RAP regenerant fusion degree detection method based on infrared spectrum

Abstract

The invention discloses an infrared spectrum-based RAP regenerant fusion degree detection method, and particularly relates to the technical field of road engineering material detection.The method comprises the following steps: constructing a reference sample set containing aged asphalt, a regenerant and mixed samples under different fusion conditions, and acquiring data through infrared spectrum acquisition and pretreatment; identifying an aged asphalt characteristic peak A, a regenerant characteristic peak R and an interaction sensitive peak I; a uniformity index UI is calculated based on a characteristic peak area ratio, an interaction index DI is calculated based on sensitive peak displacement, and a characteristic vector index FI is calculated based on difference spectrum analysis of an actual spectrum and a theoretical spectrum; the three kinds of indexes are weighted and integrated into a comprehensive fusion degree index F, correlation analysis is carried out on the comprehensive fusion degree index F and fusion conditions, and a multivariate nonlinear regression fitting model is established; and repeating the above steps for a to-be-tested RAP sample, substituting the model, and quantitatively outputting a fusion degree evaluation value and an equivalent fusion condition. According to the method, multi-dimensional quantitative evaluation of the fusion state is realized, and the detection accuracy and efficiency are improved.

Description

Technical Field

[0001] This invention relates to the field of road engineering material testing technology, and more specifically, to a method for detecting the fusion degree of RAP regenerator based on infrared spectroscopy. Background Technology

[0002] In the field of road engineering, asphalt pavements are prone to aging due to long-term exposure to traffic loads and natural environmental factors, leading to a decline in road performance. The recycling and reuse of recycled asphalt pavement (RAP) has become a core approach to achieving green and low-carbon development in road engineering and conserving asphalt resources. The degree of integration between the recycling agent and aged asphalt is a key indicator determining the road performance of RAP recycled asphalt pavements. It directly affects the core properties of recycled asphalt mixtures, such as high-temperature stability, low-temperature crack resistance, and water stability. Therefore, accurate detection of the RAP recycling agent integration degree is of significant practical engineering importance for the construction quality control, recycling process optimization, and service life assurance of recycled asphalt pavements. It has also become a research focus and technical challenge in the field of road engineering material testing.

[0003] Existing methods for testing the blending degree of RAP regenerators can meet the testing requirements, but they still have some drawbacks in practical use: The evaluation dimensions are too limited: existing methods often focus on a single aspect of characterization (such as dispersion uniformity or chemical change), and lack a comprehensive evaluation system that considers multiple perspectives such as spatial uniformity, molecular interactions, and deviations from physical mixing states. Low degree of quantification: Most methods rely on subjective judgment or indirect indicators, making it difficult to provide accurate and reproducible fusion quantification values; Weak model correlation: The failure to effectively establish a quantitative correlation model between process conditions (such as dosage, temperature, and time) and the degree of fusion limits the guiding value of the test results for actual processes; Limited detection efficiency and applicability: Some methods involve complex sample pretreatment and are time-consuming, making it difficult to meet the needs of rapid on-site detection and large-scale sample analysis.

[0004] Therefore, developing a RAP (Rapid Aging) recycling degree testing method that can directly characterize the microscopic fusion state of aged asphalt and recycling agent, achieve multi-dimensional quantitative evaluation, and has a standardized testing system has become an urgent technical problem to be solved in the field of road engineering material testing. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, the present invention provides a method for detecting the fusion degree of RAP regenerant based on infrared spectroscopy, which solves the problems mentioned in the background art through the following scheme.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for detecting the fusion degree of RAP regenerant based on infrared spectroscopy, comprising: S1: Construct a benchmark sample set, which includes aged asphalt samples, recycling agent samples, and mixed samples of aged asphalt and recycling agent under different fusion conditions, and collect and preprocess the benchmark sample set by infrared spectroscopy. S2: Identify characteristic functional group absorption peaks: Define the functional group absorption peaks that exist only in aged asphalt samples as characteristic peaks of aged asphalt. The absorption peaks of functional groups that exist only in the regenerator sample are designated as characteristic peaks of the regenerator. Functional group absorption peaks whose position or shape changes during the fusion of mixed samples are designated as interaction-sensitive peaks. ; S3: Establish a fusion index calculation model: Based on the data obtained in S1 and S2, calculate the fusion index to quantify the fusion state. The fusion index includes: S301: Characteristic peak area ratio uniformity index, based on the characteristic peaks of aged asphalt in the same mixed sample. Characteristic peaks of regenerant The area ratio standard deviation and mean were calculated to characterize the uniformity of aged asphalt and recycling agent in the sample space. S302: Characteristic peak displacement interaction index, based on sensitive peaks The actual peak position was calculated from the peak position changes of the mixed sample of aged asphalt and recycling agent, and was used to characterize the degree of intermolecular interaction. S303: Difference Spectral Eigenvector Index, obtained by performing difference spectral analysis and extracting principal component features based on the actual spectrum and physical mixing spectrum of the mixed sample, used to characterize the degree of deviation from the physical mixing state; S4: Establish a standard testing system, perform correlation analysis between the different fusion conditions described in S1 and the corresponding fusion degree index calculated in S3, and establish a fitting model; S5: Repeat S2 and S3 with the RAP of the recycled asphalt pavement to be tested and the mixed sample to calculate the fusion index, substitute it into the above fitting model, and then output the fusion evaluation value and equivalent fusion conditions in a quantitative manner.

