Traffic Infrastructure Cluster Maintenance Planning Method, Device, Equipment and Storage Medium

Abstract

The present invention is applicable to the field of maintenance planning for transportation infrastructure clusters, and provides a method, device, equipment and storage medium for maintenance planning of transportation infrastructure clusters. The method includes: dividing the transportation infrastructure cluster to obtain a bottom-layer planning layer and at least one upper-layer planning layer, and according to the solution rule of bottom-up serial solution between layers and parallel solution within the same layer, using a multi-objective optimization algorithm to solve the multi-objective maintenance planning models of the basic units in all bottom-layer planning layers and the composite units in all upper-layer planning layers, so as to obtain the objective non-dominated solution set of the maintenance planning; wherein, the optimization variable of the multi-objective maintenance planning model of the composite unit is the index of the non-dominated solution set of the constituent units. This solution improves the optimization calculation efficiency of the maintenance planning, reduces resource consumption, and systematically realizes the association of maintenance planning results with different granularities in the transportation infrastructure cluster, improving the applicability of maintenance planning problems with various granularities.

Description

Technical Field

[0001] The present invention belongs to the field of facility cluster maintenance planning, and in particular relates to a transportation infrastructure cluster maintenance planning method, device, equipment and storage medium. Background Art

[0002] The service life of transportation infrastructure is as long as several decades or even hundreds of years. The coupling of disaster factors such as environmental erosion, material aging, and long-term effects of loads, fatigue effects, and mutation effects will inevitably lead to the accumulation of structural damage and attenuation of resistance, thereby reducing the ability to resist natural disasters and even normal environmental effects, and in extreme cases causing catastrophic accidents. Therefore, in order to ensure the safety, applicability and durability of the structure, the transportation infrastructure that has been built and used urgently needs to adopt effective means to assess its safety status, repair and control damage.

[0003] However, the maintenance of transportation infrastructure during the operation period involves two major conflicting aspects: safety and economy. How to ensure the operational safety of the growing facilities with limited resources is an urgent problem to be solved.

[0004] From the current application status, most of the early planning adopted a passive maintenance strategy, that is, maintenance actions were triggered after the performance reached a certain threshold. However, with the dual influence of the increase in facility volume and limited maintenance resources, one of the main solutions is to use quantitative indicators and the evolution law of structural performance to achieve scientific maintenance decisions during the facility operation period. This solution mainly includes performance prediction and decision-making. There are two main ways to predict the evolution of structural performance. One is to predict the trend of structural performance changes through historical statistical data of structural degradation. The other is to analyze the time-varying laws of factors affecting degradation and construct a multi-factor performance degradation evolution model. Commonly used prediction methods include experimental methods, deterministic curve models (such as regression model methods, time series methods, etc.), random degradation models (such as Markov chain models, methods based on reliability theory, etc.) and deep learning models. These models can all achieve structural performance prediction in specific scenarios. Based on the law of performance evolution, planning problems involving economic indicators, safety indicators and other objectives are constructed. Finding the optimal solution through optimization algorithms is a common way to solve decision-making problems. One of the solutions is decision optimization of multi-attribute utility target values ​​based on single-objective genetic algorithms, but this method cannot match the management preferences of multiple users. Although the current multi-objective optimization solution concept solves the problem of a single recommended solution, it does not open up the data link between facility-level and network-level planning. The planning of the network level and facility level is independent. In summary, the above methods have the problems of low calculation efficiency and high resource consumption in maintenance planning optimization, and cannot systematically solve the problem of the association of maintenance planning results of different granularities in transportation infrastructure clusters. Summary of the invention

[0005] The object of the present invention is to provide a method, device, equipment and storage medium for maintenance planning of transportation infrastructure clusters, aiming to solve the problems that in the prior art, the optimization calculation efficiency of maintenance planning of transportation infrastructure clusters is low, resource consumption is high, and the association of maintenance planning results with different granularities in transportation infrastructure clusters cannot be systematically solved.

[0006] On the one hand, the present invention provides a method for maintenance planning of transportation infrastructure clusters, and the method includes:

[0007] Perform multi-level and multi-granularity division on the transportation infrastructure cluster to obtain a bottom-layer planning layer and at least one upper-layer planning layer. Among them, the planning units in the bottom-layer planning layer are basic units, and the planning units in each upper-layer planning layer are composite units. Each planning layer progresses layer by layer. For two adjacent planning layers, the composite units in the upper planning layer are a set composed of the constituent units in the lower planning layer, and the constituent units are basic units or composite units that make up the composite unit;

[0008] According to the solution rule of serial solution from bottom to top between layers and parallel solution within the same layer, use a multi-objective optimization algorithm to solve the established multi-objective maintenance planning models of all planning units to obtain the target non-dominated solution set of the maintenance planning of the transportation infrastructure cluster; among them, the optimization variable of the multi-objective maintenance planning model of the composite unit is the index of the non-dominated solution set of the constituent unit.

[0009] Optionally, the optimization variable of the multi-objective maintenance planning model of each basic unit is the maintenance measures to be taken in each time period within the preset maintenance planning cycle;

[0010] The objective functions of the multi-objective maintenance planning models of each planning unit all include a safety index, an economic index and a maintenance frequency index; among them, the safety index is the improvement amplitude of the performance after taking maintenance measures within the preset maintenance planning cycle, the economic index is the total cost of taking maintenance measures within the preset maintenance planning cycle, and the maintenance frequency index is the number of maintenance measures taken within the preset maintenance planning cycle.

