the present disclosure provides a product oil additive detection and analysis device based on near-infrared modeling, including:

the acquisition module is used for acquiring n finished oil additive samples;

the near-infrared acquisition module is used for acquiring the near-infrared spectrum of each product oil additive sample by using near-infrared equipment;

the wet chemical processing module is used for detecting the components of each product oil additive sample by using a wet chemical method to obtain a wet chemical value corresponding to the components of each product oil additive sample;

the selecting module is used for selecting a region with the maximum correlation of information quantity from the near infrared spectrum of each product oil additive sample as a near infrared spectrum region by utilizing a pre-constructed region selection neural network model;

the mathematical processing module is used for performing first derivative processing on the near infrared spectrum region of each product oil additive sample and then performing multiple linear regression processing to obtain a near infrared detection value;

the correlation processing module is used for correlating the near infrared detection value and the wet chemical value of each product oil additive sample by utilizing partial least square operation processing to obtain correlated data of each product oil additive sample;

the deleting module is used for calculating the root mean square error between the correlated near-infrared detection value and the corresponding wet-chemical value of each product oil additive sample, and deleting the correlated data of the product oil additive samples with the root mean square error larger than the error threshold;

the construction module is used for constructing a database by utilizing the correlated data of the residual product oil additive samples, and taking the database as a near infrared analysis initial model;

the cross inspection module is used for carrying out cross inspection on the near-infrared analysis initial model by using n finished oil additive samples to obtain a near-infrared analysis model;

the pretreatment module to be detected is used for acquiring a corresponding near infrared spectrum to be analyzed of the finished oil additive to be detected by using infrared equipment, and processing the finished oil additive to be detected by using a humidification method to obtain a wet chemical value to be analyzed corresponding to the component of the finished oil additive to be detected;

and the detection analysis module is used for inputting the near infrared spectrum to be analyzed corresponding to each component of the finished oil additive to be detected and the wet chemical value to be analyzed corresponding to each component into the near infrared analysis model, and processing and determining the content data of each component of the finished oil additive to be detected by using the near infrared model.

A third aspect of the present disclosure proposes an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of the first aspect when executing the program.

From the above, the near-infrared modeling-based method and device for detecting and analyzing the finished oil additive can perform combined association according to the near-infrared detection value of the finished oil additive sample and the wet chemical value of each component, further construct the corresponding database, and form the near-infrared analysis model, so that the near-infrared analysis model can be used for detecting and analyzing the content data of each component of the finished oil additive to be detected, and the accuracy of the content of each component obtained through detection is effectively improved.

Drawings

In order to more clearly illustrate the technical solutions in the present disclosure or related technologies, the drawings needed to be used in the description of the embodiments or related technologies are briefly introduced below, and it is obvious that the drawings in the following description are only embodiments of the present disclosure, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

FIG. 1 is a schematic flow chart of a near-infrared modeling-based product oil additive detection analysis method according to an embodiment of the disclosure;

FIG. 2 is a schematic structural diagram of a product oil additive detection and analysis device based on near infrared modeling according to an embodiment of the disclosure;

fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure.

Detailed Description

For the purpose of promoting a better understanding of the objects, aspects and advantages of the present disclosure, reference is made to the following detailed description taken in conjunction with the accompanying drawings.

It is to be noted that technical terms or scientific terms used in the embodiments of the present disclosure should have a general meaning as understood by those having ordinary skill in the art to which the present disclosure belongs, unless otherwise defined. The use of "first," "second," and similar terms in the embodiments of the disclosure is not intended to indicate any order, quantity, or importance, but rather to distinguish one element from another. The word "comprising" or "comprises", and the like, means that the element or item listed before the word covers the element or item listed after the word and its equivalents, but does not exclude other elements or items. The terms "connected" or "coupled" and the like are not restricted to physical or mechanical connections, but may include electrical connections, whether direct or indirect.

As shown in fig. 1, the method for detecting and analyzing the product oil additive based on near-infrared modeling provided in this embodiment specifically includes the following steps:

and S1, obtaining n finished oil additive samples.

And S2, acquiring the near infrared spectrum of each product oil additive sample by using a near infrared device.

