Abstract
The greenhouse gas (GHG) emission due to the road transportation has increased significantly since 1990. The total emissions from road transportation originates from the following phases: material production, construction, maintenance, use and demolition phases of the pavement. Life cycle assessment (LCA) of pavements suggests that for heavily congested roads, impact of the use phase can surpass the emission from the other phases. In the use phase of pavement, besides the fuel needed to propel the vehicle and overcome the air drag, extra fuel must be added to the system to compensate the rolling resistance forces. Pavement roughness is one of the key contributors to rolling resistance and thus vehicle fuel consumption. Roughness-induced fuel consumption incorporates the energy dissipation in the suspension system of vehicles and depends on road surface characteristics, dynamic properties and velocity of the vehicle.In this study, the sensitivity of roughness-induced excess fuel consumption (EFC) to all involving factors, i.e. road roughness metrics, vehicle dynamic properties and speed is investigated, and the dominant factors affecting the EFC are identified. To this end, Monte-Carlo (MC) simulation is performed by generating realizations of all input parameters according to their probability distributions and estimating the energy consumption by the mechanistic roughness-induced pavement-vehicle interaction (PVI) model. Sobol’s method—a robust analysis of variance (ANOVA)-based technique for global sensitivity analysis—is then utilized to obtain Sobol’s global sensitivity indices. It is found that roughness metrics, i.e. the International Roughness Index (IRI) and the waviness number,account for 88 − 93% of the total variations in energy dissipation and are, by far, the most influential factors. Furthermore, the order of roughness-induced PVI model is reduced by only focusing on variables with high Sobol indices and disregarding the variability of the rest of variables. Due to the impact of road undulations on the EFC and the degree of ride comfort,road roughness metrics are widely-used indices in pavement management systems for maintenance decisions. Conventional methods of roughness characterization require the longitudinal road profile, which is costly to be directly measured. The footprint of road unevenness is, however, observable in the vibration of vehicle when traveling on a rough roads. The vibration signals therefore carry important information about the dynamic properties of the vehicle as well as the surface descriptors of the road. Therefore, a probabilistic framework is developed to estimate the dynamic characteristics of vehicles and the power spectral density of the road roughness from vertical accelerations readily recorded by passengers’ smartphones. The proposed framework is a sequential two-layer inverse analysis with a mechanistic forward model based on a two-degree-of-freedom ground-vehicle model. The first layer of inverse analysis procedure is the probabilistic Bayesian inference. Results of the first layer are then used in the second layer, a regularized cost minimization problem, to further refine consistency and convergence. The proposed inverse framework is applied to several sets of recorded data where the smartphones were mounted on different locations within the vehicle cabin. It is shown that thealgorithm efficiently determines the effective dynamic properties of the vehicle and, moreimportantly, roughness metrics that are in good agreement with the observations.