[0011] Optionally, the safety index of the basic unit is determined according to the difference between the integral area of the performance improvement prediction curve after taking maintenance measures and the integral area of the initial performance prediction curve without taking maintenance measures;

[0012] The safety index of the composite unit is determined according to the weighted sum of the calculated values of the safety index in the non-dominated solutions of the constituent units.

[0013] Optionally, the performance of each of the basic units before and after maintenance measures are taken is predicted using an adaptive performance prediction model; wherein, the parameters of the adaptive performance prediction model are updated using a preset update mechanism after new performance evaluation data is obtained.

[0014] Optionally, the adaptive performance prediction model is an adaptive exponential performance degradation model based on a Bayesian update mechanism.

[0015] Optionally, the method further includes:

[0016] Obtain m current structural condition levels of each of the basic units before maintenance measures are taken and n improved structural condition levels after maintenance measures are taken;

[0017] Establish a cost-effectiveness matrix based on the maintenance measures corresponding to each group of current structural condition levels and improved structural condition levels, the utility corresponding to each maintenance measure, and the cost;

[0018] The intercept difference between the improved performance prediction curve and the initial performance prediction curve is determined according to the utility corresponding to the maintenance measures taken, and the utility corresponding to the maintenance measures taken is determined according to the cost-effectiveness matrix.

[0019] Optionally, the maintenance measures in the cost-effectiveness matrix are represented by measure codes, and the maintenance measures in the optimization variables of the multi-objective maintenance planning model of each of the basic units are represented by the measure codes;

[0020] The utility corresponding to different maintenance measures is determined according to the utility interval of the corresponding improved structural condition level and the level span before and after the maintenance measures are taken;

[0021] The cost of different maintenance measures is obtained by multiplying the corresponding scale factor by a preset cost.

[0022] On the other hand, the present invention provides a traffic infrastructure cluster maintenance planning device, the device includes:

[0023] Hierarchical division module: used to perform multi-level and multi-granularity division on the traffic infrastructure cluster to obtain a bottom-layer planning layer and at least one upper-layer planning layer, wherein the planning units in the bottom-layer planning layer are basic units, and the planning units in each upper-layer planning layer are composite units. Each planning layer progresses layer by layer. For two adjacent planning layers, the composite units in the upper planning layer are a set composed of the constituent units in the lower planning layer, and the constituent units are the basic units or composite units that make up the composite unit; and

[0024] A multi-objective optimization module is used to solve the multi-objective maintenance planning models of all established planning units by using a multi-objective optimization algorithm according to the solution rules of bottom-up serial solution between layers and parallel solution within the same layer, so as to obtain the non-dominated solution set of the objectives of the maintenance planning of the transportation infrastructure cluster; wherein, the optimization variables of the multi-objective maintenance planning model of the composite unit are the indexes of the non-dominated solution sets of the constituent units.

[0025] On the other hand, the present invention provides a transportation infrastructure cluster maintenance planning device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned method are implemented.

[0026] On the other hand, the present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method are implemented.

[0027] The present invention divides the transportation infrastructure cluster into multiple levels and multiple granularities to obtain a bottom-level planning layer and at least one upper-level planning layer. According to the solution rules of bottom-up serial solution between layers and parallel solution within the same layer, a multi-objective optimization algorithm is used to solve the multi-objective maintenance planning models of the basic units in all bottom-level planning layers and the composite units in all upper-level planning layers that have been established, so as to obtain the non-dominated solution set of the objectives of the maintenance planning of the transportation infrastructure cluster. Among them, the optimization variables of the multi-objective maintenance planning model of each composite unit are the indexes of the non-dominated solution sets of the constituent units, thereby improving the optimization calculation efficiency of the maintenance planning, reducing resource consumption, and systematically realizing the association of the maintenance planning results of different granularities in the transportation infrastructure cluster, and improving the applicability of the maintenance planning problems of various granularities. Description of the Drawings

[0028] Figure 1A is the implementation flowchart of the transportation infrastructure cluster maintenance planning method provided in Embodiment 1 of the present invention;

[0029] Figure 1B is the schematic diagram of the optimization calculation process of the transportation infrastructure cluster maintenance planning provided in Embodiment 1 of the present invention;

[0030] Figure 1C is the performance prediction curve corresponding to before and after the update of the adaptive exponential performance degradation model based on the Bayesian update mechanism provided in Embodiment 1 of the present invention;

[0031] Figure 1D is the example diagram of the performance improvement amplitude provided in Embodiment 1 of the present invention;

[0032] Figure 1EIt is an example diagram of the measure coding part of the cost-effectiveness matrix provided in the first embodiment of the present invention;

[0033] Figure 1F It is an example diagram of the optimization calculation process of the transportation infrastructure cluster provided in the first embodiment of the present invention;

[0034] Figure 2 It is a schematic structural diagram of the maintenance planning device for the transportation infrastructure cluster provided in the second embodiment of the present invention; and

[0035] Figure 3 It is a schematic structural diagram of the maintenance planning equipment for the transportation infrastructure cluster provided in the third embodiment of the present invention. Detailed implementation manners

[0036] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0037] The following describes the specific implementation of the present invention in detail with reference to specific embodiments.

[0038] Terms and definitions:

[0039] Maintenance planning: Aiming at structural safety and maintenance economy, under the constraints such as structural performance requirements and maintenance funds limitations, reasonable maintenance timing and maintenance actions are proposed to achieve the optimal allocation of maintenance funds during the planning period.