After step S2, spectral regions in the obtained near infrared spectrum in which the absorbance exceeds the light absorption threshold are deleted, and noise regions in each near infrared spectrum are deleted.

To avoid interference in the near infrared spectrum in regions of the spectrum where absorbance exceeds the absorbance threshold, they are deleted, as well as regions of the spectrum where noise is greater. This makes the near infrared spectrum more distinctive.

Since the absorption threshold corresponding to the absorbance of each spectrum may change continuously, however, one absorption threshold cannot be set every time a near infrared spectrum is acquired, which wastes time and reduces the overall efficiency.

Therefore, a large number of labeled near infrared spectrums with corresponding light absorption threshold values and noise threshold values are selected, the labeled near infrared spectrums are used for training the deep neural network, and after the training is finished, the deep neural network can be used for performing drying removal treatment on various near infrared spectrums. Because the deep neural network has the function of relearning, the identification precision can be continuously improved in the process of denoising treatment.

And S3, detecting the components of each product oil additive sample by using a wet chemical method to obtain a wet chemical value corresponding to the components of each product oil additive sample.

Wherein the wet-chemical process used comprises at least one of: liquid phase deposition, electrochemical deposition, sol-gel. The wet chemical values of all components in the finished oil additive can be accurately determined by using a wet chemical method.

And S4, selecting a region with the maximum correlation of information quantity from the near infrared spectrum of each product oil additive sample as a near infrared spectrum region by utilizing a pre-constructed region selection neural network model.

The near infrared spectrum can be selected by utilizing a pre-trained region selection neural network model, the near infrared spectrum is input into the region selection neural network model for processing, and a corresponding near infrared spectrum region is output.

And performing region identification on the corresponding near infrared spectrum by using the obtained region selection neural network model, and further more accurately screening out the spectral region with the maximum correlation of information quantity. The screening precision and accuracy are effectively improved.

And S5, performing first derivative treatment on the near infrared spectrum region of each product oil additive sample, and performing multiple linear regression treatment to obtain a near infrared detection value.

The first derivative processing is carried out for the purposes of reducing the influence of low-frequency noise and drift of the near-infrared spectrogram, improving the resolution, facilitating accurate positioning of the peak position and amplifying high-frequency noise.

And S6, correlating the near infrared detection value and the wet chemical value of each product oil additive sample by utilizing partial least squares operation processing to obtain correlated data of each product oil additive sample.

And S7, calculating the root mean square error between the near infrared detection value associated with each product oil additive sample and the corresponding wet chemical value, and deleting the associated data of the product oil additive samples with the root mean square error larger than the error threshold.

And S8, constructing a database by using the correlated data of the residual product oil additive samples, and taking the database as a near infrared analysis initial model.

And S9, performing cross inspection on the near-infrared analysis initial model by using n finished oil additive samples to obtain a near-infrared analysis model.

And S10, acquiring the corresponding near infrared spectrum to be analyzed of the finished oil additive to be detected by using infrared equipment, and processing the finished oil additive to be detected by using a humidification method to obtain the wet chemical value to be analyzed corresponding to the component of the finished oil additive to be detected.

And S11, inputting the near infrared spectrum to be analyzed of the finished oil additive to be detected and the corresponding wet chemical value to be analyzed into the near infrared analysis model, and processing by using the near infrared model to determine the content data of each component of the finished oil additive to be detected.

By the aid of the scheme, the near-infrared detection values of the product oil additive samples and the wet-chemical values of the corresponding components can be combined and correlated, and then the corresponding database is constructed to form a near-infrared analysis model, so that the near-infrared analysis model can be used for detecting and analyzing content data of each component of the product oil additive to be detected, and accuracy of content of each component obtained through detection is effectively improved.

In some embodiments, the building process of the region selection neural network model comprises:

SA1, constructing a deep neural network having an input layer, a plurality of hidden layers, and an output layer, wherein the number of hidden layers corresponds to the number of categories of the product oil additive.

And SA2, acquiring a preset number of product oil additives, acquiring the near infrared spectrum of each product oil additive by using near infrared equipment, marking the target region in the spectral region with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrum of the preset number of product oil additives after marking as a training sample.

SA3, classifying the training samples according to the categories of the finished oil additives.

