Trenchless repair material service life prediction method and system based on big data

Abstract

This invention relates to the field of underground pipeline technology, specifically to a method and system for predicting the service life of trenchless repair materials based on big data. The method identifies fluctuation peaks and starting points from comparative graphs to extract key parameters characterizing the material's aging trend. Based on the distribution characteristics of these key parameters within the test cycle, it determines service life determination nodes. Utilizing the mapping relationship between service life determination nodes and the test cycle, it outputs the predicted service life of the trenchless repair material. This invention uses feature nodes in the coordinate system to calculate delay times, calibrates node positions, and constructs a verification set composed of reference nodes. Abnormal data exceeding the comparison standard is replaced or retained, thus correcting measurement deviations and filtering abnormal data, ensuring a high degree of consistency in the test data used for service life calculation.

Description

Technical Field

[0001] This invention belongs to the field of underground pipeline technology, specifically relating to a method and system for predicting the service life of trenchless repair materials based on big data. Background Technology

[0002] Traditional open-cut repair methods have long construction cycles, significant traffic disruptions, and environmental damage, posing serious drawbacks. In contrast, trenchless repair technologies offer advantages such as high efficiency, environmental friendliness, and minimal impact on ground activities, making them the mainstream solution in pipeline repair. In trenchless repair projects, the performance of composite repair materials such as epoxy resin, vinyl ester resin, modified polyurethane elastomers, and matching curing agents directly determines the structural strength, corrosion resistance, and overall service life of the repaired pipeline.

[0003] However, existing methods for predicting the service life of materials rely heavily on historical experimental data and static empirical formulas. These data and formulas are often obtained in idealized laboratory environments, which cannot accurately reflect the complex acid and alkali corrosion, alternating loads, and temperature and humidity fluctuations of underground pipe networks, resulting in insufficient accuracy in predicting the service life of materials.

[0004] To address the above problems, this invention proposes a method and system for predicting the service life of trenchless repair materials based on big data. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, the purpose of this invention is to provide a method and system for predicting the service life of trenchless repair materials based on big data.

[0006] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: A big data-based method for predicting the service life of trenchless repair materials includes: Acquire material testing information of trenchless repair materials; respond to the fluctuation state of the raw data in the material testing information meeting the preset feature extraction conditions, determine the comparison benchmark based on the usage parameters in the raw data, and generate a comparison graph reflecting the evolution of material performance in combination with the comparison benchmark; The peak values ​​and corresponding starting points of fluctuations are identified from the comparison graphs to extract important parameters characterizing the aging trend of materials; the lifespan determination nodes are determined based on the distribution characteristics of important parameters within the test cycle; and the predicted service life of trenchless repair materials is output by utilizing the mapping relationship between the lifespan determination nodes and the test cycle.

[0007] Preferably, identifying fluctuation peaks and corresponding starting points from the comparison graphs to extract key parameters characterizing the material aging trend includes: The comparison graph is mapped to a preset coordinate system, and feature nodes in the comparison graph whose slope change rate exceeds a preset threshold are identified; the node with the largest amplitude among the feature nodes is determined as the fluctuation peak, and the fluctuation peak is traced back to the turning point where the slope tends to be stable, as the starting point; Furthermore, the region between the starting point and the peak of the fluctuation is defined as the key feature node region in order to extract important parameters.

[0008] Preferably, the lifespan determination node is determined based on the distribution characteristics of key parameters during the test cycle, including: The starting point of the test cycle is determined based on the starting point, and the ending point is determined based on the time point when the important parameters return to the preset stable range. The time span between the starting point and the ending point is calculated to obtain the test cycle. Furthermore, during the testing period, the lifetime determination node is determined based on the intersection of the evolution rate of key parameters and the preset lifetime termination threshold.

[0009] Preferably, determining the lifetime determination node based on the distribution characteristics of key parameters during the test cycle further includes: The delay time of key parameters under different test environments is obtained, and the cutoff node is corrected based on the delay time to obtain the calibration calculation result; the position of the life determination node is compensated using the calibration calculation result.

[0010] Preferably, after outputting the predicted service life of the trenchless repair material, the method further includes: Obtain a validation set of known lifetime samples and extract reference nodes from the validation set; compare the lifetime determination nodes with the reference nodes to evaluate the deviation value of the predicted lifetime. And when the deviation value exceeds the preset range, adjust the calculation weight of important parameters in the process of determining the life determination node.

