A lane-level maintenance plan generation method integrating multi-step prediction of performance indicators

Abstract

The present invention discloses a lane-level maintenance plan generation method integrating multi-step prediction of performance indicators, step 1, determining the implementation cycle of the road maintenance plan; step 2, predicting the section quality index: based on the lane-level pavement detection data for many years, using the integrated model XGBoost to perform multi-step prediction of the pavement performance index value; step 3, maintenance measure judgment: constructing a multi-type maintenance measure judgment logic, and determining the maintenance measures of the section for many years in the future based on the pavement parameters; step 4, merging the lane-level sections horizontally and vertically in turn to construct a maintenance project; step 5, predicting the project indicators; step 6, constructing a maintenance decision-making multi-objective optimization model and solving to determine the pavement lane-level maintenance plan. The application of the present invention can realize the refined maintenance of the road lane level, effectively improve the utilization efficiency of maintenance funds, and reduce maintenance costs and maintenance carbon emissions.

Description

Technical Field

[0001] The present invention relates to the technical field of road maintenance, and in particular to a method for generating a lane-level maintenance plan integrating multi-step prediction of performance indicators. Background Art

[0002] Highway maintenance refers to the upkeep and maintenance of highways. Maintenance focuses on the entire process of maintenance from the time the highway is completed and opened to traffic. Maintenance focuses on repairing damaged parts. It is an operation to keep the highway in good condition, prevent its quality from deteriorating, and provide good services to highway users. With the continuous increase in the mileage of highways in my country, the demand for highway maintenance has grown rapidly.

[0003] However, at present, the formulation of trunk road maintenance plans in many areas of my country mainly relies on manual experience or relatively simple statistical results analysis. However, it is highly subjective and the decision-making results are easily affected by the subjective judgment of decision makers. Although there is a certain experience guidance role, there is a lack of scientific basis, which can easily lead to waste of resources and increase maintenance costs, and at the same time fail to achieve the expected results. Summary of the invention

[0004] The technical problem to be solved by the present invention is to address the deficiencies of the above-mentioned prior art and to provide a lane-level maintenance plan generation method that integrates multi-step prediction of performance indicators. The lane-level maintenance plan generation method that integrates multi-step prediction of performance indicators can perform real-time refined maintenance of the roads to be maintained, avoid excessive maintenance or neglect of some areas, improve the utilization efficiency of maintenance funds, reduce the comprehensive maintenance cost, and promote the improvement of road network-level maintenance benefits.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A lane-level maintenance plan generation method integrating multi-step prediction of performance indicators includes the following steps.

[0007] Step 1: Determine the implementation period T of the road maintenance plan in years.

[0008] Step 2: Predicting the cross-section quality index, including the following steps:

[0009] Step 2-1, select the base year: select one year as the base year from the existing historical section quality indicators of the road to be maintained, and use the section quality indicator corresponding to the base year as the base section quality indicator.

[0010] Step 2-2, construct a section quality index prediction model: take the year difference between the future year and the base year, the inherent parameters of the road to be maintained and the base section quality index as input, take the section quality index of the future year as output, and construct a section quality index prediction model.

[0011] Step 2-3, prediction: Use the section quality index prediction model constructed in step 2-2, select the current year as the base year, and predict the section quality index of each section of the road to be maintained in each of the next T years.

[0012] Step 3: Determine maintenance measures: Based on the inherent parameters of the road to be maintained and the section quality indicators of each section each year in the next T years, determine the maintenance measures for each section each year within the implementation period T of the maintenance plan.

[0013] Step 4: Construct maintenance project: According to the maintenance measures determined in step 3, the lane-level section data is merged horizontally and vertically to form a maintenance project.

[0014] Step 5: Predict project indicators: Project indicators include maintenance benefits, maintenance costs and maintenance carbon emissions; for each maintenance project formed in step 4, calculate the project indicators separately.

[0015] Step 6: Construct the multi-objective optimization model MaxF for maintenance decision-making. The specific expression is:

[0016] MaxF=a*F1+b*F2+c*F3 (6-1)

[0017] In the formula, a, b, and c are model coefficients respectively; among which, a>0, b<0, c<0.