SA4, extracting training samples of the target class from the training samples of the plurality of classes.

SA5, inputting the training sample of the target class to an input layer of the deep neural network, preprocessing the training sample of the target class by using the input layer, and sending the preprocessed training sample to a hidden layer corresponding to the target class for processing.

And SA6, sending the processing result to the output layer by the hidden layer corresponding to the target class, and finishing the processing result into a corresponding spectrum region result by the output layer for outputting.

SA7, judging whether the output spectrum region result is the same as the corresponding target region mark, if so, selecting the next training sample of the target class for training, otherwise, calculating a loss function between the output spectrum region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer corresponding to the target class in the deep neural network according to the loss value until the output spectrum region result is the same as the corresponding target region mark.

And SA8, continuously repeating the steps SA5 to SA7 until all training samples of the target class are trained to obtain the trained neural network.

SA9, extracting training samples of a new target class from the training samples of the remaining classes, repeating the processes from the step SA5 to the step SA8 to train the trained neural network until all training samples of all classes are completely trained, and taking the finally obtained neural network as a region selection neural network model.

Through the above-mentioned steps, the specific implementation process corresponding to step S4 is as follows:

and S41, inputting the near infrared spectrum of the finished oil additive sample into an input layer of the region selection neural network model, and preprocessing the near infrared spectrum of the finished oil additive sample by using the input layer. Converting the near infrared spectrum of the analog signal into a digital signal, and judging the category of the near infrared spectrum of the finished oil additive sample.

And S42, the input layer sends the preprocessed near infrared spectrum to the hidden layer of the corresponding type, the hidden layer of the type is utilized for processing, and the hidden layer sends the processing result to the output layer after processing.

S43, the output layer converts the near infrared spectrum region having the largest correlation between the information amounts of the digital signals into analog signals in the near infrared spectrum region having the largest correlation between the information amounts, and outputs the analog signals together with the corresponding categories.

And S44, identifying the near infrared spectrum region with the maximum correlation of information quantity according to the near infrared spectrums of the other product oil additive samples in the steps S41 to S43, and finally obtaining the near infrared spectrum region with the maximum correlation of information quantity corresponding to each product oil additive sample.

Through the scheme, the area selection neural network model is used for identifying the near infrared spectrum area with the maximum correlation of information quantity, the identification precision and accuracy can be effectively improved compared with the prior art, and the area selection neural network model has the relearning function, so that the subsequent identification process can be continuously and automatically learned, and the identification precision and accuracy are further improved continuously.

In some embodiments, in addition to the construction scheme of the area selection neural network model in the above process, the area selection neural network model may be constructed by the following process:

SB1, obtaining the product oil additives in the preset quantity, obtaining the near infrared spectrum of each product oil additive by using near infrared equipment, marking the target area in the spectral area with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrum of the product oil additives in the preset quantity after marking as the training sample.

SB2, classifying the training samples according to the categories of the finished oil additives to obtain training samples of M categories.

SB3, M deep neural networks with input, hidden and output layers were constructed.

SB4, obtaining a training sample of a target class and a corresponding deep neural network, training the selected deep neural network by using the training sample of the target class to obtain a neural network initial model capable of identifying the spectral region with the maximum correlation of the near infrared spectrum of the corresponding class, wherein M deep neural networks are trained to obtain M neural network initial models corresponding to class labels.

Step SB4 specifically includes:

and SB41, inputting the training sample of the target class into the input layer of the corresponding deep neural network, preprocessing the training sample of the target class by using the input layer, and sending the preprocessed training sample to the hidden layer for processing.

SB42, the hidden layer sends the processing result to the output layer, and the output layer arranges the processing result into the corresponding spectrum region result for output.

SB43, judging whether the output spectrum region result is the same as the corresponding target region mark, if so, selecting the next training sample of the target class for training, otherwise, calculating a loss function between the output spectrum region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer in the deep neural network according to the loss value until the output spectrum region result is the same as the corresponding target region mark.

And SB44, repeating the steps from SB41 to SB43 until all the training samples of the target class are trained to obtain the initial model of the neural network.

SB5, integrating the M neural network initial models, adding a category judgment layer before the M neural network initial models, and adding an output result sorting layer after the M neural network initial models to obtain the region selection neural network model.

