Operating road asphalt pavement maintenance technology post-evaluation method

Abstract

The invention provides a post-evaluation method for an asphalt pavement maintenance technology of an operating road, and relates to the technical field of road maintenance evaluation. Establishing an adaptive weight calculation model based on machine learning, and calculating a weight vector of each index; performing economic benefit quantitative evaluation according to an arc tangent decay model, and determining an economic benefit-cost ratio; performing qualitative evaluation by adopting a success degree method, establishing a grading evaluation system, and forming a standardized evaluation grade table; establishing a comprehensive evaluation system, and generating a fusion evaluation matrix; calculating a comprehensive evaluation vector, constructing a comprehensive evaluation matrix, calculating a comprehensive evaluation result based on the comprehensive evaluation matrix, determining a final evaluation grade, and obtaining a final evaluation result after maintenance. According to the method, a set of complete post-maintenance evaluation method is established by constructing a multi-dimensional evaluation index system and combining a qualitative and quantitative evaluation method, standardization and normalization of the evaluation process are realized, and scientificity and reliability of an evaluation result are improved.

Description

Technical Field

[0001] The invention relates to the technical field of highway maintenance evaluation, and in particular to a post-evaluation method for maintenance technology of asphalt pavement of an operating highway. Background Art

[0002] As an important infrastructure, highways are the lifeline of economic development and an important support for social development. During use, highways will inevitably suffer from various diseases and damages, which not only affect driving safety and comfort, but also accelerate the destruction of road surface structure and shorten the service life of highways. Therefore, the use of scientific and reasonable maintenance technology to timely maintain and repair highways is of great significance to ensure the normal operation of the highway network, extend its service life and improve its service level. This can not only guide the scientific development of maintenance work, but also provide an important basis for the continuous improvement and innovation of maintenance technology.

[0003] At present, the following methods are mainly used for the evaluation of asphalt pavement maintenance: First, a single evaluation method based on the pavement technical condition index, which is mainly evaluated based on the changes of technical indicators such as PCI and RQI; second, a qualitative evaluation method based on expert experience, which is comprehensively evaluated through expert scoring or questionnaire survey; third, a benefit evaluation method based on economic indicators, which mainly considers the economic relationship between maintenance input and output. These evaluation methods have the following problems in practical application: 1. The evaluation index system is imperfect. The existing evaluation methods often pay too much attention to technical indicators or economic indicators, lack comprehensive consideration of social benefits, environmental benefits, etc., and it is difficult to fully reflect the actual effect of maintenance technology. 2. The evaluation process is highly subjective. The current evaluation work mainly relies on the experience and technical level of the decision-making team or individuals, lacks scientific quantitative evaluation standards, and the objectivity and comparability of the evaluation results are insufficient. 3. The evaluation method is not standardized enough. Due to the lack of unified industry standards and specifications, the methods and standards used by various regions in carrying out post-maintenance evaluation are inconsistent, which makes it difficult to compare the evaluation results horizontally. 4. Insufficient data support. The existing evaluation methods do not make sufficient use of the analysis of historical monitoring data and lack in-depth research on the evolution laws of pavement performance, which affects the scientificity and accuracy of the evaluation results.

[0004] Therefore, establishing a scientific and systematic post-maintenance evaluation method has important theoretical significance and practical value. Summary of the invention

[0005] In view of this, the present invention proposes a post-evaluation method for asphalt pavement maintenance technology of operating highways. By constructing a comprehensive evaluation index system including technical applicability, social benefits, environmental benefits and economic benefits, and combining quantitative and qualitative evaluation methods, the problem of excessive reliance on personal experience, lack of industry standards and scientific basis in existing post-maintenance evaluation is solved, thereby achieving standardized and scientific post-maintenance evaluation.

[0006] The technical solution of the present invention is implemented as follows: The present invention provides a post-evaluation method for maintenance technology of asphalt pavement of an operating highway, comprising: S1. Construct a post-maintenance evaluation index system, which includes technical applicability index, social benefit index, environmental benefit index and economic benefit index; S2. Establish an adaptive weight calculation model based on machine learning to calculate the weight vector of each indicator; S3. Based on the economic benefit index, the economic benefit quantitative evaluation is carried out in combination with the arctangent decay model to determine the economic benefit cost ratio; S4. Use the success method to conduct qualitative evaluation of technical applicability indicators, social benefit indicators and environmental benefit indicators, establish a hierarchical evaluation system, and form a standardized evaluation grade table; S5, establishing a comprehensive evaluation system integrating subjective and objective factors, unifying the quantitative evaluation results of step S3 and the qualitative evaluation results of step S4 under the same evaluation standard, and generating a fusion evaluation matrix; S6. Calculate the comprehensive evaluation vector based on the weight vector of each indicator and the fusion evaluation matrix, and construct a comprehensive evaluation matrix. Calculate the comprehensive evaluation result based on the comprehensive evaluation matrix, determine the final evaluation level according to the maximum membership principle, and obtain the final evaluation result after maintenance.

[0007] Based on the above scheme, preferably, the evaluation index system includes primary index and secondary index, and the primary index includes technical applicability index A 1 、Social Benefit Index A 2 、Environmental benefit index A 3 and economic benefit index A 4 ; Technical applicability index A 1 、Social Benefit Index A 2 、Environmental benefit index A 3 It is a qualitative indicator, and the economic benefit indicator A 4 is a quantitative indicator; Among them, each first-level indicator has a corresponding second-level indicator: Technical Applicability Index A 1 The road performance improvement rate A 11 and maintenance technology service life A 12 ; Social Benefit Index A 2 The local economic development impact 21 , Road capacity improvement A 22 and traffic accident rate reduction 23 ; Environmental Benefit Index A 3 The noise pollution level is A 31 、Air pollution level A 32and solid waste pollution level A 33 ; Economic Benefit Index A 4 Maintenance technology benefit cost ratio A 41 .