[0007] The technical effects and advantages of this invention are as follows: This invention identifies characteristic peak A of aged asphalt, characteristic peak R of rejuvenator, and interaction-sensitive peak I, and constructs characteristic peak area ratio uniformity index UI, characteristic peak displacement interaction index DI, and difference spectrum eigenvector index FI. It comprehensively characterizes the fusion state from three dimensions: spatial uniformity, molecular interaction, and degree of chemical fusion. Compared with the shortcomings of existing technologies that have a single evaluation dimension and rely on subjective judgment, this method realizes a quantifiable and reproducible comprehensive evaluation of the fusion state, thus improving the accuracy of detection. This invention constructs a benchmark sample set with different fusion conditions (regenerant dosage, fusion temperature, and fusion time), and performs correlation analysis and multivariate nonlinear regression modeling on the fusion conditions and fusion degree index, thereby realizing a quantitative correlation between fusion degree and process parameters and overcoming the deficiency of weak model correlation in the prior art. This invention establishes a complete testing standard from sample preparation, spectral acquisition, characteristic peak identification, index calculation to model input, and establishes a standard testing system database. This system supports rapid processing and automated analysis of batch samples, improves testing efficiency, is suitable for laboratory and field testing needs, and overcomes the problems of complex preprocessing, long time consumption, and poor applicability of traditional methods. This invention achieves its goals by setting up parallel sample testing, spectral consistency verification, validation set error control ≤5%, and model accuracy verification R. 2 With a ≥0.85, RMSE≤0.05, and MAE≤0.04, this invention ensures the reliability and repeatability of the test results. The classification of the degree of integration into excellent, good, medium, and poor further enhances the engineering practicality of the evaluation results, providing a reliable basis for the assessment and prediction of the construction quality of recycled asphalt pavement. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of the overall method structure of the present invention.

[0009] Figure 2 This is a schematic diagram of the construction of the benchmark sample set and the spectral acquisition preprocessing in S1 of the present invention.

[0010] Figure 3 This is a schematic diagram of the absorption peak structure of the functional group that identifies the features of the present invention S2.

[0011] Figure 4 This is a schematic diagram of the fusion index calculation model structure of S3 in this invention.

[0012] Figure 5 This is a schematic diagram of the standard detection system and fitting model established in S4 of the present invention.

[0013] Figure 6 This is a schematic diagram of the RAP sample detection structure of the present invention S5. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Please see Figures 1-6As shown, this embodiment of the invention provides a method for detecting the fusion degree of RAP regenerator based on infrared spectroscopy, including: S1: Construct a benchmark sample set, which includes aged asphalt samples, recycling agent samples, and mixed samples of aged asphalt and recycling agent under different fusion conditions, and collect and preprocess the benchmark sample set by infrared spectroscopy. S2: Identify characteristic functional group absorption peaks: Define the functional group absorption peaks that exist only in aged asphalt samples as characteristic peaks of aged asphalt. The absorption peaks of functional groups that exist only in the regenerator sample are designated as characteristic peaks of the regenerator. Functional group absorption peaks whose position or shape changes during the fusion of mixed samples are designated as interaction-sensitive peaks. ; S3: Establish a fusion index calculation model: Based on the data obtained in S1 and S2, calculate the fusion index to quantify the fusion state. The fusion index includes: S301: Characteristic peak area ratio uniformity index, based on the characteristic peaks of aged asphalt in the same mixed sample. Characteristic peaks of regenerant The area ratio standard deviation and mean were calculated to characterize the uniformity of aged asphalt and recycling agent in the sample space. S302: Characteristic peak displacement interaction index, based on sensitive peaks The actual peak position was calculated from the peak position changes of the mixed sample of aged asphalt and recycling agent, and was used to characterize the degree of intermolecular interaction. S303: Difference Spectral Eigenvector Index, obtained by performing difference spectral analysis and extracting principal component features based on the actual spectrum and physical mixing spectrum of the mixed sample, used to characterize the degree of deviation from the physical mixing state; S4: Establish a standard testing system, perform correlation analysis between the different fusion conditions described in S1 and the corresponding fusion degree index calculated in S3, and establish a fitting model; S5: Repeat S2 and S3 with the RAP of the recycled asphalt pavement to be tested and the mixed sample to calculate the fusion index, substitute it into the above fitting model, and then output the fusion evaluation value and equivalent fusion conditions in a quantitative manner.

[0016] In S1, a benchmark sample set is constructed, which includes aged asphalt samples, recycling agent samples, and mixed samples of aged asphalt and recycling agent under different fusion conditions. The benchmark sample set is collected and preprocessed by infrared spectroscopy. A baseline sample set was formed by standardizing the preparation of aged asphalt, recycling agents, and mixed samples under different fusion conditions. Fourier transform infrared spectroscopy was used to complete spectral acquisition and preprocessing, providing a data foundation for subsequent identification of characteristic functional group absorption peaks and calculation of fusion index.

[0017] S101: Selection and Pretreatment of Experimental Materials Road engineering base asphalt, such as 70# and 90# road petroleum asphalt, was selected as the raw material; aging treatment was carried out using a rotating thin-film oven aging method, and the aging conditions were as follows: Rotation speed control at The aging time is 5 hours, and aged asphalt base material is prepared and placed in... Store in a sealed container in a constant temperature desiccator for later use; Conventional recyclers, such as aromatic oil recyclers, waste rubber powder modified recyclers, and bio-oil-based recyclers, are selected as the base raw materials for the recyclers. Vacuum filtration is used to remove solid impurities and flocculent matter from the recyclers. The filtration screen pore size is [not specified]. After filtration, place Store in a sealed container in a constant temperature desiccator for later use; After the above raw materials are processed, basic performance indicators are tested, including: penetration, softening point, and ductility of aged asphalt; kinematic viscosity and flash point of the recycling agent. This ensures that the performance of the raw materials meets road engineering standards and avoids interference from the performance deviations of the raw materials themselves with subsequent spectral detection and fusion analysis.

