Decoupling method, device and storage medium for multi-factor coupled aging of high-altitude asphalt materials

Abstract

This invention discloses a decoupling method, apparatus, and storage medium for multi-factor coupled aging of plateau asphalt materials. The absolute baseline degradation rate R0 is determined in a standard reference environment. The independent degradation increment Δ is calculated using four factors: peeling temperature, ultraviolet radiation, air pressure, and oxygen partial pressure, through an isolated test chamber with independent air pressure and oxygen partial pressure air paths. i ; through tensor difference subtraction σ ij Extract the net contribution of the two-factor interaction; combine CCD design to capture higher-order collaborative terms σ global After mechanistic closed-loop verification, the independent contribution vector W and coupling effect matrix σ are output. This invention explicitly introduces R0 as a common background benchmark into the subtraction of each order, eliminating the systematic bias of repeated subtraction of the basic background in the traditional subtraction method; it incorporates low air pressure and low oxygen partial pressure as physically independent and adjustable factors into the plateau asphalt aging decoupling model; the two-way game mechanism of synergistic acceleration and antagonistic inhibition can be quantified, and the relative error of decoupling reconstruction does not exceed 8%.

Description

Technical Field

[0001] This invention relates to the field of aging test and evaluation technology for asphalt pavement materials, and particularly to a decoupling method, device and storage medium for multi-factor coupled aging of plateau asphalt materials. Background Technology

[0002] The Qinghai-Tibet Plateau, with an average altitude of over 4000m, is characterized by four extreme environmental features: strong ultraviolet radiation (UV intensity 1.5 to 2 times that of plains areas), low atmospheric pressure (50–70 kPa), large temperature differences (daily temperature differences exceeding 30°C), and low oxygen partial pressure (approximately 7.3–10.5 kPa). These factors work synergistically to significantly shorten the actual lifespan of asphalt pavements compared to those in plains areas. Accurately quantifying the independent and coupled contributions of each factor is a crucial technical foundation for the optimized design and lifespan prediction of pavement materials on the plateau.

[0003] The existing technology, specifically the method for predicting the low-temperature performance of aged asphalt (patent number CN121830951A), only evaluates single factors. For decoupled evaluation of multi-factor coupled aging, existing technical approaches mainly fall into three categories: orthogonal experimental design (referring to a method for calculating the contribution rate of heat, oxygen, and light to asphalt aging, publication number CN117030985A), traditional subtraction method (referring to a comprehensive evaluation method, device, medium, and product for asphalt aging degree, publication number CN118310945A), and response surface methodology (referring to an environmentally friendly road asphalt rejuvenator optimized using response surface methodology, publication number CN117747028A). However, all these methods suffer from the following technical defects: First, the orthogonal experimental design suffers from the linear additive assumption. This type of method assumes the linear additive nature of each factor, failing to capture nonlinear coupling effects, resulting in systematic bias in the estimation of factor contributions. Second, the traditional subtraction method suffers from the defect of repeatedly deducting the absolute baseline deterioration rate R0. Existing multi-factor aging subtraction methods do not explicitly introduce a baseline reference term when deducting the influence of single factors. The baseline thermo-oxygen aging amount under normal temperature and pressure is implicitly present in both single-factor and two-factor experimental groups, leading to repeated subtraction of the baseline background during difference calculations. This results in unclear physical meaning of the extracted net interactive effect, and may even lead to spurious negative values ​​that are mistakenly identified as experimental errors and discarded. Thirdly, existing high-altitude asphalt aging models suffer from dimensional deficiencies and limitations in pressure-oxygen partial pressure coupling control. Existing high-altitude asphalt aging models primarily focus on temperature and ultraviolet radiation, while pressure and oxygen partial pressure, two characteristic dimensions of high-altitude environments, are either missing or simplified as constants in most models. More importantly, existing climate aging test devices, such as the multifunctional asphalt aging test device with announcement number CN210376078U, regulate pressure and oxygen partial pressure through a shared vacuum-gas distribution loop. This means that both decrease proportionally during vacuuming, leading to a technical bias in the field that "pressure and oxygen partial pressure inevitably change together in high-altitude environments and cannot be independently controlled." This technical bias makes it impossible to quantify and identify the independent contributions of air pressure and oxygen partial pressure to asphalt aging in existing technologies, and in particular, it fails to reveal the antagonistic and inhibitory effect of low oxygen partial pressure alone on the thermal oxidative carbonylation reaction.

[0004] The above three types of technical defects together lead to a series of problems at the engineering application level, such as large deviations in the prediction of the life of plateau pavement materials, lack of quantitative basis for the screening of modifiers, and ambiguity in the determination of the time window for preventive maintenance. Summary of the Invention

[0005] The purpose of this invention is to provide a decoupling method, device and storage medium for multi-factor coupled aging of plateau asphalt materials, which solves the problems in the prior art such as the inability to truly separate multi-factor nonlinear interactions, repeated subtraction of the reference background, lack of air pressure and oxygen partial pressure dimensions and inability to independently control them, and inability to identify antagonistic inhibition effects.

[0006] This invention provides a technology that can explicitly introduce an absolute benchmark, cover four types of plateau characteristic factors, and has the ability to independently regulate air pressure and oxygen partial pressure, while taking into account both physical separation and statistical modeling.

[0007] Firstly, the present invention provides a decoupling method for multi-factor coupled aging of plateau asphalt materials, the method comprising the following steps:

[0008] S1. Absolute reference calibration: Asphalt pavement material specimens are placed in a standard reference environment including temperature T, ultraviolet radiation UV, atmospheric pressure P, and oxygen partial pressure O for aging tests, and the absolute reference deterioration rate R0 is obtained by actual measurement.

[0009] S2. Single-factor independent incremental extraction: One of the following factors—temperature T, ultraviolet irradiation UV, atmospheric pressure P, and oxygen partial pressure O—is adjusted to its set value, while the remaining factors are locked to the level of step S1. Atmospheric pressure P is controlled via a vacuum pumping circuit, and oxygen partial pressure O is controlled via a nitrogen / oxygen mixing circuit independent of the vacuum pumping circuit. These two gas paths are independent, allowing both the low-pressure-normal oxygen condition (maintaining oxygen partial pressure while only reducing gas pressure) and the normal-pressure-low oxygen condition (maintaining gas pressure while only reducing oxygen partial pressure) to be implemented independently. The four single-factor degradation rates R are measured. i Calculate the independent increment Δ i =R i -R0, i∈{T,UV,P,O};

[0010] S3. Two-factor interaction net contribution extraction: Coupled aging tests were conducted on six pairwise combinations of four environmental factors to measure the two-factor degradation rate R. ij Calculate the net interaction contribution σ ij =R ij -(R0+Δ i +Δ j Extract the net contribution of the two-factor interaction, where the subscripts i, j ∈ {T, UV, P, O} and i ≠ j; σ ij Positive values ​​indicate cooperative acceleration interaction, while negative values ​​indicate antagonistic inhibition interaction and are retained.

[0011] S4. Extraction of high-order synergistic terms across all factors: A fully coupled experiment was conducted using a four-factor, five-level central composite star design to measure the comprehensive degradation response R. total According to formula σ global =R total -(R0+ΣΔ i +Σσ ij Extract higher-order cooperative terms and solve based on the quadratic response surface. R / x i Multidimensional deterioration extreme value coordinates under the condition =0;

[0012] S5. Mechanism Closed-Loop Verification and Iteration: Verify and iterate each Δ... i Converted to apparent activation energy and compared with independent benchmarks, when the total mechanism deviation rate |ε i When |≤15% and the mean relative error of decoupling and reconstruction (MRE) ≤8%, output the independent contribution vector W and the coupling effect matrix σ; otherwise, output σ. ij Perform a reallocation and return to step S4 iteration, with no more than 20 iterations.

[0013] Through a hierarchical decoupling architecture consisting of a 0th-order absolute benchmark, a 1st-order independent increment, a 2nd-order two-factor tensor subtraction, and a 3rd-order CCD high-order coordination, the mathematical defect of repeated subtraction of the underlying background aging in traditional subtraction methods is fundamentally eliminated, making Δ i With σ ij They have independent physical meanings, and the relative error of decoupling and reconstruction is significantly reduced to within 8%. At the same time, through mechanism closed-loop verification and iterative correction mechanism, they provide dual quality assurance for engineering applications.

[0014] A further technical solution of the present invention is: the temperature of the standard reference environment in step S1 is 25±0.5℃, the air pressure is 101.3±0.3kPa, the ultraviolet radiation intensity is zero, and the oxygen partial pressure is 21.2±0.2kPa.

[0015] By clearly defining the specific parameters of the standard reference environment, we can ensure that the absolute baseline degradation rate R0 has good repeatability and universality, providing a unified and stable background subtraction baseline for subsequent subtraction operations of each order.

[0016] A further technical solution of the present invention is: in step S2, a single factor is adjusted to a set value, wherein the set value is an extreme level, and the extreme level is set in the following manner:

[0017] (a) Pure heat group: temperature is -30℃ to +40℃, and the cycle is 12h;

[0018] (b) Pure UV group: UV irradiance of 300–700 W / m², wavelength range of 280–400 nm, and main peak wavelength of 340 nm;

[0019] (c) Pure low-pressure group: The cabin pressure is adjusted to 50–70 kPa, and the oxygen partial pressure is compensated to 21.2 ± 0.2 kPa through the nitrogen / oxygen mixing gas distribution circuit.

[0020] (d) Pure hypoxic group: The cabin pressure was maintained at 101.3±0.3 kPa, and the oxygen partial pressure was adjusted to 7.3–10.5 kPa by nitrogen dilution;

[0021] The aging time for each single factor was no less than 100 hours. At least one of the following was used as the characterization index of degradation response: dynamic shear rheometer (DSR) complex modulus aging index CAI and Fourier transform infrared spectroscopy (FTIR) carbonyl index increment ΔCI.

[0022] Four types of single-factor independent test conditions covering extreme plateau environments are presented. In particular, the independent gas path realizes the decoupled conditions of "low pressure-normal oxygen" and "normal pressure-low oxygen" that cannot be achieved by existing devices. This allows the independent contributions of gas pressure and oxygen partial pressure ΔP and ΔO to be quantified and identified separately, breaking the technical prejudice in this field that "gas pressure and oxygen partial pressure must be co-tuned".

