CN-122016696-A - Chemical new material quality detection method and system based on artificial intelligence

CN122016696ACN 122016696 ACN122016696 ACN 122016696ACN-122016696-A

Abstract

The invention relates to the technical field of artificial intelligence detection, in particular to a chemical new material quality detection method and system based on artificial intelligence, comprising the following steps: the method comprises the steps of collecting an infrared absorption spectrum of a polyolefin copolymer, extracting absorption peak parameters, constructing a spectrum section index, matching a standard interval to judge functional groups, identifying structural change classification characteristics, constructing spectrum section structural sequence and label mapping, training an artificial intelligent model, deploying the artificial intelligent model in a system, realizing automatic identification and prediction of a new sample from a spectrum section structure to the functional group composition and quality state, and outputting a material quality detection result set.

Inventors

Lan Shuhui
TANG GUOJIN
REN PENGWEI
JIN QIYANG
WEI YAJUAN

Assignees

中国检验认证集团广西有限公司
广西中检检测技术服务有限公司

Dates

Publication Date: 20260512
Application Date: 20260113

Claims (10)

1. The quality detection method of the new chemical material based on artificial intelligence is characterized by comprising the following steps of: S1, collecting absorption spectrum data of polyolefin copolymer in an infrared band, extracting wavelength, peak intensity and bandwidth of an absorption peak, sorting and arranging spectral band information according to wavelength, establishing a spectral band index, and generating a spectral response characteristic table; S2, extracting the peak position and bandwidth of a spectrum segment according to the spectrum response characteristic table, and combining the standard absorption range of the aromatic C=C, the hydroxyl O-H and the hydrocarbon C-H bond to complete spectrum segment matching and functional group attribution and generate a functional group segment labeling information set; S3, calling the functional group section marking information set, identifying functional group changes among spectrum sections, extracting absorbance differences and half-width changes, identifying structural change fragments by combining peak shape trend, classifying the change types, and generating a spectrum section structural difference characteristic set; S4, based on the spectrum segment structure difference feature set, summarizing the functional group type, spectrum shape parameters and variation trend, constructing a spectrum segment structure sequence index, mapping with a material class label, and generating a spectrum input feature sequence file; S5, according to the spectrum input characteristic sequence file, combining the functional group distribution and the response mode, matching the structural characteristics and the quality standard, judging the structural integrity of the material and the composition state of the functional groups, and outputting a quality detection result set of the new chemical material; s6, training an artificial intelligent model based on the spectrum input characteristic sequence file and the label data of the existing sample, realizing automatic identification and prediction from a spectrum structure to a functional group composition and quality state, deploying the trained model in a detection system, receiving the spectrum sequence input of a new sample, and outputting a structure integrity and quality grade judgment result.
2. The method for detecting the quality of the new chemical material based on the artificial intelligence according to claim 1, wherein the spectrum response characteristic table comprises a spectrum segment index, a peak position parameter, a peak intensity parameter, a bandwidth parameter and a normalized absorbance, the functional group section labeling information set comprises a matching section identifier, a functional group type label, a standard range number and section annotation, the spectrum segment structure difference characteristic set comprises a functional group change position, an absorbance difference quantity, a half-width change quantity and a change type classification, the spectrum input characteristic sequence file comprises a structure sequence number, a functional group distribution sequence, a spectrum shape parameter quantity, a change trend index and a sample class mapping table, and the chemical material quality detection result set comprises a structure integrity judgment, a composition state evaluation, a quality standard matching conclusion and a sample qualification conclusion.
3. The artificial intelligence-based new chemical material quality detection method according to claim 1, wherein the specific steps of S1 are as follows: S101, acquiring absorption spectrum data of a polyolefin copolymer material to be detected in an infrared band range, recording a corresponding relation between light intensity and wavelength, detecting the change condition of absorbance at the wavelength, calculating an absorption value according to the relation between the light intensity and the absorbance, and finishing to form spectrum curve data of continuous bands to generate an absorption spectrum data set; S102, based on the absorption spectrum data set, identifying the maximum point position of an absorption peak in a wave band, calculating the peak position wavelength and the corresponding peak intensity, determining bandwidth parameters according to the peak width range, arranging the peak parameters according to the wavelength sequence, and generating an absorption peak parameter sequence; and S103, extracting band absorbance data according to the absorption peak parameter sequence, carrying out normalization processing on absorbance, adjusting peak intensity values to a unified standard range, establishing an index list comprising peak positions, peak intensity, bandwidth and normalized absorbance, integrating the index list into spectrum response data, and generating a spectrum response characteristic table.