[0018] F1 is the maintenance benefit value of all maintenance projects.

[0019] F2 is the maintenance carbon emission value of all maintenance projects.

[0020] F3 is the maintenance cost value of all maintenance items.

[0021] F1, F2 and F3 are all variables X ijt Function of X ijt It takes the value of 0 or 1. In the tth year of the maintenance plan implementation cycle T, when the i-th maintenance project implements the j-th type of maintenance measures determined in step 3, the value is 1, otherwise, the value is 0.

[0022] By solving MaxF, we can obtain the maintenance year of the jth maintenance measure determined in step 3 for each maintenance project within the maintenance plan implementation period T, and form a long-term road maintenance plan.

[0023] In step 2-1, the section quality indicators include the pavement quality index PQI, the pavement damage index PCI, the pavement ride quality index RQI and the damage rate DR.

[0024] The specific method for selecting the base year in step 2-1 includes the following steps.

[0025] Step 2-1A, select several groups of section quality index data for N consecutive years, match the section quality index data for N years based on the road name, driving direction, lane number and start and end pile number information, and form several groups of time series data sets of section quality index data for N years.

[0026] Step 2-1B: Delete the abnormal data of indicator changes in the section quality indicator data time series data set, and form a new section quality indicator data time series data set with each group being no less than T years.

[0027] Step 2-1C: Select the year with the shortest time in each group of the new section quality index data time series data set as the base year.

[0028] In step 2-2, the inherent parameters of the road to be maintained include route code, driving direction, lane, city name, county name, road section technical grade, road section pavement type, terrain, roadbed width, pavement width, number of lanes and design speed.

[0029] In step 3, the maintenance measures include no maintenance, preventive maintenance, structural maintenance and functional maintenance; the specific determination method of the maintenance measures includes the following steps:

[0030] Step 3-1: Section not maintained: When the road to be maintained meets any of the following conditions A, B, and C, it is judged as not maintained, specifically:

[0031] A. Road age ≤ 3 years.

[0032] B. Road age > 3 years and PQI ≥ 90.

[0033] C. Road age>3 years, 80≤PQI<90, road grade is national highway, road traffic load grade is non-heavy load and RQI≥92.

[0034] Step 3-2, preventive maintenance: When the age of the road to be maintained is greater than 3 years, 80≤PQI<90, the road grade is a national highway, the road traffic load grade is non-heavy load and 80<RQI<92; at this time, it is determined that the road to be maintained should be subject to preventive maintenance.

[0035] Step 3-3, structural maintenance: When the road to be maintained meets any of the following conditions A, B, and C, it is judged as structural maintenance, specifically:

[0036] A. Road age > 3 years and PQI < 70.

[0037] B. Road age>3 years, 70≤PQI<80, road grade is non-national highway, road traffic load grade is non-heavy load and RQI<70.

[0038] C. Road age>3 years, 70≤PQI<80, road grade is national highway, road traffic load grade is heavy load, RQI<75.

[0039] Step 3-4: Functional maintenance: For the remaining sections that do not meet the requirements of steps 3-1 to 3-3, functional maintenance shall be performed.

[0040] In step 4, the horizontal merging method is: select all lane data of the same starting and ending pile number section in the same direction of the same road, and select the most serious maintenance measure in each lane as the unified maintenance measure for the section; the vertical merging method is: select adjacent sections with the same maintenance measures and merge them.

[0041] In step 5, suppose that in the tth year within the maintenance plan implementation period T, the ith maintenance project has m sections, then the maintenance benefit A of all sections of the ith maintenance project is it , maintenance cost it and maintain carbon emissions it The calculation formulas are:

[0042]

[0043]

[0044]

[0045] in:

[0046]

[0047] d itk = l itk × itk ×DR itk (5-5)

[0048] In the formula, A itk It is the maintenance benefit of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0049] c j is the maintenance cost per unit area when the jth type of maintenance measure is adopted, a known quantity.

[0050] e j is the carbon emission per unit area when the jth type of conservation measure is adopted, a known quantity.

[0051] d itk It is the maintenance area of ​​the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0052] l itkIt is the length of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0053] w itk It is the width of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0054] DR itk is the damage rate of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0055] PCI1, PCI2, PCI t-1 、PCI T They are the PCI values ​​in the first year, the second year, the t-1 year, and the T year when the kth section of the ith maintenance project is not maintained within T years.

