Emulsified asphalt demulsification speed grading method

Abstract

The invention relates to the technical field of road engineering material detection, in particular to an emulsified asphalt demulsification speed grading method which comprises the following steps: acquiring torque time sequence data in an emulsified asphalt demulsification process, and drawing a torque curve; the demulsification process is divided into an acceleration period and a stationary period, a torque curve second derivative extreme point serves as an acceleration period end point, linear regression is conducted on data of the two stages, and geometric phase change characteristic parameters are extracted; obtaining a current environment parameter, and matching a grading threshold value database according to the current environment parameter to obtain a threshold value interval of the geometric phase change characteristic parameter of each grade of emulsified asphalt under the current environment parameter; based on the geometric phase change characteristic parameters and the threshold interval, the asphalt type is judged; demulsification process stages are divided through torque curve second derivative extreme points, geometric phase change characteristic parameters are extracted, the subjectivity of manual interpretation is avoided, meanwhile, a grading threshold database is constructed, the grading result is made to accord with the actual construction environment, and an objective and repeatable grading technology system is provided.

Description

Technical Field

[0001] This invention relates to the technical field of road engineering material testing, and in particular to a method for classifying the demulsification rate of emulsified asphalt. Background Technology

[0002] Emulsified asphalt, as an environmentally friendly road construction material, forms a stable emulsion by mixing asphalt and water under the action of an emulsifier. Its demulsification rate directly affects the paving performance and pavement forming quality. The demulsification process is essentially a rheological transformation process in which the emulsion viscosity gradually increases as water evaporates and asphalt particles aggregate. Accurately classifying the demulsification rate (such as fast cracking, medium cracking, and slow cracking) is crucial for optimizing construction technology and selecting materials.

[0003] With the development of intelligent sensing technology, some studies have attempted to collect viscosity change signals during the demulsification process using torque sensors, but existing methods have the following limitations: Feature extraction is crude: only a single parameter (such as maximum torque value, demulsification time) is used for classification, without effectively dividing the stage features of the demulsification process, resulting in insufficient distinguishability between different demulsification types; Lack of geometric phase variable quantification analysis: The morphological changes of the torque curve during demulsification contain rich rheological information, but existing technologies have not converted these geometric features into quantifiable characteristic parameters, making it difficult to construct an objective hierarchical index system; Environmental parameters are not effectively modeled: Demulsification rate is significantly affected by the coupling of environmental parameters, but existing methods have not established a correlation mechanism between environmental parameters and demulsification characteristic parameters, which limits the engineering application scope of the detection methods.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the present invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for classifying the demulsification rate of emulsified asphalt. This method divides the demulsification process into stages by identifying the extreme points of the second derivative of the torque curve, extracts geometric phase transition characteristic parameters to avoid the subjectivity of manual interpretation, and simultaneously constructs a classification threshold database to establish a mapping relationship between the environment and characteristic parameters. This ensures that the classification results conform to the actual construction environment, providing an objective and repeatable classification technology system.

[0006] The present invention provides a method for classifying the demulsification rate of emulsified asphalt, comprising: Collect torque time-series data during the demulsification process of emulsified asphalt and plot the torque curve; The demulsification process is divided into an acceleration phase and a steady phase. The extreme point of the second derivative of the torque curve is taken as the end point of the acceleration phase. Linear regression is performed on the data of the two phases to extract the geometric phase transition characteristic parameters. Obtain the current environmental parameters, match the pre-built graded threshold database with the current environmental parameters, and obtain the threshold range of geometric phase change characteristic parameters of each grade of emulsified asphalt under the current environmental parameters; Asphalt type is determined based on geometric phase transformation characteristic parameters and threshold ranges.

[0007] As a preferred embodiment of the present invention, the geometric phase transition characteristic parameters include: the difference ratio of the slopes of the regression lines during the acceleration period and the stationary period, and the cosine value of the angle between the vectors formed by the inflection point of the maximum slope and the end point of the acceleration period.

[0008] As a preferred embodiment of the present invention, the method for calculating the difference ratio includes: Linear regression was performed on the torque data during the acceleration period to obtain the slope k1, and linear regression was performed on the torque data during the stationary period to obtain the slope k2. The difference ratio was calculated using the following formula: ; Where δ is the difference ratio and φ is a minimum quantity to prevent the denominator from being zero.