The judging layer is connected with the input layer of each neural network initial model, and the finishing layer is connected with the output layer of each neural network initial model.

Through the above-mentioned steps, the specific implementation process corresponding to step S4 is as follows:

s41', inputting the near infrared spectrum of the product oil additive sample into a judgment layer, and determining the category of the product oil additive sample by the judgment layer.

S42', the judgment layer transmits the near infrared spectrum of the product oil additive sample to the input layer of the neural network initial model of the corresponding category, the input layer of the neural network initial model is utilized to preprocess the near infrared spectrum, the analog signal of the near infrared spectrum is converted into a digital signal, and the digital signal is transmitted to the hidden layer to be processed.

S43', the hidden layer sends the processing result to the output layer, and since the processing result is a digital signal, the output layer converts the near infrared spectrum region where the correlation of the information amount of the digital signal is the maximum into an analog signal of the near infrared spectrum region where the correlation of the information amount is the maximum, and outputs the analog signal to the finishing layer.

And S44', the finishing layer integrates the analog signal of the near infrared spectrum region with the category of the neural network initial model which is correspondingly processed, and outputs the integration result.

Through the scheme, a plurality of trained neural network initial models of all classes are integrated to form a large area selection neural network model, so that the identification process of all classes can be ensured not to be disordered, the probability of identification errors is reduced, and the identification effect is improved. Because the area selection neural network model has the function of relearning, autonomous learning can be continuously carried out in the subsequent identification process, and the identification precision and accuracy are further continuously improved.

In addition, if the new type of product oil additive exists, the new type of product oil additive can be added into the area selection neural network model after being correspondingly trained into the neural network initial model, so that the identification types of the area selection neural network model are increased.

In some embodiments, the multiple linear regression processing procedure in step S5 specifically includes:

and S51, multiplying at least one principal component spectrum in the near-infrared spectrum region after the first derivative processing by the corresponding principal component score coefficient to obtain corresponding principal component spectrum data, and constructing at least one principal component matrix according to the at least one principal component spectrum data.

S52, at least one main linear function is constructed according to the at least one principal component matrix, and linear regression processing is carried out on the at least one main linear function to obtain a corresponding near-infrared detection value.

In the scheme, effective information of a sample spectrum is fully extracted by analyzing main components of a near infrared spectrum, the problem of linear correlation is solved, and the internal relation between a spectrum matrix and a sample component matrix is considered, so that a near infrared analysis model constructed based on the method is more stable, and the method is suitable for a complex analysis system.

In some embodiments, the operation process of the partial least squares method in step S6 specifically includes:

and constructing a corresponding spectrum matrix according to the near infrared detection value, constructing a main wet chemical matrix according to a main wet chemical value in the wet chemical values, correlating the spectrum matrix and the main wet chemical matrix, and calculating a prediction linear relation function of the two matrixes.

Step S8, specifically, it is: and (4) constructing a database by using the predicted linear relation functions obtained by the residual product oil additive samples, and taking the database as a near infrared analysis initial model.

In some embodiments, the finished oil additive comprises at least one of:

diesel oil cetane number improver (isooctyl nitrate), diesel oil pour point depressant, diesel oil stabilizer, combustion promoter, diesel oil combustion-supporting smoke suppressor, antiknock agent, gasoline antioxidant, methanol gasoline additive, emulsified oil, water-soluble oil, emulsified liquid, saponified oil and saponified liquid.

The finished oil additive may be a petroleum additive including at least one of:

detergents, dispersants, oxidation and corrosion inhibitors, extreme pressure agents, anti-wear agents, oiliness agents, friction modifiers, antioxidants, metal deactivators, viscosity index improvers, rust inhibitors, corrosion inhibitors, pour point depressants, anti-foaming agents, emulsifiers, anti-emulsifiers, and anti-mold agents.

The specific finished oil additive can also be other additives, and can be selected according to actual needs, and is not particularly limited herein.

In some embodiments, step S9 specifically includes:

s9-1, sorting N finished oil additive samples N₁，N₂……N_nSelecting N from N finished oil additive samples₁The finished oil additive sample is used as a test sample, and the rest of the finished oil additive sample is used as a modeling sample.