[0018] S102: Hierarchical preparation of the benchmark sample set The baseline sample set is divided into a pure sample group and a mixed sample group. The pure sample group consists of aged asphalt samples and recycling agent samples, while the mixed sample group consists of mixed samples of aged asphalt and recycling agent under different fusion conditions. Parallel samples are prepared for each group, with at least 3 parallel samples for each type. Specific preparation methods include: Preparation of aged asphalt samples: Take the treated aged asphalt base material, heat it to a molten state, and take... Molten aged asphalt was uniformly coated onto an infrared spectroscopy sample plate and cooled to room temperature to obtain an aged asphalt sample. Preparation of regenerant samples: Take the treated regenerant base material The regenerant sample is obtained by directly and uniformly coating it onto an infrared spectroscopy detection sample plate; Preparation of mixed samples under different fusion conditions: Using aged asphalt as the base material, gradient fusion conditions were set with recycling agent dosage, fusion temperature, and fusion time as variables. Mixed samples were prepared using a high-temperature, high-speed shear method. The gradient and mixed sample preparation included: Variable gradient settings: The amount of recycling agent is based on the quality of aged asphalt, and eight gradients are set: 3%, 4%, 5%, 6%, 7%, 8%, 9%, and 10%. The fusion temperature can be set to six gradients: 130℃, 140℃, 150℃, 160℃, 170℃, and 180℃. The fusion time was set at six gradients: 10 min, 20 min, 30 min, 40 min, 50 min, and 60 min. These gradients were combined to form different fusion conditions. Preparation of mixed samples: Weigh aged asphalt and recycling agent according to the fusion conditions. Heat the aged asphalt to the fusion temperature and hold for 10 minutes. After adding the recycling agent, place it in a high-temperature shear mill and set the shear rate to [value missing]. Shearing is continued at the fusion temperature until the preset fusion time is reached, and the sample is immediately removed after shearing. The mixed sample was uniformly coated onto an infrared spectroscopy detection sample slide and cooled to room temperature to obtain the mixed sample. After all mixed samples are prepared, each sample is numbered, and the number includes information on fusion temperature, fusion time, and regenerator dosage to facilitate subsequent data correlation and analysis.

[0019] S103: Infrared Spectrum Acquisition Infrared spectra of all samples in the reference sample set were acquired using the attenuated total reflectance (ATR) method, and detection was performed using a Fourier transform infrared spectrometer. Specific detection parameters and operations included: Spectrometer warm-up: Power on the Fourier transform infrared spectrometer to allow it to warm up and reach a stable operating state. Control the ambient temperature of the detection environment to [temperature value missing]. Humidity ≤60% to avoid the impact of temperature and humidity changes on spectral detection results; Background scan: A background scan is performed on the spectrometer in a sample-free state, with a scan wavenumber range of [range missing]. The resolution is set to The number of scans was set to 32 to complete the background spectrum acquisition and subtraction, which is used to eliminate spectral interference from the environment and the instrument itself; Sample scanning: Place the prepared reference sample slide on the ATR detection stage of the spectrometer, ensuring it is in close contact with the ATR crystal surface, and scan using the same parameters as the background scan described above. wave number, The infrared spectra of the samples were scanned at a resolution of 32 scans. Each type of sample was scanned 3 times in parallel, and the original spectral data of each scan were recorded. The average spectrum is calculated from the raw spectral data to obtain the infrared spectral curve of the sample. All sample spectral data are stored by number to form a raw database of infrared spectra of the benchmark sample set.

[0020] S104: Infrared spectral preprocessing The spectral curves in the original infrared spectral database of the benchmark sample set are subjected to standardization preprocessing to eliminate the influence of non-target factors such as baseline drift, noise interference, and light scattering during the spectral detection process. This preprocessing includes: Baseline correction: The adaptive iterative reweighted penalized least squares method is used to correct the baseline of the original spectral curve. The correction parameters are set to 50 iterations and 0.01 weighting coefficient to eliminate the baseline drift and tilt of the spectral curve and make the spectral baseline tend to be horizontal. Spectral smoothing: The Savitzky-Golay smoothing method is used to smooth the spectral curve after baseline correction. The smoothing window size is set to 7 points and the polynomial order is 2 to eliminate random noise in spectral detection and retain the characteristic peak shape and peak position information of the spectral curve. Vector normalization: This involves performing vector normalization on the smoothed spectral curve, mapping the spectral data to... The range eliminates the differences in spectral intensity caused by physical factors such as sample coating amount and film thickness, making the spectral data of different samples comparable; Wavenumber calibration: Using the spectrometer's built-in standard peak as a reference, wavenumber calibration is performed on the normalized spectral curve, with the calibration error controlled within [value missing]. This ensures accurate wavenumber localization of the absorption peaks of characteristic functional groups.

[0021] The final data processing involves converting the format of the preprocessed spectral data, extracting the corresponding wavenumber and absorbance data pairs, and forming a standardized reference sample set infrared spectral database for subsequent identification and analysis of S2 characteristic functional group absorption peaks.

[0022] In S2, the identification of characteristic functional group absorption peaks refers to the absorption peaks of functional groups that exist only in aged asphalt samples, which are then designated as characteristic peaks of aged asphalt. The absorption peaks of functional groups that exist only in the regenerator sample are designated as characteristic peaks of the regenerator. Functional group absorption peaks whose position or shape changes during the fusion of mixed samples are designated as interaction-sensitive peaks. ; S201: Reference Spectral Data Retrieval and Consistency Verification From the standardized reference sample set infrared spectral database obtained from S104, wavenumber and absorbance spectral data of aged asphalt samples, recycling agent samples, and mixed samples under different fusion conditions were retrieved according to sample type. The corresponding information such as the sample number and fusion conditions (recycling agent dosage, fusion temperature, fusion time) were also retrieved. Consistency verification is performed on the spectral data of parallel samples of the same type. The spectral correlation coefficient between samples of the same type is calculated using a spectral similarity algorithm. A similarity of ≥95% is set as the pass standard. For abnormal samples with a similarity of <95%, their spectral data are removed and the sample preparation, spectral acquisition and preprocessing steps of S1 are re-executed. The spectral data of the same type of samples that passed the verification were averaged to obtain the average spectral curves of aged asphalt samples, the average spectral curves of rejuvenator samples, and the average spectral curves of mixed samples under different fusion conditions. These were used to eliminate random errors in parallel samples and serve as the benchmark curves for subsequent characteristic peak identification.