[0023] A further technical solution of the present invention is: the net contribution σ of the two-factor interaction in step S3 is... ij After significance testing via ANOVA, data were categorized and included in subsequent calculations according to the following rules: when the significance level p < 0.05, all data were included with a weight of 1.0; when 0.05 ≤ p < 0.10, a portion was included with a weight of 0.5; when p ≥ 0.10, σ0.05 was set to 0.05. ij =0 indicates exclusion; the significance test for the variance analysis was performed using Bootstrap resampling 5000 times, and the 95% confidence interval was given using the quantile method; for σ ij Antagonistic inhibition interaction terms with a value less than 0 are retained as negative values ​​and included in the weighting calculation when the significance is true.

[0024] By using hierarchical weighting and Bootstrap resampling, the errors caused by simply discarding negative values ​​or subjectively setting thresholds in traditional methods are avoided, thus improving σ. ij The sign can directly reflect the interaction property (positive value for synergistic acceleration, negative value for antagonistic inhibition), thus quantifying for the first time the antagonistic inhibition effect of low oxygen partial pressure on thermo-oxidative aging, an effect that has long been discarded as experimental error in existing technologies.

[0025] A further technical solution of the present invention is: in step S4, the four-factor, five-level, central composite star point design uses the extreme levels of each factor as the axis point, employs non-equidistant practical coding, repeats the center point at least 6 times, and has a total of at least 30 experimental groups; the five-level coding parameters are:

[0026] (a) Temperature T corresponds to five levels: -30 / -10 / 10 / 25 / 40℃;

[0027] (b) Ultraviolet irradiance I UV The corresponding levels are 200 / 300 / 500 / 700 / 800 W / m²;

[0028] (c) The air pressure P corresponds to five levels of 105 / 95 / 75 / 55 / 45 kPa;

[0029] (d) The partial pressure of oxygen (O) corresponds to five levels of 22.0 / 19.0 / 16.0 / 13.0 / 10.0 kPa;

[0030] The fitted quadratic response surface model must simultaneously satisfy the following conditions: coefficient of determination R² ≥ 0.95 and fit test p > 0.05.

[0031] The four-factor, five-level central composite design (CCD) was used to theoretically increase the number of full-factor experimental groups from 5. 4 =625 groups were reduced to 30 groups, a reduction of more than 95%. At the same time, the effectiveness of the response surface model was guaranteed by the dual criteria of R²≥0.95 and lack of fit test p>0.05, realizing the complementarity of statistical modeling and physical decoupling of high-dimensional nonlinear coupling effects.

[0032] A further technical solution of the present invention is: the dual threshold for convergence determination in step S5 is MRE≤8% and all |ε i |≤15%; When any threshold is exceeded, the corresponding two-factor interaction contribution σ will be... ij The components are redistributed according to the direction of deviation, and the process returns to step S4 to re-execute the central composite star point physical test. The maximum number of iterations is 20. If the iteration reaches the upper limit and still fails to converge, it is determined that the aging mechanism of the asphalt specimen is not applicable to the quadratic response surface model of this method, and an inapplicability warning is output.

[0033] A further technical solution of the present invention is: the output independent contribution vector W in step S5 is normalized according to the following formula: W i =Δ i / Σ k |Δ k |, where i, k∈{T, UV, P, O};

[0034] It further outputs the corrected contribution vector after considering the amortization of the two-factor coupling term. , i =C i / Σ k |C k |, where C i =Δ i +0.5×Σ j≠i σ ij ·w ij w ij The significance level weights in step S3.

[0035] By normalizing the contribution vector W and coupling the modified contribution vector The study quantitatively presents the independent contribution of each environmental factor to the overall deterioration and the comprehensive contribution after considering the interaction, providing a clear numerical basis for the screening of modifiers for plateau pavement materials, formulation optimization, and prediction of worst-case conditions. The weights are allowed to take negative values, thus preserving the physical meaning of antagonistic factors.

[0036] A further technical solution of the present invention includes step S6: based on the modified Arrhenius extended equation, combined with the independent contribution vector W or the modified contribution vector In addition to meteorological observation sequences, it outputs equivalent service life prediction curves and preventive maintenance time windows.

[0037] It achieves a complete closed loop from indoor decoupled testing to actual engineering life prediction, and can directly output equivalent service life, preventive maintenance time window and structural layer thickness compensation, serving the durability design and maintenance decisions of asphalt pavements in high-altitude and cold regions.

[0038] A further technical solution of the present invention is: based on the modified Arrhenius extended equation: lnt (eq, i) =A i +B i / T+C i ·ln(I UV / I0)+D i ·ln(P / P0)+E i ·ln(O / O0), where the subscript i∈{T, UV, P, O}, and the coefficient A i To E i Least squares fitting was performed using the single-factor data obtained in step S2 combined with the intermediate-level supplementary experiment; the independent contribution vector W obtained in step S5 was used to approximate the equivalent accelerated aging time t of the four types of factors. (eq, i) Weighted summation yields the overall equivalent acceleration factor k. acc Input a multi-year meteorological observation sequence for the target road section. The meteorological observation sequence includes at least five parameters: annual average temperature, annual average total radiation, annual average air pressure, annual average oxygen partial pressure, and daily average temperature difference. Output the equivalent service life prediction curve and preventive maintenance time window for the asphalt pavement of the road section.

[0039] A further technical solution of the present invention is: the method is applicable to both asphalt binder specimens and asphalt mixture specimens; a quantitative correlation is established between the decoupling results of the binder and the decoupling results of the macroscopic properties of the mixture through a scale conversion factor κ, wherein the calculation formula of the scale conversion factor κ is: κ=(VMA / 100)·(1+K a )·(h film / h ref ), where VMA is the void fraction of the aggregate, in percentage; Ka h is the adsorption coefficient of the aggregate, determined by the specific surface area of ​​the aggregate and the amount of bitumen adsorbed; film The average asphalt film thickness is expressed in μm; h ref The reference film thickness is set to 10 μm; the equivalent independent contribution vector W of the mixture. mix =κ·W, Equivalent coupling effect matrix of mixture σ mix =κ·σ.

[0040] The κ factor was used to establish a quantitative correlation between the decoupling results of micro-binders and the macro-asphalt mixture pavement performance, enabling this method to be extended from the material level to structural engineering applications, with the average relative deviation controlled within 10%.

[0041] Secondly, the present invention provides a decoupling device for implementing the method, characterized in that: it includes a single-factor isolation test chamber, a multi-factor coupled superposition control system, and a degradation feature acquisition and decoupling calculation unit; the data acquisition port of the degradation feature acquisition and decoupling calculation unit is communicatively connected to the degradation response monitoring device of the single-factor isolation test chamber, and the degradation feature acquisition and decoupling calculation unit is communicatively connected to the multi-factor coupled superposition control system; the single-factor isolation test chamber includes a pressure control module and an oxygen partial pressure control module, and the gas paths of the two are independent of each other, enabling the test chamber to independently control the oxygen partial pressure at any pressure level; the degradation feature acquisition and decoupling calculation unit includes a tensor difference subtraction operation module and a quadratic response surface fitting module, and performs subtraction operations of each order with R0 as a common background benchmark.

[0042] By adopting a physically independent structural design for the gas pressure circuit and the oxygen partial pressure circuit, the technical bias of "co-adjustment of gas pressure and oxygen partial pressure" in the existing climate aging chamber is broken. It can independently realize the switching of four working conditions: low pressure-normal oxygen, low oxygen-normal pressure, low pressure-low oxygen, and normal pressure-normal oxygen. This provides a physically independent device basis for the hierarchical decoupled mathematical architecture to independently control the single-factor test conditions.

[0043] A further technical solution of the present invention is: the single-factor isolation test chamber includes:

[0044] The main body of the cabin has a double-layer stainless steel structure, with the inner wall surface sprayed with a low UV reflectance paint film and the interlayer filled with polyurethane insulation material.

[0045] The temperature control module includes a cascaded refrigeration circuit and a PTC heating circuit, with a temperature control accuracy of ±0.5℃. Temperature sensors are arranged at the four vertices of the cabin diagonally.

[0046] The ultraviolet irradiation module includes an array of at least four medium-pressure mercury lamps or mercury-xenon lamps, with the lamp array positioned 150±5 mm from the sample surface and an ultraviolet irradiation uniformity of at least 92%.

[0047] The air pressure control module consists of a vacuum pump, an electromagnetic proportional valve, and a thin-film capacitor barometer forming a three-stage closed loop, with an air pressure control accuracy of ±0.3 kPa. The vacuum pump is connected to the main body of the cabin through an air extraction pipeline.

[0048] The oxygen partial pressure control module includes a nitrogen / oxygen mixing gas path, an electrochemical oxygen concentration meter, and a proportional mass flow controller, with an oxygen partial pressure control accuracy of ±0.2 kPa. The gas path of the oxygen partial pressure control module and the gas pressure control module are physically independent of each other, enabling the test chamber to independently control the oxygen partial pressure at any gas pressure level.

[0049] Each module is equipped with an independent switching circuit and parameter locking mechanism. The parameter locking mechanism is implemented by the state machine of the programmable logic controller (PLC) so that the parameters of any module remain at the standard reference environment level after entering the locked state.

[0050] A further technical solution of the present invention is: the multi-factor coupling superposition control system includes a factor superposition controller and a state monitoring unit. The factor superposition controller supports independent opening, locking, and coupling switching of four types of modules: temperature, ultraviolet, air pressure, and oxygen partial pressure. The module switching response time does not exceed 30 seconds.

[0051] A further technical solution of the present invention is: the degradation feature acquisition and decoupling calculation unit includes: a data acquisition port, which is communicatively connected to the degradation response monitoring equipment of the single-factor isolation test chamber; and a tensor difference subtraction operation module, which, based on R0 and R2 received by the data acquisition port... i and R ij According to formula σ ij =R ij -(R0+Δ i +Δ j The calculation is performed with R0 as a common background benchmark; the quadratic response surface fitting module performs quadratic polynomial least squares fitting based on the experimental data designed by the central composite star point and calculates the multidimensional degradation extreme value coordinates; the tensor difference subtraction operation module and the quadratic response surface fitting module are both deployed in the data processor of the degradation feature acquisition and decoupling calculation unit.