[0008] Based on the above solution, preferably, step S2 includes: S21. Based on the evaluation index system, the initial judgment matrix is ​​constructed at two levels: Technical Applicability Index A 1 、Social Benefit Index A 2 、Environmental benefit index A 3 and economic benefit index A 4 The four first-level indicators are compared in pairs to construct a first-level indicator judgment matrix; Compare the secondary indicators under each primary indicator in pairs and construct multiple secondary indicator judgment matrices; S22. Collect evaluation data of historical maintenance projects, including expert scoring data and actual effect data, as training samples for the BP neural network, wherein the expert scoring data includes the relative importance judgment values ​​of indicators at all levels in the historical projects; the actual effect data includes the objective indicators of the road performance improvement rate, service life, and economic benefits of the historical projects; S23, construct a three-layer BP neural network, take the expert scoring data as input and the actual effect data as output, and train to obtain a weight optimization model; S24, input the indicator characteristics of each evaluation indicator corresponding to the new evaluation project into the weight optimization model, and dynamically adjust the primary indicator judgment matrix and the secondary indicator judgment matrix according to the model output results; S25, based on the adjusted judgment matrix, first calculate the maximum characteristic root and eigenvector of the primary indicator judgment matrix; then calculate the maximum characteristic root and eigenvector of each group of secondary indicator judgment matrices respectively; S26. Normalize the first-level indicator feature vector to obtain the first-level indicator weight vector; normalize each group of second-level indicator feature vectors respectively to obtain the corresponding second-level indicator weight vector.

[0009] On the basis of the above scheme, preferably, in step S23, when training the BP neural network, an adaptive optimization mechanism and an adaptive correction rule are used to update the network weights: , , Where t represents the current time step, is the learning rate of the current time step, is the basic learning rate, is the time attenuation coefficient, is the period adjustment coefficient, is the length of the training cycle; Update network weights for the next time step, is the network weight at the current time step, is the comprehensive relevance weight, is the error gradient, is the adaptive momentum factor, , is the initial momentum coefficient, is the gradient adjustment parameter, is the last weight change.

[0010] Based on the above solution, preferably, step S3 includes: S31, obtaining pavement performance index data of the road section to be evaluated before and after maintenance, wherein the pavement performance index includes pavement damage condition index PCI, pavement ride quality index RQI, pavement rutting depth index RDI, pavement bounce index PBI, pavement wear index PWI, pavement skid resistance index SRI, and pavement structure strength index PSSI; S32. Based on the pavement performance index data, an inverse tangent decay model is established to perform nonlinear curve fitting on the pavement performance index PPI: , In the formula, is the initial performance index value, The service life of the road surface. The elastic parameter of the equation is the ratio of the minimum standard value of the performance maintenance to the initial value of the performance index; , are the equation fitting parameters, The size of reflects the corresponding service life when the performance index decays to the maintenance standard, which is called the pavement life factor. The size of can achieve different decay forms, which is called the pavement decay mode factor; S33, according to the pavement performance index PPI calculated in step S32, calculate the area A formed by the pavement performance change curve before maintenance and the minimum maintenance standard value 0 , and calculate the area A of the curve after curing j : , , In the formula, is the decay curve function of performance index before maintenance, is the decay curve function of performance index after maintenance, The minimum standard value for maintenance is The time for implementing maintenance technology j, The time when the PPI curve drops to the lowest standard value before maintenance. It is the time when the PPI curve drops to the lowest standard value after maintenance; S34, according to the calculation result value of step S33, calculate the standardized benefit SB of the performance indicator PPI j (PPI): , S35. Standardized benefit SB based on the performance indicator PPI j (PPI) result value, calculate the pavement maintenance benefit index PBI j : , In the formula, is the weight coefficient of different pavement performance indicators; S36, according to the pavement maintenance benefit index PBI j The result value is used to calculate the benefit-cost ratio (BCR) of road maintenance technology. j : , In the formula, is the average life cycle cost of the maintenance technology.

[0011] Based on the above solution, preferably, step S4 includes: S41. Establish the success quantification standard of the qualitative evaluation secondary indicators, and divide the evaluation values ​​of each qualitative evaluation secondary indicator into five levels: excellent (0.9, 1.0], good (0.8, 0.9], average (0.7, 0.8], poor (0.6, 0.7], and bad (0, 0.6]; S42. Organize experts to score each secondary indicator of qualitative evaluation, use the Delphi method to collect and organize expert opinions, and form an expert scoring matrix ,in, is the score of the xth expert on the yth qualitative evaluation secondary indicator, m is the number of experts, and n is the number of qualitative evaluation secondary indicators; S43. Calculate the expert consistency coefficient C of each qualitative evaluation secondary indicator y : , in, is the scoring standard deviation of the yth qualitative evaluation secondary indicator, is the average score of the yth qualitative evaluation secondary indicator; S44, when the expert consistency coefficient If yes, take the result of this round of evaluation; otherwise, return to step S42 to continue to seek expert opinions; S45. Calculate the success score S of each qualitative evaluation secondary indicator y : , in, is the weight coefficient of the x-th expert; S46. Calculate the comprehensive success scores of the qualitative indicators A1, A2, and A3 according to the weight values ​​of the secondary indicators; S47, determining the evaluation grades of indicators at all levels according to the success score and the evaluation grade standard established in step S41, and forming a standardized evaluation grade table.

[0012] Based on the above solution, preferably, step S5 includes: S51. Establish evaluation standards for qualitative and quantitative indicators: Qualitative indicators use a standardized rating scale established using the success method; Quantitative indicators are based on the calculated benefit-cost ratio Divided into five levels: Excellent ,good ,generally , poor ,Difference ; S52. Constructing a qualitative indicator evaluation matrix ,in, For the The membership degree of a first-level qualitative indicator in the jth evaluation level, is the number of first-level qualitative evaluation indicators, and 5 is the number of evaluation levels of qualitative indicators; S53. Constructing a quantitative indicator evaluation matrix ,in, For the The membership degree of a quantitative indicator in the jth evaluation level, is the number of quantitative indicators, and 5 is the number of evaluation levels of the quantitative indicators; S54. Using a weighted fusion method, the qualitative indicator evaluation matrix Q and the quantitative indicator evaluation matrix P are fused to generate a fused evaluation matrix R.