[0023] S202: Characteristic peaks of aged asphalt Identification and calibration The average spectral curves of the aged asphalt samples and the average spectral curves of the recycling agent samples obtained from S201 were imported into the spectral analysis platform for analysis across the entire wavenumber range. Point-by-point comparison, with the comparison parameters being the absorbance value corresponding to the wavenumber and the peak shape characteristics; Set a threshold for determining the absence of characteristic absorption peaks: if the absorbance value of the regenerator sample in a certain wavenumber range is ≤0.01 and there are no obvious peak protrusions, shoulder peaks or other characteristic peak shapes, it is determined that the regenerator sample in that wavenumber range has no corresponding characteristic absorption peaks. Mark the wavenumber intervals in which aged asphalt samples have characteristic absorption peaks (absorbance value > 0.01, peak shape complete, wave position clear), and the rejuvenator samples have no characteristic absorption peaks in the corresponding wavenumber intervals. Record the wavenumber of the characteristic peak position, peak shape parameters, and corresponding functional group type. The characteristic absorption peaks of the marked aged asphalt were numbered sequentially and denoted as follows: ,in It is a positive integer; for example, the carbonyl group produced by oxidation of aged asphalt. The peak range of stretching vibration is ( ), sulfoxide The peak range of stretching vibration is ( All of these are attributed to characteristic peaks of aged asphalt. ; Establish characteristic peaks of aged asphalt The database is used to input the peak number, peak position, wavenumber range, peak shape parameters, and corresponding functional group information for each characteristic peak, thus completing the characteristic peak data processing. The calibration.

[0024] S203: Characteristic peaks of regenerant Identification and calibration The average spectral curves of the rejuvenator samples and the average spectral curves of the aged asphalt samples obtained from S201 were imported into the spectral analysis software. Point-by-point comparisons were performed across the entire wavenumber range, with the comparison parameters being absorbance values ​​and peak shape characteristics. Using the threshold for determining the absence of characteristic absorption peaks (absorbance value ≤ 0.01 and no obvious peak shape) from S202, the wavenumber intervals that meet the criteria of having obvious characteristic absorption peaks in the rejuvenator sample and having no characteristic absorption peaks in the corresponding wavenumber intervals are marked. Record the characteristic peak positions, peak shape parameters, and corresponding functional group types for the marked wavenumber intervals; for example: the aromatic ring skeleton vibration peaks of aromatic hydrocarbons in regenerators ( The peak range of the CH stretching vibration of long-chain alkanes is ( All of these are attributed to characteristic peaks of regenerants. ; The characteristic absorption peaks of the labeled regenerant are numbered sequentially and denoted as follows: ,in It is a positive integer; Establish characteristic peaks of regenerant The database is used to input the peak number, peak position, wavenumber range, peak shape parameters, and corresponding functional group information for each characteristic peak, thus completing the characteristic peak data processing. The calibration.

[0025] S204: Interaction-sensitive peak Identification and calibration Constructing theoretical physical mixing spectral curves: Based on the mass ratio of aged asphalt to recycling agent under different fusion conditions, the average spectral curve of the aged asphalt sample obtained from S201 is linearly superimposed with the average spectral curve of the recycling agent sample to obtain the theoretical physical mixing spectral curves of the mixed samples under different fusion conditions. This curve characterizes the spectral features of the simple physical mixing of aged asphalt and recycling agent when there is no intermolecular interaction. Spectral curve comparison: The actual average spectral curves of the mixed samples under each fusion condition were imported into the spectral analysis platform along with the theoretical physical mixing spectral curves of the corresponding blending ratios. Point-by-point comparisons were performed across the entire wavenumber range, with a focus on screening characteristic peaks of aged asphalt. Characteristic peaks of regenerant The corresponding wavenumber range, while also taking into account other absorption peaks within the entire wavenumber range.

[0026] Quantitative determination of peak position and peak shape changes: A quantifiable threshold for determining changes is set; if any threshold is met, the characteristic peak is considered to have changed. Specific thresholds include: Peak position variation: The deviation between the peak position of the characteristic peak in the actual spectral curve and the peak position of the corresponding characteristic peak in the theoretical physical mixing spectral curve. ; Peak shape variation: the rate of change of the peak height of the characteristic peak in the actual spectral curve. Or, visual changes in peak shape may occur, such as peak splitting, peak merging, shoulder peak formation / disappearance, peak broadening / narrowing, etc.

[0027] Sensitive peak Labeling and Numbering: Screen and label absorption peaks that meet the above-mentioned change judgment thresholds, and record their wavenumber range, peak position after change, peak position change amplitude, peak shape change type and amplitude, and corresponding functional group type. This absorption peak is the interaction sensitive peak. ; For all sensitive peaks Number them in order, and denote them as follows: ,in Positive integers, interaction-sensitive peaks It can originate from the characteristic peaks of aged asphalt. Characteristic peaks of regenerant It can also originate from the characteristic absorption peaks newly generated after aged asphalt is mixed with recycling agent; Establish interaction sensitive peaks The database is used to input the number of each sensitive peak, its original position, its changed position, the magnitude of the change, the characteristics of the peak shape change, and the corresponding functional group information to complete the sensitive peak data processing. The calibration.