[0040] Basic unit: In multi-granularity maintenance planning, the basic unit refers to the smallest and indivisible planning unit, and the planning result of the basic unit is the basis for the planning result of the composite unit.

[0041] Composite unit: In multi-granularity maintenance planning, the composite unit refers to a meaningful set composed of basic units or composite units with smaller granularity. The planning result of the composite unit is the synthesis of the planning results of the planning units it contains.

[0042] Structural condition grade: It refers to the structural technical condition grade evaluated according to the regular inspection data or the structural comprehensive evaluation grade obtained by integrating multi-source data.

[0043] Non-inferior solution: The result of a multi-objective optimization problem is a set of non-inferior solutions. For any non-inferior solution, no possible solution that is better in all optimization objectives can be found.

[0044] Non - dominated solution set: In multi - objective optimization, the non - dominated solution set (Pareto Optimal Set, also known as the Pareto front) refers to a set of solutions where each solution cannot be further optimized without sacrificing other objectives. In the non - dominated solution set, any improvement to a solution will result in the degradation of at least one objective. Therefore, these solutions form a balance among all objectives.

[0045] Example 1:

[0046] Figure 1A The implementation process of the traffic infrastructure cluster maintenance and management planning method provided by the first embodiment of the present invention is shown. For the sake of convenience, only the parts related to the first embodiment of the present invention are shown and are described in detail as follows:

[0047] In step S101, the traffic infrastructure cluster is divided at multiple levels and with multiple granularities to obtain a bottom - layer planning layer and at least one upper - layer planning layer. Among them, the planning units in the bottom - layer planning layer are basic units, and the planning units in each upper - layer planning layer are composite units. Each planning layer progresses layer by layer. For two adjacent planning layers, the composite unit in the upper - layer planning layer is a set composed of the constituent units in the lower - layer planning layer. The constituent unit is a basic unit or a composite unit that constitutes the composite unit.

[0048] The embodiments of the present invention are applicable to the maintenance and management planning of traffic infrastructure clusters. In the embodiments of the present invention, the traffic infrastructure cluster can be divided into multiple levels, specifically including a bottom - layer planning layer and an upper - layer planning layer. Different planning layers are composed of planning units with different granularities. Among them, the planning units in the bottom - layer planning layer are defined as basic units, and the planning units in the upper - layer planning layer are defined as composite units, that is, the planning unit is a basic unit or a composite unit. In two adjacent planning layers, the composite unit in the upper - layer planning layer is a set composed of the constituent units in the lower - layer planning layer, and the constituent unit is a basic unit or a composite unit that constitutes the composite unit. For example, in the maintenance and management planning of a bridge group, components such as main girders and bridge piers are respectively used as basic units, and a certain bridge containing the above - mentioned components or a bridge cluster containing multiple bridges can be used as a composite unit. The maintenance and management planning result of each composite unit can be the optimal combination obtained by screening and integrating the planning results of the planning objects it contains.

[0049] In step S102, according to the solution rule of serial solution from bottom to top between layers and parallel solution within the same layer, a multi - objective optimization algorithm is used to solve the established multi - objective maintenance and management planning models of all planning units to obtain the target non - dominated solution set of the traffic infrastructure cluster maintenance and management planning; among them, the optimization variable of the multi - objective maintenance and management planning model of each composite unit is the index of the non - dominated solution set of the constituent units.

[0050] In the embodiments of the present invention, the optimization variables of the multi-objective maintenance planning model of each composite unit are the indices of the non-dominated solution sets of all constituent units of the composite unit, and the target non-dominated solution set is the non-dominated solution set corresponding to the top-level composite unit. The multi-objective optimization algorithm can adopt multi-objective optimization algorithms such as multi-objective genetic algorithms and particle swarm optimization algorithms, and this embodiment does not make specific limitations thereon.

[0051] As an example, as Figure 1B shown, the transportation infrastructure cluster includes a bottom planning layer and upper planning layers L1 and L2. The bottom planning layer contains multiple basic units A 111 , A 112 , A 113 ..., the upper planning layer L1 contains composite units A 11 ..., A 1N , the upper planning layer L2 contains composite unit A 1 , composite unit A 11 contains basic units A 111 , A 112 , A 113 ..., composite unit A 1N is the set composed of basic units A 1N1 , A 1N2 , A 1N3 ..., composite unit A 1 is the set composed of composite units A 11 ..., A 1N . When solving the multi-objective maintenance planning models of the established planning units, first, the multi-objective maintenance planning models of each basic unit A 111 , A 112 , A 113 ... are optimized and solved in parallel. Then, the multi-objective maintenance planning models of the composite units A 11 ..., A 1N in the upper planning layer L1 are optimized and solved in parallel. Finally, the multi-objective maintenance planning model of composite unit A 1 is optimized and solved to obtain the target non-dominated solution set. Among them, the optimization variables of the multi-objective maintenance planning model of composite unit A 11 are the indices of the non-dominated solution sets of the multi-objective maintenance planning models of basic units A 111 , A 112 , A 113 ..., and the optimization variables of the multi-objective maintenance planning model of composite unit A 1N are the indices of the non-dominated solution sets of the multi-objective maintenance planning models of basic units A 1N1 , A 1N2 , A 1N3 ..., and composite unit A 1The optimization variable of the multi-objective maintenance planning model is the composite unit A 11 ,..., A 1N The index of the non-dominated solution set of the multi-objective maintenance planning model, and the objective non-dominated solution set is for the top-level composite unit A 1 The non-dominated solution set obtained by solving the multi-objective maintenance planning model.