[0009] As a preferred embodiment of the present invention, the method for determining the inflection point of the maximum slope includes: Calculate the first derivative sequence dT / dt of the torque time-series data during the acceleration period, identify the maximum value in the first derivative sequence, and determine the value corresponding to (t). p T p ) represents the coordinates of the inflection point with the maximum slope, where t p For time, T p This is the torque value.

[0010] As a preferred embodiment of the present invention, the method for calculating the cosine value includes: The coordinates of the end of the acceleration period are (t) e T e Based on the coordinates of the end of the acceleration period and the coordinates of the inflection point of the maximum slope, the vector is obtained as a = (t e -t p ,T e -T p ); The formula for calculating the cosine value is: cosθ = (a × b) / |a|; Where b is the reference vector, b = (1, 0).

[0011] As a preferred embodiment of the present invention, the method for constructing a hierarchical threshold database includes: Within the preset environmental parameter range, material intrinsic parameters such as asphalt grade, emulsifier parameters, and emulsion solid content are incorporated. A multi-factor coupled test is designed using the Box-Behnken response surface method. Torque data is collected and geometric phase transition characteristic parameters are calculated to construct a multi-dimensional dataset. Engineering sensitivity weights were assigned to the geometric phase transition characteristic parameters, and a benchmark threshold range was defined using a weighted fuzzy C-means clustering algorithm. The threshold calibration model was then established by combining Bayesian inference model with expert experience and field verification data. An incremental learning module is set up to iteratively optimize the threshold range through construction verification samples, thereby completing the dynamic construction of the database.

[0012] As a preferred embodiment of the present invention, when performing linear regression on the two-stage data: The data selection interval for the acceleration period is as follows: calculate the total duration from the start time of data collection to the end time of the acceleration period, and then take the first 80-90% of the total duration; The range of data selected for the stable period is 15-30% from the end of the acceleration period to the total time of demulsification, where the total time of demulsification is from the start time of data collection to the end time when the torque value stabilizes. The least squares method was used to fit the acceleration period data and the stationary period data respectively.

[0013] As a preferred embodiment of the present invention, the acquisition start time point is the moment when the torque value first exceeds the baseline value nσ, where σ is the noise standard deviation of the data acquisition system and n is a preset multiple. The time point at which the torque value stabilizes is the first time point in which the torque change rate is less than 0.05 N·m / s within m consecutive sampling periods, where m is a preset number.

[0014] As a preferred embodiment of the present invention, the sampling frequency of the torque timing data is 10~50Hz.

[0015] As a preferred embodiment of the present invention, the environmental parameters include ambient temperature, humidity and pressure.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1) The present invention divides the demulsification process into an acceleration period and a steady period, and uses the extreme point of the second derivative of the torque curve as the basis for stage division. Linear regression is performed on the data of the two stages respectively to extract geometric phase transition characteristic parameters such as the difference ratio of the regression line slope between the acceleration period and the steady period, and the cosine value of the angle between the vector formed by the inflection point of the maximum slope and the end point of the acceleration period. From the perspectives of the rate of change in the time dimension and the characteristics of the curve shape in the spatial dimension, the demulsification process is finely characterized. Compared with a single parameter, it can more comprehensively and accurately reflect the essential differences of different demulsification types and improve the ability to distinguish between different demulsification speed types such as fast cracking, medium cracking, and slow cracking. 2) This invention quantifies the abrupt change in the demulsification viscosity growth rate by using the difference ratio, captures the directional change of key turning points in the torque curve using the cosine value, and transforms the torque curve shape into quantifiable geometric phase transition characteristic parameters. At the same time, it optimizes the data intervals of the acceleration period and the steady period and uses the least squares method for fitting to improve the accuracy of the characteristic parameters. On this basis, it incorporates the coupling model of intrinsic material parameters and environmental parameters, defines the benchmark threshold by weighted fuzzy C-means clustering and corrects it by combining it with a Bayesian inference model, and realizes dynamic optimization of the database by using an incremental learning module. It constructs an objective and scientific classification index system that takes into account the coupling influence of multiple parameters, and provides a more reliable quantitative standard for the classification of demulsification rate of emulsified asphalt that is more in line with engineering practice. Attached Figure Description

