Development of a multi-dimensional life cycle analysis framework towards sustainable pavement management on project and network levels

Abstract

Pavements act as an essential component of civil infrastructure for supporting transportation, economic development, and improvement of life quality. With ever-increasing road mileages and emerging functional requirements, simple procedures and cumulative but unmethodical personnel experience that worked previously are no longer able to manage the continuous expansion of pavement networks. In addition, conventional approaches to maintaining huge pavement network in satisfactory condition inevitably result in considerable budget and environmental burdens. This exacerbates tensions between the multi-dimensional pillars of sustainability (environment, society, and economics) which decision-makers must balance. Consequently, it is a critical part for developing an effective and efficient pavement management system. The research in this dissertation proposes a methodological decision-support framework for sustainable pavement management and implements it in several emerging green pavement technologies (e.g. in-place recycling, rubberized asphalt, warm mix asphalt, and low-noise porous asphalt pavement). The entire framework comprises two general decision levels, namely project level and network level. Although the two decision levels differ in their system boundaries, they share the same ultimate principles oriented towards the selection of optimal alternatives among competitors. The sustainability targets identified through this study are destined to not merely achieve the best pavement utility, but also to minimize the influences on ecology that result from each decision made. To realize sustainable pavement management on two decision-making levels, several sustainable evaluation, integration, and optimization techniques have been developed and interconnected in the framework. At the project level, after the life-cycle pavement-related sustainability indicators (e.g. environmental impact, cost and performance) have been identified and evaluated, the integration method to support final identification of alternatives was firstly realized by developing the single-dimensional integration (i.e., cost-benefit integration), and further improved by applying the concept of eco-efficiency as the multi-dimensional integration (i.e., eco-efficiency integration). Uncertainty analysis was subsequently incorporated in the evaluation procedures to increase assessment reliability and avoid the likelihood of misunderstanding. The usability and capability of the above-mentioned methods were verified by practical comparisons of competitive pavement rehabilitation techniques and asphalt materials. At the network level, decisions are required to consider both the degree of consumption and functional improvement as they relate to the adoptable projects, and the pavement function deterioration rates varying by time and space. As two critical components for network-level decision-making, the long-term performance prediction model and the multi-objective optimization model have been investigated and developed by two machine learning techniques (i.e., artificial neural network and support vector machine) and a multi-objective optimization algorithm (i.e., genetic algorithm). This research employed pavement acoustic performance as the functional performance to test the feasibility and applicability of the proposed modelling methods and provide complementary reference for decision-makers as well. The analysis outcomes at both the project and network levels could contribute to finding approaches to facilitate improvements in sustainable pavement management decision-making in three ways. First, multi-disciplinary cooperation in quantitative evaluation, sustainable integration, uncertainty analysis, and multi-objective planning increases the approach capability and versatility in addressing complicated problems. Second, the innovative applications in the emerging pavement technologies could contribute as a quantitative and informative reference for agencies, pave a more sustainable and efficient way to improve current decision-making tools, and lay a rational basis for future resource allocation and policy planning. Finally, the successful implementation of the proposed decision-making framework in pavement infrastructure management in this study indicates its potentials for application to other types of civil infrastructure.