[0052] A further technical solution of the present invention is as follows: the single-factor isolation test chamber further includes a crosstalk real-time monitoring and closed-loop compensation system; the system includes a sensor array, which consists of no less than 2 oxygen concentration sensors, no less than 2 ultraviolet irradiance sensors, no less than 4 K-type thermocouples, and no less than 2 capacitive thin-film barometers, and the sampling frequency of all sensors is no less than 2Hz; when the measured value of any non-target factor deviates from the standard reference environmental level by more than ±2%, the closed-loop compensation action of the corresponding module is automatically triggered, and the time period is marked as pollution data in the monitoring data in the chamber and removed.

[0053] A further technical solution of the present invention is as follows: the gas extraction pipeline of the gas pressure control module and the gas distribution pipeline of the oxygen partial pressure control module are respectively arranged at different ports of the single-factor isolation test chamber, and an independent one-way valve, mass flow controller and proportional solenoid valve are provided between them; the device performs nitrogen dilution gas distribution through the independent gas distribution pipeline, so that the device can independently realize the following four working conditions:

[0054] (a) Low-pressure-normal oxygen condition: The gas pressure is adjusted to 50–70 kPa, and oxygen is simultaneously supplemented through the independent gas distribution path to maintain the oxygen partial pressure at 21.2 ± 0.2 kPa;

[0055] (b) Low oxygen-normal pressure condition: The gas pressure is maintained at 101.3±0.3kPa, and the oxygen partial pressure is reduced to 7.3–10.5kPa by nitrogen dilution through the independent gas distribution path.

[0056] (c) Low-pressure-low-oxygen synchronous operating condition: both air pressure and oxygen partial pressure are adjusted to extreme plateau levels;

[0057] (d) Normal pressure-normal oxygen reference condition: Both air pressure and oxygen partial pressure are maintained at the standard reference level;

[0058] The switching response time between the four operating conditions does not exceed 30 seconds, and the measured deviation of air pressure and oxygen partial pressure under each operating condition does not exceed ±2%.

[0059] A further technical solution of the present invention is: the degradation feature acquisition and decoupling calculation unit further includes a mechanism closed-loop verification module; the mechanism closed-loop verification module converts each Δ output by the tensor difference subtraction operation module into a single Δ value. i Converted to apparent activation energy E (a, i) The computer compares the deviation rate ε with a preset independent physical reference parameter library. i When all |ε iWhen |≤15% and mean relative error MRE≤8%, the mechanism closed-loop verification module outputs a decoupling convergence signal to the multi-factor coupled superposition control system, triggering the system to terminate the test; otherwise, it outputs an iteration command to the multi-factor coupled superposition control system, triggering the repeated execution of the system execution center composite star point test.

[0060] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it can receive data obtained from physical experiments conducted by the decoupling device, perform decoupling calculations in the method, and then send the calculation results back to the control system of the decoupling device to trigger or terminate the iteration of the physical experiment.

[0061] It achieves automated linkage between physical experiments and decoupled calculations. When the deviation of the mechanism closed-loop verification exceeds the threshold, it can automatically trigger the next round of physical experiment iteration without manual intervention, thereby improving the decoupling efficiency and reliability, and ensuring that the convergence can be achieved in no more than 20 iterations.

[0062] A further technical solution of the present invention is: a computer-readable storage medium storing a computer program thereon, the computer program being communicatively connected to the degradation characteristic acquisition and decoupling calculation unit in the plateau asphalt multi-factor coupled aging decoupling device; when the computer program is executed by a processor, it receives the baseline degradation rate R0 and the single-factor degradation rate R obtained by the degradation characteristic acquisition and decoupling calculation unit from the physical aging test in steps S1 to S4. i Two-factor degradation rate R ij Given the CCD full-factor response dataset, perform the following steps:

[0063] According to the formula Δ i =R i -R0 is used to calculate the independent increment of a single factor; according to the formula σ ij =R ij -(R0+Δ i +Δ j Calculate the net contribution of the two-factor interaction and perform the ANOVA significance classification; perform quadratic response surface fitting according to the CCD parameters, and calculate the result according to formula σ. global =R total -[R0+ΣΔ i +Σσ ij Extract higher-order cooperative terms; perform the mechanism closed-loop convergence determination; output the independent contribution vector W and the coupling effect matrix σ to the multi-factor coupled superposition control system, and the multi-factor coupled superposition control system triggers or terminates the physical test iteration of step S4 accordingly.

[0064] The key technologies of this invention are: (1) explicitly incorporating the absolute benchmark deterioration rate R0 into the subtraction operation of each order, eliminating the systematic bias of repeated subtraction of the basic background in the traditional subtraction method; (2) the air pressure and oxygen partial pressure are controlled separately by independent air paths, realizing for the first time in the asphalt aging test the decoupling conditions that cannot be implemented by the two existing co-regulation devices of low pressure-normal oxygen and normal pressure-low oxygen; (3) the negative antagonistic interaction term is retained and quantified after the significance test is established, instead of being discarded as test error; (4) the mechanism closed-loop verification triggers the automatic iterative correction of the physical test, forming a dual convergence criterion of data self-consistency and mechanism self-consistency.

[0065] The beneficial effects of this invention are: (1) Eliminating systematic biases from the mathematical foundation. By explicitly incorporating R0 into the subtraction calculations of each order, the mathematical defect of repeated subtraction of the basic background aging in the traditional subtraction method is avoided, thus making Δ i With σ ij They have independent physical meanings, and the relative error of decoupling and reconstruction is greatly reduced.

[0066] (2) Accurately identify the two-way game of co-acceleration and antagonistic inhibition. Tensor difference subtraction σ ij The sign of the symbol directly reflects the interaction property, with positive values ​​indicating synergistic acceleration and negative values ​​indicating antagonistic inhibition. Based on this, the present invention quantifies the antagonistic inhibition effect of low oxygen partial pressure on thermo-oxidative aging, with typical values ​​ranging from -0.0014 to -0.0021 (see data from Stage 3 of Example 1 for details). This effect has long been discarded as experimental error in the prior art.

[0067] (3) Physical independent control of air pressure and oxygen partial pressure. Through the design of a dedicated independent air circuit, the present invention realizes the physical independent control of air pressure and oxygen partial pressure during the aging of asphalt at high altitudes, so that the independent contributions of the two (typical values ​​ΔP=+0.0026 synergistic, ΔO2=-0.0021 antagonistic) can be quantified and identified separately, breaking the technical prejudice of "co-regulation of air pressure and oxygen partial pressure" in this field.

[0068] (4) Physical stripping and statistical modeling complement each other. For low-order (0-2 order) experiments, isolated physical subtraction is used to ensure that the mechanism is clear; for high-order (3 order and above) experiments, CCD response surface fitting is introduced to avoid combinatorial explosion in high-dimensional factorial experiments, and the number of experimental groups is reduced from the theoretical 5. 4 =625 groups were reduced to 30 groups, a reduction of 95%.

[0069] (5) Mechanism closed-loop verification improves reliability. By cross-comparing the decoupled dynamic parameters with independent benchmark parameters, physical test iterative correction is triggered when the deviation exceeds 15%, providing dual quality assurance for engineering applications. In the first embodiment of the present invention, convergence is achieved in only 3 iterations.

[0070] (6) Full-process engineering transformation path. A closed loop is formed from benchmark calibration to life prediction, which can directly output the preventive maintenance time window and the structural layer thickness compensation amount, serving the durability design of asphalt pavement in high-altitude and cold regions. Attached Figure Description

[0071] Figure 1 This is the overall computational logic architecture diagram of the layered decoupling method provided by the present invention;

[0072] Figure 2 This is a three-dimensional cross-sectional structural diagram of the single-factor isolation test chamber of the present invention.

[0073] Reference numerals in the attached figures: 1. Chamber body, 2. Upper port, 3. Lower port, 4. Ultraviolet radiation module, 5. Temperature control coil, 6. Barometer, 7. Oxygen concentration meter, 8. Sample platform. Detailed Implementation

[0074] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0075] A decoupling method for the multi-factor coupled aging effects of asphalt pavement materials in high-altitude areas includes five core steps:

[0076] S1, Absolute reference calibration stage: The absolute reference degradation rate R0 was measured in a standard reference environment with a temperature of 25±0.5℃, an air pressure of 101.3±0.3kPa, an ultraviolet radiation intensity of zero, and an oxygen partial pressure of 21.2±0.2kPa. This R0 was used as the common background subtraction reference for all test groups in subsequent S2 to S4.

[0077] S2. Single-factor independent incremental physical stripping stage: Any one of the four factors—temperature (T), ultraviolet radiation (UV), atmospheric pressure (P), and oxygen partial pressure (O)—was individually adjusted to the extreme plateau level, while the other three were locked at the reference level. The deterioration rates of the four single-factor groups were measured and calculated according to Δ... i =R i -R0 extracts the independent degradation increment of each factor; in particular, the extreme level control of the gas pressure P is achieved through the vacuum pumping circuit, and the extreme level control of the oxygen partial pressure O is achieved through the nitrogen / oxygen mixing circuit independent of the vacuum pumping circuit. The two are two independent factors that participate in the entire test process of S2, S3 and S4.

[0078] S3, the two-factor interaction net contribution extraction stage, involves performing a two-factor coupled aging test on six pairwise combinations of four factors, according to σ. ij =Rij -(R0+Δ i +Δ j Extract the net contribution of tensor difference subtraction interaction, positive values ​​are judged as co-acceleration, and negative values ​​are judged as antagonistic inhibition and are retained;

[0079] S4. In the stage of extracting high-order synergistic terms across all factors, a four-factor, five-level central composite star design was used to execute a multi-level fully coupled experiment, according to σ... global =R total -(R0+ΣΔ i +Σσ ij Extract higher-order cooperative terms and solve for multidimensional degradation extremum coordinates based on the quadratic response surface;

[0080] S5, Mechanism Closed-Loop Verification Stage, will verify each Δ i Converted to apparent activation energy and compared with independent physical benchmarks, when all |ε i Convergence is determined when |≤15% and the mean relative error of decoupling and reconstruction (MRE)≤8%, and the independent contribution vector W and the coupling effect matrix σ are output; if convergence is not achieved, return to S4 to re-execute the central composite star point physical experiment, with no more than 20 iterations.