[0052] Considering that the maintenance planning is essentially an optimization problem, the structural safety and maintenance economy can be balanced by adjusting the maintenance timing and measures. However, different users have different preferences for safety and economy. To meet the needs of different users and solve the problems of low computational efficiency and weak scientificity of the maintenance decision-making for transportation infrastructure clusters, optionally, the optimization variable of the multi-objective maintenance planning model for each basic unit is the maintenance measures to be taken in each period within the preset maintenance planning cycle. The objective functions of the multi-objective maintenance planning models for each basic unit and each composite unit all include safety indicators, economic indicators, and maintenance frequency indicators. That is, a multi-objective optimization algorithm with structural safety, maintenance economy, and maintenance frequency as optimization objectives is adopted in the maintenance planning. The maintenance planning result is a set of maintenance strategies that cannot be further optimized simultaneously in terms of structural safety, maintenance economy, and maintenance frequency. Thus, through the strategy of multi-objective hierarchical optimization, multiple optimization objectives such as structural performance, maintenance cost, and maintenance frequency are comprehensively considered, reducing the number of optimization variables, improving the computational efficiency, making the business logic smoother, and enhancing the business meaning and scientificity of the optimization result. Among them, the safety indicator can be the improvement amplitude of the performance after taking maintenance measures within the preset maintenance planning cycle, the economic indicator can be the total cost of taking maintenance measures within the preset maintenance planning cycle, and the maintenance frequency indicator can be the number of maintenance measures taken within the preset maintenance planning cycle. In specific implementation, the safety indicator of the basic unit can be determined according to the utility difference before and after taking measures, and the safety indicator of the composite unit can be the sum of the calculated values of the corresponding safety indicators in the non-dominated solutions of all constituent units of the composite unit; the economic indicator of the basic unit can be the sum of the costs of all maintenance measures taken, and the economic indicator of the composite unit can be the sum of the calculated values of the economic indicators in the non-dominated solutions of all constituent units of the composite unit; the maintenance frequency indicator of the basic unit can be the sum of the number of all maintenance measures taken, and the maintenance frequency indicator of the composite unit can be the sum of the calculated values of the maintenance frequency indicators in the non-dominated solutions of all constituent units of the composite unit.

[0053] Optionally, the performance of each basic unit before and after maintenance measures are taken is predicted using an adaptive performance prediction model. By updating the parameters of the adaptive performance prediction model, the model can gradually approximate the true law of performance degradation of each basic unit, effectively reducing the dependence of the adaptive performance prediction model on historical data and improving the accuracy of the prediction results. Among them, the parameters of the adaptive performance prediction model are updated using a preset update mechanism after obtaining new performance evaluation data. Optionally, the update mechanism is Bayesian update.

[0054] Optionally, the adaptive performance prediction model is an adaptive exponential performance degradation model based on the Bayesian update mechanism, so as to realize the adaptive update of the adaptive performance prediction model through the Bayesian update technology to fit the performance degradation law of the current structure. In specific implementation, the application of this model can involve two steps: initialization and adaptive update. Taking civil engineering as an example, by analyzing the performance evolution law of a large number of structures in civil engineering, it can be found that the performance of the structure degrades slowly in the early stage and the degradation rate gradually accelerates with the increase of the operation time. This law conforms to the evolution law of the exponential model. The expression of the adaptive exponential performance degradation model is as follows:

[0055]

[0056] Where the parameter represents the intercept, and the parameters and represent the random variables of the exponential term parameters, represents the random error, with a mean of 0 and a variance of .

[0057] When the target cluster facility is just put into operation, the accumulation of historical evaluation data is insufficient. To make the model available, the model parameters can be initialized through the initial performance IP and expected service life EL of the structure, without considering the random error term of the model. Assuming the model parameters , the core parameters of the model are calculated as follows:

[0058]

[0059]

[0060] In the formula represents the lower limit of the structure performance. To make the model have the ability of adaptive update, the Bayesian update technology can be used. Each time new evaluation information is added, the parameters of the model can be updated, and the above model is transformed into:

[0061]

[0062] Let , , then the above equation is transformed into:

[0063]

[0064] According to the Bayesian update theory, when there is new structural state information, that is known, the updated parameters and , namely:

[0065]

[0066] According to Equation , ; assume that follows a joint normal distribution, and its prior mean and variance are , and , , and the correlation coefficient is . After derivation, the posterior mean, variance, and correlation coefficient are obtained as:

[0067]

[0068]

[0069]

[0070] In the formula represents the posterior variance of the parameter , represents the posterior variance of the parameter , represents the posterior mean of the parameter , represents the posterior mean of the parameter , represents the posterior correlation coefficient of the parameter . The process of model initialization and update is as shown in Figure 1C . It can be seen from Figure 1C that compared with the initial performance curve, the updated performance curve fits the measured points of the structural performance, that is, it fits the performance degradation law of the current structure.

[0071] Optionally, the safety index of the basic unit is determined according to the difference between the integral area of the improved performance prediction curve after taking maintenance measures and the integral area of the initial performance prediction curve without taking maintenance measures, so as to calculate the safety index through the performance prediction curve.

[0072] Optionally, the safety index of the composite unit is determined according to the weighted sum of the calculated values of the safety indices in the non-inferior solutions of the constituent units to achieve the calculation of the safety index of the composite unit. In specific implementation, considering the different importance of each constituent unit, the weighting of safety for different constituent units is different.