[0056] RQI1, RQI2, RQI t-1 、RQI T They are the RQI values ​​in the first year, the second year, the t-1 year, and the T year when the kth section of the ith maintenance project is not maintained within T years.

[0057] PQI1, PQI2, PQI t-1 、PQI T They are the PQI values ​​in the first year, the second year, the t-1 year, and the T year when the kth section of the ith maintenance project is not maintained within T years.

[0058] PCI'0, PCI' T-t They are the PCI values ​​of the kth section of the ith maintenance project in the future T year and the future t year when it remains in a natural decay state after being maintained in the tth year.

[0059] RQI'0, RQI' T-t They are the RQI values ​​of the kth section of the ith maintenance project in the future T year and the future t year when it remains in a natural decay state after being maintained in the tth year.

[0060] PQI'0, PQI' T-t are the PQI values ​​of the kth section of the ith maintenance project in the future T year and the future t year after the kth section is maintained in the natural decay state in the tth year.

[0061] In step 6, the calculation formulas for F1, F2 and F3 are:

[0062]

[0063]

[0064]

[0065] In the formula, w i is the priority weight of the i-th maintenance project, which is assigned according to the road grade and pavement damage.

[0066] In step 6, F1, F2 and F3 must also meet the following constraints:

[0067] F3≤C (6-5)

[0068] PCI′≥PCI m (6-6)

[0069] RQI′≥RQI m (6-7)

[0070] PQI′≥PQI m (6-8)

[0071]

[0072]

[0073] Where C is the upper limit of maintenance funds.

[0074] x ijtk It takes the value of 0 or 1. In the tth year within the maintenance plan implementation cycle T, when the kth section of the ith maintenance project implements the jth type of maintenance measures determined in step 3, the value is 1; otherwise, the value is 0.

[0075] PCI′ is PCI1,...,PCI t-1 , PCI'0, ..., PCI' T-t The minimum value in .

[0076] RQI′ is RQI1, ..., RQI t-1 , RQI'0, ..., RQI' T-t The minimum value in .

[0077] PQI′ is PQI1, ..., PQI t-1 , PQI'0, ..., PQI' T-t The minimum value in .

[0078] PCI m 、RQI m and PQI m They are the minimum acceptable levels of PCI, RQI and PQI for the kth section of the ith maintenance project within the decision-making cycle of T years.

[0079] L and U are the lower and upper limits of maintenance mileage in a single year respectively.

[0080] The present invention has the following beneficial effects:

[0081] The present invention first uses the integrated model XGBoost to perform multi-step predictions on the pavement performance index values ​​based on many years of lane-level pavement detection data, then constructs a multi-type maintenance measure judgment logic, determines the maintenance measures for the sections in the future years based on the pavement parameters, then merges the lane-level sections horizontally and vertically in turn, and finally determines the pavement lane-level maintenance plan by constructing and solving a multi-objective optimization model. The application of the present invention can better plan and implement corresponding maintenance measures, and perform real-time and refined maintenance; on the other hand, it can also avoid excessive maintenance or neglect of some areas, improve the efficiency of maintenance funds, reduce comprehensive maintenance costs, and promote the improvement of road network-level maintenance benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 A flow chart of a lane-level maintenance plan generation method integrating multi-step prediction of performance indicators of the present invention is shown.

[0083] Figure 2 A flow chart of the maintenance measure determination method in the present invention is shown. DETAILED DESCRIPTION

[0084] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific preferred embodiments.

[0085] like Figure 1 As shown, a lane-level maintenance plan generation method integrating multi-step prediction of performance indicators includes the following steps.

[0086] Step 1: Determine the implementation period T of the road maintenance plan in years.

[0087] Step 2: Predicting the cross-section quality index, including the following steps:

[0088] Step 2-1, select the base year: select one year as the base year from the existing historical section quality indicators of the road to be maintained, and use the section quality indicator corresponding to the base year as the base section quality indicator.

[0089] The above-mentioned section quality indicators preferably include pavement quality index PQI, pavement damage index PCI, pavement ride quality index RQI and damage rate DR, etc.