[0081] Furthermore, the four types of single-factor extreme level parameters in step S2 include the temperature cycling parameters of the pure heat group, the irradiation parameters of the pure UV group, and the air pressure / oxygen partial pressure parameters of the pure low pressure group and the pure low oxygen group, and the aging time is limited to not less than 100h. The degradation response characterization index adopts at least one of DSR complex modulus aging index CAI and FTIR carbonyl index increment ΔCI.

[0082] Furthermore, in step S3, the ANOVA significance ranking weighting rules are as follows: p < 0.05, weight 1.0; 0.05 ≤ p < 0.10, weight 0.5; p ≥ 0.10, weight zero; Bootstrap resampling is used to give a 95% confidence interval after 5000 iterations; σ... ij Antagonistic inhibition terms with a value less than 0 retain negative values ​​when significance is established.

[0083] Furthermore, in step S4, the CCD parameters are: four factors and five levels, the axis points are taken as the extreme levels of each factor (non-equidistant practical coding), the center point is repeated no less than 6 times, and the total number of experimental groups is no less than 30; the specific values ​​of the five levels of the four factors of temperature cycling amplitude, ultraviolet irradiance, air pressure, and oxygen partial pressure are clearly given; the response surface fitting must meet the requirements of R²≥0.95 and the lack of fit test p>0.05.

[0084] Furthermore, the dual convergence threshold MRE in step S5 is ≤ 8% and |ε i |≤15%, and σ when out of tolerance ijThe redistribution strategy and the 20-iteration limit, as well as the inapplicable warning mechanism when the iteration limit still does not converge.

[0085] Furthermore, step S6, service life prediction engineering application, is added. Based on the modified Arrhenius extended equation combined with weight vectors and meteorological observation sequences, the equivalent service life prediction curve and preventive maintenance time window are output.

[0086] Furthermore, a scale conversion factor κ mechanism is added to make this method applicable to both asphalt binder specimens and mixture specimens, and to establish a quantitative correlation between microscopic and macroscopic decoupling results.

[0087] Furthermore, a decoupling device for implementing the above method includes three main parts: a single-factor isolation test chamber, a multi-factor coupled superposition control system, and a degradation feature acquisition and decoupling calculation unit. In particular, the gas paths of the gas pressure control module and the oxygen partial pressure control module in the single-factor isolation test chamber are independent of each other, enabling the test chamber to independently control the oxygen partial pressure at any gas pressure level. The degradation feature acquisition and decoupling calculation unit includes a tensor difference subtraction operation module and a quadratic response surface fitting module, which perform subtraction operations of each order with R0 as a common background reference.

[0088] Furthermore, a crosstalk real-time monitoring and closed-loop compensation system is added, which uses a sensor array to detect deviations of non-target factors in real time and triggers closed-loop compensation and contamination data removal.

[0089] Furthermore, the port distribution and control elements of the gas pressure path and oxygen partial pressure path are added, enabling the device to independently switch between four operating conditions: low pressure-normal oxygen, low oxygen-normal pressure, low pressure-low oxygen synchronization, and normal pressure-normal oxygen reference. The response time is no more than 30 seconds, and the measured deviation of each operating condition is no more than ±2%.

[0090] Furthermore, a mechanism closed-loop verification module is added to realize automatic linkage between the decoupled convergence / iteration signal and the multi-factor coupled superposition control system.

[0091] Furthermore, a computer-readable storage medium, on which a computer program is stored, is communicatively connected to the degradation feature acquisition and decoupling calculation unit in the aforementioned decoupling device. After receiving physical test data from the device, the program performs decoupling calculations and transmits the results back to the device control system to trigger or terminate the physical test iteration.

[0092] This invention firstly constructs a hierarchical decoupled mathematical framework consisting of a 0th-order absolute baseline, a 1st-order independent increment, a 2nd-order two-factor tensor difference subtraction, and a 3rd-order CCD high-order synergy. Let the performance degradation response value of an asphalt specimen under any environmental state E be R(E). Define the standard reference state E0 = (T0, UV0, P0, O0), corresponding to the absolute baseline R0 = R(E0). A perturbation operator P is introduced for each environmental factor i. iIf the temperature is deviated from the baseline to the extreme level of the plateau, the single-factor degradation response is R. i =R(P i E0), two-factor R ij =R(P j P i E0), with all factors being R total =R(P T , P UV , P, O, E0).

[0093] Define first-order independent increment Δ i =R i -R0 represents the independent influence of factor i relative to the baseline; the net contribution of the second-order interaction is defined as: σ ij =R ij -(R0+Δ i +Δ j )=R ij -R i -R j +R 0。 This is the tensor difference subtraction formula, which is essentially a second-order mixed deviation integral of the response function with respect to the perturbations of the two factors, and can be written as σ ij ≈ ²R / ( P i P j )·ΔP i ·ΔP j This formula, by explicitly incorporating R0, eliminates the double-subtracted R0 term found in traditional difference subtraction, thus ensuring the physical purity of the interaction terms. Third-order and higher-order cooperative terms are defined as σ. global =R total -(R0+ΣΔ i +Σσ ij The physical meaning of is a high-dimensional nonlinear cooperative quantity that cannot be explained by the superposition of zero-order, first-order and second-order linear quantities.

[0094] The second aspect of this invention is that, at the device level, it breaks through the technical bias of existing climate aging chambers that rely on "co-regulation of air pressure and oxygen partial pressure" by using an independent gas path system for air pressure and oxygen partial pressure. This provides physically independently controllable single-factor experimental conditions for the aforementioned hierarchical decoupled mathematical architecture. Specifically, the air pressure control module is implemented through a vacuum pumping circuit, while the oxygen partial pressure control module is implemented through a nitrogen / oxygen mixing circuit independent of the vacuum pumping circuit. The two circuits are physically independent of each other. This design enables the device to achieve four independent operating condition switching modes: (a) low-pressure-normal oxygen condition (only air pressure is reduced, and oxygen is simultaneously supplemented to maintain normal oxygen partial pressure); (b) low oxygen-normal pressure condition (only oxygen partial pressure is reduced, and normal pressure is maintained); (c) low-pressure-low oxygen synchronous condition (both decrease simultaneously, corresponding to the actual plateau environment); and (d) normal pressure-normal oxygen reference condition (standard reference environment). Conditions (a) and (b) are not achievable with existing commercial climate aging test equipment. Existing equipment uses a shared vacuum-gas distribution circuit for both gas pressure and oxygen partial pressure, causing the oxygen partial pressure to decrease proportionally during vacuuming, creating an inseparable coupling between them. This device is not "statically isolated" but rather "dynamically steady-state": partial pressure = total pressure × oxygen mole fraction. Maintaining low-pressure to constant oxygen levels relies on the vacuum pump continuously pumping gas to the target value while the gas distribution circuit continuously replenishes oxygen-rich nitrogen-oxygen mixture based on real-time feedback from the oxygen concentration meter and via a mass flow controller, thus replenishing the removed oxygen and dynamically stabilizing the oxygen partial pressure at 21.2 kPa. Essentially, it is an open-flow steady-state process of simultaneous pumping and replenishment (similar to maintaining a constant water level while simultaneously releasing and replenishing water).

[0095] The five-level coding for environmental factors in this specification is consistent with the aforementioned five-level coding parameters, denoted as -2 / -1 / 0 / +1 / +2, corresponding to: temperature T=-30 / -10 / +10 / +25 / +40℃; ultraviolet irradiance I... UV =200 / 300 / 500 / 700 / 800W / m 2 Air pressure P = 105 / 95 / 75 / 55 / 45 kPa; oxygen partial pressure O = 22.0 / 19.0 / 16.0 / 13.0 / 10.0 kPa. The coding direction of air pressure and oxygen partial pressure is consistent with the direction of plateau deterioration (code +2 corresponds to the lowest air pressure / lowest oxygen partial pressure).

[0096] Example 1:

[0097] SBS modified asphalt four-factor coupled aging process decoupling test

[0098] 1.1 Materials, Specimens, and Deterioration Response Indicators

[0099] This embodiment uses SBS (styrene-butadiene-styrene triblock copolymer) modified asphalt with a performance grade of PG76-28. The SBS content is 4.5% of the relative asphalt mass, the penetration is 58×0.1mm (25℃, 100g, 5s), the softening point is 76℃ (ring and ball method), and the ductility at 5℃ is 38cm.

[0100] Following ASTM D7175-2015 standards, molten asphalt was cast into DSR standard thin-film specimens with a diameter of 25 mm and a thickness of 1 mm, totaling 84 specimens for all stages S1 to S5. Additionally, AC-13 (continuous dense gradation, nominal maximum particle size 13.2 mm) gradation mixtures were prepared according to asphalt mixture gradation design standards, and Marshall specimens with a diameter of 101.6 mm and a height of 63.5 mm were formed for the scale conversion verification in Example 3. All specimens were left to stand at room temperature in the dark for 24 hours after preparation to eliminate residual stress.

[0101] In this embodiment, the carbonyl index increment ΔCI of FTIR (Fourier Transform Infrared Spectroscopy) is selected as the main response index, which is defined as 1700 cm⁻¹. -1 The absorption peak area of ​​the carbonyl group (C=O) is 1460 cm⁻¹ -1 The change in the ratio of the methyl (CH3) reference peak area before and after aging, i.e.:

[0102] ΔCI=(A C=O / A CH3 ) 老化后 -(A C=O / A CH3 ) 老化前 (1)

[0103] Simultaneous measurement of DSR complex modulus aging index CAI (= 老化后 / 老化前 ) and ductility retention rate were used as auxiliary verification indicators. FTIR scan parameters: resolution 4cm -1 The number of scans was 32; 5 measurement points were randomly selected on the surface of each specimen, and the mean value was taken as the ΔCI value of the specimen; 6 specimens were taken in each group of tests, totaling 30 measurement points, and the results were reported as mean ± standard deviation.

[0104] 1.2 Experimental setup and parameter monitoring

[0105] The core structure and parameters of the single-factor isolation test chamber described above in this invention are as follows:

[0106] (1) Cabin body 1: Double-layer 316L stainless steel shell (wall thickness 3mm) + 50mm polyurethane insulation interlayer, inner wall sprayed with low UV reflectivity paint film (reflectivity ≤5% in 280~400nm band), cabin working space 500mm×400mm×400mm.