[0028] In S3, the fusion index calculation model is established: based on the data obtained in S1 and S2, a fusion index that quantifies the fusion state is calculated, and the fusion index includes: S301: Characteristic peak area ratio uniformity index, based on the characteristic peaks of aged asphalt in the same mixed sample. Characteristic peaks of regenerant The area ratio standard deviation and mean were calculated to characterize the uniformity of aged asphalt and recycling agent in the sample space. The uniformity index of the characteristic peak area ratio is denoted as... The range of values ​​is , A value closer to 1 indicates better spatial uniformity. The specific calculation process includes: Multi-region spectral acquisition of the same mixed sample: Take the same mixed sample slide prepared in S102, randomly select 5 non-overlapping spatial detection points in the effective detection area of ​​the mixed sample, with a detection point spacing of ≥2mm, and collect the standardized spectral curves of each detection point according to the infrared spectral acquisition parameters in S103 to obtain 5 spatial region spectral curves of the mixed sample. Characteristic peak area calculation: Calculate the characteristic peaks of all aged asphalt in the spectral curves of the five spatial regions respectively. Total area Characteristic peaks of all regenerants Total area ,in For the space area Characteristic peaks ( The sum of the peak areas of ) For the space area Characteristic peaks ( The sum of the peak areas; Area ratio calculation: For each spatial detection point, calculate the area ratio of its characteristic peaks. ,in This corresponds to 5 detection points; Statistical parameter calculation: Calculate the arithmetic mean of the five area ratios. and standard deviation Calculation formula: Quantitative calculation of uniformity index: To ensure that the index value is positively correlated with uniformity, the characteristic peak area ratio uniformity index is calculated by subtracting the relative standard deviation from 1. The calculation formula is as follows: ; Index validity check, if Approaching 0, the sample underwent multi-region spectral acquisition and calculation again; the calculated values ​​were then used to... The value is associated with the fusion conditions and sample number of the corresponding mixed sample and stored to complete the sample fusion. Calculation of the index.

[0029] S302: Characteristic peak displacement interaction index, based on sensitive peaks The actual peak position was calculated from the peak position changes of the mixed sample of aged asphalt and recycling agent, and was used to characterize the degree of intermolecular interaction. The characteristic peak displacement interaction index is denoted as... ,in A larger value indicates a more significant intermolecular interaction; the specific characteristic peak shift interaction index is denoted as... The calculations include: Sensitive peak Peak position and shape parameter extraction, and extraction of interaction-sensitive peaks of mixed samples under various fusion conditions. actual peak position Sensitive peaks in the corresponding theoretical physical mixed spectrum Theoretical peak position theory Simultaneously extract each sensitive peak Peak height change rate ,in For actual peak height, The theoretical peak height; Peak position offset calculation: for each sensitive peak Calculate the absolute peak offset absolute peak offset This reflects the degree to which intermolecular interactions affect functional group vibrations; Peak shape variation weighting coefficient assignment: To reflect the synergistic effect of peak position variation and peak shape variation, the weighting coefficient is determined based on the peak height variation rate. For each sensitive peak Assign weight coefficients The weighting coefficients are set according to the contribution of actual peak shape changes to the characterization of molecular interactions, and the specific assignment criteria are as follows: like like ; like The weighting coefficients are assigned values ​​that cover the entire range from no significant change in peak shape to a significant change, ensuring the accuracy of the index in characterizing intermolecular interactions. Quantitative calculation of displacement interaction index: The displacement interaction index of characteristic peaks is calculated using a weighted summation method, which includes all sensitive peaks. The product of the absolute peak offset and the corresponding weighting coefficient, divided by the sum of all weighting coefficients, is calculated using the following formula: in Interaction-sensitive peak Total quantity ( ); Exponential standardization: This process applies to all mixed samples. Values ​​are mapped to extreme value normalization method. The interval eliminates the difference in exponential values ​​caused by different numbers of sensitive peaks. The standardized formula is: in For all mixed samples Minimum value For all mixed samples Maximum value; after standardization It is associated with and stored in relation to the corresponding sample information to complete the calculation of the DI index for that sample; S303: Difference Spectral Eigenvector Index, obtained by performing difference spectral analysis and extracting principal component features based on the actual spectrum and physical mixing spectrum of the mixed sample, used to characterize the degree of deviation from the physical mixing state; Let the difference spectral eigenvector index be denoted as... ,in A larger value indicates a more significant deviation from a purely physical mixing state in the sample, and a higher degree of chemical integration between the aged asphalt and the recycling agent. Specific calculations include: Differential spectrum preprocessing involves baseline correction and Savitzky-Golay smoothing of the differential spectrum curves according to the S104 method to eliminate random noise in the differential spectrum and retain effective feature information. The preprocessed differential spectrum provides basic data for subsequent principal component analysis. Principal Component Analysis (PCA) of Difference Spectrum Data: PCA was performed on the preprocessed difference spectrum data of all mixed samples. During the analysis, the data was centered, principal components were extracted, and the contribution rate of each principal component was calculated. The contribution rates were then selected. The first two principal components As a core feature, the first two principal components can reflect 95% of the effective information of the difference spectrum, meeting the requirements for feature characterization. Principal component eigenvalue extraction: Extracting eigenvalues ​​from the difference spectrum of each mixture sample. eigenvalues ​​on As the spatial coordinates of the principal components, it directly reflects the characteristic differences in the difference spectrum; The difference spectrum eigenvector index is quantitatively calculated using the eigenvalues ​​of two principal components. Using the coordinate axes, the modulus is calculated as the exponent of the difference spectrum eigenvector. The calculation formula is: The calculated The fusion conditions between the value and the corresponding sample are used to complete the calculation of the FI index of the sample.