[0017] Figure 1 This is a schematic flowchart of the method for classifying the demulsification rate of emulsified asphalt according to the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0019] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0020] Secondly, the term "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0021] Reference Figure 1 This embodiment provides a method for classifying the demulsification rate of emulsified asphalt, including: S1 collects torque time-series data during the demulsification process of emulsified asphalt and plots torque curves; more specifically, a high-precision torque sensor is used to monitor the change in stirring resistance during the demulsification process of emulsified asphalt in real time, and the collected time-series data is transmitted to the data processing system. After filtering and noise reduction of the raw data, a continuous curve is plotted with time as the horizontal axis and torque value as the vertical axis to obtain the torque curve. S2 divides the demulsification process into an acceleration phase and a steady phase. The extreme point of the second derivative of the torque curve is taken as the end point of the acceleration phase. Linear regression is performed on the data of the two phases to extract the geometric phase transition characteristic parameters. More specifically, during the demulsification process of emulsified asphalt, the change in the torque curve reflects the transformation of the internal structure of the emulsion. During the acceleration period, as water evaporates and asphalt particles aggregate, the viscosity of the emulsion increases rapidly, and the torque rise rate gradually accelerates. This is manifested by the increasing first derivative of the torque curve and the second derivative being greater than zero. When a certain point is reached, the changes in the internal structure of the emulsion tend to stabilize, and the torque rise rate begins to slow down. At this time, the second derivative of the torque curve shows an extreme point. This extreme point marks the transition of the emulsion from the acceleration period of rapid structural change to a relatively stable plateau period. Therefore, taking this as the end point of the acceleration period can accurately divide the different stages of the demulsification process, providing a reasonable stage division basis for subsequent linear regression and extraction of geometric phase transition characteristic parameters. This helps to more accurately describe and analyze the rheological characteristics of the demulsification process. The above piecewise linearization process not only simplifies the complex nonlinear demulsification process into quantifiable geometric features, but also provides a mathematical basis for constructing an objective hierarchical index system by comparing the changing trends at different stages, effectively solving the problem that traditional methods cannot accurately capture the stage characteristics of the demulsification process. S3 acquires the current environmental parameters and matches them with a pre-built grading threshold database to obtain the threshold ranges of geometric phase change characteristic parameters of each grade of emulsified asphalt under the current environmental parameters. The purpose of this step is to acquire the current environmental parameters and match them with the pre-built grading threshold database to provide a quantitative basis for determining the demulsification level of emulsified asphalt. Environmental parameters affect the demulsification process and related characteristic parameters of emulsified asphalt. By establishing a grading threshold database, the threshold ranges of geometric phase change characteristic parameters of each grade of emulsified asphalt under different environmental parameters are clearly defined. This allows for accurate determination of the corresponding judgment criteria based on the current actual environmental conditions. In practical applications, these precise threshold ranges can be used to objectively and accurately classify the demulsification level of emulsified asphalt, which helps to improve the scientificity and reliability of the evaluation of the demulsification performance of emulsified asphalt. S4 determines the asphalt type based on geometric phase transition characteristic parameters and threshold ranges; More specifically, this step compares and analyzes the geometric phase change characteristic parameters obtained from actual measurements with the corresponding threshold ranges. If the geometric phase change characteristic parameters fall within the threshold range of a certain grade of emulsified asphalt, it can be determined that the emulsified asphalt belongs to this type.

[0022] This invention transforms demulsification process data collected by a high-precision torque sensor into a torque curve. Combined with a stage division based on the extreme points of the second derivative, it accurately captures the acceleration and plateau characteristics of emulsion viscosity changes. Geometric phase transition characteristic parameters are then extracted through linear regression, transforming the complex demulsification rheological process into quantifiable indicators. Simultaneously, by constructing a tiered threshold database under multiple environmental parameters, the influence of environmental factors on demulsification is transformed into specific threshold ranges, linking with the geometric phase transition characteristic parameters. Finally, by comparing actual parameters with the threshold ranges under corresponding environments, accurate determination of asphalt type is achieved. This method avoids the subjectivity of traditional manual observation and solves the problem that single-parameter analysis cannot reflect the essence of the demulsification process. It improves the accuracy, environmental adaptability, and engineering practicality of emulsified asphalt demulsification rate grading, providing a reliable quantitative basis for construction material selection and process optimization.