[0013] Based on the above solution, preferably, the expression of the fusion evaluation matrix R is as follows: , in: , , In the formula, is the qualitative evaluation weight, is the quantitative evaluation weight, and ; The synergy coefficient is used to characterize the synergy effect of qualitative evaluation and quantitative evaluation; i is the first-level indicator number, and j is the evaluation level number.

[0014] Based on the above solution, preferably, step S6 includes: S61. The final evaluation level after maintenance is divided into five levels: very successful, relatively successful, basically successful, unsuccessful and very unsuccessful, and the corresponding evaluation level characteristic values ​​are 1.0, 0.8, 0.6, 0.4 and 0.2; S62. Based on the weight vector W of the primary index and the fusion evaluation matrix R, calculate the comprehensive evaluation vector V: , In the formula, W is the weight vector of the first-level index, R is the fusion evaluation matrix, is the comprehensive evaluation vector, v j represents the evaluation value of the jth evaluation level; S63. Construct a comprehensive evaluation matrix G: , In the formula, H is the evaluation level feature matrix, ; S64. Calculate the comprehensive evaluation result Z: , In the formula, is the characteristic value of the jth evaluation level; S65. Calculate the normalized membership of each evaluation level: , In the formula, is the normalized membership of the jth evaluation level; S66, according to the maximum membership principle, select The grade corresponding to the maximum value is taken as the final evaluation grade.

[0015] On the basis of the above scheme, preferably, when the difference between the degrees of membership of two evaluation levels is less than 0.1, the lower level is taken as the final evaluation level.

[0016] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention establishes a complete post-maintenance evaluation method by constructing a multi-dimensional evaluation index system including technical applicability, social benefits, environmental benefits and economic benefits, and combining qualitative and quantitative evaluation methods, thereby achieving standardization and regularization of the evaluation process and improving the scientificity and reliability of the evaluation results; (2) The present invention adopts a BP neural network adaptive weight calculation model, and realizes the dynamic adjustment of the evaluation index weights through the training of historical maintenance project data and the adaptive optimization mechanism, thus overcoming the problem that the weight determination in the traditional hierarchical analysis method is too dependent on the subjective judgment of experts, making the weight calculation results more objective; (3) The present invention introduces the inverse tangent decay model to perform nonlinear curve fitting on the pavement performance index, and quantifies the maintenance effect through the benefit area method, thereby achieving an accurate evaluation of the economic benefits and improving the accuracy of the quantitative evaluation; (4) The present invention establishes a fusion mechanism for qualitative and quantitative indicators. By introducing a weighted fusion method and a synergy coefficient, it achieves effective integration of different types of evaluation indicators, solves the problem of difficulty in unifying qualitative and quantitative indicators in traditional evaluation methods, and improves the credibility of comprehensive evaluation results. (5) The present invention determines the final evaluation level based on the maximum membership principle and sets a judgment rule for similar membership, which effectively avoids the ambiguity of the evaluation results and makes the evaluation results clearer and more reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 is a flow chart of a method according to an embodiment of the present invention; Figure 2 This is a fitting diagram of the PCI inverse tangent decay curve of the road section evaluated before maintenance according to an embodiment of the present invention; Figure 3 This is a fitting diagram of the RQI inverse tangent decay curve of the road section evaluated before maintenance according to an embodiment of the present invention; Figure 4 This is a fitting diagram of the PCI inverse tangent decay curve of the road section evaluated after maintenance according to an embodiment of the present invention; Figure 5 This is a fitting diagram of the RQI inverse tangent decay curve of the road section evaluated after maintenance according to an embodiment of the present invention; Figure 6 It is a performance indicator change curve and benefit area schematic diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] like Figure 1 As shown, the present invention provides a post-evaluation method for maintenance technology of asphalt pavement of an operating highway, comprising: S1. Construct a post-maintenance evaluation index system, which includes technical applicability index, social benefit index, environmental benefit index and economic benefit index; S2. Establish an adaptive weight calculation model based on machine learning to calculate the weight vector of each indicator; S3. Based on the economic benefit index, the economic benefit quantitative evaluation is carried out in combination with the arctangent decay model to determine the economic benefit cost ratio; S4. Use the success method to conduct qualitative evaluation of technical applicability indicators, social benefit indicators and environmental benefit indicators, establish a hierarchical evaluation system, and form a standardized evaluation grade table; S5, establishing a comprehensive evaluation system integrating subjective and objective factors, unifying the quantitative evaluation results of step S3 and the qualitative evaluation results of step S4 under the same evaluation standard, and generating a fusion evaluation matrix; S6. Calculate the comprehensive evaluation vector based on the weight vector of each indicator and the fusion evaluation matrix, and construct a comprehensive evaluation matrix. Calculate the comprehensive evaluation result based on the comprehensive evaluation matrix, determine the final evaluation level according to the maximum membership principle, and obtain the final evaluation result after maintenance.

[0021] The post-maintenance technology evaluation method referred to in the present invention is to use the evaluation method proposed in the present invention to evaluate the asphalt pavement of an operating highway after the maintenance technology is implemented.

[0022] Specifically, in one embodiment of the present invention, first, a post-maintenance evaluation index system is constructed. The evaluation index system includes primary indicators and secondary indicators. The primary indicators include technical applicability indicators A 1 、Social Benefit Index A 2 、Environmental benefit index A 3 and economic benefit index A 4 ; Technical applicability index A 1 、Social Benefit Index A 2 、Environmental benefit index A 3 It is a qualitative indicator, and the economic benefit indicator A 4 is a quantitative indicator; Among them, each first-level indicator has a corresponding second-level indicator: Technical Applicability Index A 1 The road performance improvement rate A 11 and maintenance technology service life A 12 ; Social Benefit Index A 2 The local economic development impact 21 , Road capacity improvement A 22 and traffic accident rate reduction 23 ; Environmental Benefit Index A 3 The noise pollution level is A 31 、Air pollution level A 32 and solid waste pollution level A 33 ; Economic Benefit Index A 4 Maintenance technology benefit cost ratio A 41 .