4. The artificial intelligence based quality detection method for new chemical materials according to claim 3, wherein the specific steps of S2 are as follows: S201, extracting peak position wavelength and bandwidth information in a spectrum section based on the spectrum response characteristic table, recording the corresponding relation between the peak position and the bandwidth in the spectrum section, and sorting the spectrum section structure information according to the wavelength sequence to generate a peak position bandwidth parameter list; S202, according to the peak bandwidth parameter list, calling a standard absorption wavelength range of an aromatic carbon-carbon double bond stretching vibration interval, a hydroxyl group stretching vibration interval and a hydrocarbon single bond stretching vibration interval, judging whether the peak falls into the standard interval, acquiring the corresponding condition of a spectrum section and a functional section, and generating a functional section matching judging result; And S203, identifying the wavelength, the bandwidth and the type of the functional group which fall into the spectrum section of the section according to the functional section matching judgment result, constructing the matching relation between the type of the functional group and the position of the spectrum section, integrating the functional group section information which the spectrum section belongs to, and generating a functional group section labeling information set.
5. The artificial intelligence based new chemical material quality detection method according to claim 4, wherein the specific steps of S3 are as follows: S301, calling the functional group section marking information set, sequentially comparing the types of the functional groups in adjacent spectral bands, identifying the position where the type of the functional group changes, recording the spectral band index information corresponding to the change, and generating a functional group conversion position set; S302, extracting the absorbance difference and the half-width change of adjacent spectral bands according to the functional group conversion position set, judging whether the dual conditions of the absorbance difference and the peak shape change are met, marking the spectral band positions meeting the conditions, and generating a spectral band response change characteristic value; s303, according to the spectrum response change characteristic value, carrying out combination judgment on the absorbance difference direction and the half-width change trend, verifying the structure change type corresponding to the spectrum, and finishing the matching relation between the structure change and the spectrum position to generate a spectrum structure difference characteristic set.
6. The artificial intelligence based new chemical material quality detection method according to claim 5, wherein the specific steps of S4 are as follows: S401, extracting a functional group type, an absorbance value and a half-width parameter corresponding to a spectrum based on the spectrum structure difference feature set, and sorting the combination relationship between the functional group mark and the spectrum shape parameter in the same spectrum according to the change trend to generate a spectrum structure combination trend; S402, calling the spectrum segment structure combination trend value, numbering the functional group type and the spectrum shape parameter combination index position of the spectrum segment, establishing a combination index sequence according to the spectrum segment arrangement sequence, numbering continuous structure combinations, and generating a structure sequence filing index value; S403, according to the structure sequence archiving index value, matching the content of the class label recorded in the original material sample, establishing a corresponding table of the structure sequence and the corresponding label, integrating the index sequence information of the corresponding table and the spectrum segment combination, and generating a spectrum input characteristic sequence file.
7. The artificial intelligence based quality detection method for new chemical materials according to claim 6, wherein the specific steps of S5 are as follows: S501, calling the spectrum input characteristic sequence file, extracting functional group identifiers and response characteristics corresponding to spectrum structural sequences in a material sample, constructing a functional group change path by combining spectrum shape parameter distribution positions, and generating a functional group change trend coefficient according to response change directions and position continuity; S502, according to the functional group change trend coefficient, comparing a functional group composition sequence defined in a quality standard with a spectral response characteristic interval, identifying a difference index inconsistent with a standard range, screening a response characteristic deviation item, and generating a functional group matching deviation amount; S503, judging the structure formation state in the sample according to the matching deviation amount of the functional groups and combining the structural continuity standard of the functional group arrangement integrity and the spectrum, recording response abnormal information and structural stability indexes corresponding to the sample, and generating a quality detection result set of the new chemical material.