And S9-2, utilizing the modeling sample to construct the near infrared analysis initial model again on the constructed near infrared analysis initial model according to the scheme of the steps S2 to S8.

And S9-3, acquiring near infrared spectra corresponding to each component of the test sample by using infrared equipment, processing the test sample by using a humidification method to obtain wet chemical values corresponding to each component, inputting the wet chemical values into the reconstructed near infrared analysis initial model for processing, outputting content data of each component of the test sample, comparing the output content data of each component of the test sample with real data of the test sample, and outputting the accuracy of the test.

S9-4, sequentially selecting the next product oil additive sample as a test sample according to the sequence, taking the rest product oil additive samples as modeling samples, and repeating the schemes of the steps S9-2 and S9-3 until the n product oil additive samples are respectively taken as test samples to obtain the accuracy rate p of corresponding test₁，p₂……p_n。

S9-5, calculating the accuracy rate p of the test₁，p₂……p_nAnd judging whether P is larger than or equal to P, if so, taking the finally obtained near-infrared analysis initial model as a near-infrared model, otherwise, obtaining a new finished oil additive sample again, repeating the processes from S9-1 to S9-4 by using the new finished oil additive sample until the obtained P is larger than or equal to P, and taking the finally obtained near-infrared analysis initial model as the near-infrared model, wherein P is an accuracy threshold value.

In the scheme, the obtained near-infrared analysis initial model is tested in a cross-checking mode, and each product oil additive sample can be guaranteed to be used for testing and modeling, so that a model with relatively high accuracy can be built even if the number of the samples is limited, and the modeling efficiency is improved.

In some embodiments, after step S9, the method further comprises:

and S9', obtaining a product oil additive samples as a correction set, obtaining b product oil additive samples as a verification set, constructing a near-infrared analysis verification model on the constructed near-infrared analysis model again by using the product oil additive samples in the correction set according to the scheme from the step S2 to the step S9, sequentially inputting the product oil additive samples in the verification set into the near-infrared analysis verification model to output a verification result, comparing the verification result with the real result of the product oil additive samples in the verification set, and determining the verification accuracy.

And S9 ', judging whether the verification accuracy exceeds a verification threshold value, if so, taking the obtained near-infrared analysis verification model as a final near-infrared analysis model, otherwise, obtaining a new correction set and a new verification set, repeating the scheme of the step S9' until the obtained verification accuracy exceeds the verification threshold value, and taking the obtained near-infrared analysis verification model as the final near-infrared analysis model.

In the above scheme, in order to further ensure the accuracy of the near-infrared model, the correction set is used to construct a corresponding near-infrared verification model on the basis of the near-infrared verification model obtained in the above step S9-5, and then the sample in the verification set is used to verify the near-infrared verification model.

Further, the near-infrared verification model may be further verified by performing the above-described steps S9' to S9 "with the calibration set as the verification set and the verification set as the correction set.

And the near-infrared verification model after verification is used as a final near-infrared analysis model, so that the analysis precision of the finally obtained near-infrared analysis model is further improved.

It should be noted that the method of the embodiments of the present disclosure may be executed by a single device, such as a computer or a server. The method of the embodiment can also be applied to a distributed scene and completed by the mutual cooperation of a plurality of devices. In such a distributed scenario, one of the devices may only perform one or more steps of the method of the embodiments of the present disclosure, and the devices may interact with each other to complete the method.

It should be noted that the above describes some embodiments of the disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments described above and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

Based on the same inventive concept, corresponding to the method of any embodiment, the disclosure also provides a finished oil additive detection and analysis device based on near infrared modeling.