[0023] Specifically, the evaluation index system is shown in Table 1: Table 1 Evaluation index system for highway asphalt pavement maintenance Specifically, in one embodiment of the present invention, step S2 includes: S21. Based on the evaluation index system, the initial judgment matrix is ​​constructed at two levels: Technical Applicability Index A 1 、Social Benefit Index A 2 、Environmental benefit index A 3 and economic benefit index A 4 The four first-level indicators are compared in pairs to construct a first-level indicator judgment matrix; Compare the secondary indicators under each primary indicator in pairs and construct multiple secondary indicator judgment matrices; S22. Collect evaluation data of historical maintenance projects, including expert scoring data and actual effect data, as training samples for the BP neural network, wherein the expert scoring data includes the relative importance judgment values ​​of indicators at all levels in the historical projects; the actual effect data includes the objective indicators of the road performance improvement rate, service life, and economic benefits of the historical projects; S23, construct a three-layer BP neural network, take the expert scoring data as input and the actual effect data as output, and train to obtain a weight optimization model; S24, input the indicator characteristics of each evaluation indicator corresponding to the new evaluation project into the weight optimization model, and dynamically adjust the primary indicator judgment matrix and the secondary indicator judgment matrix according to the model output results; S25, based on the adjusted judgment matrix, first calculate the maximum characteristic root and eigenvector of the primary indicator judgment matrix; then calculate the maximum characteristic root and eigenvector of each group of secondary indicator judgment matrices respectively; S26. Normalize the first-level indicator feature vector to obtain the first-level indicator weight vector; normalize each group of second-level indicator feature vectors respectively to obtain the corresponding second-level indicator weight vector.

[0024] In this embodiment, when training the BP neural network, an adaptive optimization mechanism and an adaptive correction rule are used to update the network weights: , , Where t represents the current time step, is the learning rate of the current time step, is the basic learning rate, is the time attenuation coefficient, is the period adjustment coefficient, is the length of the training cycle; Update network weights for the next time step, is the network weight at the current time step, is the comprehensive relevance weight, is the error gradient, is the adaptive momentum factor, , is the initial momentum coefficient, is the gradient adjustment parameter, is the last weight change.

[0025] In one example, the implementation process of step S2 is as follows: Construction of the first-level indicator judgment matrix: , Construct the secondary indicator judgment matrix: , Collect training sample data. In this example, the expert rating data collected is: History Project 1: A 1 :A 2 =3:1; A 1 :A 3 =2:1; A 1 :A 4 =4:1; A 2 :A 3 =1:2; A 2 :A 4 =2:1; A 3 :A 4 =3:1; A11 :A 12 =2:1; A 21 :A 22 :A 23 =1:2:3; A 31 :A 32 :A 33 =3:2:1; actual effect data: road performance improvement rate: 85%; service life: 5 years; traffic capacity improvement: 25%; accident rate reduction: 30%; noise reduction: 12dB; air quality improvement: 15%; energy saving: 20%; benefit-cost ratio: 1.8.

[0026] History Project 2: A 1 :A 2 =2.5:1; A 1 :A 3 =1.5:1; A 1 :A 4 =3.8:1; A 2 :A 3 =1:2.1; A 2 :A 4 =1.9:1; A 3 :A 4 =2.8:1; A 11 :A 12 =1.5:1; A 21 :A 22 :A 23 =1:2:2.5; A 31 :A 32 :A 33 =2.5:1.5:1; actual effect data: road performance improvement rate: 78%; service life: 4 years; traffic capacity improvement: 20%; accident rate reduction: 25%; noise reduction: 10dB; air quality improvement: 12%; energy saving: 18%; benefit-cost ratio: 1.5.

[0027] Construct a BP neural network with the following structure: input layer: set to 13 nodes according to the actual number of indicators (4 first-level indicator comparison values ​​+ 9 second-level indicator comparison values); hidden layer 1: 16 nodes; hidden layer 2: 10 nodes; output layer: weight value.

[0028] The training parameters are set to: basic learning rate η 0 =0.1; learning rate decay coefficient λ=0.01; cycle adjustment coefficient β=0.2; training cycle length T=1000.

[0029] According to the output results of the BP network, the first-level indicator judgment matrix is ​​adjusted: , Calculate the characteristic roots and eigenvectors, the first-level indicator judgment matrix: the largest characteristic root λ max =4.0123; eigenvector p=[0.4978,0.0717,0.2753,0.1552].

[0030] Perform consistency test: CI=(λ max -n) / (n-1)=(4.0123-4) / (4-1)=0.0041; RI is obtained by looking up in Table 2; CR=CI / RI<0.1. The judgment matrix passes the consistency test.

[0031] Table 2 RI standard values Degree 1 2 3 4 5 6 7 8 RI 0 0 0.52 0.89 1.12 1.26 1.36 1.41 First-level indicator weight vector W: W=[0.4978,0.1717,0.2753,0.0552]; Second-level indicator weight vector w i :Technical applicability (A 1 )Next: w 1 =[0.6667,0.3333], social benefit (A 2 )Next: w 2 =[0.1634,0.2970,0.5396], environmental benefit (A 3 )Next: w 3 =[0.5396,0.2970,0.1634], economic benefit (A 4 )Next: w 4 =[1.0000].

[0032] Specifically, in one embodiment of the present invention, step S3 includes: S31, obtaining pavement performance index data of the road section to be evaluated before and after maintenance, wherein the pavement performance index includes pavement damage condition index PCI, pavement ride quality index RQI, pavement rutting depth index RDI, pavement bounce index PBI, pavement wear index PWI, pavement skid resistance index SRI, and pavement structure strength index PSSI; S32. Based on the pavement performance index data, an inverse tangent decay model is established to perform nonlinear curve fitting on the pavement performance index PPI: , In the formula, is the initial performance index value, The service life of the road surface. is the elastic parameter of the equation, which is the ratio of the minimum standard value of the performance maintenance to the initial value of the performance index; , are the equation fitting parameters, The size of reflects the corresponding service life when the performance index decays to the maintenance standard, which is called the pavement life factor. The size of can achieve different decay forms, which is called the pavement decay mode factor; S33, according to the pavement performance index PPI calculated in step S32, calculate the area A formed by the pavement performance change curve before maintenance and the minimum maintenance standard value 0 , and calculate the area A of the curve after curing j : , , In the formula, is the decay curve function of performance index before maintenance, is the decay curve function of performance index after maintenance, The minimum standard value for maintenance is The time for implementing maintenance technology j, The time when the PPI curve drops to the lowest standard value before maintenance. It is the time when the PPI curve drops to the lowest standard value after maintenance; S34, according to the calculation result value of step S33, calculate the standardized benefit SB of the performance indicator PPI j (PPI): , S35. Standardized benefit SB based on the performance indicator PPI j (PPI) result value, calculate the pavement maintenance benefit index PBI j : , In the formula, is the weight coefficient of different performance indicators; S36, according to the pavement maintenance benefit index PBI j The result value is used to calculate the benefit-cost ratio (BCR) of road maintenance technology. j : , In the formula, is the average life cycle cost of the maintenance technology.