[0030] S304: Integration and Standardization of the Integration Index Calculation Model The three indices calculated from S301, S302, and S303 are normalized and weighted to construct a standardized fusion index calculation model. This model enables a comprehensive quantitative characterization of the fusion state of aged asphalt and recycling agents. Simultaneously, the model parameters are standardized and stored, providing a calculation basis for the correlation analysis of S4. Specifically, this includes: Single-index normalization: Since the original value ranges of the three types of indices are different, first normalize them... ( ), ( ), (Non-negative) Mapped uniformly to the extreme value normalization method The interval is used to obtain the normalized exponent. , , The normalization formula is: in The original index value, The normalized exponent value, These are the minimum and maximum values ​​of the index among all mixed samples, respectively. After normalization, the values ​​of the three indices are all positively correlated with the degree of fusion, and the closer the value is to 1, the better the fusion effect. Based on the essential characteristics of the fusion of aged asphalt and recycling agents, Pearson correlation analysis was used to determine the weighting coefficients of three types of normalized indices for characterizing the degree of fusion. Spatial homogeneity is the basis of fusion, intermolecular interactions are the core of fusion, and deviation from the physical mixing state is the result of fusion. Based on the correlation analysis results, the weighting coefficients were set as follows: Weight , Weight , Weight The weighted summation method was used to calculate the comprehensive integration index. As the core quantitative indicator of the fusion state, the calculation formula is: Comprehensive Integration Index The range of values ​​is , The closer the value is to 1, the better the overall integration of the aged asphalt and the recycling agent. The core components of the fusion index calculation model include four parts: data input, single index calculation module, normalization module, and weighted integration module. The data input consists of the standardized infrared spectrum curve of the sample and the characteristic peak information of S2 calibration. The single index calculation module is the index calculation of S301-S303. The normalization module is the index normalization. The weighted integration module is the weighted summation and weight coefficient. The fusion conditions, original spectral data, characteristic peak parameters, single index values, and overall fusion index of all mixed samples are included. Perform associated storage to form a complete data chain of fusion conditions, spectral features, and fusion degree index; Randomly select 10% of the mixed samples with different fusion conditions as the validation set, and recalculate the comprehensive fusion index according to the steps in S3. The error between the validation set calculation results and the original calculation results is ≤5%, indicating that the fusion index calculation model has good stability and accuracy and can be used to establish the standard detection system for S4.

[0031] In S4, the establishment of a standard detection system involves performing correlation analysis between the different fusion conditions described in S1 and the corresponding fusion degree index calculated in S3 to establish a fitting model. Based on the multi-gradient fusion conditions determined in S1 and the fusion index data calculated in S3, a detection system for quantitative evaluation of fusion degree and back-calculation of equivalent fusion conditions is established through correlation analysis between fusion conditions and fusion index, and by constructing a fitting model. This provides a unified judgment standard and calculation basis for the fusion degree detection of RAP samples of recycled asphalt pavement to be tested, specifically including: S401: Construction and Preprocessing of Related Datasets Based on the comprehensive fusion index of S304, retrieve the fusion condition parameters and fusion index parameters of the mixed samples; The fusion condition parameters: regenerant dosage ( ), fusion temperature ( ), fusion time ( The fusion index parameter is: characteristic peak area ratio uniformity index. Characteristic peak displacement interaction index Difference spectral eigenvector index and comprehensive integration index ; The fusion condition parameters are used as the set of independent variables. ( =Regenerator dosage, =Fusion temperature, =fusion time), with the fusion degree index parameter as the dependent variable set. ( , , , ), matching according to sample number to construct a dataset associated with fusion conditions and fusion degree index; Outlier detection was performed on the associated dataset. Box plots were used to remove outlier samples whose values ​​exceeded 1.5 times the interquartile range. The dataset after outlier removal was then standardized, and the independent variables were... Dependent variable Mapped to Intervals are used to eliminate the impact of dimensional differences on correlation analysis and model fitting.

[0032] S402: Correlation Analysis between Fusion Conditions and Fusion Degree Index For the standardized association dataset, the Pearson correlation coefficient method was used to calculate the set of independent variables. With the set of dependent variables The correlation coefficients of each index were used to clarify the degree of influence and correlation (positive / negative correlation) of each fusion condition on the single and comprehensive fusion indexes, and to screen the fusion conditions that have a significant impact on the fusion index. Partial correlation analysis was used to eliminate the interaction effects between fusion conditions, and the correlation between fusion conditions and fusion index was verified. A correlation coefficient with an absolute value ≥ 0.6 was set as a significant correlation, and the significantly correlated fusion conditions were used as independent variables in the fitted model.

[0033] S403: Construction of Fitting Model for Fusion Conditions and Fusion Degree Index Based on the correlation analysis, a multivariate nonlinear regression model was selected as the fitting model, with significantly correlated fusion conditions as independent variables, to construct a model that combines the index and the comprehensive fusion degree index. The fitted model for the dependent variable; The preprocessed associated dataset is sorted by The data is randomly divided into training and test sets. The training set data is used to solve for and train the fitting model, determine the regression coefficients and constant term parameters of the model, and define the expression of the master fitting model as follows: ; To address the range of values ​​for the fusion conditions, value constraints were set for the independent variables of the fitted model, limiting the range of values ​​for regenerant dosage, fusion temperature, and fusion time to be consistent with the gradient settings in S102, thus ensuring the applicability of the model.

[0034] S404: Accuracy Verification and Classification of Fitted Models The accuracy of the constructed fitting model is verified using test set data, and the model is calculated. The coefficient of determination between the predicted and actual values Root mean square error Mean absolute error ,set up , , 4. These are the criteria for model qualification; models that meet these criteria are considered valid fit models. If the model does not meet the qualification standard, the stepwise regression method is used to optimize the model, remove insignificant influencing factors, and refit the model. Based on the comprehensive integration index The range of values Based on application requirements, the degree of integration is classified into levels, including: For superior integration, 0.8 represents good fusion. 0.6 represents intermediate integration. 0.4 represents differential fusion.

[0035] S405: Integration of Standard Testing System By using the validated fusion conditions and fusion index fitting model, model parameters, accuracy indicators, and independent variable value constraints as conditions, and integrating the sample preparation standard of S1, the characteristic peak identification standard of S2, and the fusion index calculation standard of S3, a standardized testing system for the fusion degree of aged asphalt and recycling agent is formed. Construct a standard testing system database, input the system's model formulas, parameter thresholds, and fusion degree classification standards, store the original data of correlation analysis, and model training and validation results, so as to achieve unified management and rapid retrieval of data and models; Establish a standard testing system, clarify the requirements for the entire process from sample testing and index calculation to model input and result output, and ensure that those skilled in the art can complete the fusion degree test and the reverse deduction of equivalent fusion conditions in accordance with the specifications, thus completing the establishment of the standard testing system.