[0023] In some embodiments of the present invention, the geometric phase transition characteristic parameters include: the difference ratio of the slopes of the regression lines during the acceleration phase and the stationary phase, and the cosine value of the angle between the vectors formed by the inflection point of the maximum slope and the end point of the acceleration phase. The difference ratio, by quantifying the relative change in the slope of the two stages, directly reflects the degree of abrupt change in the viscosity growth rate during demulsification, highlighting the essential differences in the structural formation rate of different types of asphalt. Meanwhile, the cosine value of the vector angle captures the directional change of the torque curve at key inflection points from a geometric perspective, quantifying the severity of viscosity change and the speed of structural transformation during demulsification. The combination of the two not only covers the temporal dimension characteristics of the demulsification process but also captures the spatial dimension characteristics of the curve shape, forming a complementary and independent parameter system. This enhances the distinguishability of different demulsification speed types while avoiding the limitations of a single parameter being affected by environmental fluctuations, providing a core basis for constructing an objective and universally applicable grading index system.

[0024] In some embodiments of the present invention, the method for calculating the difference ratio includes: Linear regression was performed on the torque data during the acceleration period to obtain the slope k1, and linear regression was performed on the torque data during the stationary period to obtain the slope k2. The difference ratio was calculated using the following formula: ; Where δ is the difference ratio, and φ is a minimum quantity to prevent the denominator from being zero; |k1-k2| represents the absolute difference between the slope k1 of the acceleration period and the slope k2 of the stationary period. It directly reflects the change in the viscosity growth rate of the emulsion in the two stages of the demulsification process. The larger the difference, the more significant the change in viscosity growth rate from the acceleration period to the stationary period. This means taking the larger of the absolute values ​​of k1 and k2, and adding a small amount φ. On the one hand, this avoids the calculation error of the denominator being zero when both k1 and k2 are zero. On the other hand, it normalizes the difference in the numerator, eliminates the influence of the slope dimension, and makes the difference ratio a dimensionless relative value, which is more convenient for comparison under different conditions. The difference ratio is calculated using the above formula to quantify the degree of slope difference between the acceleration period and the steady period, accurately reflecting the rate change characteristics of the demulsification process. If the difference ratio approaches 1, it indicates that the slope difference between the two stages is significant (e.g., for fast-cracking asphalt, the viscosity increases sharply during the acceleration period and the growth rate drops sharply during the steady period); if the difference ratio approaches 0, the changes between the two stages tend to be slower (e.g., for slow-cracking asphalt).

[0025] In some embodiments of the present invention, the method for determining the inflection point of maximum slope includes: Calculate the first derivative sequence dT / dt of the torque time-series data during the acceleration period, identify the maximum value in the first derivative sequence, and determine the value corresponding to (t). p T p ) represents the coordinates of the inflection point with the maximum slope, where t p For time, T p This is the torque value; More specifically, for the torque time-series data during the acceleration period, a numerical differentiation method (such as the central difference method) is used to calculate its first derivative sequence dT / dt. This sequence characterizes the rate of change of torque with time during the acceleration period. Then, the sequence is traversed to find the maximum value. The corresponding time and torque value constitute the coordinates of the inflection point of the maximum slope (t). p T p The maximum slope inflection point is the moment when the torque change rate is fastest during the demulsification acceleration period, marking the peak of the viscosity increase rate caused by asphalt particle aggregation. Its coordinates provide key data for subsequent calculation of geometric parameters such as the cosine value of the vector angle. It accurately depicts the key inflection characteristics of the demulsification process from the perspective of curve morphology, reflects the intensity of the dynamic changes in demulsification, and provides an important basis for distinguishing different demulsification rate types such as fast cracking, medium cracking, and slow cracking. It effectively makes up for the shortcomings of existing technologies in not fully utilizing the geometric characteristics of the demulsification process, and improves the accuracy of classification and the depth of mining of demulsification rheological information.

[0026] In some embodiments of the present invention, the method for calculating the cosine value includes: The coordinates of the end of the acceleration period are (t) e T e Based on the coordinates of the end of the acceleration period and the coordinates of the inflection point of the maximum slope, the vector is obtained as a = (t e -t p ,T e -T p ); The formula for calculating the cosine value is: cosθ = (a × b) / |a|; Where b is the reference vector, b = (1, 0).