8. The method for detecting the quality of the new chemical material based on artificial intelligence according to claim 1, wherein the polyolefin copolymer is a high molecular compound formed by polymerization reaction of olefin monomers and one or more comonomers, and the type of the high molecular compound comprises ethylene-propylene copolymer, ethylene-butene copolymer and ethylene-hexene copolymer, and has the characteristics of adjustable crystallinity and high chemical stability; the infrared band range refers to an electromagnetic wave wavelength range used in spectrum detection, is within two-point five micrometers to twenty-five micrometers, and is used for measuring the molecular vibration absorption characteristic of the new chemical material under differential light energy; The peak intensity refers to the maximum absorbance value at the corresponding wavelength of the absorption peak in the spectrum, and the parameter is directly obtained by a spectrophotometer.
9. The method for detecting the quality of the new chemical material based on artificial intelligence according to claim 1, wherein the standard absorption range is a typical absorption wavelength range of a functional group obtained by statistics according to a published infrared spectrum or ultraviolet-visible spectrum database, and a telescopic vibration absorption region comprising an aromatic carbon-carbon double bond is positioned in a range of one thousand four hundred to one thousand six hundred wavenumbers; the half-width change refers to the change condition of a width value corresponding to a half of the maximum absorbance of a spectrum peak; The structure change segment refers to a spectrum region generated between adjacent spectrum segments due to the attribution change of the functional group; the spectral shape parameter is a spectral index for describing the morphological characteristics of the absorption peak, and comprises absorbance, half-width, peak position offset and peak symmetry factors; The quality standard is a quality judgment basis established based on the structural integrity, component uniformity and spectrum distribution rule of the new chemical material, and a structural comparison template is formed according to standard samples or industry detection specifications; And constructing a training data set based on the sample label, training a convolutional neural network model to realize the prediction relation from the spectrum input characteristic sequence to the quality grade, and deploying the model in a detection system for predicting the quality of a new sample.
10. An artificial intelligence based quality detection system for new chemical materials, wherein the system is used for realizing the artificial intelligence based quality detection method for new chemical materials according to any one of claims 1 to 9, and the system comprises: The spectrum acquisition analysis module acquires absorption spectrum data of the polyolefin copolymer material in an infrared band, extracts absorption peak positions, absorbance values and half-width parameters of a spectrum band, sorts all spectrum band information according to a wavelength sequence, and generates a spectrum response characteristic table after normalization processing of the absorbance data; the response characteristic extraction module invokes the spectrum response characteristic table, extracts peak position and bandwidth information of a spectrum segment, matches the spectrum segment with a standard functional group telescopic vibration range according to a wavelength interval, establishes a corresponding relation between a matching result and a functional group type, and generates a functional group segment labeling information set; the functional group identification module calls the functional group section marking information set, identifies the change positions of the functional group types between adjacent spectrum sections, extracts the absorbance difference and the half-width change condition of the change sections, judges the occurrence positions and the change types of the structural change according to the change characteristics of the absorption response parameters, and generates a spectrum section structural difference characteristic set; The structural difference extraction module invokes the spectral band structural difference feature set, sorts the functional group type, spectral shape parameter and response change trend corresponding to the spectral band, establishes the mapping relation between the spectral band sequence and the material sample class label, and generates a spectral input feature sequence file; The quality state judging module calls the spectrum input characteristic sequence file, judges the structural integrity and the functional group composition state of the sample according to the functional group distribution and the response change mode in the spectrum structural sequence, and generates a chemical new material quality detection result set; The artificial intelligent reasoning module calls the trained AI recognition model, inputs the spectrum input characteristic sequence file, outputs quality grade prediction, structural integrity score and abnormal risk prompt corresponding to the material sample, and achieves the intelligent judging function of the detection system.