[0090] The specific method for selecting the above base year includes the following steps.

[0091] Step 2-1A, select several groups of N consecutive years of section quality index data, match several groups of N years of section quality index data (including PQI, PCI, RQI, DR) based on the road name, driving direction, lane number and starting and ending pile number information, and form several groups of N years of section quality index data time series data sets. Wherein, N≥T, preferably N=T, and can also be N=T+1.

[0092] Step 2-1B: Delete the abnormal data of indicator changes in the section quality indicator data time series data set to form several groups of new section quality indicator data time series data sets of no less than T years.

[0093] The above indicators change abnormally, for example: the cross-sectional data of the indicators in the PQI, PCI and RQI time series data sets increase within N = T + 1 years, and the cross-sectional data of the indicators in the DR time series data set decrease within T + 1 years.

[0094] Step 2-1C, select the year with the shortest time from each group in the new section quality index data time series data set as the base year. Preferably, the specific steps are: respectively match the performance values ​​of each section in each year in the remaining data of the PQI, PCI, RQI and DR time series data sets, and at the same time identify the base year of the year with a smaller section detection time in the matching results, and the larger future year, whose corresponding data are the base year data and the future year data, and calculate the year difference between the future year and the base year in each matching result.

[0095] Step 2-2, construct a section quality index prediction model: take the year difference between the future year and the base year, the inherent parameters of the road to be maintained and the base section quality index as input, take the section quality index of the future year as output, and construct a section quality index prediction model.

[0096] The inherent parameters of the above-mentioned road to be maintained preferably include route code, driving direction, lane, city name, county name, road section technical grade, road section pavement type, terrain, roadbed width, pavement width, number of lanes and design speed, etc.

[0097] 80% of the data were randomly selected from the PQI, PCI, RQI and DR training data sets as the training set, 10% as the test set, and 10% as the validation set to train the PQI prediction model, PCI prediction model, RQI prediction model and DR prediction model respectively.

[0098] Step 2-3, prediction: Use the section quality index prediction model constructed in step 2-2, select the current year as the base year, and predict the section quality index of each section of the road to be maintained in each of the next T years.

[0099] Step 3: Determine maintenance measures: Based on the inherent parameters of the road to be maintained and the section quality indicators of each section each year in the next T years, determine the maintenance measures for each section each year within the implementation period T of the maintenance plan.

[0100] like Figure 2 As shown, maintenance measures include no maintenance, preventive maintenance, structural maintenance and functional maintenance; the specific method for determining maintenance measures includes the following steps.

[0101] Step 3-1: Section not maintained: When the road to be maintained meets any of the following conditions A, B, and C, it is judged as not maintained, specifically:

[0102] A. Road age ≤ 3 years.

[0103] B. Road age > 3 years and PQI ≥ 90.

[0104] C. Road age>3 years, 80≤PQI<90, road grade is national highway, road traffic load grade is non-heavy load and RQI≥92.

[0105] Step 3-2, preventive maintenance: When the age of the road to be maintained is greater than 3 years, 80≤PQI<90, the road grade is a national highway, the road traffic load grade is non-heavy load and 80<RQI<92; at this time, it is determined that the road to be maintained should be subject to preventive maintenance.

[0106] Step 3-3, structural maintenance: When the road to be maintained meets any of the following conditions A, B, and C, it is judged as structural maintenance, specifically:

[0107] A. Road age > 3 years and PQI < 70.

[0108] B. Road age>3 years, 70≤PQI<80, road grade is non-national highway, road traffic load grade is non-heavy load and RQI<70.

[0109] C. Road age>3 years, 70≤PQI<80, road grade is national highway, road traffic load grade is heavy load, RQI<75.

[0110] Step 3-4: Functional maintenance: For the remaining sections that do not meet the requirements of steps 3-1 to 3-3, functional maintenance shall be performed.

[0111] This step 3 only determines the type of maintenance measures for each year within the maintenance plan implementation cycle T, but the specific year in which the corresponding maintenance measures are implemented within the maintenance plan implementation cycle T is determined in step 6 of the present invention.