[0107] (2) Temperature control module: The dual-loop configuration of cascade refrigeration compression circuit (lower temperature limit -30℃) and PTC heating circuit (upper temperature limit +80℃) is used to achieve heat exchange in the cabin through temperature control coil 5. Four K-type thermocouples are arranged at the four vertices of the cabin diagonal. The temperature control accuracy is ±0.5℃ and the temperature uniformity is ≤1.0℃.

[0108] (3) Ultraviolet radiation module 4: An array of 6 medium-pressure mercury lamps, 150 mm from the sample surface, with an irradiance uniformity of 92.4% (verified by 9-point measurement throughout the chamber), a spectral range of 280~400 nm, a main peak wavelength of 340 nm, and an adjustable irradiance range of 100~1000 W / m. 2 ;

[0109] (4) Pressure control module: The rotary vane vacuum pump (ultimate vacuum 0.01kPa) is connected to the upper port 2 of the chamber through the air extraction pipeline. The electromagnetic proportional valve and the thin film capacitor barometer 6 form a three-stage closed loop. The pressure control accuracy is ±0.3kPa and the control range is 40~110kPa.

[0110] (5) Oxygen partial pressure control module: The nitrogen / oxygen mixed gas distribution path is connected to the cabin through an independent gas distribution port (arranged at the lower port 3 of the cabin, which is completely separated from the upper port 2 exhaust pipeline in terms of space and control). The gas distribution pipeline is independently equipped with a one-way valve, a proportional solenoid valve and a mass flow controller. The electrochemical oxygen concentration meter 7 provides real-time feedback. The oxygen partial pressure control accuracy is ±0.2kPa and the control range is 5.0~22.0kPa.

[0111] (6) PLC state machine control: After any module enters the locked state, its parameters are automatically maintained at the standard reference environment level (temperature 25℃, air pressure 101.3kPa, ultraviolet radiation intensity is zero, oxygen partial pressure 21.2kPa), and the working condition switching response time is ≤30s.

[0112] Two additional UV irradiators, one laser oxygen concentration meter, and one capacitive thin-film barometer are installed inside the chamber as redundant calibration sensors, with a sampling frequency of 5Hz, and the data is fed back to the PID closed-loop control system in real time. When the measured value of any non-target factor deviates from its reference level by more than ±2%, the closed-loop compensation action of the corresponding module is immediately triggered, and the corresponding time period is marked as polluted data and removed.

[0113] 1.3 Stage 1—Absolute Baseline Deterioration Rate R Calibration (Order 0)

[0114] Six SBS-modified bitumen DSR thin film specimens were aged for 120 hours in a standard reference environment (constant temperature 25±0.5℃, air pressure 101.3±0.3kPa, complete darkness without UV, oxygen partial pressure 21.2±0.2kPa). This 120-hour duration was calculated based on the equivalent aging time of ASTM D2872-2012 RTFOT to ensure that R0 is within the first-order linear region of reaction kinetics, avoiding nonlinear distortion in the zero-order reaction region and saturation region.

[0115] Preliminary experiments (including four time points: 24h, 60h, 120h, and 180h) verified that the slope of the carbonyl index-time curve at 120h corresponds to the first-order rate constant k0 = 4.32 × 10⁻⁶. -5 h -1 The relative deviation of the predicted slope is ≤3%, which meets the first-order linear interval criterion (deviation threshold 5%).

[0116] ΔCI ​​was measured for each specimen before and after aging using the FTIR scanning parameters described above, yielding R0 = 0.0052 ± 0.0003 (n = 6, 30 measurement points). This value was used as the common background subtraction benchmark for all subsequent test groups in stages S2 to S4.

[0117] 1.4 Stage 2—Single-Factor Independent Incremental Physical Stripping (Order 1)

[0118] Four sets of single-factor extreme aging tests were performed sequentially, with 6 specimens in each set and an aging time of 120 hours. The other three factors were strictly controlled at the standard reference level. The specific test conditions and measured results are as follows.

[0119] (a) Pure thermal group: Temperature T cycles sinusoidally between -30°C and +40°C, with a period of 12 hours, for a total of 10 complete cycles, totaling 120 hours; other factors are strictly locked (atmospheric pressure 101.3±0.3 kPa, oxygen partial pressure 21.2±0.2 kPa, ultraviolet radiation intensity zero). Actual measurement: R T =0.0187±0.0008, Δ T =R T -R0=+0.0135

[0120] (b) Pure UV group: constant temperature 25℃, air pressure 101.3±0.3kPa, oxygen partial pressure 21.2±0.2kPa, applied at 500W / m 2 Continuous ultraviolet irradiation (weight 4, center point horizontal, wavelength 280–400 nm, main peak 340 nm). Measured: R UV =0.0156±0.0006, Δ UV =+0.0104

[0121] (c) Pure Low-Pressure Group (Low-Pressure-Normal Oxygen Condition): Temperature is constant at 25°C, no ultraviolet radiation, and the chamber pressure is adjusted to P=55±0.3kPa via vacuum extraction; simultaneously, oxygen is supplemented through an independent oxygen partial pressure distribution gas path, maintaining the oxygen partial pressure at a reference level of 21.2±0.2kPa. This independent achievement of the low-pressure-normal oxygen condition directly reflects the independent gas path design of this invention, which is impossible to achieve in existing commercial climate aging chambers sharing a vacuum-gas distribution circuit. Actual measurement: R P =0.0078±0.0004, Δ P =+0.0026

[0122] (d) Pure hypoxic group (atmospheric pressure-hypoxic condition): Temperature constant at 25℃, no ultraviolet radiation, chamber pressure maintained at 101.3±0.3 kPa, oxygen partial pressure reduced to O=10.0±0.2 kPa via nitrogen dilution through an independent gas distribution path. This atmospheric pressure-hypoxic condition is also unique to the device of this invention. Actual measurement: R O =0.0031±0.0003, Δ O =-0.0021. Δ O A negative value indicates that low oxygen partial pressure alone has an independent inhibitory effect on the carbonylation reaction, i.e., the antagonistic inhibition effect described in this invention. The chemical mechanism is as follows: a decrease in oxygen partial pressure leads to a reduction in the rate of hydroperoxide (ROOH) formation, thereby inhibiting the subsequent chain reaction of β-fracture to form carbonyl groups; furthermore, the diffusion coefficient of oxygen in asphalt at an oxygen partial pressure of 10.0 kPa is approximately 47% lower than at 21.2 kPa (estimated by Effick's law D∝O), further limiting the effective oxygen supply concentration for the oxidation reaction.

[0123] The independent implementation of the above two sets of operating conditions (c) and (d) relies on the structural design of the device of the present invention, in which the gas pressure circuit and the oxygen partial pressure circuit are independent. Existing commercial climate aging chambers share a vacuum-gas distribution circuit for gas pressure and oxygen partial pressure. When vacuuming, the oxygen partial pressure will inevitably decrease proportionally and synchronously, which cannot achieve the above decoupled operating conditions, and therefore cannot quantify Δ separately. P With Δ O Independent contributions.

[0124] Table 1. Summary of Single-Factor Independent Incremental Amounts in Stage 2 of Example 1 (SBS Modified Asphalt)

[0125]

[0126] 1.5 Stage 3—Two-factor tensor difference minus interaction net contribution (2nd order)

[0127] Two-factor coupled aging tests were conducted according to group C(4, 2)=6, with 6 specimens in each group and an aging time of 120 hours. The two factors not involved were strictly locked at the reference level. The net contribution of the two-factor interaction σ... ij Calculated using the tensor difference subtraction formula as described in this invention:

[0128] σ ij =R ij -(R0+Δ i +Δ j )=R ij -R i -R j +R0(2)

[0129] Positive values ​​were identified as co-accelerating interactions, while negative values ​​were identified as antagonistic or inhibitory interactions and were retained when significance was established. The significance test employed Bootstrap resampling 5000 times, and σ was calculated according to the ANOVA significance grading rules. ij Inclusion weight w ij : p<0.05 take w ij =1.0; 0.05≤p<0.10, take w ij =0.5; p≥0.10 take w ij =0 (zeroed out). Specific parameters and measured results are shown in Table 2.

[0130] Table 2. Results of six groups of two-factor coupled tests in Stage 3 of Example 1 (SBS modified asphalt)

[0131]

[0132] After weighted inclusion, the total contribution of the two-factor interaction term in this stage is: Σσ ij ·w ij =0.0121+0.0028+(-0.0014)+0.0019×0.5+0+0=+0.01445 (3)

[0133] Item (c) in the table above clearly reveals an antagonistic-inhibitory interaction (σ) between temperature and hypoxia. T (O=-0.0014, p=0.041<0.05, significant), its physical meaning is: under the environment of thermal oxidation induced by temperature cycling, low oxygen partial pressure further inhibits the carbonylation reaction by limiting the formation of ROOH intermediates. This effect is manifested as an inseparable negative experimental error in existing technologies using gas pressure-oxygen partial pressure co-regulation devices, and is therefore discarded. This invention retains and quantifies it as an independent physical mechanism through independent gas path design.

[0134] 1.6 Stage 4—Central Composite Star Point Design for High-Order Collaborative Extraction (3rd Order)

[0135] A four-factor, five-level central composite star design (CCD) was adopted, with the pivot point representing the extreme levels of each factor (non-equidistant practical coding), and the pivot point being repeated 6 times, for a total of 30 experimental groups. Specifically, it consisted of 16 factorial points (2... 4=16) + 8 axis points (2×4=8) + 6 center point repetitions. The five-level codes of each factor and their corresponding physical parameter levels are shown in Table 3.

[0136] Table 3. Parameters corresponding to the CCD four-factor five-level coding (Example 1)

[0137]

[0138] The coding directions for both air pressure and oxygen partial pressure are consistent with the direction of plateau deterioration (code +2 corresponds to the most severe plateau working conditions); temperature is coded in ascending order of value, with no single deterioration direction (high temperature and low temperature correspond to two different mechanisms: accelerated thermal oxidation and brittle damage, respectively).

[0139] In the 30 groups of tests, each group had 6 specimens, and the aging time was 120h (when temperature factors were involved, a sinusoidal cycle was performed between the T set value and T-10℃, with a cycle of 12h, to keep the thermal stress action mode consistent with the single-factor pure heat group); the measured mean value of ΔCI ranged from 0.0068 to 0.0752.