[0073] Further optionally, the intercept of the improved performance prediction curve is different from that of the initial performance prediction curve. The intercept difference between the improved performance prediction curve and the initial performance prediction curve is determined according to the difference between the utilities before and after the maintenance measures are taken. The utility after the maintenance measures are taken is determined according to the established cost-effectiveness matrix, so as to achieve the calculation of the safety index through the improved utility after the maintenance measures are taken. In specific implementation, the current structural condition can be input into the adaptive performance prediction model to obtain the initial performance prediction curve without taking maintenance measures. The adaptive performance prediction model after taking maintenance measures can be obtained by changing the parameter in the adaptive performance prediction model that represents the intercept of the corresponding curve of the model. The parameter representing the intercept of the corresponding performance prediction curve of the model can be determined according to the difference between the utility value after taking maintenance measures and the utility value before taking maintenance measures. As an example, as Figure 1D shown, the intercept of the performance prediction curve changes after the maintenance measures are executed, and the corresponding gray area is the amplitude of performance improvement.

[0074] When establishing the above-mentioned cost-effectiveness matrix, considering that it is difficult to accurately define the cost and utility of the maintenance measures, due to the influence of factors such as the form of the basic unit structure and the environment where it is located, it is difficult to accurately quantify the cost and utility of specific maintenance measures to achieve the goal of application and implementation. And the specific details of the maintenance measures are not the focus of the plan. Optionally, m current structural condition levels of each basic unit without taking maintenance measures and n improved structural condition levels after taking corresponding maintenance measures are obtained. According to the maintenance measures corresponding to each group of current structural condition levels and improved structural condition levels, the utilities and costs corresponding to each maintenance measure, a cost-effectiveness matrix is established. This cost-effectiveness matrix abstracts the maintenance measures as the improvement of the structural condition level, which is convenient for accurately defining the cost and utility of the maintenance measures and is not affected by the specific structural working environment and measure type, meeting the goals of medium- and long-term plans that should focus on capital allocation, maintenance timing, and maintaining the structural safety performance. Among them, each element in the cost-effectiveness matrix can be a three-dimensional vector matrix including the maintenance measure, the utility corresponding to the maintenance measure, and the cost, or the cost-effectiveness matrix includes a maintenance measure matrix, a utility matrix, and a cost matrix respectively.

[0075] Further optionally, the maintenance measures in the cost-effectiveness matrix are represented by measure codes, and the maintenance measures in the optimization variables of the multi-objective maintenance planning model for each basic unit are represented by measure codes, so as to improve the optimization calculation efficiency of the maintenance planning while solving the problem of quantifying the effectiveness corresponding to the maintenance measures. In specific implementation, sequential coding can be performed on the measures corresponding to each group of current structural condition levels and the levels for improving the structural condition. As an example, as Figure 1E shown, the current structural condition level and the level for improving the structural condition are each divided into 5 levels. The measure code corresponding to the improvement from the current structural condition level 1 to the improved structural condition level 1 is 1, the measure code corresponding to the improvement from the current structural condition level 2 to the improved structural condition level 1 is 2, and so on. 15 maintenance measures correspond to 15 codes.

[0076] The effectiveness corresponding to different maintenance measures can be determined according to the increase value of the effectiveness after adopting the maintenance measure, or can be determined according to the effectiveness value corresponding to the improved structural condition level. Optionally, the effectiveness corresponding to different maintenance measures is determined according to the effectiveness interval corresponding to the improved structural condition level and the level span before and after taking the maintenance measure, so as to improve the accuracy of the quantified data while realizing the quantification of the effectiveness. In specific implementation, each improved structural condition level can correspond to an effectiveness interval. If the structural condition level is improved to the same level, the effectiveness corresponding to the maintenance measure is the first preset value within the effectiveness interval corresponding to the improved structural condition level. If the condition level is improved across levels, the effectiveness corresponding to the maintenance measure is the second preset value within the effectiveness interval corresponding to the improved structural condition level. Taking a highway bridge as the maintenance object, referring to the correspondence between the grades and scores in the "Technical Condition Assessment Standard for Highway Bridges" JTG TH21-2011, for category 1 [95, 100], category 2 [80, 95), category 3 [60, 80), category 4 [40, 60), category 5 [0, 40). For the improvement to the same level, it is considered that after this maintenance measure, the structural effectiveness will reach the upper limit of this level. For example, when improving from the current structural state level 1 to the improved structural condition level 1, the effectiveness after improvement is 100. For the improvement across levels, it is considered that after this maintenance measure, the structural effectiveness will reach the middle of the effectiveness interval corresponding to this level. For example, when improving from the current structural state level 2 to the improved structural condition level 1, the effectiveness after improvement is 97.5.

[0077] Optionally, the costs of different maintenance measures are obtained by multiplying the corresponding proportional factor by the preset cost. In specific implementation, the cost values of each maintenance measure are determined according to the costs of each maintenance measure. Taking civil engineering as an example, it can be obtained by multiplying the construction and installation costs of each structure by the corresponding proportional factor. It should be noted here that the costs in the cost-effectiveness matrix can be represented by specific cost values or by the proportional factors for measure charging. This embodiment does not make specific limitations.