[0112] Step 4: Construct maintenance project: According to the maintenance measures determined in step 3, the lane-level section data is merged horizontally and vertically to form a maintenance project.

[0113] The above-mentioned lateral merging method is preferably: selecting all lane data of the same starting and ending pile number section on the same road and in the same direction, and selecting the most serious maintenance measure in each lane as the unified maintenance measure for the section.

[0114] The above-mentioned method of longitudinal merging is preferably: adjacent sections with the same maintenance measures are selected and merged.

[0115] Step 5: Predict project indicators: Project indicators include maintenance benefits, maintenance costs and maintenance carbon emissions; for each maintenance project formed in step 4, calculate the project indicators separately.

[0116] Assume that in the tth year of the maintenance plan implementation cycle T, the ith maintenance project has m sections, then the maintenance benefit A of all sections of the ith maintenance project is it , maintenance cost it and maintain carbon emissions it The calculation formulas are:

[0117]

[0118]

[0119]

[0120] in:

[0121] A itk =(PCI1+...+PCI t-1 +PCI'0+...+PCI' T-t )-(PCI1+PCI2+...+PCI t-1 +...+PCI T )+(RQI1+...+RQI t-1 +RQI'0+...+RQI' T-t )-(RQI1+RQI2+...+RQI t-1 +...+RQI T )+(PQI1+...+PQI t-1 +PQI'0+...+PQI' T-t )-(PQI1+PQI2+...+PQI t-1 +...+PQI T ) (5-4)

[0122] d itk = l itk × itk ×DR itk (5-5)

[0123] In the formula, A itk It is the maintenance benefit of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0124] c j is the maintenance cost per unit area when the jth type of maintenance measure is adopted, a known quantity.

[0125] e j is the carbon emission per unit area when the jth type of conservation measure is adopted, a known quantity.

[0126] d itk It is the maintenance area of ​​the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0127] l itk It is the length of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0128] w itk It is the width of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0129] DR itk is the damage rate of the kth section of the ith maintenance project in the tth year within the implementation period T of the maintenance plan.

[0130] PCI1, PCI2, PCI t-1 、PCI T They are the PCI values ​​in the first year, the second year, the t-1 year, and the T year when the kth section of the ith maintenance project is not maintained within T years.

[0131] RQI1, RQI2, RQI t-1 、RQI T They are the RQI values ​​in the first year, the second year, the t-1 year, and the T year when the kth section of the ith maintenance project is not maintained within T years.

[0132] PQI1, PQI2, PQI t-1 、PQI T They are the PQI values ​​in the first year, the second year, the t-1 year, and the T year when the kth section of the ith maintenance project is not maintained within T years.

[0133] PCI'0, PCI' T-t They are the PCI values ​​of the kth section of the ith maintenance project in the future T year and the future t year when it remains in a natural decay state after being maintained in the tth year.

[0134] RQI'0, RQI'T-t They are the RQI values ​​of the kth section of the ith maintenance project in the future T year and the future t year when it remains in a natural decay state after being maintained in the tth year.

[0135] PQI'0, PQI' T-t They are the PQI values ​​of the kth section of the ith maintenance project in the future T year and the future t year when it remains in a natural decay state after being maintained in the tth year.

[0136] The above PCI'0, PCI' T-t , RQI'0, RQI' T-t , RQI'0, RQI' T-t The calculation method includes the following steps.

[0137] Step 5-1: Calculate the section quality index after maintenance in the tth year

[0138] When there is no maintenance in the tth year, the section quality index prediction model in step 2 is used to predict the section quality indicators when there is no maintenance in the tth year: PQI0, PCI0 and RQI0.

[0139] When the maintenance measures determined in step 3 are adopted for maintenance in the tth year, the corresponding section quality indicators before decay after maintenance in the tth year are: PQIa, PCIa and RQIa. The specific calculation method is shown in the following table:

[0140] Project Type QUR PCIa QUR Preventive repair <![CDATA[0.6*PCI0+0.4*RQI0]]> 95 <![CDATA[RQI0]]> Functional repair <![CDATA[0.6*PCI0+0.4*RQI0]]> 100 92 Structural repairs <![CDATA[0.6*PCI0+0.4*RQI0]]> 100 95

[0141] Step 5-2: Predict PCI'0, RQI'0, RQI'0

[0142] The tth year is taken as the base year, and the section quality indicators PQIa, PCIa and RQIa before decay after maintenance in the tth year are taken as the base section quality indicators; the section quality indicator prediction model in step 2 is used to predict the section quality indicators PCI'0, RQI'0, RQI'0 after natural decay in the future T year.