Referring to fig. 2, the near infrared modeling-based product oil additive detection and analysis device includes:

an obtaining module 201, configured to obtain n product oil additive samples;

a near-infrared obtaining module 202, configured to obtain a near-infrared spectrum of each product oil additive sample by using a near-infrared device;

the wet chemical processing module 203 is configured to detect components of each product oil additive sample by using a wet chemical method, and obtain a wet chemical value corresponding to the component of each product oil additive sample;

a selecting module 204, configured to select, from the near infrared spectrum of each product oil additive sample, a region with a largest correlation of information amount as a near infrared spectrum region by using a pre-constructed region selection neural network model;

the mathematical processing module 205 is configured to perform first derivative processing on the near infrared spectrum region of each product oil additive sample, and then perform multiple linear regression processing to obtain a near infrared detection value;

a correlation processing module 206, configured to correlate the near-infrared detection value and the wet chemical value of each product oil additive sample by using partial least squares operation to obtain correlated data of each product oil additive sample;

a deleting module 207, configured to calculate a root mean square error between the correlated near-infrared detection value and the corresponding wet-chemical value of each product oil additive sample, and delete correlated data of product oil additive samples whose root mean square errors are greater than an error threshold;

a construction module 208 for constructing a database using the correlated data of the remaining product oil additive samples, the database being used as an initial model for near-infrared analysis;

the cross inspection module 209 is used for performing cross inspection on the near-infrared analysis initial model by using n finished oil additive samples to obtain a near-infrared analysis model;

the pretreatment module 210 to be detected is used for acquiring a corresponding near infrared spectrum to be analyzed of the finished oil additive to be detected by using infrared equipment, and processing the finished oil additive to be detected by using a humidification method to obtain a wet chemical value to be analyzed corresponding to the component of the finished oil additive to be detected;

the detection analysis module 211 is configured to input the near-infrared spectrum to be analyzed corresponding to each component of the finished oil additive to be detected and the wet chemical value to be analyzed corresponding to each component into the near-infrared analysis model, and determine content data of each component of the finished oil additive to be detected by processing using the near-infrared analysis model.

In some embodiments, the apparatus further comprises a first training module for constructing the region selection neural network model.

The first training module is specifically configured to:

constructing a deep neural network with an input layer, a plurality of hidden layers and an output layer, wherein the number of the hidden layers corresponds to the category number of the product oil additives;

the method comprises the steps of obtaining a preset number of finished oil additives, obtaining near infrared spectrums of the finished oil additives by using near infrared equipment, marking a target area in a spectrum area with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrums of the preset number of marked finished oil additives as training samples;

classifying the training samples according to the category of the finished oil additive;

extracting training samples of a target class from a plurality of classes of training samples;

inputting the training sample of the target class to an input layer of the deep neural network, preprocessing the training sample of the target class by using the input layer, and sending the preprocessed training sample to a hidden layer corresponding to the target class for processing;

the hidden layer corresponding to the target class sends the processing result to the output layer, and the output layer arranges the processing result into a corresponding spectrum region result to be output;

judging whether the output spectrum region result is the same as the corresponding target region mark, if so, selecting a next training sample of the target class for training, otherwise, calculating a loss function between the output spectrum region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer corresponding to the target class in the deep neural network according to the loss value until the output spectrum region result is the same as the corresponding target region mark;

continuously repeating the steps until all the training samples of the target class are trained to obtain a trained neural network;

and extracting a training sample of a new target class from the training samples of the rest classes, training the trained neural network repeatedly until all training samples of all classes are trained, and taking the finally obtained neural network as a region selection neural network model.

In some embodiments, the apparatus further comprises a second training module for constructing the region selection neural network model.

The second training module is specifically configured to:

the method comprises the steps of obtaining a preset number of finished oil additives, obtaining near infrared spectrums of the finished oil additives by using near infrared equipment, marking target areas in spectrum areas with the maximum correlation in each near infrared spectrum, and taking the near infrared spectrums of the preset number of marked finished oil additives as training samples.

And classifying the training samples according to the categories of the finished oil additives to obtain training samples of M categories.

M deep neural networks with an input layer, a hidden layer and an output layer are constructed.

The method comprises the steps of obtaining a training sample of a target class and a corresponding deep neural network, training the selected deep neural network by using the training sample of the target class to obtain a neural network initial model capable of identifying a spectrum region with the maximum correlation on near infrared spectra of corresponding classes, wherein M deep neural networks are trained to obtain M neural network initial models corresponding to class labels.

And inputting the training samples of the target class to an input layer of a corresponding deep neural network, preprocessing the training samples of the target class by using the input layer, and sending the preprocessed training samples to a hidden layer for processing.

The hidden layer sends the processing result to the output layer, and the output layer arranges the processing result into a corresponding spectrum region result to be output.