[0011] This invention also discloses a trenchless repair material service life prediction system based on big data, comprising: The information acquisition module is used to acquire material test information of trenchless repair materials and extract raw data containing usage parameters from it. The feature construction module is used to respond to the acquisition results of the information acquisition module, determine the comparison benchmark based on the dosage parameters, and generate a comparison graph that reflects the evolution of material properties. The node decision module is used to identify fluctuation peaks and starting points from the comparison graphs to extract important parameters characterizing the aging trend of materials, and to determine the life determination node based on the distribution characteristics of important parameters during the test cycle. The life prediction module is used to execute life prediction logic based on the life determination node to output the predicted life of trenchless repair materials. It also includes a feedback calibration module, which is used to evaluate and optimize the output of the lifetime prediction module by comparing the lifetime determination node with the reference node in the calibration set.

[0012] Preferably, the node decision module is configured to: determine the start and end nodes of the test cycle, and calculate the evolution rate of important parameters within the test cycle; In addition, by combining the evolution rate with the preset material failure criteria, the coordinates of the lifetime determination node on the time axis are located.

[0013] Preferably, when generating comparison graphs, the feature construction module is specifically used to: establish a differential mapping relationship between the original data and the comparison benchmark; Furthermore, the differential mapping relationship is visualized in a coordinate system to form a continuous curve characterizing the material properties as a function of test time.

[0014] Preferably, identifying fluctuation peaks and corresponding starting points from the comparison graphs to extract key parameters characterizing the material aging trend includes: The comparison graph is mapped to a preset coordinate system, and feature nodes in the comparison graph whose slope change rate exceeds a preset threshold are identified; the node with the largest amplitude among the feature nodes is determined as the fluctuation peak, and the fluctuation peak is traced back to the turning point where the slope tends to be stable, as the starting point; Furthermore, the region between the starting point and the peak of the fluctuation is defined as the key feature node region in order to extract important parameters.

[0015] Preferably, the feedback calibration module is configured as follows: Obtain a validation set of known lifetime samples and extract reference nodes from the validation set; compare the lifetime determination nodes with the reference nodes to evaluate the deviation value of the predicted lifetime. And when the deviation value exceeds the preset range, adjust the calculation weight of important parameters in the process of determining the life determination node.

[0016] Beneficial effects

[0017] 1. This invention obtains the dosage parameters, evolution rate, and offset parameters of trenchless repair materials and converts them into graphical comparison benchmarks and comparison graphs to identify fluctuation peaks and their starting points, so as to lock in important parameters. This enables the extraction of key features affecting material performance from complex raw data, solving the prediction bias problem caused by failure to reflect parameter correlation or omission of key indicators during the analysis process.

[0018] 2. This invention utilizes feature nodes in the coordinate system to calculate the delay time, calibrates and measures the node positions, and constructs a verification set composed of reference nodes. Abnormal data that exceeds the comparison standard is replaced or retained, thereby correcting measurement deviations and filtering abnormal data, ensuring that the test data involved in the life calculation have a high degree of consistency.

[0019] 3. This invention optimizes the calculation weights in real time based on the fluctuation range of important parameters and combines the prediction results with the life determination node, so that the prediction model can keenly capture the performance degradation trend of trenchless repair materials in actual service, thus making the final output predicted service life more consistent with the actual wear and tear of the material. Attached Figure Description

[0020] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a system module diagram of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely for explaining the invention and are not intended to limit the scope of protection of the invention.

[0022] Example 1

[0023] Please see Figure 1 As shown, this embodiment provides a method for predicting the service life of trenchless repair materials based on big data. The method includes the following steps: Acquire material testing information related to trenchless repair materials and use it as raw data. This raw data includes dosage parameters reflecting the proportion of material components, evolution rate parameters reflecting the speed of material reaction or degradation, and offset parameters reflecting changes in the physical morphology of the material.

[0024] Trenchless repair materials are mainly composed of epoxy resin, vinyl ester resin, modified polyurethane elastomer, and curing agent. Epoxy resin serves as the main substrate to provide structural strength, vinyl ester resin is used to enhance chemical corrosion resistance, modified polyurethane elastomer is used to improve the material's flexibility and impact resistance, and curing agent is used to control the rate and extent of the crosslinking reaction.