[0140] Table 4. Measured ΔCI at some representative test points of CCD in Example 1

[0141]

[0142] A quadratic polynomial response surface model was used to fit 30 sets of measured ΔCI:

[0143] R(x) = β0 + Σ i β i ·x i +Σ i β ii ·x i² +Σ i<j β ij ·x i ·x j (4)

[0144] In the formula, subscripts i, j ∈ {T, UV, P, O}, x i The values ​​are coded for each factor. The coefficients obtained by least-squares fitting are shown in Table 5.

[0145] Table 5. Fitting coefficients of the CCD quadratic response surface in Example 1

[0146]

[0147] Model determination coefficient R 2 =0.967, adjust R 2 adj =0.941, the lack of fit test p=0.213>0.05, and the specified R0.05 is satisfied. 2The response surface model is effective based on the dual criteria of ≥0.95 and p>0.05 for lack of fit.

[0148] Four types of factors are simultaneously applied to extreme levels (T=+40℃, I UV =800W / m 2 P=45kPa, O=10.0kPa (i.e., all codes are +2 or equivalent extreme combinations), measured results show: R total =0.0738

[0149] σ is calculated according to the higher-order collaborative term extraction formula described in this invention. global :

[0150] σ global =R total -[R0+ΣΔ i +Σσ ij ·w ij (5)

[0151] Substitute the data:

[0152] ΣΔ i =+0.0135+0.0104+0.0026+(-0.0021)=+0.0244

[0153] Σσ ij ·w ij =+0.01445

[0154] σ global =0.0738-(0.0052+0.0244+0.01445)=+0.02975

[0155] σ global R total The 40.3% contribution indicates a significant nonlinear higher-order synergistic effect in the system, which cannot be explained by the 0th to 2nd order linear superposition. This higher-order contribution reflects both the real nonlinear mechanical-chemical coupling and is incorporated into the decoupling loop as a statistical compensation term of the 0th to 2nd order model residuals.

[0156] Solve the first-order partial derivative conditions based on the above response surface model. R / x i =0, multidimensional degradation extreme coordinates (within the five-level envelope): T=+38.5℃, I UV =782W / m 2 P = 48.5 kPa, O = 10.2 kPa; corresponding predicted degradation response R max =0.0761.

[0157] These extreme coordinates can serve as a benchmark for predicting the most unfavorable aging conditions and can be used to set targets for structural layer anti-aging design and modifier screening tests.

[0158] 1.7 Stage 5—Mechanism Closed-Loop Verification and Decoupling Weight Matrix Output

[0159] The Δ obtained in stage 2 i Converted to the corresponding apparent activation energy E (a , i) Compared with an independent physical reference parameter library, the computer-generated deviation rate is:

[0160] ε i =(E a,i -E a,i,ref ) / E a,i,ref (6)

[0161] Where E a , i k is obtained by fitting the first-order reaction kinetics of ΔCI as a function of time. i k0 is the first-order reaction rate constant under the reference environment (4.32 × 10⁻⁶). -5 h -1 (estimated by R0 / 120h).

[0162] Apparent activation energies of each factor According to the Arrhenius two-point formula, the first-order reaction rate constant is derived from the reference environment and the single-factor extreme environment. , Reverse calculation:

[0163] E a,i =-R·ln( / / (1 / T i,eff -1 / (7)

[0164] Where: R is the universal gas constant (8.314 J·mol⁻¹) -1 ·K -1 ); =298.15K (reference ambient absolute temperature); T i,eff T is the Arrhenius equivalent absolute temperature under the action of factor i alone. For temperature cycling conditions (pure heat group -30~+40℃ sinusoidal cycle), T i,eff Solve it consistently using equation (7a):

[0165] T i,eff =- / [R·ln (7a)

[0166] Where T(t) is the instantaneous absolute temperature of the temperature cycling procedure. It is the time average within the period.

[0167] For the two non-chemical activation energy terms, gas pressure and oxygen partial pressure, the diffusion coefficient ratio of oxygen in asphalt was used as an independent physical benchmark for cross-validation. The validation results are shown in Table 6.

[0168] Table 6. Results of Mechanism Closed-Loop Verification in Phase 5 of Example 1

[0169]

[0170] The absolute values ​​of the deviation rates for all four mechanisms are ≤15%, passing the mechanism closure verification. Overall reconstruction verification: using the obtained Δ... i With σ ij Refactoring R total Predicted value R recon =R0+ΣΔ i +Σσ ij +σ global =0.0738, relative error with the measured value is 0% (due to σ). global (Definition refers to the difference term); the mean relative error (MRE) of the decoupling reconstruction of 30 CCD factorial points is 6.4%, which meets the ≤8% threshold. In this embodiment, convergence occurred in 3 iterations, which did not reach the upper limit of 20 iterations.

[0171] The output independent contribution vector W is normalized according to the following formula (the denominator takes |Δ i The sum of | to ensure W i (The algebraic sum is not distorted by the presence of negative terms).

[0172] W i =Δ i / Σ k |Δ k |, where i, k∈{T, UV, P, O} (8)

[0173] Substitute the data: Σ k |Δ k |=0.0135+0.0104+0.0026+0.0021=0.0286, therefore: W=[W T, W UV, W P, W O =[+0.472, +0.364, +0.091, -0.073]

[0174] W i Negative values ​​are allowed to characterize antagonistic and inhibitory factors (oxygen partial pressure in this example); their algebraic sum is not equal to 1, |W i The sum of these is 1.

[0175] The contribution vector is corrected after considering the amortization of the two-factor coupling term. The specific allocation rule is as follows: for each σ ij ·w ij Distribute the weights equally (0.5 each) to i and j. i and j Insignificant terms (w) ij =0) No allocation. First calculate the coupling-extended contribution C of each factor. i :

[0176] C i =Δ i +0.5×Σ j≠i σ ij ·w ij (9a)

[0177] Renormalization:

[0178] i =C i / Σ k |C k | (9b)

[0179] Substitute the data from this embodiment and calculate item by item:

[0180] C T =+0.0135+0.5×(0.0121+0.0028+(-0.0014))=+0.02025

[0181] C UV =+0.0104+0.5×(0.0121+0.0019×0.5+0)=+0.016925

[0182] C P =+0.0026+0.5×(0.0028+0.0019×0.5+0)=+0.004475

[0183] C O =-0.0021+0.5×((-0.0014)+0+0)=-0.0028

[0184] Σ k |C k = 0.02025 + 0.016925 + 0.004475 + 0.0028 = 0.04445

[0185] Therefore, the corrected contribution vector is obtained:

[0186] =[ T, UV, P, O =[+0.456, +0.381, +0.101, -0.063]

[0187] | i The sum of | is approximately 1.001 ≈ 1.0, and the normalization check passes. The coupling effect matrix σ is shown in Table 2. W and The difference reflects the correction of the relative contributions of each factor after the two-factor coupling term is allocated: the temperature term, due to its strong synergy with UV and low pressure and weak antagonism with hypoxia, has a lower proportion of its coupling extension contribution (W). T =0.472→ T =0.456); the proportion of ultraviolet terms increased slightly due to the partial inclusion of UV-low pressure synergy (W UV =0.364→ UV =0.381); the pressure term, due to its synergistic effect with temperature and ultraviolet radiation, saw a significant increase in its proportion (W). P =0.091→ P =0.101); the antagonistic effect of the oxygen partial pressure term is through σ T,O The absolute contribution increases after allocation, but the proportion decreases after normalization (|W O |=0.073→| O |=0.063).

[0188] 1.8 Stage 6—Life Prediction Engineering Applications

[0189] Using the modified Arrhenius extended equation:

[0190] ln(t eq,i )=A i +B i / T+C i ·ln(I UV / I UV,0 )+D i ·ln(P / P0)+E i ·ln(O / O ,0 (10)

[0191] In the formula, the subscript i ∈ {T, UV, P, O}; T is the absolute temperature (K); I UV ,0,P0,O,0 represent the annual average total ultraviolet radiation (I0) of the plain reference environment, respectively. UV, 0=4500MJ·m -2 ·a -1 ), atmospheric pressure (P0=101.3kPa) and oxygen partial pressure (O ,0 =21.2 kPa); Reference absolute temperature T0 = 293.15 K (20℃). Various coefficients A i ~E i The single-factor data obtained from Phase 2, combined with eight supplementary densification experiments at the intermediate levels of the five-level coding (-1, 0, +1), were obtained through nonlinear least squares global fitting. The specific values ​​are shown in Table 7.

[0192] Table 7. Fitting coefficients of the modified Arrhenius extended equation in Example 1

[0193]

[0194] Physical meaning of each coefficient: B i =E a,i / R (R is a universal gas constant), typical asphalt aging system B i The value should be in the range of 2000~12000K; C i The exponent characterizing the nonlinear response of ultraviolet photon flux to this term (typically 0.5~1.2); D i The power-law exponent characterizing the effect of pressure on evaporation-diffusion coupling; E i The correction index characterizes the effect of oxygen partial pressure on the order of the carbonylation reaction. All coefficients obtained in this example fall within the typical range reported in the literature, meeting the requirements of physical interpretability.

[0195] The data was taken from the annual average observation series of the Lhasa meteorological station over the past 10 years (2014-2023): annual average temperature 7.6℃ (T=280.75K), annual average total ultraviolet radiation 7200 MJ·m -2 ·a -1 (I) UV / I UV The average annual atmospheric pressure is 65.3 kPa (P / P0 = 0.645), the average annual oxygen partial pressure is 13.7 kPa (O / O,0 = 0.646), and the average daily temperature difference is 15.8℃. Substituting the above parameters into equation (10), the equivalent acceleration factor r of each factor is calculated. i =t eq, i, 0 / t eq, i :

[0196] r T =1.16 (Temperature cycles exacerbate aging, but this is partially offset by lower average annual temperatures)

[0197] r UV =2.20 (UV intensity at high altitude is 1.6 times stronger, superimposed with nonlinear response)

[0198] r P =1.45 (Low pressure accelerates evaporation and mass transport)

[0199] r O =0.55 (Low oxygen partial pressure inhibits carbonylation, antagonistic, r<1)

[0200] According to the modified contribution vector Weighted summation (for negative weight factors) O (Retaining algebraic notation, the physical meaning is that the overall contribution of this factor is to slow down aging.)