[0036] In S5, the test sample of the recycled asphalt pavement RAP and the mixed sample are repeatedly executed in S2 and S3 to calculate the fusion index. The index is then substituted into the above fitting model to quantitatively output the fusion evaluation value and the equivalent fusion condition. Based on the standardized testing system established in S4, after standardizing the RAP samples of the recycled asphalt pavement to be tested, the absorption peak identification of characteristic functional groups in S2 and the calculation of the fusion index in S3 are repeated. The obtained fusion index is substituted into the fitting model constructed in S4, and finally the fusion evaluation value and equivalent fusion conditions of the RAP sample to be tested are quantitatively output, so as to achieve accurate determination of the fusion status of the engineering RAP sample, specifically including: S501: Pretreatment of RAP Samples and Preparation of Detection Samples Samples of the recycled asphalt pavement (RAP) to be tested were taken, and more than three representative sampling points were selected. After removing surface impurities and aggregates, the recycled asphalt components in the RAP were extracted to obtain the basic raw material of the recycled asphalt to be tested. Referring to the sample pretreatment and preparation standards of S101 and S102, the above-mentioned recycled asphalt base raw materials to be tested were heated, melted, and impurities were removed to prepare test samples that meet the requirements of infrared spectroscopy detection. At the same time, three parallel samples were prepared to ensure the representativeness of the samples. Following the infrared spectral acquisition and preprocessing parameters in S103 and S104, the infrared spectral acquisition and standardized preprocessing of the parallel samples of the RAP sample to be tested were completed, and the standardized average spectral curve of the sample to be tested was obtained.

[0037] S502: Repeat step S2 to identify the absorption peaks of characteristic functional groups in the sample to be tested. The characteristic functional group absorption peak database constructed in S2 is retrieved as a reference standard. The standardized average spectral curve of the RAP sample to be tested is used to complete the identification and calibration of the characteristic functional group absorption peaks according to the operation procedures from S201 to S204. Complete the characteristic peaks of aged asphalt in the test sample Characteristic peaks of regenerant Interaction sensitive peak The identification, numbering, and parameter recording of the peaks are ensured to be consistent with the S2 standard, so as to avoid identification deviations from affecting subsequent index calculations.

[0038] S503: Repeat step S3 to calculate the fusion index of the sample to be tested. Based on the standardized spectral data of the sample to be tested and the identified characteristic peak parameters, and referring to the calculation specifications and formulas of S301 to S303, the calculation process of the fusion index in S3 is repeated. Calculate the uniformity index of the characteristic peak area ratio of the test samples sequentially. Characteristic peak displacement interaction index ( ), difference spectral eigenvector index ( The comprehensive fusion index of the test sample was calculated using the weighted integration method in S304. Complete the index calculation and average value processing of parallel samples to eliminate random errors.

[0039] S504: Substitute the input into the fitting model to output the fusion degree evaluation value and the equivalent fusion condition. Retrieve the fusion condition and fusion degree index master fitting model constructed in S403 and the fusion degree level classification standard in S404, and combine the comprehensive fusion degree index of the test sample calculated in S503. And the values ​​of each index, substituted into the master fitting model ; The model calculates and quantitatively outputs the fusion degree evaluation value of the RAP sample to be tested, and clarifies its fusion degree level as excellent, good, medium, or poor by referring to the S404 fusion degree level classification standard. Based on the reverse derivation function of the fitted model, and combined with the correlation between the fusion index and the fusion conditions, the equivalent fusion conditions corresponding to the RAP sample to be tested are derived, namely the recycling agent dosage, fusion temperature, and fusion time parameters that are consistent with the fusion state of the sample to be tested, providing data support for subsequent recycled asphalt pavement repair and process optimization.

[0040] S505: Verification and Archiving of Test Results The consistency of the test results of the three parallel samples was verified to ensure that the relative error of the fusion degree evaluation value was ≤5% and the deviation of the equivalent fusion condition parameters was within a reasonable range. After the verification was qualified, the average value was taken as the final test result. If the verification fails, the sample to be tested is prepared again, the spectrum is collected, the characteristic peaks are identified and the index is calculated again, until the results meet the consistency requirements. The detection data, characteristic peak parameters, fusion index, fusion evaluation value, equivalent fusion conditions, and verification results of the RAP samples to be tested are uniformly entered into the S405 standard testing system database to complete the archiving of test results, which facilitates subsequent querying and traceability.

[0041] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting the blending degree of RAP regenerant based on infrared spectroscopy, characterized in that, include: S1: Construct a benchmark sample set, which includes aged asphalt samples, recycling agent samples, and mixed samples of aged asphalt and recycling agent under different fusion conditions, and collect and preprocess the benchmark sample set by infrared spectroscopy. S2: Identify characteristic functional group absorption peaks: Define the functional group absorption peaks that exist only in aged asphalt samples as characteristic peaks of aged asphalt. The absorption peaks of functional groups that exist only in the regenerator sample are designated as characteristic peaks of the regenerator. Functional group absorption peaks whose position or shape changes during the fusion of mixed samples are designated as interaction-sensitive peaks. ; S3: Establish a fusion index calculation model: Based on the data obtained in S1 and S2, calculate the fusion index to quantify the fusion state. The fusion index includes: S301: Characteristic peak area ratio uniformity index, based on the characteristic peaks of aged asphalt in the same mixed sample. Characteristic peaks of regenerant The area ratio standard deviation and mean were calculated to characterize the uniformity of aged asphalt and recycling agent in the sample space. S302: Characteristic peak displacement interaction index, based on sensitive peaks The actual peak position was calculated from the peak position changes of the mixed sample of aged asphalt and recycling agent, and was used to characterize the degree of intermolecular interaction. S303: Difference Spectral Eigenvector Index, obtained by performing difference spectral analysis and extracting principal component features based on the actual spectrum and physical mixing spectrum of the mixed sample, used to characterize the degree of deviation from the physical mixing state; S4: Establish a standard testing system, perform correlation analysis between the different fusion conditions described in S1 and the corresponding fusion degree index calculated in S3, and establish a fitting model; S5: Repeat S2 and S3 with the RAP of the recycled asphalt pavement to be tested and the mixed sample to calculate the fusion index, substitute it into the above fitting model, and then output the fusion evaluation value and equivalent fusion conditions in a quantitative manner.

2. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, S1 specifically includes: S101: The selection of test raw materials includes the preparation of basic raw materials for aged asphalt and basic raw materials for recycling agents; S102: The hierarchical preparation of the benchmark sample set, including the preparation of aged asphalt samples, recycling agent samples, and mixed samples with different fusion conditions using recycling agent dosage, fusion temperature, and fusion time as gradient variables; S103: Infrared spectral acquisition, using the attenuated total reflection method to acquire infrared spectra of all samples in the reference sample set; S104: Infrared spectral preprocessing, which performs baseline correction, smoothing, vector normalization and wavenumber calibration on the acquired spectral data.

3. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, S2 specifically includes: S201: Reference spectral data retrieval and consistency verification. Retrieve spectral data and related information, verify using a spectral similarity algorithm, set a similarity of ≥95% as qualified, and average the spectral data of qualified samples. S202: Characteristic peaks of aged asphalt The identification and calibration of characteristic absorption peaks were carried out by comparing the average spectral curves of aged asphalt and recycling agents. A threshold of ≤0.01 absorbance value with no obvious peak shape was set as the criterion for the absence of characteristic absorption peaks. The unique characteristic absorption peaks of aged asphalt were marked and numbered, thus establishing a characteristic peak system. database; S203: Characteristic peaks of regenerant For identification and calibration, the threshold for determining the absence of characteristic absorption peaks is used, while the unique characteristic absorption peaks of the regenerant are marked and numbered to establish a characteristic peak system. database; S204: Interaction-sensitive peak The identification and calibration of the peak position deviation and peak height change rate are used to calculate the theoretical physical mixed spectrum curve by linear superposition. The actual and theoretical spectrum curves are compared, and thresholds for judging peak shape and peak position changes are set. Absorption peaks that meet the thresholds are marked as sensitive peaks. Number them and establish sensitive peaks database.

4. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, The calculation of the uniformity index of the characteristic peak area ratio of S301 includes: Characteristic peak area ratio uniformity index The calculation involved selecting five non-overlapping detection points with a spacing of ≥2mm within the effective detection area of ​​the same mixed sample, collecting spectra, and calculating the characteristic peaks of aged asphalt at each point. Total area Characteristic peaks of regenerant Total area To obtain the area ratio Calculate 5 arithmetic mean and standard deviation ,according to Calculate the index, The range of values ​​is .

5. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, The calculation of the S302 characteristic peak displacement interaction index includes: Extracting sensitive peaks actual peak position Theoretical peak position Calculate the absolute peak offset Extracting the peak height change rate And based on the peak height change rate For each sensitive peak Assign weight coefficients ,according to Calculate the characteristic peak displacement interaction index Then, using the extreme value normalization method, Mapped to The interval is standardized .

6. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, The calculation of the difference spectral eigenvector index in S303 includes: Baseline correction and Savitzky-Golay smoothing were performed on the difference spectral curves of the actual and theoretical physical mixed spectra of the mixed samples. The preprocessed difference spectral data were then centered before PCA principal component analysis was performed, selecting the top two principal components with a contribution rate ≥ 95%. Extract principal component eigenvalues ,according to Calculate the difference spectral eigenvector index. The value of represents a deviation from a purely physical mixed state.

7. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, The fusion index in S3 includes: Will , , Mapped uniformly to the extreme value normalization method. The interval is used to obtain the normalized exponent. Based on the correlation analysis results, the weighting coefficients are set as follows: ,according to Calculate the comprehensive integration index ,in The range of values ​​is A value close to 1 indicates a good fusion state; 10% of the mixed samples are randomly selected as the validation set, and the error between the validation set calculation result and the original result is ≤5%.

8. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, S4 specifically includes: S401: Construct a dataset relating fusion conditions and fusion degree index, using regenerator dosage, fusion temperature, and fusion time as the set of independent variables. ,by , , , For the set of dependent variables Outliers were removed and standardized using a box plot method. S402: The Pearson correlation coefficient method combined with partial correlation analysis was used for association analysis. The absolute value of the correlation coefficient ≥ 0.6 was set as a significant correlation, and the fusion conditions for significant correlation were selected as independent variables for the fitting model. S403: Construct a multivariate nonlinear regression fitting model, divide the associated dataset into training and test sets in a 7:3 ratio, train the model and set constraints on the values ​​of independent variables; S404: Verify model accuracy and set the coefficient of determination. Root mean square error Mean absolute error As a qualification standard, according to For superior integration, For good-level integration, For intermediate integration, Differential fusion is classified into fusion degree levels; S405: Integrate standards for sample preparation, characteristic peak identification, index calculation, and fitting models to build a standard testing system database and formulate full-process testing specifications.

9. The method for detecting the blending degree of RAP regenerant based on infrared spectroscopy according to claim 1, characterized in that, S5 specifically includes: S501: Pretreatment and preparation of RAP samples to be tested. Select more than 3 representative sampling points, extract recycled asphalt components and prepare 3 parallel test samples. Complete spectral acquisition and pretreatment according to parameters S103 and S104. S502: Using the feature peak database constructed in S2 as a reference, repeat the S2 process to complete the feature peaks of the sample to be tested. Identification and calibration; S503: Repeat step S3 to calculate the test sample. , , and comprehensive integration index And take the average value of the parallel sample results; S504: Substitute the fusion index into the fitting model, quantitatively output the fusion evaluation value and determine the fusion level, and back-calculate the equivalent regenerator dosage, fusion temperature and fusion time. S505: Verify the results of parallel sample tests to ensure that the relative error of the fusion evaluation value is ≤5%. After verification, all test data shall be entered into the standard test system database for archiving. If the data fails, the test shall be repeated.

Images