[0033] In the present invention, the pavement performance index PPI refers to a comprehensive performance parameter used to evaluate the effect of pavement maintenance technology, which is obtained by fitting a nonlinear attenuation model based on basic performance indicators such as the pavement damage index (PCI) and the pavement ride quality index (RQI). Each basic performance indicator can obtain a corresponding PPI value, and these PPI values ​​are finally weighted and integrated through weight coefficients to calculate the pavement maintenance benefit index.

[0034] Specifically, The acquisition method is: obtain pavement performance index data (such as PCI, RQI, etc.) through actual testing. For each performance index, the initial test value before maintenance is taken as the index. , this initial value will be used as the benchmark value for subsequent PPI attenuation calculations.

[0035] In one example, PCI and RQI were used as indicator data for the pavement performance index PPI. The PCI, RQI and inspection data of the road section to be evaluated for at least two years before and after maintenance were collected. The road section to be evaluated was selected as the main up lane of the Guangshao Expressway, with the starting and ending pile numbers of K1986+920~K1988+080. Micro-surfacing preventive maintenance technology was adopted for this section in 2017.

[0036] Based on the principles of integrity and effectiveness of test data, the PCI and RQI data of the road surface several years before and after maintenance were obtained, as shown in Table 3 below.

[0037] Table 3 PCI and RQI data of the road sections to be evaluated over the years years 2014 2015 2016 2017 (before maintenance) 2017 (after maintenance) 2018 2019 2020 2021 2022 PCI 94.5 92.8 90.0 87.2 98.8 97.6 94.3 92.2 89.5 86.1 RQI 93.2 91.4 89.3 88.1 95.8 93.1 92.1 91.0 88.6 87.9 See also Figure 2-Figure 5 The inverse tangent decay model was used to fit the PCI and RQI data of the pavement before and after maintenance of the evaluated section. Through a survey of the maintenance department of the Guangshao Expressway, it was found that the minimum standard values ​​of PCI and RQI for preventive maintenance of the highway were 85 and 82, respectively, that is, λ before maintenance in the PCI decay equation was 0.85, and λ after maintenance was 0.86; λ before maintenance in the RQI decay equation was 0.82, and λ after maintenance was 0.856.

[0038] Specifically, according to the decay equation before maintenance, the benefit area A of PCI without preventive maintenance technology is 0 (PCI), which is calculated as: , According to the decay equation before maintenance, the benefit area A of RQI when preventive maintenance technology is not adopted is 0 (RQI), which is calculated as: , See also Figure 6 , the increased benefit area A of PCI after preventive maintenance j (PCI), which is calculated as: , The increased benefit area of ​​RQI after preventive maintenance is A j (RQI), which is calculated as: , Specifically, according to the benefit area A of PCI when preventive maintenance technology is not adopted 0 (PCI) and the increased benefit area of ​​PCI after preventive care A j The calculation results of (PCI) are used to calculate the standardized benefit SB using the performance indicator PPI j (PPI), which is calculated as follows: , According to the benefit area A of RQI when preventive maintenance technology is not adopted 0 (RQI) and the increased benefit area A of RQI after preventive maintenance j The calculation results of (RQI) are used to calculate the standardized benefit SB using the performance indicator PPI j (RQI), which is calculated as: , Construct the judgment matrix of the two indicators PCI and RQI as follows: , Calculate the maximum eigenvalue of the judgment matrix as λ max =2, the corresponding eigenvector is P=(0.89,0.45) T , after normalization, the weight vector is W=(0.67,0.33) T , that is, the benefit weight coefficients of PCI and RQI are 0.67 and 0.33 respectively, so the maintenance benefit index PBI j , and its calculation formula is: , Through the investigation of the maintenance department of Guangshao Expressway, the average life cycle cost of micro-surfacing preventive maintenance technology is j 155 yuan / m 2 , then the benefit-cost ratio of this preventive maintenance technology is BCR j , and its calculation formula is: , BCR j The size of reflects the economic benefits of adopting maintenance technology j.

[0039] Specifically, in one embodiment of the present invention, step S4 includes: S41. Establish the success quantification standard of the qualitative evaluation secondary indicators, and divide the evaluation values ​​of each qualitative evaluation secondary indicator into five levels: excellent (0.9, 1.0], good (0.8, 0.9], average (0.7, 0.8], poor (0.6, 0.7], and bad (0, 0.6]; S42. Organize experts to score each secondary indicator of qualitative evaluation, use the Delphi method to collect and organize expert opinions, and form an expert scoring matrix ,in, is the score of the xth expert on the yth qualitative evaluation secondary indicator, m is the number of experts, and n is the number of qualitative evaluation secondary indicators; S43. Calculate the expert consistency coefficient C of each qualitative evaluation secondary indicator y : , in, is the scoring standard deviation of the yth qualitative evaluation secondary indicator, is the average score of the yth qualitative evaluation secondary indicator; S44, when the expert consistency coefficient If yes, take the result of this round of evaluation; otherwise, return to step S42 to continue to seek expert opinions; S45. Calculate the success score S of each qualitative evaluation secondary indicator y : , in, is the weight coefficient of the x-th expert; S46. Calculate the comprehensive success scores of the qualitative indicators A1, A2, and A3 according to the weight values ​​of the secondary indicators; S47, determining the evaluation grades of indicators at all levels according to the success score and the evaluation grade standard established in step S41, and forming a standardized evaluation grade table.