[0143] Step 5-3: Predict PCI' T-t , RQI' T-t , RQI' T-t

[0144] Using a similar method to step 5-2, take the tth year as the base year, and the section quality indicators PQIa, PCIa and RQIa before decay after maintenance in the tth year as the base section quality indicators; use the section quality indicator prediction model in step 2 to predict the section quality indicator PCI' after natural decay in the tth year in the future T-t , RQI' T-t , RQI'T-t .

[0145] Step 6: Construct the multi-objective optimization model MaxF for maintenance decision-making. The specific expression is:

[0146] MaxF=a*F1+b*F2+c*F3 (6-1)

[0147] In the formula, a, b, and c are model coefficients respectively; among them, a>0, b<0, and c<0. When solving, they are first converted to the same order of magnitude, and then the weight coefficients involved are solved using the conventional AHP hierarchical analysis method.

[0148] F1 is the maintenance benefit value of all maintenance projects.

[0149] F2 is the maintenance carbon emission value of all maintenance projects.

[0150] F3 is the maintenance cost value of all maintenance items.

[0151] F1, F2 and F3 are all variables X ijt Function of X ijt It takes the value of 0 or 1. In the tth year of the maintenance plan implementation cycle T, when the i-th maintenance project implements the j-th type of maintenance measures determined in step 3, the value is 1, otherwise, the value is 0.

[0152] By solving MaxF, we can obtain the maintenance year of the jth maintenance measure determined in step 3 for each maintenance project within the maintenance plan implementation period T, and form a long-term road maintenance plan.

[0153] The calculation formulas for the above F1, F2 and F3 are:

[0154]

[0155]

[0156]

[0157] In the formula, w i is the priority weight of the ith maintenance item. According to the road grade, damage and other characteristics of the section, a certain maintenance priority weight is assigned to the section, including:

[0158] ① Road grade additional weight. The higher the road grade, the higher the road grade additional weight.

[0159] ② Additional weight of road surface damage: The higher the damage rate in the project, the higher the additional weight of road surface damage.

[0160] The final priority weight is calculated by adding the basic weight to the grade additional weight and the road damage additional weight, as follows:

[0161] Priority weight = basic weight + additional weight for road grade + additional weight for road damage

[0162] The above F1, F2 and F3 must also meet the following constraints:

[0163] F3≤C (6-5)

[0164] PCI′≥PCI m (6-6)

[0165] RQI′≥RQI m (6-7)

[0166] PQI′≥PQI m (6-8)

[0167]

[0168]

[0169] Where C is the upper limit of maintenance funds.

[0170] x ijtk It takes the value of 0 or 1. In the tth year within the maintenance plan implementation cycle T, when the kth section of the ith maintenance project implements the jth type of maintenance measures determined in step 3, the value is 1; otherwise, the value is 0.

[0171] PCI′ is PCI1,...,PCI t-1 , PCI'0, ..., PCI' T-t The minimum value in .

[0172] RQI′ is RQI1, ..., RQI t-1 , RQI'0, ..., RQI' T-t The minimum value in .

[0173] PQI′ is PQI1, ..., PQI t-1 , PQI'0, ..., PQI' T-t The minimum value in .

[0174] PCI m 、RQI m and PQI m They are the minimum acceptable levels of PCI, RQI and PQI for the kth section of the ith maintenance project within the decision-making cycle of T years.

[0175] L and U are the lower and upper limits of maintenance mileage in a single year respectively.

[0176] The preferred embodiments of the present invention are described in detail above; however, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all belong to the protection scope of the present invention.