[0078] Figure 1FIt is an example diagram of the optimization calculation process for the hierarchical maintenance planning of a transportation infrastructure cluster. Figure 1F The upper-middle planning layer only contains one composite unit. The following combines Figure 1F as shown, and explains the solution process of the multi-objective maintenance planning model for basic units and composite units:

[0079] The purpose of the maintenance planning for basic units is to determine the types of maintenance measures and the implementation timing, and then determine the costs during the planning period and the performance changes of the structure. The maintenance planning problem at this level is transformed into an optimization problem of which measures to implement each year during the planning period. Assuming that each basic unit will only implement one measure in a certain year, a multi-objective genetic algorithm is used to optimize and obtain the non-dominated solution set under multiple objectives of safety, economy, and the number of maintenance times under certain constraints. The algorithm is executed according to the following steps:

[0080] Step 1: Determine the planning period and optimization variables of each basic unit such as the main girder and the cable-stayed cable system. The planning period generally can be taken in the range of 10 to 20 years. The maintenance measures within each year can select any integer from 0 to 15, where 0 means no measures are implemented in that year, and 1 to 15 are the above measure codes. During initialization, randomly generate the measure codes for each year to realize the initialization of individuals.

[0081] Step 2: Calculate the objective functions of each individual. When calculating, a multi-objective function is used. Among them, safety is defined as the amplitude of performance improvement after taking maintenance measures; economy is defined as the cost of maintenance measures. Obtain the cost of a certain measure according to the cost-effectiveness matrix, and then calculate the total cost during the planning period; the number of maintenance times is defined as the total number of maintenance measures implemented during the planning period.

[0082] Step 3: Judge whether there are unreasonable situations in the chromosomes of individuals. If there are situations that do not meet the constraints (such as the minimum performance constraint) or the measures taken are unreasonable (such as the current structure condition level is 2, and the measure taken is from level 3 to level 1), when there are unreasonable situations, correct the individuals in a timely manner.

[0083] Step 4: Take the corrected multiple individuals as the initial population. For the iterative processes of the second generation and subsequent generations, it is necessary to combine the excellent individuals of the parent generation and the individuals of the offspring to form a new population.

[0084] Step 5: Perform a fast non-dominated sorting on the population. The fast non-dominated sorting is used to calculate the Pareto front under multiple objectives. Compare the performance of individual P and individual Q under multiple objectives. If all objectives of P are not weaker than Q, it is said that P dominates Q, and vice versa, Q dominates P. Form the first layer of the front with individuals whose domination number is 0. After removing the first layer of the front, iterate in turn to obtain the fronts of each layer.

[0085] Step 6: Calculate the crowding degree of each individual (i.e., the sum of the crowding distances of multiple objectives). The crowding distance is defined as the distance between the target individual and its adjacent individuals under a certain objective, and the formula is as follows:

[0086]

[0087] where represents the crowding distance of the i-th individual under the m-th objective, represents the (i + 1)-th individual, and represent the boundary individuals, represents the value of the m-th objective of the (i + 1)-th individual.

[0088] Step 7: Select excellent individuals to perform crossover to generate offspring individuals, and set a certain mutation rate to mutate some individuals to generate an offspring population.

[0089] Step 8: Merge the parent population and the offspring population, and repeat Steps 2 to 7 until the number of iterations meets the set limit, then output the non-dominated solution set of the Pareto front.

[0090] The planning of the composite unit is based on the non-dominated solution set output by its constituent units. Optimized variables are constructed, and the multi-objective genetic algorithm is called to calculate the non-dominated solution set for the planning of the composite unit. The multi-objective genetic algorithm for the composite unit is executed according to the following steps:

[0091] Step 1: Determine the optimized variables for the maintenance planning object of the composite unit. The maintenance planning period of the composite unit is the same as that of the basic unit. The optimized variables take the indices of the non-dominated solution sets of all basic units included in the composite unit. For example, the first value of the optimized variable of the composite unit takes the index of the non-dominated solution set of the main beam of the basic unit, and the value range is [0, set_len - 1], where set_len is the length of the non-dominated solution set, and all values are integers. And so on, the optimized variables of the composite unit are obtained.

[0092] Step 2: Calculate the objective function of the individual. The objectives of the composite unit are the same as those of the basic unit, both controlling the objectives of safety, economy, and maintenance frequency. The economy and maintenance frequency are the sum of the corresponding objectives of the non-dominated solutions of the constituent units of the composite unit. Since the importance of different basic units is different, the weighting of safety for different basic units varies. Taking a bridge as an example, referring to the part and component weights in the "Technical Condition Assessment Standard for Highway Bridges" JTG TH21-2011, the weight values of each basic unit are calculated. For example, the weight of the superstructure is 0.4, the weight of the main beam is 0.25. If the upper-layer node is a facility, then its weight is 0.4×0.25 = 0.1, and so on. For the weights of each facility at the regional road network level, the weights can be determined by comprehensively considering various factors such as the importance of the facility and the vulnerability of the structure, and then the safety score is calculated.

[0093] Steps 3 to 8 are the same as above. Finally, the maintenance planning scheme of the composite unit is output. The output of the scheme is a non-dominated solution set, providing multiple alternative solutions for different users. If the facility volume of the regional road network is large, on this basis, the region can be divided into pieces for optimization, and then the facilities within a piece of the region are divided, and finally a four-layer optimization scheme of basic unit i.e., component node → facility → piece of region → region is formed. This hierarchical scheme has strong scalability, high calculation efficiency, and reasonable results, and has good application value.