[0025] Furthermore, the dosage parameters are a set of numerical values ​​reflecting the mass or volume percentage of each component in the trenchless repair material, used to characterize the initial proportioning state of the composite material. Specifically, the dosage parameters cover the dosage of epoxy resin, vinyl ester resin, modified polyurethane elastomer, and curing agent.

[0026] The raw data is standardized and the processed numerical sequence is mapped into a graphical format as a comparison benchmark. In other words, the standardized material test information is mapped to a reference trajectory in the coordinate system, which serves as an idealized standard for evaluating the evolution of material properties.

[0027] Furthermore, the offset parameter is a metric that characterizes the deviation of the material from a preset standard position in spatial dimension or geometry, and is used to measure the tightness of the bond between the material and the pipe wall or the uniformity of thickness.

[0028] Rate parameters are physical quantities used to describe how quickly a material releases energy during the curing process or how fast a chemical medium penetrates it in service environments.

[0029] Simultaneously, the raw data of trenchless repair materials under actual operation or simulated normal use conditions are collected and compared, and the results are output in graphic format as comparison graphics. These are real-time performance curves generated based on the raw data collected under actual working conditions or simulated environments, used for geometric morphological difference analysis with the comparison benchmark.

[0030] In practical applications, the comparison benchmark and the comparison graph are represented by coordinate axes. The horizontal axis of the coordinate system represents the acquisition time or sample sequence of the original data, and the vertical axis represents the range of data values. The comparison benchmark and the comparison graph form two curves reflecting the performance trend in the coordinate system. By analyzing the geometric shape of the curves, the intersection area of ​​the two curves in the evolution process is identified, and the position where the rate of curvature change near the intersection point exceeds a preset threshold is set as a feature node, representing the key moment when the material's physical properties undergo a stage transformation. The preset threshold is preferably 0.5.

[0031] The location of curvature changes reflects the abrupt change in material properties under specific environments or time points, such as the transition from the elastic stage to the plastic deformation stage. From multiple characteristic nodes, the point with the largest value amplitude is selected as the key characteristic node, representing the extreme state of material property fluctuations.

[0032] Specifically, key feature nodes are the coordinate points where the vertical axis value reaches its extreme value among multiple feature nodes, used to pinpoint the most vulnerable state of material property fluctuations.

[0033] By performing morphological comparisons of the comparative graphs and the comparison benchmark, the peak values ​​of the fluctuations in the waveforms are identified, which are the maximum points where the data deviates from the baseline and the local derivative is zero, reflecting the instantaneous intensity of performance fluctuations.

[0034] Determine the starting point corresponding to each fluctuation peak, and set the original data value corresponding to the starting point as an important parameter as a quantitative indicator for evaluating the degree of material damage.

[0035] In the above process, the time difference between the key feature nodes and the standard node time determined under the laboratory standard environment is identified, and this time difference is used as the delay time to quantify the hysteresis effect of the actual working conditions on the material performance response.

[0036] The delay time reflects the difference in response lag or performance degradation rate between actual working conditions such as complex temperature and humidity environments of underground pipe networks and standard laboratory environments. The delay time is used to perform translation correction analysis on other feature nodes to obtain the corresponding calibration calculation results, that is, the corrected data sequence obtained after time axis translation compensation of feature nodes using the delay time.

[0037] The calibration calculation specifically includes: using the delay time as the compensation amount, calculating the corrected position of each feature node on the time axis to ensure that all feature data are comparable under a unified time reference, performing feedback processing on the obtained correction data, and outputting the adjusted analysis sequence, thereby eliminating the impact of systematic measurement errors caused by environmental differences on lifetime prediction.

[0038] The process involves classifying and organizing multiple important parameters of the same category. This includes determining the physical attribute category of the parameter, aggregating similar data, selecting a preset benchmark or historical average value as a reference node in each group of data, and pairing each important parameter in the group with its reference node to form a validation set for performing data consistency checks and outlier removal.

[0039] The deviation value is obtained by subtracting the reference node value from the value of each important parameter in the verification set. This deviation value is then compared with a preset tolerance standard. If the deviation value is less than the tolerance standard, the data is considered to be within a reasonable measurement fluctuation range and is retained. If the deviation value is not less than the tolerance standard, the data is determined to be an outlier. To ensure the stability of subsequent data processing, the outlier data is replaced with the reference node value. The acquisition or processing time corresponding to the important parameter is used as the end point of the test cycle, and the initial test time corresponding to the important parameter is used as the start point of the test cycle. The time span between the start and end points is set as the test cycle.