Claims

1. A lane-level maintenance plan generation method integrating multi-step prediction of performance indicators, characterized by: The steps include: Step 1: Determine the implementation period T of the road maintenance plan in years; Step 2: Predicting the cross-section quality index, including the following steps: Step 2-1, select the base year: select one year as the base year from the existing historical section quality indicators of the road to be maintained, and use the section quality indicators corresponding to the base year as the base section quality indicators; the section quality indicators include the pavement quality index PQI, the pavement damage index PCI, the pavement ride quality index RQI and the damage rate DR; Step 2-2, constructing a section quality index prediction model: taking the year difference between the future year and the base year, the inherent parameters of the road to be maintained and the base section quality index as input, and taking the section quality index of the future year as output, constructing a section quality index prediction model; the inherent parameters of the road to be maintained include route code, driving direction, lane, city name, county name, road section technical grade, road section pavement type, terrain, roadbed width, pavement width, number of lanes and design speed; Step 2-3, prediction: using the section quality index prediction model constructed in step 2-2, and selecting the current year as the base year, the section quality index of each section of the road to be maintained is predicted in each of the next T years; Step 3: Determine maintenance measures: Based on the inherent parameters of the road to be maintained and the section quality indicators of each section in each year in the next T years, determine the maintenance measures for each section in each year within the maintenance plan implementation period T; maintenance measures include no maintenance, preventive maintenance, structural maintenance and functional maintenance; Step 4: Construct a maintenance project: According to the maintenance measures determined in step 3, the lane-level section data is merged horizontally and vertically each year to form a maintenance project. Step 5: Predict project indicators: Project indicators include maintenance benefits, maintenance costs and maintenance carbon emissions. For each maintenance project formed in step 4, calculate the project indicators separately. Suppose that in the tth year within the maintenance plan implementation cycle T, the ith maintenance project has m sections, then the maintenance benefit A of all sections of the ith maintenance project is it , maintenance cost it and maintain carbon emissions it The calculation formulas are: in: d itk l itk ×w itk ×DR itk (5-5) In the formula, A itk is the maintenance benefit of the kth section of the ith maintenance project in the tth year within the maintenance plan implementation cycle T; c j is the maintenance cost per unit area when the jth type of maintenance measure is adopted, a known quantity; e j is the carbon emission per unit area when the jth type of conservation measure is adopted, a known quantity; d itk is the maintenance area of ​​the kth section of the ith maintenance project in the tth year within the maintenance plan implementation cycle T; l itk is the length of the kth section of the ith maintenance project in the tth year within the maintenance plan implementation cycle T; w itk is the width of the kth section of the ith maintenance project in the tth year within the maintenance plan implementation cycle T; DR itk is the damage rate of the kth section of the ith maintenance project in the tth year within the maintenance plan implementation period T; PCI1, PCI2, PCI t-1 、PCI T They are the PCI values ​​of the first year, the second year, the t-1 year, and the T year in the future when the kth section of the i-th maintenance project is not maintained within T years; RQI1, RQI2, RQI t-1 、RQI T They are the RQI values ​​of the first year, the second year, the t-1 year, and the T year when the kth section of the i-th maintenance project is not maintained within T years; PQI1, PQI2, PQI t-1 、PQI T They are the PQI values ​​of the first year, the second year, the t-1 year, and the T year in the future when the kth section of the i-th maintenance project is not maintained within T years; PCI'0, PCI' T-t are the PCI values ​​of the kth section of the ith maintenance project in the future T year and the future t year after the kth section is maintained in the natural decay state in the tth year; RQI0', RQI' T-t are the RQI values ​​of the kth section of the ith maintenance project in the future T year and the future t year after the kth section is maintained in the natural decay state in the tth year; PQI'0, PQI' T-t are the PQI values ​​of the kth section of the ith maintenance project in the future T year and the future t year after the kth section is maintained in the natural decay state in the tth year; Step 6: Construct the multi-objective optimization model MaxF for maintenance decision-making. The specific expression is: MaxF=a*F1+b*F2+c*F3 (6-1) In the formula, a, b, c are model coefficients respectively; among which, a>0, b<0, c<0; F1 is the maintenance benefit value of all maintenance projects; F2 is the maintenance carbon emission value of all maintenance projects; F3 is the maintenance cost value of all maintenance items; F1, F2 and F3 are all variables X ijt Function of X ijt The value is 0 or 1. In the tth year of the maintenance plan implementation cycle T, when the i-th maintenance project implements the j-th type of maintenance measures determined in step 3, the value is 1, otherwise, the value is 0; By solving MaxF, we can obtain the maintenance year of the jth maintenance measure determined in step 3 for each maintenance project within the maintenance plan implementation period T, and form a long-term road maintenance plan.