[0040] Specifically, in one example, the evaluation values ​​of each qualitative evaluation secondary indicator are first divided into five levels: excellent (i.e. very successful) (0.9, 1.0], good (i.e. relatively successful) (0.8, 0.9], average (i.e. basically successful) (0.7, 0.8], poor (i.e. unsuccessful) (0.6, 0.7], and bad (i.e. very unsuccessful) (0, 0.6]). Then an expert group is established to select 7-9 experts with rich maintenance experience, including maintenance engineers, material experts, construction experts and other different fields. The Delphi method is used for multiple rounds of anonymous scoring. Experts score each qualitative indicator according to quantitative standards to form an expert scoring matrix D.

[0041] In this example, the experts scored according to the evaluation rating table shown in Table 4-Table 6: Table 4 Evaluation index table of technical applicability after maintenance Table 5 Evaluation level of social benefit evaluation index after maintenance Table 6 Evaluation level of environmental benefit evaluation index after maintenance Afterwards, the evaluation results of the experts are summarized, and the grade characteristic values ​​corresponding to the evaluation of each expert are quantified in step S41, and the success scores of the secondary indicators are calculated in combination with the weights of each expert. Then, according to the weight vectors of each secondary indicator assigned in step S2, the weight values ​​of the specific secondary indicators are determined to calculate the comprehensive success scores of the primary indicators. According to the success scores, the evaluation grades of the indicators at all levels are determined by comparing them with the evaluation grade standards established in step S41 to form a standardized evaluation grade table.

[0042] Specifically, in one embodiment of the present invention, step S5 includes: S51. Establish evaluation standards for qualitative and quantitative indicators: Qualitative indicators use a standardized rating scale established using the success method; Quantitative indicators are based on the calculated benefit-cost ratio Divided into five levels: Excellent ,good ,generally , poor ,Difference ; S52. Constructing a qualitative indicator evaluation matrix ,in, For the The membership degree of a first-level qualitative indicator in the jth evaluation level, is the number of first-level qualitative evaluation indicators, and 5 is the number of evaluation levels of qualitative indicators; S53. Constructing a quantitative indicator evaluation matrix ,in, For the The membership degree of a quantitative indicator in the jth evaluation level is the number of quantitative indicators, and 5 is the number of evaluation levels of the quantitative indicators; S54. Using a weighted fusion method, the qualitative indicator evaluation matrix Q and the quantitative indicator evaluation matrix P are fused to generate a fused evaluation matrix R.

[0043] The expression of the fusion evaluation matrix R is as follows: , in: , , In the formula, is the qualitative evaluation weight, is the quantitative evaluation weight, and ; is the synergy coefficient, which is used to characterize the synergy effect of qualitative evaluation and quantitative evaluation; i is the first-level indicator number, and j is the evaluation level number.

[0044] Specifically, in one example, the constructed qualitative indicator evaluation matrix Q is as follows: , Among them, the first row of the matrix Q is A 1 The second row is the A level of each evaluation level of the indicator. 2 The third row is the A level of each indicator. 3 The degree of membership of each evaluation level of the indicator.

[0045] The constructed quantitative indicator evaluation matrix P is as follows: , The matrix P represents the degree of membership of each evaluation level of the A4 index.

[0046] After that, the fusion evaluation matrix R is calculated. When calculating, the value is , specifically ,Pick , , when fusion calculation is performed, the first element r 11 For example, the calculation formula is as follows: , The final fusion evaluation matrix R is as follows: , According to the membership distribution of the secondary indicators under each primary indicator in the fusion evaluation matrix R, it can be analyzed that: Technical applicability index (A 1 )Analysis: Membership distribution: [0.318, 0.424, 0.208, 0.104, 0.000], the highest membership is "relatively successful" (0.424), and the membership of "very successful" is also relatively high (0.318), with the largest weight and the greatest impact on the final result. Conclusion: The technical applicability performance is good and has met the expected technical requirements.

[0047] Social benefit index (A 2)Analysis: Membership distribution: [0.212, 0.524, 0.208, 0.104, 0.000], "relatively successful" has the highest membership (0.524), with obvious advantages, and other levels have relatively low membership, small weights, and limited influence. Conclusion: The social benefits are outstanding.

[0048] Environmental benefit index (A 3 ) Analysis: Membership distribution: [0.424, 0.318, 0.208, 0.104, 0.000], "very successful" has the highest membership (0.424), and "relatively successful" has the second highest membership (0.318). The weight is moderate and has a certain influence. Conclusion: The environmental benefit performance is excellent and has reached a relatively high level.

[0049] Economic benefit index (A 4 )Analysis: Membership distribution: [0.300, 0.400, 0.200, 0.100, 0.000], the distribution of each level is relatively balanced, "relatively successful" has the highest membership (0.400), the smallest weight, and the smallest impact. Conclusion: The economic benefits are stable and have reached a good level overall.

[0050] Specifically, in one embodiment of the present invention, step S6 includes: S61. The final evaluation level after maintenance is divided into five levels: very successful, relatively successful, basically successful, unsuccessful and very unsuccessful, and the corresponding evaluation level characteristic values ​​are 1.0, 0.8, 0.6, 0.4 and 0.2; S62. Based on the weight vector W of the primary index and the fusion evaluation matrix R, calculate the comprehensive evaluation vector V: , In the formula, W is the weight vector of the first-level index, R is the fusion evaluation matrix, is the comprehensive evaluation vector, v j represents the evaluation value of the jth evaluation level; S63. Construct a comprehensive evaluation matrix G: , In the formula, H is the evaluation level feature matrix, ; S64. Calculate the comprehensive evaluation result Z: , In the formula, is the characteristic value of the jth evaluation level; S65. Calculate the normalized membership of each evaluation level: , In the formula, is the normalized membership of the jth evaluation level; S66, according to the maximum membership principle, select The grade corresponding to the maximum value is taken as the final evaluation grade.

[0051] Among them, when the difference in membership between two evaluation levels is less than 0.1, the lower level is taken as the final evaluation level.