And judging whether the output spectral region result is the same as the corresponding target region mark, if so, selecting the next training sample of the target class for training, otherwise, calculating a loss function between the output spectral region result and the corresponding target region mark, calculating a loss value by using the loss function, and adjusting the parameters of the hidden layer in the deep neural network according to the loss value until the output spectral region result is the same as the corresponding target region mark.

And continuously repeating the training process until all the training samples of the target class are trained to obtain the neural network initial model.

Integrating the M neural network initial models, adding a category judgment layer in front of the M neural network initial models, and adding an output result sorting layer behind the M neural network initial models to further obtain the region selection neural network model.

In some embodiments, the cross-checking module 209 is specifically configured to:

ranking N finished oil additive samples N₁，N₂……N_nSelecting N from N finished oil additive samples₁The finished oil additive sample is used as a test sample, and the rest of the finished oil additive sample is used as a modeling sample.

And building a near infrared analysis initial model again on the built near infrared analysis initial model by using the modeling sample according to the scheme of the steps S2 to S8.

The method comprises the steps of acquiring near infrared spectra corresponding to various components of a test sample by using infrared equipment, processing the test sample by using a humidification method to obtain test wet chemical values corresponding to the various components, inputting the test wet chemical values into a reconstructed near infrared analysis initial model for processing, outputting content data of the various components of the test sample, and comparing the output content data of the various components of the test sample with real data of the test sample to obtain test accuracy.

Sequentially selecting the next product oil additive sample as a test sample according to the sequence, taking the rest product oil additive samples as modeling samples, and repeating the schemes of the steps S9-2 and S9-3 until n product oil additive samples are respectively taken as test samples to obtain the accuracy rate p of corresponding test₁，p₂……p_n。

Calculating the accuracy p of the test₁，p₂……p_nAnd judging whether P is larger than or equal to P, if so, taking the finally obtained near-infrared analysis initial model as a near-infrared model, otherwise, obtaining a new finished oil additive sample again, repeating the processes from S9-1 to S9-4 by using the new finished oil additive sample until the obtained P is larger than or equal to P, and taking the finally obtained near-infrared analysis initial model as the near-infrared model, wherein P is an accuracy threshold value.

In some embodiments, the apparatus further comprises: the verification module is specifically configured to:

and a product oil additive samples are obtained as a correction set, b product oil additive samples are obtained as a verification set, the product oil additive samples in the correction set are utilized to construct a near-infrared analysis verification model on the constructed near-infrared analysis model again according to the scheme, the product oil additive samples in the verification set are sequentially input into the near-infrared analysis verification model to output verification results, the verification results are compared with the real results of the product oil additive samples in the verification set, and the verification accuracy is determined.

And judging whether the verification accuracy rate exceeds a verification threshold value, if so, taking the obtained near-infrared analysis verification model as a final near-infrared analysis model, otherwise, obtaining a new correction set and a new verification set, repeating the scheme until the obtained verification accuracy rate exceeds the verification threshold value, and taking the obtained near-infrared analysis verification model as the final near-infrared analysis model.

In some embodiments, the mathematical processing module 205 is specifically configured to:

and multiplying at least one principal component spectrum in the near-infrared spectrum region after the first derivative processing by the corresponding principal component score coefficient to obtain corresponding principal component spectrum data, and constructing at least one principal component matrix according to the at least one principal component spectrum data.

And constructing at least one main linear function according to the at least one principal component matrix, and performing linear regression processing on the at least one main linear function to obtain a corresponding near-infrared detection value.

In some embodiments, the association processing module 206 is specifically configured to:

and constructing a corresponding spectrum matrix according to the near infrared detection value, constructing a main wet chemical matrix according to a main wet chemical value in the wet chemical values, correlating the spectrum matrix and the main wet chemical matrix, and calculating a prediction linear relation function of the two matrixes.

The construction module 208 is further configured to construct a database using the predicted linear relationship functions obtained from the remaining product oil additive samples, and the database is used as an initial model for near-infrared analysis.

For convenience of description, the above devices are described as being divided into various modules by functions, and are described separately. Of course, the functionality of the various modules may be implemented in the same one or more software and/or hardware implementations of the present disclosure.

