| With the increasingly complete construction of China’s road network infrastructure,the maintenance tasks are severe.Scientific,efficient,and sustainable road network infrastructure maintenance intelligent decision system is the key to guarantee the smoothness and safety of the national road network.However,the life cycle assessment(LCA)of maintenance decision objectives is not comprehensive,and the evaluation system of infrastructure mainte-nance and operation stage is insufficient.The maintenance decision method is highly subjective,and the intelligent level of maintenance decision needs to be improved.The scope of maintenance decision is mainly project level sin-gle facilities,and the optimization method of road network level maintenance decision needs to be further studied.In view of the deficiency of the existing research,this paper relies on the national key research and development plan“Road Infrastructure Life–cycle Operation Support Technology”(2018YFB1600300 & 2018YFB1600304),aiming at improving the evaluation system of life-cycle maintenance and operation stage,improving the intelligent level of maintenance decision,and realizing the global optimization of road network-level and project-level infrastructure,and makes the following research:Firstly,aiming at the problem of incomplete LCA of maintenance-decision objectives,by quantitative analysis of related factors of maintenance activities and transportation,the evaluation method of road network infrastructure life cycle maintenance and operation stage was put forward.Based on the existing LCA framework and research foundation,the assessment objectives of maintenance and operation stage were defined.Aiming at the problem that the LCA was limited to maintenance construction,the interference mode of maintenance construction on traf-fic operation was analyzed.An evaluation model of maintenance management and traffic operation factors during maintenance construction was developed.By adopting simulation analysis and numerical simulation,an evaluation model was put forward to evaluate traffic operation factors under the coupling effect of the evolution of infrastructure performance and traffic conditions in the post-maintenance operation stage.The evaluation method can comprehen-sively evaluate the economic and environmental factors in the whole life cycle of maintenance management and traffic operation based on economies of scale,which provided an evaluation basis for maintenance decision.Secondly,maintenance decision needs to accurately predict the combination of infrastructure diseases.In order to improve the accuracy of maintenance decision,a road network foundation performance state representation and prediction method was proposed.Aiming at the problem of inaccurate prediction of long-term development state of disease combination,a probabilistic prediction model of facility performance state is proposed based on continuous time and state space process by using disease direct evaluation index to weaken artificial weighted interference.Aim-ing at the inaccuracy of residual life prediction caused by the associated development of multi-disease combination and the original defects of data censoring,a numerical test environment was established by combining Monte Carlo simulation and Copula analysis,and a residual life prediction model based on historical performance development was developed by defining the failure mode of facilities.The results indicate that the prediction accuracy of the model was improved by 12%,and the probabilistic prediction results of the performance state of road network level and project level facilities can be provided at the same time,which provides the disease combination development law for maintenance decision.Then,aiming at providing project-level maintenance decision support for the road network level and reducing the subjective experience dependence of maintenance decision methods.This study took the intelligent formulation of maintenance opportunity-maintenance mode as the core component,and the intelligent formulation and opti-mization method of project-level infrastructure maintenance strategy was put forward.To provide data support for maintenance decision,the method of road network infrastructure division and the structured data of maintenance decision were determined.In order to fully mine historical maintenance methods to expand alternative schemes and realize intelligent recommendation of maintenance methods,K–Median + topic–sensitive Page Rank algorithm were utilized to mine the matching historical maintenance information,and an intelligent recommendation model of maintenance methods under similar historical conditions of target infrastructure was put forward and verified.By studying the multi-maintenance cycle of performance improvement and development mode after maintenance,an intelligent formulation method of project-level infrastructure maintenance strategy was put forward.Aiming at the trigger limit of fixed threshold maintenance,reinforcement learning model was introduced to solve the optimal matching of dynamic maintenance opportunity–wide-area maintenance mode combination.An intelligent optimiza-tion model of project-level maintenance strategy was developed,which realizes the efficient optimization solution of maintenance decision.Case analysis showed that the intelligent recommendation accuracy rate of maintenance mode reached 90%,and the optimization ability was improved by 8%.Finally,a road network–project-level double-layer joint intelligent maintenance decision model was innova-tively developed in order to break through the limitation of single project-level facilities in the scope of maintenance decision,further optimize the road network-level maintenance decision,and link the changes of project-level de-cision planning simultaneously.According to the characteristics of double-layer joint decision,a bi-level planning model was constructed,with the project-level intelligent maintenance decision method as the lower model.The steady-state genetic algorithm was used to solve the upper-level road network model,thus realizing the global opti-mization of road network infrastructure in the maintenance planning period.On this basis,the influence of various factors on the decision results were further explored.Case analysis showed that the decision method had obvious advantages over the traditional maintenance decision process in terms of service life loss and cost-benefit control.This paper realized the intelligent maintenance decision and optimization of the maintenance planning period of road network infrastructure with the optimization goal of LCA by investigating the intelligent maintenance decision method of the whole life cycle of road network infrastructure.The research method provides a new technical means for the full life cycle assessment of road network infrastructure.The research results preliminarily construct an intelligent maintenance decision system for road network infrastructure.The research conclusions have certain guiding significance for the theory and practice of road network infrastructure maintenance management. |