[0201] k acc =Σ i i ·r i =0.456×1.16+0.381×2.20+0.101×1.45+(-0.063)×0.55≈1.48 (11)

[0202] Using a 12-year service life as a benchmark for plain asphalt, the equivalent service life of this SBS modified asphalt in the Lhasa section is predicted to be: L pred =L0 / k acc =12 / 1.48≈8.1 years

[0203] Based on this, the following recommendations are made regarding the preventative maintenance time window and structural compensation:

[0204] (1) First preventive maintenance window: 5.8 to 6.2 years (when the equivalent service life reaches 71 to 77%), it is recommended to apply fog seal treatment, with a reference cost of RMB 18 to 25 per m. 2 ;

[0205] (2) Second preventive maintenance window: 7.3 years (when the equivalent service life reaches 90%), it is recommended to apply micro-surfacing, with a reference cost of 35~50 yuan / m. 2 ;

[0206] (3) Structural compensation suggestion: Add a 5-8mm UV sacrificial layer to the upper layer design. It is recommended to use high aromatic content modified asphalt or add 0.3-0.6% hindered amine light stabilizer (HALS).

[0207] Example 2: Comparative Verification and Method Universality of 70# Base Asphalt

[0208] To verify the universality of the method of this invention for different asphalt types, and the ability of the decoupled output contribution vector to distinguish modifiers, this embodiment uses 70# Grade A road petroleum asphalt (penetration 68×0.1mm, softening point 48℃, ductility at 5℃ 18cm), and follows the same specimen preparation, testing equipment, CCD design, significance criterion, and decoupling calculation process as in Example 1, executing the entire process from S1 to S5. The measured results at each stage are as follows.

[0209] 2.1 Measured results of Phase 1 and Phase 2 (0th order + 1st order)

[0210] R0=0.0061±0.0004 (n=6).

[0211] Table 8 Summary of Single-Factor Independent Increments in Stage 2 of Example 2 (70# Base Asphalt)

[0212]

[0213] 2.2 Stage 3 Two-Factor Interaction (Second Order)

[0214] Table 9 Results of the six-group two-factor coupled test in Stage 2 of Example 2 (70# base asphalt)

[0215]

[0216] Final result:

[0217] Σσ ij ·w ij =0.0168+0.0034+(-0.0017)+0.0022×0.5+0+0=+0.0196.

[0218] 2.3 Stage 4CCD Higher-Order Synergy (3rd Order)

[0219] Response surface model determination coefficient R 2 =0.958, lack of fit test p=0.167>0.05, both satisfying the criteria. The measured R0 was obtained when all four factors were applied simultaneously to extreme levels. total =0.0880; Calculate:

[0220] ΣΔ i =+0.0156+0.0142+0.0031+(-0.0018)=+0.0311

[0221] σ global =0.0880-(0.0061+0.0311+0.0196)=+0.0312

[0222] 2.4 Stage 5 Mechanism Closed-Loop Verification

[0223] ε T =-3.5%, ε UV =-6.1%, ε P =-9.3%, ε O =+12.4%, all four terms are |ε i |≤15%; The mean relative error (MRE) of the decoupling reconstruction of 30 CCD factorial points is 7.2%≤8%, which is verified by mechanism closure. This embodiment converges after 4 iterations.

[0224] 2.5 Contribution Vector and Comparative Analysis

[0225] Normalize according to equation (8): Σ|Δ i |=0.0156+0.0142+0.0031+0.0018=0.0347, therefore:

[0226] W 70 #=[+0.450,+0.409,+0.089,-0.052]

[0227] Calculate the corrected contribution vector according to equations (9a) and (9b). 70 #:

[0228] C T =+0.0156+0.5×(0.0168+0.0034+(-0.0017))=+0.02485

[0229] C UV =+0.0142+0.5×(0.0168+0.0022×0.5+0)=+0.02315

[0230] C P =+0.0031+0.5×(0.0034+0.0022×0.5+0)=+0.00535

[0231] C O =-0.0018+0.5×((-0.0017)+0+0)=-0.00265

[0232] Σ|C k |=0.0560

[0233] Therefore, we can conclude that:

[0234] 70 #=[+0.444, +0.413, +0.096, -0.047]

[0235] Table 10 Comparison of decoupling results between SBS modified asphalt and 70# base asphalt

[0236]

[0237] Results Analysis: (1) In terms of absolute degradation increment, the degradation increment of SBS modified asphalt was less than that of 70# base asphalt under all four factors, with the UV term showing the most significant improvement (26.8%), verifying the anti-UV aging mechanism of SBS modifier; (2) In terms of normalized contribution, the proportion of UV term decreased while the proportion of temperature term increased in the SBS system, indicating that temperature factor became the dominant aging mechanism after SBS modification; (3) The absolute value of the antagonistic term, SBS (0.063), was greater than that of 70# (0.047), reflecting that the SBS system is more sensitive to oxygen concentration and has a greater relative benefit in the low-oxygen environment of the plateau. The above comparison results confirm the universality of the method of this invention for different asphalt types, and also prove that the contribution vector of the decoupled output has sufficient engineering discrimination and can be used as a quantitative basis for modifier screening and formulation optimization.

[0238] Example 3: Correlation Verification of Scale Conversion Factor κ and Asphalt Mixture

[0239] To establish a quantitative correlation between the decoupling results of pure asphalt binder (Example 1) and the macroscopic properties of asphalt mixture pavement, this example introduces a scale conversion factor κ. This correlation is achieved by analyzing the results obtained in Example 1. The mixture is weighted by κ with the σ matrix to obtain the equivalent decoupled output of the mixture, which is then verified using independent macroscopic performance test data.

[0240] 3.1 Calculation of parameters and κ of the mixture specimen

[0241] In this embodiment, AC-13 graded Marshall specimens (density 2.456 g·cm³) were prepared using the same batch of SBS modified asphalt as in Example 1. -3 Key parameters were determined according to JTG E20-2025 "Test Procedures for Asphalt and Asphalt Mixtures in Highway Engineering":

[0242] VMA (Void Ratio) = 14.2%;

[0243] K a (Mineral adsorption coefficient, determined by the aggregate specific surface area and bitumen adsorption capacity) = 0.58

[0244] h film (Average asphalt film thickness) = 8.3 μm

[0245] h ref (Reference film thickness) = 10.0 μm

[0246] The formula for calculating the scale conversion factor as defined is as follows:

[0247] κ=(VMA / 100)·(1+K a)·(h film / h ref (12)

[0248] Substitute the data:

[0249] κ=(14.2 / 100)×(1+0.58)×(8.3 / 10.0)=0.142×1.58×0.83=0.1862

[0250] 3.2 Equivalent Decoupling Output of Mixture

[0251] According to W mix =κ·W、σ mix =κ·σ, the result obtained in Example 1 Multiply the σ matrix by κ respectively: mix =κ· SBS =0.1862×[+0.456, +0.381, +0.101, -0.063]=[+0.0849, +0.0709, +0.0188, -0.0117]σ mix (Select σ) T, UV For example): σ T, UV, mix =0.1862×(+0.0121)=+0.00225.

[0252] 3.3 Independent macroscopic performance test verification

[0253] The same four-factor CCD test (30 groups, specimen number × 3, average value of parallel tests) was performed on Marshall specimens with the same gradation as in Example 1 using a Hamburg rutting apparatus (60℃ immersion, 20,000 wheel load cycles, rutting depth increment as response index) and SCB (semi-circular bending) low-temperature fracture toughness test (-10℃, fracture energy as response index). The measured macroscopic contribution vector of macroscopic performance degradation was obtained after normalization and compared with the κ bridging prediction value of the present invention. The results are shown in Table 11.

[0254] Table 11 Comparison of κ bridging predicted values ​​and independent macroscopic test measured values ​​in Example 3

[0255]

[0256] The average relative deviation was 9.9%, which is consistent with the expected engineering target of ≤10%. This verifies that the κ bridging method can effectively establish a quantitative correlation between the decoupling results of micro-cement binder and macro-asphalt mixture, providing a quantitative pathway for the transformation of the method of this invention from the material level to the structural level engineering application.

[0257] Error source analysis and control:

[0258] The potential error sources of the method of this invention mainly include four categories, and the contribution of each type of error and the control measures are as follows:

[0259] (1) Crosstalk of non-target factors inside the cabin, that is, the deviation of non-target factors is suppressed to within ±2% by the sensor array and closed-loop feedback control of the device of the present invention, and the corresponding decoupling error contribution is ≤1.2%;

[0260] (2) FTIR baseline drift and peak area integration error, i.e., averaged by at least 6 specimens per group, 5 measurement points per specimen (30 measurement points in total), and 1460cm -1 The correction for the methyl internal standard reference was controlled to be within 0.8%.

[0261] (3) The subjectivity of ANOVA significance threshold selection, i.e., giving a 95% confidence interval by resampling 5000 times using Bootstrap, and using graded weights (w ij =1.0 / 0.5 / 0) replaces binarization judgment, reducing the subjective influence of a single threshold, with a contribution of ≤1.5%;

[0262] (4) The extrapolation error of the extreme value extrapolation interval of the CCD model is reduced to ≤3.1% by expanding the axis point to the extreme level of each factor and repeating it 6 times with the center point.

[0263] The expected relative error of the overall decoupling and reconstruction after the sum of the above four types of errors (approximate by the square root of the sum of squares) is √(1.2²+0.8²+1.5²+6.5²)=6.8%≤8%, which is consistent with the MRE≤8% convergence threshold set in this invention.

[0264] Other optional implementation methods

[0265] (1) Replacement of deterioration response index: In addition to the FTIR carbonyl index increment ΔCI, the deterioration response index of this invention can be replaced by any index that can quantitatively characterize the degree of asphalt aging, such as the DSR complex modulus aging index CAI, the BBR (bending beam rheometer) creep stiffness modulus S (60s), and the low temperature fracture strain energy release rate. The replacement does not affect the mathematical framework and beneficial effects of the method of this invention.