[0052] Specifically, in one example, the evaluation rating system is established as follows: Very successful: eigenvalue = 1.0; relatively successful: eigenvalue = 0.8; basically successful: eigenvalue = 0.6; unsuccessful: eigenvalue = 0.4; very unsuccessful: eigenvalue = 0.2.

[0053] Evaluation level feature matrix .

[0054] The first-level indicator weight vector W: W=[0.4978, 0.1717, 0.2753, 0.0552]; the fusion evaluation matrix R is as follows: , According to the formula Calculate the comprehensive evaluation vector: , Calculation process: The first component: 0.4978×0.318+0.1717×0.212+0.2753×0.424+0.0552×0.300=0.3316; The second component: 0.4978×0.424+0.1717×0.524+0.2753×0.318+0.0552×0.400=0.4089; The third component: 0.4978×0.208+0.1717×0.208+0.2753×0.208+0.0552×0.200=0.2075; The fourth component: 0.4978×0.104+0.1717×0.104+0.2753×0.104+0.0552×0.100=0.1037; The fifth component: 0.4978×0.000+0.1717×0.000+0.2753×0.000+0.0552×0.000=0.0000.

[0055] V=[0.3316,0.4089,0.2075,0.1037,0.0000].

[0056] Construct a comprehensive evaluation matrix G: , Calculate the comprehensive evaluation result Z: , Calculate the normalized membership: , μ 1 =0.3316 / 1.0517=0.3153 (very successful); μ 2 =0.4089 / 1.0517=0.3888 (relatively successful); μ 3 =0.2075 / 1.0517=0.1973 (basically successful); μ 4 =0.1037 / 1.0517=0.0986 (unsuccessful); μ 5 =0.0000 / 1.0517=0.0000 (very unsuccessful).

[0057] Membership difference check: μ 1 and μ 2 The difference: , since the difference is less than 0.1, and μ 2 >μ 1 , a lower grade should be taken. Final evaluation grade: relatively successful.

[0058] From the calculation results we can see that: The comprehensive evaluation result Z is 0.8247, the membership degree of "relatively successful" is the highest (0.3888), and the membership degree of "very successful" is the second (0.3153), and the difference between the two is less than 0.1. Since the difference in membership between adjacent levels is less than 0.1, it is necessary to adopt the conservative principle and take the lower level "relatively successful" as the final evaluation level.

[0059] Conclusion: The final evaluation grade of this maintenance technology is “relatively successful”, which reflects that this maintenance technology has a good implementation effect.

[0060] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A post-evaluation method for maintenance technology of asphalt pavement of an operating highway, characterized in that: include: S1. Construct a post-maintenance evaluation index system, which includes technical applicability index, social benefit index, environmental benefit index and economic benefit index; S2. Establish an adaptive weight calculation model based on machine learning to calculate the weight vector of each indicator; S3. Based on the economic benefit index, the economic benefit quantitative evaluation is carried out in combination with the arctangent decay model to determine the economic benefit cost ratio; S4. Use the success method to conduct qualitative evaluation of technical applicability indicators, social benefit indicators and environmental benefit indicators, establish a hierarchical evaluation system, and form a standardized evaluation grade table; S5, establishing a comprehensive evaluation system integrating subjective and objective factors, unifying the quantitative evaluation results of step S3 and the qualitative evaluation results of step S4 under the same evaluation standard, and generating a fusion evaluation matrix; S6. Calculate the comprehensive evaluation vector based on the weight vector of each indicator and the fusion evaluation matrix, and construct a comprehensive evaluation matrix. Calculate the comprehensive evaluation result based on the comprehensive evaluation matrix, determine the final evaluation level according to the maximum membership principle, and obtain the final evaluation result after maintenance.

2. The post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 1, characterized in that: The evaluation index system includes primary indicators and secondary indicators. The primary indicators include technical applicability index A1, social benefit index A2, environmental benefit index A3 and economic benefit index A4. Technical applicability index A1, social benefit index A2 and environmental benefit index A3 are qualitative indicators, and economic benefit index A4 is a quantitative indicator. Among them, each first-level indicator has a corresponding second-level indicator: Technical applicability index A1 includes pavement performance improvement rate A 11 and maintenance technology service life A 12 ; Social benefit indicator A2 includes local economic development impact A 21 , Road capacity improvement A 22 and traffic accident rate reduction 23 ; Environmental benefit index A3 includes noise pollution level A 31 、Air pollution level A 32 and solid waste pollution level A 33 ; Economic benefit index A4 includes maintenance technology benefit-cost ratio A 41 .

3. A post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 2, characterized in that: Step S2 includes: S21. Based on the evaluation index system, the initial judgment matrix is ​​constructed at two levels: The four first-level indicators of technical applicability index A1, social benefit index A2, environmental benefit index A3 and economic benefit index A4 are compared in pairs to construct a first-level indicator judgment matrix; Compare the secondary indicators under each primary indicator in pairs and construct multiple secondary indicator judgment matrices; S22. Collect evaluation data of historical maintenance projects, including expert scoring data and actual effect data, as training samples for the BP neural network, wherein the expert scoring data includes the relative importance judgment values ​​of indicators at all levels in the historical projects; the actual effect data includes the objective indicators of the road performance improvement rate, service life, and economic benefits of the historical projects; S23, construct a three-layer BP neural network, take the expert scoring data as input and the actual effect data as output, and train to obtain a weight optimization model; S24, input the indicator characteristics of each evaluation indicator corresponding to the new evaluation project into the weight optimization model, and dynamically adjust the primary indicator judgment matrix and the secondary indicator judgment matrix according to the model output results; S25, based on the adjusted judgment matrix, first calculate the maximum characteristic root and eigenvector of the primary indicator judgment matrix; then calculate the maximum characteristic root and eigenvector of each group of secondary indicator judgment matrices respectively; S26. Normalize the first-level indicator feature vector to obtain the first-level indicator weight vector; normalize each group of second-level indicator feature vectors respectively to obtain the corresponding second-level indicator weight vector.