[0027] Using the extreme point of the second derivative of the torque curve as the end of the acceleration period, the coordinates of the maximum slope inflection point are determined according to the above method. Vector a is calculated based on the coordinates of these two points, with its abscissa being the difference between the end time of the acceleration period and the maximum slope inflection time (t). e -t p The vertical axis represents the difference in torque values ​​(T). e -T p The reference vector b is set as a horizontal unit vector (1,0) to the right. The cosine of the angle between a and b is calculated using the vector dot product formula. The numerator is the dot product of vectors a and b, and the denominator is the magnitude of vector a. Thus, the dimensionless value cosθ, which reflects the degree of the angle between vector a and the horizontal direction, is obtained. The cosine value is selected at the end of the acceleration period (the extreme point of the second derivative) and the turning point of the maximum slope (the maximum point of the first derivative). The purpose is to accurately capture the key features of the torque curve shape change by quantifying the geometric spatial relationship between the peak of the growth rate and the end of the acceleration phase during the demulsification process. The former reflects the most intense activation state of asphalt particle aggregation, while the latter marks the critical boundary of the transition from rapid acceleration to stability during demulsification. The cosine value of the angle between the vector formed by the two points and the horizontal reference vector can transform the steepness of the curve into a dimensionless quantitative index, which complements the difference ratio. A multi-feature system is constructed from the dimensions of dynamic turning point within the stage and rate difference between stages, respectively. The closer cosθ is to 1, the more horizontal the line connecting the two points is, reflecting a smoother transition in torque growth rate during the demulsification acceleration period; conversely, the smaller cosθ is, the steeper the line is, indicating a rapid change in torque growth rate in a short period of time.

[0028] In some embodiments of the present invention, the method for constructing a hierarchical threshold database includes: Within the preset environmental parameter range, S1 incorporates the intrinsic material parameters of asphalt grade, emulsifier parameters, and emulsion solid content. A multi-factor coupled test is designed using the Box-Behnken response surface method. Torque data is collected and geometric phase transition characteristic parameters are calculated to construct a multi-dimensional dataset. The specific process is as follows: An environmental parameter range of 10~50℃, relative humidity 30~90%, and environmental pressure 90~110kPa is preset. Simultaneously, intrinsic material parameters such as asphalt grade, emulsifier type and dosage, and emulsion solid content are included. A multi-factor coupled environmental-material test matrix is ​​designed using the Box-Behnken response surface methodology, covering emulsified asphalt samples at standard demulsification levels of fast, medium, and slow cracking. Torque time-series data for the entire demulsification process are collected for each test group sample at a sampling frequency of 10~50Hz. After filtering and noise reduction, geometric phase transition characteristic parameters are calculated. Environmental parameters, intrinsic material parameters, geometric phase transition characteristic parameters, and corresponding demulsification level labels are associated to construct a standardized multi-dimensional dataset, eliminating dimensional differences between different parameters. S2 assigns engineering sensitivity weights to the geometric phase transition characteristic parameters, uses a weighted fuzzy C-means clustering algorithm to define the baseline threshold range, and establishes an environment-material coupled threshold calibration model by combining Bayesian inference model with expert experience and field verification data. The specific process is as follows: Based on the engineering characteristics of emulsified asphalt demulsification, engineering sensitivity weights are assigned to the geometric phase change characteristic parameters. Using the weighted geometric phase change characteristic parameters as input, a weighted fuzzy C-means clustering algorithm is used to perform cluster analysis on the multi-dimensional dataset to define the benchmark threshold range for each demulsification level. Then, through a Bayesian inference model, the grading experience of road engineering experts is used as the prior probability, and the on-site paving construction performance verification data is used as the posterior probability to quantitatively correct the benchmark threshold range. Finally, with environmental parameters and intrinsic material parameters as independent variables and the upper and lower limits of the geometric phase change characteristic parameter thresholds as dependent variables, an environment-material coupled threshold calibration model is constructed to realize the threshold interpolation calculation for external working conditions of the test matrix. S3 is equipped with an incremental learning module, which iteratively optimizes the threshold range through construction verification samples to complete the dynamic construction of the database; The specific process is as follows: An incremental learning module is configured for the grading threshold database. The module has built-in effective sample screening rules, which only include emulsified asphalt test samples that are complete, verified by on-site construction performance, and have no outliers. At the same time, an iteration trigger condition is set. When the cumulative number of newly added effective samples reaches a preset value or the matching degree between the database grading results and the on-site verification results is lower than a preset threshold, the module automatically includes the newly added samples into the multi-dimensional dataset, re-executes the weighted fuzzy C-means clustering analysis, and updates the threshold range of each demulsification level and the fitting parameters of the environment-material coupling threshold calibration model. Through continuous iteration of on-site verification samples, the dynamic construction and optimization of the database are completed, thereby improving the engineering adaptability and grading accuracy of the database.