Description

Chemical new material quality detection method and system based on artificial intelligence Technical Field The invention relates to the technical field of artificial intelligence detection, in particular to a chemical new material quality detection method and system based on artificial intelligence. Background The technical field of artificial intelligent detection relates to the intelligent recognition, judgment and evaluation of materials, products or systems by utilizing an artificial intelligent algorithm, belongs to the automatic detection category of researching or analyzing physical or chemical properties of materials, mainly comprises an intelligent recognition method based on models such as machine learning, deep learning, neural networks and the like, and is combined with means such as image processing, spectrum analysis, pattern recognition and the like, the method has the advantages that automatic analysis and judgment of detection data are realized, the artificial intelligent detection technology has important application in the aspects of material quality control, manufacturing process optimization, automatic detection systems and the like, and systematic characteristics are represented by taking algorithm driving as a core, taking data acquisition and training as a basis and taking model output as a result form, and the whole process from sample acquisition, feature extraction to result judgment is covered. The quality detection method and system of the novel chemical material are characterized in that in the production or application process of various novel chemical materials, manual or semi-automatic detection means are adopted to ensure that physical properties, chemical stability or other key indexes of the novel chemical materials reach the standard, the conventional detection is dependent on a spectrum instrument, chromatographic equipment or a mechanical test device and the like, detection parameters are obtained by combining specific steps of a chemical analysis method, an infrared or ultraviolet absorption detection method, a thermal analysis method, a tensile test and the like, and then interpretation and recording are carried out by a professional technician, a great amount of manual operation and experience judgment are needed in the implementation process of the method, and the data processing is dependent on a statistical method or set threshold judgment, so that the detection of the micro defects and the identification of the complex materials is limited. In the prior art, the traditional means such as a spectrum instrument, chromatographic equipment and mechanical test are relied on in the material detection process, the detection flow is dominated by manual operation, the detection parameters are required to be manually interpreted by professional technicians, the mode relies on experience judgment in the aspect of identifying complex chemical structures and microscopic material differences, the information extraction dimension is single, the structural change details of the materials are difficult to fully reveal, the set threshold value or statistical rule judgment is mostly adopted when multidimensional spectrum data are processed, the influence of external interference factors is easy to cause, the stability of the detection result is poor, obvious limitations exist in the aspects of identification precision and quantitative analysis capability, and particularly when the detection requirements of diversified chemical new materials are met, a unified and systematic identification path and a standardized modeling mechanism are lacked. Disclosure of Invention In order to solve the technical problems in the prior art, the embodiment of the invention provides a chemical new material quality detection method based on artificial intelligence, which comprises the following steps: In order to achieve the purpose, the invention adopts the following technical scheme that the quality detection method of the new chemical material based on artificial intelligence comprises the following steps: S1, collecting absorption spectrum data of polyolefin copolymer in an infrared band, extracting wavelength, peak intensity and bandwidth of an absorption peak, sorting and arranging spectral band information according to wavelength, establishing a spectral band index, and generating a spectral response characteristic table; S2, extracting the peak position and bandwidth of a spectrum segment according to the spectrum response characteristic table, and combining the standard absorption range of the aromatic C=C, the hydroxyl O-H and the hydrocarbon C-H bond to complete spectrum segment matching and functional group attribution and generate a functional group segment labeling information set; S3, calling the functional group section marking information set, identifying functional group changes among spectrum sections, extracting absorbance differences and half-width changes, identifying structural