[0266] (2) Implementation by altitude zone classification, that is, the single-factor extreme level mentioned in step S2 of this invention can be selected in five levels according to the specific target altitude zone - for areas with an altitude of 3000m, P=75kPa (code 0) and O=13.0kPa (code +1) can be set; for areas with an altitude of 4000m, P=65kPa (between code 0 and +1) and O=11.5kPa (between code +1 and +2) can be set; for areas with an altitude of 5000m and above, P=55kPa (code +1) and O=10.0kPa (code +2) can be set; the remaining steps remain unchanged.

[0267] (3) CCD design degradation: When the number of specimens is limited, the CCD design of this invention can be degraded to a Box-Behnken design (BBD), reducing the number of test groups from 30 to 27. Correspondingly, the threshold for the coefficient of determination of the response surface model can be relaxed to R. 2 ≥0.92.

[0268] (4) Material type expansion: In addition to asphalt binders and asphalt mixtures, the method of the present invention can also be extended to the multi-factor decoupling evaluation of other building materials such as cement concrete, waterproof membranes, and road marking paint in plateau environments after appropriate parameter adjustments. Specific adjustments should include changing the temperature level of the standard reference environment (e.g., 20℃ for cement concrete), changing the main response index (e.g., compressive strength retention rate for cement concrete), and refitting the κ factor.

[0269] This invention is analyzed from different feature dimensions: Benchmark subtraction method: In existing technologies, R0 is not explicitly introduced, and the basic background is repeatedly subtracted. In this invention, R0 is explicitly included as a common background benchmark in each order of difference subtraction. Gas pressure control method: In existing technologies, vacuum and gas distribution share a common loop, and a decrease in gas pressure inevitably leads to a decrease in oxygen partial pressure. In this invention, the gas pressure and oxygen partial pressure gas paths are physically independent, and four operating conditions can be switched independently. Independent control of oxygen partial pressure: Existing technologies cannot achieve this. In this invention, nitrogen / oxygen gas distribution is independently controlled, enabling normal pressure-low oxygen operating conditions. Antagonistic term identification: In existing technologies, negative values ​​are discarded as errors. In this invention, negative values ​​are retained after significance testing to identify the antagonistic inhibition mechanism. Mechanism closure: Existing technologies use unidirectional fitting without a verification triggering mechanism. In this invention, activation energy cross-comparison triggers physical experimental iterations, with a maximum of 20 iterations. Decoupling error: Existing technologies have an error greater than 15%, while this invention does not exceed 8%. Applicable scope: Existing technologies only cover conventional plain environments, while this invention covers the entire altitude range of plateaus with gas pressures of 50–105 kPa and oxygen partial pressures of 6.5–22.0 kPa.

[0270] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A decoupling method for multi-factor coupled aging of plateau asphalt materials, characterized by: The method includes the following steps: S1. Absolute reference calibration: Asphalt pavement material specimens are placed in a standard reference environment including temperature T, ultraviolet radiation UV, atmospheric pressure P, and oxygen partial pressure O for aging tests, and the absolute reference deterioration rate R0 is obtained by actual measurement. S2. Single-factor independent incremental extraction: One of the following factors—temperature T, ultraviolet irradiation UV, atmospheric pressure P, and oxygen partial pressure O—is adjusted to its set value, while the remaining factors are locked to the level of step S1. Atmospheric pressure P is controlled via a vacuum pumping circuit, and oxygen partial pressure O is controlled via a nitrogen / oxygen mixing circuit independent of the vacuum pumping circuit. These two gas paths are independent, allowing both the low-pressure-normal oxygen condition (maintaining oxygen partial pressure while only reducing gas pressure) and the normal-pressure-low oxygen condition (maintaining gas pressure while only reducing oxygen partial pressure) to be implemented independently. The four single-factor degradation rates R are measured. i Calculate the independent increment Δ i =R i -R0, i∈{T,UV,P,O}; S3. Two-factor interaction net contribution extraction: Coupled aging tests were conducted on six pairwise combinations of four environmental factors to measure the two-factor degradation rate R. ij Calculate the net interaction contribution σ ij =R ij -(R0+Δ i +Δ j Extract the net contribution of the two-factor interaction, where the subscripts i, j ∈ {T, UV, P, O} and i ≠ j; σ ij Positive values ​​indicate cooperative acceleration interaction, while negative values ​​indicate antagonistic inhibition interaction and are retained. S4. Extraction of high-order synergistic terms across all factors: A fully coupled experiment was conducted using a four-factor, five-level central composite star design to measure the comprehensive degradation response R. total According to formula σ global =R total -(R0+ΣΔ i +Σσ ij Extract higher-order cooperative terms and solve based on the quadratic response surface. R / x i Multidimensional deterioration extreme value coordinates under the condition =0; S5. Mechanism Closed-Loop Verification and Iteration: Verify and iterate each Δ... i Converted to apparent activation energy and compared with independent benchmarks, when the total mechanism deviation rate |ε i When |≤15% and the mean relative error of decoupling and reconstruction (MRE) ≤8%, output the independent contribution vector W and the coupling effect matrix σ; otherwise, output σ. ij Perform a reallocation and return to step S4 to re-execute the central composite star point physical experiment, with no more than 20 iterations.

2. The decoupling method for multi-factor coupled aging of plateau asphalt materials according to claim 1, characterized in that: The standard reference environment in step S1 has a temperature of 25±0.5℃, an air pressure of 101.3±0.3kPa, an ultraviolet radiation intensity of zero, and an oxygen partial pressure of 21.2±0.2kPa.

3. The decoupling method for multi-factor coupled aging of plateau asphalt materials according to claim 1, characterized in that: In step S2, the single factor is adjusted to a set value, which is an extreme level. The extreme level is set as follows: (a) Pure heat group: temperature is -30℃ to +40℃, and the cycle is 12h; (b) Pure UV group: UV irradiance of 300–700 W / m², wavelength range of 280–400 nm, and main peak wavelength of 340 nm; (c) Pure low-pressure group: The cabin pressure is adjusted to 50–70 kPa, and the oxygen partial pressure is compensated to 21.2 ± 0.2 kPa through the nitrogen / oxygen mixing gas distribution circuit. (d) Pure hypoxic group: The cabin pressure was maintained at 101.3±0.3 kPa, and the oxygen partial pressure was adjusted to 7.3-10.5 kPa by nitrogen dilution; The aging time for each single factor is not less than 100 hours, and at least one of the following is used as the characterization index of degradation response: dynamic shear rheological complex modulus aging index CAI and Fourier transform infrared spectral carbonyl index increment ΔCI.

4. The decoupling method for multi-factor coupled aging of plateau asphalt materials according to claim 1, characterized in that: In step S3, the net contribution σ of the two-factor interaction ij After significance testing by analysis of variance, the data were classified and included in subsequent calculations according to the following rules: when the significance level p < 0.05, all data were included with a weight of 1.0; when 0.05 ≤ p < 0.10, some data were included with a weight of 0.

5. When p ≥ 0.10, let σ ij =0 indicates exclusion; the significance test for the variance analysis was performed using Bootstrap resampling 5000 times, and the 95% confidence interval was given using the quantile method; for σ ij Antagonistic inhibition interaction terms with a value less than 0 are retained as negative values ​​and included in the weighting calculation when the significance is true.

5. The decoupling method for multi-factor coupled aging of plateau asphalt materials according to claim 1, characterized in that: In step S4, the four-factor, five-level central composite star point design uses the extreme levels of each factor as the pivot point, employs non-equidistant practical coding, repeats the pivot point at least 6 times, and has a total of at least 30 experimental groups; the five-level coding parameters are: (a) Temperature T corresponds to five levels: -30 / -10 / 10 / 25 / 40℃; (b) Ultraviolet irradiance I UV The corresponding levels are 200 / 300 / 500 / 700 / 800 W / m²; (c) The air pressure P corresponds to five levels of 105 / 95 / 75 / 55 / 45 kPa; (d) The partial pressure of oxygen (O) corresponds to five levels of 22.0 / 19.0 / 16.0 / 13.0 / 10.0 kPa; The fitted quadratic response surface model must simultaneously satisfy the following conditions: coefficient of determination R² ≥ 0.95 and fit test p > 0.

05.

6. The decoupling method for multi-factor coupled aging of plateau asphalt materials according to claim 4, characterized in that: The output independent contribution vector W in step S5 is normalized according to the following formula: W i =Δ i / Σ k |Δ k | where i, k∈{T, UV, P,O}; It further outputs the corrected contribution vector after considering the amortization of the two-factor coupling term. , i =C i / Σ k |C k |, where C i =Δ i +0.5×Σ j≠i σ ij ·w ij w ij The significance level weights in step S3.

7. The decoupling method for multi-factor coupled aging of plateau asphalt materials according to claim 1, characterized in that: It also includes step S6: based on the modified Arrhenius extended equation, combined with the independent contribution vector W or the modified contribution vector In addition to meteorological observation sequences, it outputs equivalent service life prediction curves and preventive maintenance time windows.

8. The decoupling method for multi-factor coupled aging of plateau asphalt materials according to claim 1, characterized in that: A scale conversion factor κ is also introduced to convert the decoupling results of asphalt binder specimens into equivalent decoupling outputs of asphalt mixture specimens, where κ = (VMA / 100)·(1+K) a )·(h film / h ref VMA is the void fraction of the mineral aggregate, K a h is the adsorption coefficient of the mineral. film h represents the average asphalt film thickness. ref The reference film thickness.

9. A decoupling device for implementing the method according to any one of claims 1-8, characterized in that: The system includes a single-factor isolation test chamber, a multi-factor coupled superposition control system, and a degradation feature acquisition and decoupling calculation unit. The data acquisition port of the degradation feature acquisition and decoupling calculation unit is communicatively connected to the degradation response monitoring equipment of the single-factor isolation test chamber, and the degradation feature acquisition and decoupling calculation unit is also communicatively connected to the multi-factor coupled superposition control system. The single-factor isolation test chamber includes a pressure control module and an oxygen partial pressure control module, and their gas paths are independent of each other, enabling the test chamber to independently control the oxygen partial pressure at any pressure level. The degradation feature acquisition and decoupling calculation unit includes a tensor difference subtraction operation module and a quadratic response surface fitting module, and performs subtraction operations of each order using R0 as a common background benchmark.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it can receive data obtained from physical experiments by the decoupling device of claim 9, perform decoupling calculations in the method of any one of claims 1-8, and then send the calculation results back to the control system of the decoupling device to trigger or terminate the iteration of the physical experiment.

Images