[0029] In some embodiments of the present invention, when performing linear regression on the two-stage data: The data selection interval for the acceleration period is as follows: calculate the total duration from the start time of data collection to the end time of the acceleration period, and then take the first 80-90% of the total duration. The acceleration period is the stage in which torque rises rapidly during the demulsification process, but near the end point, the slope may fluctuate due to the emulsion starting to stabilize. Selecting the first 80-90% of the duration can cover the pure acceleration stage where the linear increase in torque is most significant, while also retaining enough data for reliable linear regression. For example, if the total duration of the acceleration period is 100 seconds, then take the data from the first 80-90 seconds and exclude the slope decay that may occur in the last 10-20 seconds due to particle aggregation approaching saturation. The selected range for the steady-state period data is 15-30% from the end of the acceleration period to the total demulsification time, where the total demulsification time is from the start time of data collection to the end time when the torque value stabilizes. The steady-state period is the stage where the torque growth tends to be slow, the emulsion particle aggregation is basically completed, and the viscosity is close to stable. Selecting 15-30% of the time after the end of the acceleration period can avoid the short-term fluctuations that may remain at the end of the acceleration period, while capturing the linear growth characteristics of the steady-state phase. The least squares method was used to fit the acceleration and stationary data respectively. The least squares method finds the optimal linear fitting equation by minimizing the sum of squares of the vertical distances between the data points and the fitted line. In the acceleration and stationary data, this method can filter out high-frequency noise and highlight the trend characteristics of torque changes over time.

[0030] In some embodiments of the present invention, the data acquisition start point is the moment when the torque value first exceeds the baseline value nσ, where σ is the noise standard deviation of the data acquisition system and n is a preset multiple. By analyzing the baseline value and noise standard deviation of the initial data, an adaptive threshold is dynamically set to accurately filter environmental noise and sensor interference, ensuring that effective data acquisition is only initiated when the actual torque change signal caused by demulsification occurs. This avoids the subjectivity of manually setting the threshold and enables the start point determination to automatically adapt to the noise level under different working conditions, providing a reliable time reference for the subsequent stage division of the demulsification process and the calculation of geometric phase transition characteristic parameters. The termination point for torque value stabilization is the first time point within m consecutive sampling periods where the torque change rate is less than 0.05 N·m / s, where m is a preset number. The above method can accurately capture the natural end time of the demulsification process, ensuring that the collected data completely covers the entire process from the start of demulsification to structural stabilization. It provides a reliable time boundary for the scientific division of the steady-state data interval, enabling the steady-state slope calculation to truly reflect the rheological characteristics of the emulsion after stabilization, and avoiding the introduction of invalid noise or omission of key data due to the deviation of the termination point.

[0031] In some embodiments of the present invention, the sampling frequency of torque timing data is 10~50Hz. This frequency range ensures that 10~50 torque data points are collected per second during the demulsification process, which can capture the details of the rapid torque changes during the acceleration period, while avoiding high-frequency noise or computational redundancy introduced by excessively high frequencies.

[0032] In some embodiments of the present invention, environmental parameters include ambient temperature, humidity, and pressure. Increased temperature accelerates the evaporation of water in the emulsion and the aggregation of asphalt particles, increasing the slope of torque rise during the demulsification acceleration period and shortening the time from demulsification initiation to the stable period. Increased humidity may slow down the failure rate of the emulsifier due to the increased water vapor content in the environment, resulting in a slower change in torque during the stable period and prolonging the demulsification completion time. Pressure changes indirectly alter the dynamic process of asphalt particle collision and aggregation by affecting the phase equilibrium of the emulsion system. For example, a high-pressure environment may inhibit the expansion of microbubbles and slow down the formation rate of the demulsification structure. Under the combined effect of these three factors, the torque time-series curve characteristics of the demulsification process will exhibit regular changes. Therefore, incorporating these factors into the classification system allows for the capture of the mapping relationship between environmental variables and demulsification characteristics through orthogonal experimental modeling, solving the problem of poor environmental adaptability in the prior art. This makes the classification results more consistent with the demulsification behavior under actual working conditions, providing accurate environmentally sensitive detection basis for engineering applications.