[0040] This test cycle, together with the corresponding key parameters, constitutes a complete set of test data.

[0041] Collect all test data from multiple consecutive test cycles and execute a repeatability judgment logic: determine whether any important parameter exhibits multiple abnormal value fluctuations within the same test cycle.

[0042] If any important parameter fluctuates multiple times, it is determined that the important parameter is unstable. This usually means that the internal structure of the material has undergone repeated physical fatigue or chemical degradation in a short period of time. If the proportion of such abnormal fluctuations in the total number of monitoring times is greater than the preset percentage standard value, it is determined that the trenchless repair material has entered the performance deterioration stage and is in an abnormal state. The time point of this test cycle is taken as the life determination node.

[0043] If any important parameter does not fluctuate multiple times during the test period, it indicates that the material performance is still in a stable operating period. At this time, the end time of the test period is extended to the next period, and a new monitoring end time is set, and continuous real-time monitoring is carried out.

[0044] Based on the determined lifespan determination nodes, the predicted lifespan of trenchless repair materials is calculated and output. The specific process includes: Obtain the preset material design reference life, the current life determination node time, and quantitative information of important parameters. Assign corresponding calculation weights to the above three, and set a dynamic weight adjustment mechanism before performing the final numerical aggregation processing: Real-time monitoring of important parameters with dynamic changes is performed. If the fluctuations of all important parameters are within a safe range, their initial calculation weight allocation is maintained to ensure the stability of the prediction results. If the change in certain important parameters exceeds the preset fluctuation threshold, the calculation weight of the important parameter will be automatically increased, and the calculation weight of other important parameters will be reduced proportionally to highlight the impact of key damage factors on lifespan.

[0045] The fluctuation threshold is a preset threshold for the change in parameters used to trigger weight adjustments, representing the safety boundary for material performance fluctuations.

[0046] Furthermore, the weight adjustment logic can be specifically viewed as a calculation model that dynamically allocates weights based on the fluctuation range of important parameters, and its specific definition is as follows:

[0047] It should be noted that, Represented by natural constant An exponential function with base 0. This indicates the operation of finding the maximum value.

[0048] In the formula, This represents the adjusted weight, meaning the weight allocated to the first element after dynamic adjustment. The final weights of the key parameters are calculated. This indicates the real-time fluctuation range, and its meaning is the first... The real-time changes of several key parameters during the current monitoring period; This represents the fluctuation threshold, which is the parameter fluctuation threshold that triggers weight penalty or gain. This represents the sensitivity of the adjustment, which is a constant indicating the degree to which the control weight changes drastically with fluctuations. This represents the initial weight, which is the weight pre-assigned to the first weight before the weight adjustment logic is triggered. The baseline weights of the key parameters; This indicates the total number of parameters, meaning the total number of important parameters involved in the lifetime prediction calculation. The subscript indicates the total number of parameters. Indicates the first One important parameter.

[0049] Subscript This represents the summation index, used to iterate through all important parameters involved in the calculation, with a summation range of... arrive .

[0050] The reference life, life determination node, and important parameter information are weighted and summed according to the adjusted calculation weights to obtain the final predicted life of the trenchless repair material.

[0051] Example 2

[0052] Please see Figure 2 As shown, this embodiment provides a trenchless repair material service life prediction system based on big data, including the following modules: Information collection module: The system is configured to acquire material testing information of trenchless repair materials and extract raw data containing dosage parameters. This material testing information includes, but is not limited to, mechanical property data, chemical composition variation data, and physical property data. The preferred mechanical property data are tensile strength, elongation at break, and elastic modulus; the chemical composition variation data includes the content of aging products and changes in functional groups; the physical property data includes density, hardness, and permeability; and the preferred environmental exposure conditions are temperature, humidity, ultraviolet radiation intensity, and chemical medium concentration.

[0053] Raw data are the unprocessed data obtained directly from sensors or testing instruments in these test information, which includes dosage parameters for subsequent analysis, such as the amount of material consumed during the test, the applied stress / strain values, or the amount of specific additives used.

[0054] Feature building module: It is configured to respond to the data acquisition results from the information acquisition module, determine the comparison benchmark based on the dosage parameters, and generate a comparison graph reflecting the evolution of material properties.