2. The lane-level maintenance plan generation method integrating multi-step prediction of performance indicators according to claim 1 is characterized by: The specific method for selecting the base year in step 2-1 includes the following steps: Step 2-1A, select several groups of N consecutive years of section quality index data, match the N years of section quality index data based on the road name, driving direction, lane number and start and end pile number information of the road surface, and form several groups of N years of section quality index data time series data sets respectively; Step 2-1B, delete the abnormal data of indicator changes in the section quality indicator data time series data set, and form a new section quality indicator data time series data set with each group not less than T years; Step 2-1C: Select the year with the shortest time in each group of the new section quality index data time series data set as the base year.

3. The lane-level maintenance plan generation method integrating multi-step prediction of performance indicators according to claim 1 is characterized by: In step 4, the horizontal merging method is: select all lane data of the same starting and ending pile number section in the same direction of the same road, and select the most serious maintenance measure in each lane as the unified maintenance measure for the section; the vertical merging method is: select adjacent sections with the same maintenance measures and merge them.

4. The lane-level maintenance plan generation method integrating multi-step prediction of performance indicators according to claim 1 is characterized in that: In step 6, the calculation formulas for F1, F2 and F3 are: In the formula, w i is the priority weight of the i-th maintenance project, which is assigned according to the road grade and pavement damage.

5. The lane-level maintenance plan generation method integrating multi-step prediction of performance indicators according to claim 4 is characterized in that: In step 6, F1, F2 and F3 must also meet the following constraints: F3≤C (6-5) PCI′≥PCI m (6-6) RQI′≥RQI m (6-7) PQI′≥PQI m (6-8) In the formula, C is the upper limit of maintenance funds; x ijtk The value is 0 or 1. In the tth year of the maintenance plan implementation cycle T, when the kth section of the ith maintenance project implements the jth type of maintenance measures determined in step 3, the value is 1; otherwise, the value is 0. PCI′ is PCI1,...,PCI t-1 , PCI'0, ..., PCI' T-t The minimum value in ; RQI′ is RQI1, ..., RQI t-1 , RQI'0, ..., RQI' T-t The minimum value in ; PQI′ is PQI1, ..., PQI t-1 , PQI'0, ..., PQI' T-t The minimum value in ; PCI m 、RQI m and PQI m are the minimum acceptable levels of PCI, RQI and PQI for the kth section of the ith maintenance project within the decision cycle T years, respectively; L and U are the lower and upper limits of maintenance mileage in a single year respectively.

6. The lane-level maintenance plan generation method integrating multi-step prediction of performance indicators according to claim 1 is characterized by: In step 3, the specific method for determining maintenance measures includes the following steps: Step 3-1: Section not maintained: When the road to be maintained meets any of the following conditions A, B, and C, it is judged as not maintained, specifically: A. Road age ≤ 3 years; B. Road age > 3 years and PQI ≥ 90; C. Road age>3 years, 80≤PQI<90, road grade is national highway, road traffic load grade is non-heavy load and RQI≥92; Step 3-2, preventive maintenance: When the road age of the road to be maintained is>3 years, 80≤PQI<90, road grade is national highway, road traffic load grade is non-heavy load and 80<RQI<92; at this time, it is determined that the road to be maintained should be subject to preventive maintenance; Step 3-3, structural maintenance: When the road to be maintained meets any of the following conditions A, B, and C, it is judged as structural maintenance, specifically: A. Road age > 3 years and PQI < 70; B. Road age>3 years, 70≤PQI<80, road grade is not a national highway, road traffic load grade is not heavy load and RQI<70; C. Road age>3 years, 70≤PQI<80, road grade is national highway, road traffic load grade is heavy load, RQI<75; Step 3-4: Functional maintenance: For the remaining sections that do not meet the requirements of steps 3-1 to 3-3, functional maintenance shall be performed.

Images