[0033] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for classifying the demulsification rate of emulsified asphalt, characterized in that, include: Collect torque time-series data during the demulsification process of emulsified asphalt and plot the torque curve; The demulsification process is divided into an acceleration phase and a steady phase. The extreme point of the second derivative of the torque curve is taken as the end point of the acceleration phase. Linear regression is performed on the data of the two phases to extract the geometric phase transition characteristic parameters. Obtain the current environmental parameters, and match the current environmental parameters with a pre-constructed graded threshold database to obtain the threshold range of the geometric phase change characteristic parameters of each grade of emulsified asphalt under the current environmental parameters; The asphalt type is determined based on the geometric phase transition characteristic parameters and the threshold range.

2. The method for classifying the demulsification rate of emulsified asphalt as described in claim 1, characterized in that, The geometric phase transition characteristic parameters include: the difference ratio of the slopes of the regression lines during the acceleration period and the stationary period, and the cosine value of the angle between the vector formed by the inflection point of the maximum slope and the end point of the acceleration period.

3. The method for classifying the demulsification rate of emulsified asphalt as described in claim 2, characterized in that, The method for calculating the difference ratio includes: Linear regression was performed on the acceleration period torque data to obtain the slope k1, and linear regression was performed on the stationary period torque data to obtain the slope k2. The difference ratio was calculated using the following formula: ； Where δ is the difference ratio and φ is a minimum quantity to prevent the denominator from being zero.

4. The method for classifying the demulsification rate of emulsified asphalt as described in claim 2, characterized in that, The method for determining the inflection point of maximum slope includes: Calculate the first derivative sequence dT / dt of the torque time-series data during the acceleration period, identify the maximum value in the first derivative sequence, and determine the value corresponding to (t). p T p ) represents the coordinates of the inflection point with the maximum slope, where t p For time, T p This is the torque value.

5. The method for classifying the demulsification rate of emulsified asphalt as described in claim 4, characterized in that, The method for calculating the cosine value includes: The coordinates of the acceleration period endpoint are (t) e T e Based on the coordinates of the acceleration period endpoint and the coordinates of the maximum slope inflection point, the vector is obtained as a = (t e -t p ,T e -T p ); The formula for calculating the cosine value is: cosθ = (a × b) / |a|; Where b is the reference vector, b = (1, 0).

6. The method for classifying the demulsification rate of emulsified asphalt as described in claim 1, characterized in that, The method for constructing the hierarchical threshold database includes: Within the preset environmental parameter range, material intrinsic parameters such as asphalt grade, emulsifier parameters, and emulsion solid content are incorporated. A multi-factor coupled test is designed using the Box-Behnken response surface method. Torque data is collected and geometric phase transition characteristic parameters are calculated to construct a multi-dimensional dataset. Engineering sensitivity weights were assigned to the geometric phase transition characteristic parameters, and a benchmark threshold range was defined using a weighted fuzzy C-means clustering algorithm. The threshold calibration model was then established by combining Bayesian inference model with expert experience and field verification data. An incremental learning module is set up to iteratively optimize the threshold range through construction verification samples, thereby completing the dynamic construction of the database.

7. The method for classifying the demulsification rate of emulsified asphalt as described in claim 3, characterized in that, When performing linear regression on the two-stage data respectively: The selected interval for the acceleration period data is as follows: calculate the total duration from the start time of data collection to the end time of the acceleration period, and then take the first 80-90% of the total duration; The range selected for the steady-state period data is 15-30% of the total demulsification time from the end of the acceleration period to the end of the total demulsification time, where the total demulsification time is from the start time of data collection to the end time when the torque value stabilizes. The least squares method was used to fit the acceleration period data and the stationary period data, respectively.

8. The method for classifying the demulsification rate of emulsified asphalt as described in claim 7, characterized in that, The acquisition start time point is the moment when the torque value first exceeds the baseline value nσ, where σ is the noise standard deviation of the data acquisition system and n is a preset multiple; The time point at which the torque value stabilizes is the first time point in which the torque change rate is less than 0.05 N·m / s within m consecutive sampling periods, where m is a preset number.

9. The method for classifying the demulsification rate of emulsified asphalt as described in claim 1, characterized in that, The sampling frequency of the torque timing data is 10~50Hz.

10. The method for classifying the demulsification rate of emulsified asphalt as described in claim 6, characterized in that, The environmental parameters include ambient temperature, humidity, and pressure.

Images