[0055] When the information acquisition module obtains the raw data from the material testing information, and the fluctuation state of the raw data meets the preset feature extraction conditions, that is, when the data fluctuation amplitude exceeds a specific threshold or the data change trend shows a specific pattern, the feature construction module is activated. Specifically: The comparison benchmark is determined based on the dosage parameters in the original data. For example, the dosage parameters of the material in its initial state are selected as the benchmark, or the dosage parameters in a specific test stage are selected as the benchmark. Establish a mapping relationship between the original data and the comparison benchmark, such as calculating the percentage change, absolute difference, or normalized ratio of the original data relative to the benchmark; The differential mapping relationship is visualized in a coordinate system to form a continuous curve characterizing the material properties as a function of test time, which serves as a comparative graph and intuitively shows the evolution trend of material properties with time or test conditions.

[0056] Node decision module: It is configured to identify fluctuation peaks and starting points from the comparison graphs and determine the lifespan determination nodes based on the distribution characteristics of key parameters during the test cycle.

[0057] In the specific execution process, the node decision module maps the comparison graph to a preset coordinate system and identifies feature nodes in the comparison graph whose rate of curvature change exceeds a preset threshold. These feature nodes usually represent moments when material properties change significantly. The point with the largest vertical axis value among the feature nodes is determined as the fluctuation peak, and the fluctuation peak is traced back to the turning point where the slope tends to be stable as the starting point. The region between the starting point and the peak of the fluctuation is defined as the key feature node region, which is used to extract important parameters characterizing the aging trend of materials, such as the rate of material performance degradation, the amplitude of performance fluctuation, or the time when a specific performance index reaches the warning value.

[0058] When determining the lifespan determination node, the node decision module determines the starting node of the test cycle based on the starting point and the ending node based on the time point when the important parameters return to the preset stable range; the time span between the starting node and the ending node is calculated to obtain the test cycle. During the test cycle, the module calculates the evolution rate of key parameters; and combines the intersection of the evolution rate with the preset material failure criterion, namely the life end threshold, to locate the coordinates of the life determination node on the time axis.

[0059] In addition, during the process of determining the lifespan determination node, this module can also obtain the delay time of important parameters under different test environments, and correct the cutoff node based on the delay time to obtain the calibration calculation result; and use the calibration calculation result to compensate for the position of the lifespan determination node to improve the accuracy of lifespan determination.

[0060] Lifetime prediction module: It is configured to perform lifetime prediction logic based on lifetime determination nodes to output the predicted lifetime of trenchless repair materials.

[0061] In the specific execution process, the life prediction module receives the life determination node determined by the node decision module. This module uses the mapping relationship between the life determination node and the test cycle to calculate and output the predicted service life of the trenchless repair material through a preset prediction model based on aging kinetics model, empirical regression model or machine learning model. The mapping relationship can be a function, lookup table or trained prediction model, which associates the aging characteristics of the material in the test cycle with the actual service life.

[0062] Feedback calibration module: It is configured to evaluate and optimize the output of the lifetime prediction module by comparing the lifetime determination node with the reference node in the verification set.

[0063] After the lifetime prediction module outputs the predicted lifetime of trenchless repair materials, the feedback calibration module obtains a verification set of samples with known lifetimes and extracts reference nodes from the verification set. The verification set contains samples of trenchless repair materials with known actual lifetimes, and their reference nodes represent the actual lifetimes or key aging characteristic points of these samples. This module compares the lifespan determination node with the reference node to evaluate the deviation value of the predicted lifespan. When the deviation value exceeds the preset range, the feedback calibration module will adjust the calculation weight of important parameters in the process of determining the lifespan determination node, or adjust the parameters of the prediction model used in the lifespan prediction module, thereby optimizing the performance of the entire prediction system and improving the accuracy of subsequent predictions.

[0064] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for predicting the service life of trenchless rehabilitation materials based on big data, characterized in that, include: Obtain material testing information for trenchless repair materials; In response to the fluctuation state of the raw data in the material testing information meeting the preset feature extraction conditions, a comparison benchmark is determined based on the dosage parameters in the raw data, and a comparison graph reflecting the evolution of material performance is generated in combination with the comparison benchmark. Identify the peak values ​​and corresponding starting points of fluctuations from the comparative graphs to extract important parameters characterizing the aging trend of materials; Based on the distribution characteristics of key parameters during the test cycle, determine the lifespan determination node; Furthermore, by utilizing the mapping relationship between life determination nodes and test cycles, the predicted lifespan of trenchless repair materials is output.