The device of the above embodiment is used for implementing the corresponding near-infrared modeling-based product oil additive detection and analysis method in any one of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiment, which are not described herein again.

Based on the same inventive concept, corresponding to the method of any embodiment, the present disclosure further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the program, the method for detecting and analyzing the product oil additive based on the near-infrared modeling according to any embodiment is implemented.

Fig. 3 is a schematic diagram illustrating a more specific hardware structure of an electronic device according to this embodiment, where the electronic device may include: a processor 1010, a memory 1020, an input/output interface 1030, a communication interface 1040, and a bus 1050. Wherein the processor 1010, memory 1020, input/output interface 1030, and communication interface 1040 are communicatively coupled to each other within the device via bus 1050.

The processor 1010 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits, and is configured to execute related programs to implement the technical solutions provided in the embodiments of the present disclosure.

The Memory 1020 may be implemented in the form of a ROM (Read Only Memory), a RAM (random access Memory), a static storage device, a dynamic storage device, or the like. The memory 1020 may store an operating system and other application programs, and when the technical solution provided by the embodiments of the present specification is implemented by software or firmware, the relevant program codes are stored in the memory 1020 and called to be executed by the processor 1010.

The input/output interface 1030 is used for connecting an input/output module to input and output information. The i/o module may be configured as a component in a device (not shown) or may be external to the device to provide a corresponding function. The input devices may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output devices may include a display, a speaker, a vibrator, an indicator light, etc.

The communication interface 1040 is used for connecting a communication module (not shown in the drawings) to implement communication interaction between the present apparatus and other apparatuses. The communication module can realize communication in a wired mode (such as USB, network cable and the like) and also can realize communication in a wireless mode (such as mobile network, WIFI, Bluetooth and the like).

Bus 1050 includes a path that transfers information between various components of the device, such as processor 1010, memory 1020, input/output interface 1030, and communication interface 1040.

It should be noted that although the above-mentioned device only shows the processor 1010, the memory 1020, the input/output interface 1030, the communication interface 1040 and the bus 1050, in a specific implementation, the device may also include other components necessary for normal operation. In addition, those skilled in the art will appreciate that the above-described apparatus may also include only those components necessary to implement the embodiments of the present description, and not necessarily all of the components shown in the figures.

The electronic device of the above embodiment is used to implement the corresponding near-infrared modeling-based product oil additive detection and analysis method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiment, which are not described herein again.

Based on the same inventive concept, corresponding to any of the above-described embodiment methods, the present disclosure also provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the near-infrared modeling-based product oil additive detection analysis method according to any of the above embodiments.

Computer-readable media of the present embodiments, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the product oil additive detection and analysis method based on near infrared modeling according to any of the above embodiments, and have the beneficial effects of corresponding method embodiments, which are not described herein again.

Those of ordinary skill in the art will understand that: the discussion of any embodiment above is meant to be exemplary only, and is not intended to intimate that the scope of the disclosure, including the claims, is limited to these examples; within the idea of the present disclosure, also technical features in the above embodiments or in different embodiments may be combined, steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present disclosure as described above, which are not provided in detail for the sake of brevity.

In addition, well-known power/ground connections to Integrated Circuit (IC) chips and other components may or may not be shown in the provided figures for simplicity of illustration and discussion, and so as not to obscure the embodiments of the disclosure. Furthermore, devices may be shown in block diagram form in order to avoid obscuring embodiments of the present disclosure, and this also takes into account the fact that specifics with respect to implementation of such block diagram devices are highly dependent upon the platform within which the embodiments of the present disclosure are to be implemented (i.e., specifics should be well within purview of one skilled in the art). Where specific details (e.g., circuits) are set forth in order to describe example embodiments of the disclosure, it should be apparent to one skilled in the art that the embodiments of the disclosure can be practiced without, or with variation of, these specific details. Accordingly, the description is to be regarded as illustrative instead of restrictive.

While the present disclosure has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic ram (dram)) may use the discussed embodiments.

The disclosed embodiments are intended to embrace all such alternatives, modifications and variances which fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalents, improvements, and the like that may be made within the spirit and principles of the embodiments of the disclosure are intended to be included within the scope of the disclosure.