4. A post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 3, characterized in that: In step S23, when training the BP neural network, the network weights are updated using an adaptive optimization mechanism and an adaptive correction rule: ， ， Where t represents the current time step, is the learning rate of the current time step, is the basic learning rate, is the time attenuation coefficient, is the period adjustment coefficient, is the length of the training cycle; Update network weights for the next time step, is the network weight at the current time step, is the comprehensive relevance weight, is the error gradient, is the adaptive momentum factor, , is the initial momentum coefficient, is the gradient adjustment parameter, is the last weight change.

5. The post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 2, characterized in that: Step S3 includes: S31, obtaining pavement performance index data of the road section to be evaluated before and after maintenance, wherein the pavement performance index includes pavement damage condition index PCI, pavement ride quality index RQI, pavement rutting depth index RDI, pavement bounce index PBI, pavement wear index PWI, pavement skid resistance index SRI, and pavement structure strength index PSSI; S32. Based on the pavement performance index data, an inverse tangent decay model is established to perform nonlinear curve fitting on the pavement performance index PPI: ， In the formula, is the initial performance index value, The service life of the road surface. is the elastic parameter of the equation, which is the ratio of the minimum standard value of the performance maintenance to the initial value of the performance index; , are the equation fitting parameters, The size of reflects the corresponding service life when the performance index decays to the maintenance standard, which is called the pavement life factor. The size of can achieve different decay forms, which is called the pavement decay mode factor; S33, according to the pavement performance index PPI calculated in step S32, calculate the area A0 formed by the pavement performance change curve before maintenance and the minimum maintenance standard value, and calculate the increased area A of the curve after maintenance j : ， ， In the formula, is the decay curve function of performance index before maintenance, is the decay curve function of performance index after maintenance, The minimum maintenance standard value is The time for implementing maintenance technology j, The time when the PPI curve drops to the lowest standard value before maintenance. It is the time when the PPI curve drops to the lowest standard value after maintenance; S34, according to the calculation result value of step S33, calculate the standardized benefit SB of the performance indicator PPI j (PPI): ， S35. Standardized benefit SB based on the performance indicator PPI j (PPI) result value, calculate the pavement maintenance benefit index PBI j : ， In the formula, is the weight coefficient of different pavement performance indicators; S36, according to the pavement maintenance benefit index PBI j The result value is used to calculate the benefit-cost ratio (BCR) of road maintenance technology. j : ， In the formula, is the average life cycle cost of the maintenance technology.

6. A post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 5, characterized in that: Step S4 includes: S41. Establish the success quantification standard of the qualitative evaluation secondary indicators, and divide the evaluation values ​​of each qualitative evaluation secondary indicator into five levels: excellent (0.9, 1.0], good (0.8, 0.9], average (0.7, 0.8], poor (0.6, 0.7], and bad (0, 0.6]; S42. Organize experts to score each secondary indicator of qualitative evaluation, use the Delphi method to collect and organize expert opinions, and form an expert scoring matrix ,in, is the score of the xth expert on the yth qualitative evaluation secondary indicator, m is the number of experts, and n is the number of qualitative evaluation secondary indicators; S43. Calculate the expert consistency coefficient C of each qualitative evaluation secondary indicator y : ， in, is the scoring standard deviation of the yth qualitative evaluation secondary indicator, is the average score of the yth qualitative evaluation secondary indicator; S44, when the expert consistency coefficient If yes, take the result of this round of evaluation; otherwise, return to step S42 to continue to seek expert opinions; S45. Calculate the success score S of each qualitative evaluation secondary indicator y : ， in, is the weight coefficient of the x-th expert; S46. Calculate the comprehensive success scores of the qualitative indicators A1, A2, and A3 according to the weight values ​​of the secondary indicators; S47, determining the evaluation grades of indicators at all levels according to the success score and the evaluation grade standard established in step S41, and forming a standardized evaluation grade table.

7. A post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 6, characterized in that: Step S5 includes: S51. Establish evaluation standards for qualitative and quantitative indicators: Qualitative indicators use a standardized rating scale established using the success method; Quantitative indicators are based on the calculated benefit-cost ratio Divided into five levels: Excellent ,good ,generally , poor ,Difference ; S52. Constructing a qualitative indicator evaluation matrix ,in, For the The membership degree of a first-level qualitative indicator in the jth evaluation level, is the number of first-level qualitative evaluation indicators, and 5 is the number of evaluation levels of qualitative indicators; S53. Constructing a quantitative indicator evaluation matrix ,in, For the The membership degree of a quantitative indicator in the jth evaluation level, is the number of quantitative indicators, and 5 is the number of evaluation levels of the quantitative indicators; S54. Using a weighted fusion method, the qualitative indicator evaluation matrix Q and the quantitative indicator evaluation matrix P are fused to generate a fused evaluation matrix R.

8. A post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 7, characterized in that: The expression of the fusion evaluation matrix R is as follows: ， in: ， ， In the formula, is the qualitative evaluation weight, is the quantitative evaluation weight, and ; is the synergy coefficient, which is used to characterize the synergy effect of qualitative evaluation and quantitative evaluation; i is the first-level indicator number, and j is the evaluation level number.

9. The post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 2, characterized in that: Step S6 includes: S61. The final evaluation level after maintenance is divided into five levels: very successful, relatively successful, basically successful, unsuccessful and very unsuccessful, and the corresponding evaluation level characteristic values ​​are 1.0, 0.8, 0.6, 0.4 and 0.2; S62. Based on the weight vector W of the primary index and the fusion evaluation matrix R, calculate the comprehensive evaluation vector V: ， In the formula, W is the weight vector of the first-level index, R is the fusion evaluation matrix, is the comprehensive evaluation vector, v j represents the evaluation value of the jth evaluation level; S63. Construct a comprehensive evaluation matrix G: ， In the formula, H is the evaluation level feature matrix, ; S64. Calculate the comprehensive evaluation result Z: ， In the formula, is the characteristic value of the jth evaluation level; S65. Calculate the normalized membership of each evaluation level: ， In the formula, is the normalized membership of the jth evaluation level; S66. According to the maximum membership principle, select The grade corresponding to the maximum value is taken as the final evaluation grade.

10. A post-evaluation method for maintenance technology of asphalt pavement of an operating highway as claimed in claim 9, characterized in that: When the difference in membership between two evaluation levels is less than 0.1, the lower level is taken as the final evaluation level.

Images