2. The method for predicting the service life of trenchless repair materials based on big data according to claim 1, characterized in that, Identifying fluctuation peaks and their corresponding starting points from comparative graphs to extract key parameters characterizing material aging trends includes: The comparison graph is mapped to a preset coordinate system, and feature nodes in the comparison graph whose slope change rate exceeds a preset threshold are identified; the node with the largest amplitude among the feature nodes is determined as the fluctuation peak, and the fluctuation peak is traced back to the turning point where the slope tends to be stable, as the starting point; Furthermore, the region between the starting point and the peak of the fluctuation is defined as the key feature node region in order to extract important parameters.

3. The method for predicting the service life of trenchless repair materials based on big data according to claim 1, characterized in that, Based on the distribution characteristics of key parameters during the test cycle, the lifespan determination nodes are determined as follows: The starting point of the test cycle is determined based on the starting point, and the ending point is determined based on the time point when the important parameters return to the preset stable range. The time span between the starting point and the ending point is calculated to obtain the test cycle. Furthermore, during the testing period, the lifetime determination node is determined based on the intersection of the evolution rate of key parameters and the preset lifetime termination threshold.

4. The method for predicting the service life of trenchless repair materials based on big data according to claim 3, characterized in that, Based on the distribution characteristics of key parameters during the test cycle, determining the lifespan determination node also includes: The delay time of key parameters under different test environments is obtained, and the cutoff node is corrected based on the delay time to obtain the calibration calculation result; the position of the life determination node is compensated using the calibration calculation result.

5. The method for predicting the service life of trenchless repair materials based on big data according to claim 1, characterized in that, After outputting the predicted service life of the trenchless repair material, the method further includes: Obtain a validation set of known lifetime samples and extract reference nodes from the validation set; compare the lifetime determination nodes with the reference nodes to evaluate the deviation value of the predicted lifetime. And when the deviation value exceeds the preset range, adjust the calculation weight of important parameters in the process of determining the life determination node.

6. A trenchless repair material service life prediction system based on big data, characterized in that, include: The information acquisition module is used to acquire material test information of trenchless repair materials and extract raw data containing usage parameters from it. The feature construction module is used to respond to the acquisition results of the information acquisition module, determine the comparison benchmark based on the dosage parameters, and generate a comparison graph that reflects the evolution of material properties. The node decision module is used to identify fluctuation peaks and starting points from the comparison graphs to extract important parameters characterizing the aging trend of materials, and to determine the life determination node based on the distribution characteristics of important parameters during the test cycle. The life prediction module is used to execute life prediction logic based on the life determination node to output the predicted life of trenchless repair materials. It also includes a feedback calibration module, which is used to evaluate and optimize the output of the lifetime prediction module by comparing the lifetime determination node with the reference node in the calibration set.

7. The trenchless repair material service life prediction system based on big data according to claim 6, characterized in that, The node decision module is configured as follows: Determine the start and end points of the test cycle, and calculate the evolution rate of key parameters within the test cycle; In addition, by combining the evolution rate with the preset material failure criteria, the coordinates of the lifetime determination node on the time axis are located.

8. The trenchless repair material service life prediction system based on big data according to claim 6, characterized in that, The feature building module is configured as follows: Establish a mapping relationship between the original data and the comparison benchmark; Furthermore, the differential mapping relationship is visualized in a coordinate system to form a continuous curve characterizing the material properties as a function of test time.

9. The trenchless repair material service life prediction system based on big data according to claim 6, characterized in that, Identifying fluctuation peaks and their corresponding starting points from comparative graphs to extract important parameters characterizing material aging trends includes: mapping the comparative graphs onto a preset coordinate system and identifying feature nodes in the comparative graphs whose slope change rate exceeds a preset threshold; determining the node with the largest amplitude among the feature nodes as the fluctuation peak, and tracing the fluctuation peak back to the turning point where the slope tends to stabilize as the starting point; Furthermore, the region between the starting point and the peak of the fluctuation is defined as the key feature node region in order to extract important parameters.

10. The trenchless repair material service life prediction system based on big data according to claim 6, characterized in that, The feedback calibration module is configured as follows: Obtain a validation set of known lifetime samples and extract reference nodes from the validation set; compare the lifetime determination nodes with the reference nodes to evaluate the deviation value of the predicted lifetime. And when the deviation value exceeds the preset range, adjust the calculation weight of important parameters in the process of determining the life determination node.

Images