[0094] In an embodiment of the present invention, by performing multi-level and multi-granularity partitioning on a traffic infrastructure cluster, a bottom-layer planning layer and at least one upper-layer planning layer are obtained. According to the solution rule of serial solution from bottom to top between layers and parallel solution within the same layer, a multi-objective optimization algorithm is used to solve the multi-objective maintenance planning model of the basic units in all bottom-layer planning layers and the composite units in all upper-layer planning layers that have been established, so as to obtain the non-dominated solution set of the objectives of the traffic infrastructure cluster maintenance planning. Among them, the optimization variable of the multi-objective maintenance planning model of each composite unit is the index of the non-dominated solution set of the constituent units, thereby improving the optimization calculation efficiency of the maintenance planning, reducing resource consumption, and systematically realizing the association of maintenance planning results with different granularities in the traffic infrastructure cluster, and improving the applicability of maintenance planning problems with various granularities; in the maintenance planning, a multi-objective optimization algorithm with structural safety and maintenance economy as optimization objectives is adopted, and the optimization result is a set of maintenance strategies that cannot be further optimized simultaneously in terms of both structural safety and maintenance economy; the performance prediction model in this solution adopts an adaptive performance prediction model based on a preset update mechanism. Through the update of model parameters, the model can continuously approximate the true law of the performance degradation of each unit, and the model has little dependence on historical data and high prediction accuracy; the cost-effectiveness matrix adopted in this solution abstracts the maintenance measures as the improvement of the structural condition level, solves the problem that it is difficult to accurately define the cost and utility of the measures, and the cost and utility of the maintenance measures are not affected by the specific unit working environment and measure type, meeting the goals of medium- and long-term planning that should focus on capital allocation, maintenance timing, and maintaining structural safety performance.

[0095] Embodiment 2:

[0096] Figure 2 The structure of the traffic infrastructure cluster maintenance planning device provided in Embodiment 2 of the present invention is shown. For the convenience of description, only the parts related to Embodiment 2 of the present invention are shown, including:

[0097] The hierarchical partitioning module 21: is used to perform multi-level and multi-granularity partitioning on a traffic infrastructure cluster to obtain a bottom-layer planning layer and at least one upper-layer planning layer. Among them, the planning units in the bottom-layer planning layer are basic units, and the planning units in each upper-layer planning layer are composite units. Each planning layer progresses layer by layer. For two adjacent planning layers, the composite units in the upper planning layer are a set composed of the constituent units in the lower planning layer, and the constituent units are the basic units or composite units that make up the composite unit; and

[0098] The multi-objective optimization module 22 is configured to solve the multi-objective maintenance planning models of all the established planning units by using a multi-objective optimization algorithm according to the solution rules of bottom-up serial solution between layers and parallel solution within the same layer, so as to obtain the non-dominated solution set of the objectives of the maintenance planning of the transportation infrastructure cluster; wherein, the optimization variables of the multi-objective maintenance planning model of the composite unit are the indexes of the non-dominated solution sets of the constituent units.

[0099] In the embodiments of the present invention, each module of the transportation infrastructure cluster maintenance planning device can be implemented by corresponding hardware or software modules. Each module can be an independent software or hardware module, or can be integrated into a software or hardware module, which is not used to limit the present invention here. The specific implementation manners of each module of the transportation infrastructure cluster maintenance planning device can refer to the description of the foregoing method embodiments, and will not be elaborated here.

[0100] Embodiment Three:

[0101] Figure 3 The structure of the transportation infrastructure cluster maintenance planning device provided in Embodiment Three of the present invention is shown. For the sake of convenience of description, only the parts related to the embodiments of the present invention are shown.

[0102] The transportation infrastructure cluster maintenance planning device 3 in the embodiments of the present invention includes a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, the steps in the foregoing method embodiments are implemented, such as Figure 1A the steps S101 to S102 shown. Alternatively, when the processor 30 executes the computer program 32, the functions of each module in the foregoing device embodiments are implemented, such as Figure 2 the functions of the modules 21 to 22 shown.

[0103] In the embodiments of the present invention, by performing multi-level and multi-granularity partitioning on the transportation infrastructure cluster, a bottom-layer planning layer and at least one upper-layer planning layer are obtained. According to the solution rules of bottom-up serial solution between layers and parallel solution within the same layer, a multi-objective optimization algorithm is used to solve the multi-objective maintenance planning models of the basic units in all the bottom-layer planning layers and the composite units in all the upper-layer planning layers that have been established, so as to obtain the non-dominated solution set of the objectives of the maintenance planning of the transportation infrastructure cluster. Among them, the optimization variables of the multi-objective maintenance planning model of each composite unit are the indexes of the non-dominated solution sets of the constituent units, thereby improving the optimization calculation efficiency of the maintenance planning, reducing resource consumption, and systematically realizing the association of the maintenance planning results of different granularities in the transportation infrastructure cluster, and improving the applicability of the maintenance planning problems of various granularities.

[0104] The traffic infrastructure cluster maintenance and planning device in the embodiments of the present invention can be a computer, a server, etc. The steps implemented when the processor 30 in the traffic infrastructure cluster maintenance and planning device 3 executes the computer program 32 to implement the traffic infrastructure cluster maintenance and planning method can refer to the description of the foregoing method embodiments and will not be elaborated here.

[0105] Embodiment 4:

[0106] In the embodiments of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the foregoing method embodiments are implemented. For example, Figure 1A the steps S101 to S102 shown. Alternatively, when the computer program is executed by a processor, the functions of each module in the foregoing device embodiments are implemented. For example Figure 2 the functions of the units 21 to 22 shown.

[0107] In the embodiments of the present invention, by performing multi-level and multi-granularity partitioning on the traffic infrastructure cluster, a bottom-layer planning layer and at least one upper-layer planning layer are obtained. According to the solution rule of serial solution from bottom to top between layers and parallel solution within the same layer, a multi-objective optimization algorithm is used to solve the multi-objective maintenance and planning model of the basic units in all bottom-layer planning layers and the composite units in all upper-layer planning layers that have been established, and a non-dominated solution set of the objectives of the traffic infrastructure cluster maintenance and planning is obtained. Among them, the optimization variable of the multi-objective maintenance and planning model of each composite unit is the index of the non-dominated solution set of the constituent units, thereby improving the optimization calculation efficiency of the maintenance and planning, reducing resource consumption, and systematically realizing the association of maintenance and planning results with different granularities in the traffic infrastructure cluster, and improving the applicability of maintenance and planning problems with various granularities.

[0108] The computer-readable storage medium in the embodiments of the present invention can include any entity or device, recording medium that can carry computer program code, such as memories such as ROM / RAM, disks, optical discs, flash memories, etc.

[0109] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A transportation infrastructure cluster maintenance planning method, characterized in that: The method comprises: Carry out multi-level and multi-granular division of the traffic infrastructure cluster to obtain a bottom planning layer and at least one upper planning layer, wherein the planning unit in the bottom planning layer is a basic unit, and the planning unit in each upper planning layer is a composite unit. Each planning layer is progressive, and in two adjacent planning layers, the composite unit of the upper planning layer is a set of constituent units of the lower planning layer, and the constituent unit is a basic unit or a composite unit constituting the composite unit; According to the solving rules of bottom-up serial solving between layers and parallel solving within the same layer, a multi-objective optimization algorithm is used to solve the multi-objective maintenance planning model of all established planning units to obtain the target non-inferior solution set of the transportation infrastructure cluster maintenance planning; wherein the optimization variable of the multi-objective maintenance planning model of the composite unit is the index of the non-inferior solution set of the constituent unit.

2. The method according to claim 1, characterized in that The optimization variables of the multi-objective maintenance planning model of each basic unit are the maintenance measures to be taken in each period within the preset maintenance planning cycle; The objective functions of the multi-objective maintenance planning models of each of the planning units include a safety index, an economic index and a maintenance frequency index; wherein the safety index is the extent of performance improvement after maintenance measures are taken within the preset maintenance planning cycle, the economic index is the total cost of maintenance measures taken within the preset maintenance planning cycle, and the maintenance frequency index is the number of maintenance measures taken within the preset maintenance planning cycle.

3. The method according to claim 2, characterized in that The safety index of the basic unit is determined according to the difference between the integral area of ​​the improved performance prediction curve after the maintenance measures are taken and the integral area of ​​the initial performance prediction curve without taking the maintenance measures; The safety index of the composite unit is determined according to the weighted sum of the calculated values ​​of the safety indexes in the non-inferior solutions of the constituent units.

4. The method according to claim 3, characterized in that The performance of each basic unit before and after taking maintenance measures is predicted using an adaptive performance prediction model; wherein the parameters of the adaptive performance prediction model are updated using a preset update mechanism after obtaining new performance evaluation data.

5. The method according to claim 4, characterized in that The adaptive performance prediction model is an adaptive exponential performance degradation model based on a Bayesian update mechanism.

6. The method according to claim 3, characterized in that The method further comprises: Obtaining m current structural condition levels of each of the basic units before taking maintenance measures and n improved structural condition levels after taking maintenance measures; Establish a cost-effectiveness matrix based on the maintenance measures corresponding to each group of current structural condition levels and improved structural condition levels, the utility and cost corresponding to each maintenance measure; The intercept difference between the improved performance prediction curve and the initial performance prediction curve is determined according to the utility corresponding to the taken maintenance measures, and the utility corresponding to the taken maintenance measures is determined according to the cost-effectiveness matrix.

7. The method according to claim 6, characterized in that The maintenance measures in the cost-effectiveness matrix are represented by measure codes, and the maintenance measures in the optimization variables of the multi-objective maintenance planning model of each basic unit are represented by the measure codes; The utility corresponding to different maintenance measures is determined according to the utility interval of the corresponding improvement of the structural condition level and the level span before and after the maintenance measures are taken; The costs of different maintenance measures are obtained by multiplying the corresponding proportional factors by the preset costs.

8. A transportation infrastructure cluster maintenance planning device, characterized in that: The device comprises: A hierarchical division module: used to divide the traffic infrastructure cluster into multiple levels and multiple granularities to obtain a bottom planning layer and at least one upper planning layer, wherein the planning unit in the bottom planning layer is a basic unit, and the planning unit in each upper planning layer is a composite unit. Each planning layer is progressive, and for two adjacent planning layers, the composite unit of the upper planning layer is a set of constituent units of the lower planning layer, and the constituent unit is a basic unit or a composite unit constituting the composite unit; and The multi-objective optimization module is used to solve the multi-objective maintenance planning model of all established planning units according to the solution rules of bottom-up serial solution between layers and parallel solution within the same layer, and adopt a multi-objective optimization algorithm to obtain the target non-inferior solution set of the transportation infrastructure cluster maintenance planning; wherein the optimization variable of the multi-objective maintenance planning model of the composite unit is the index of the non-inferior solution set of the constituent unit.

9. A transportation infrastructure cluster maintenance planning device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images