Presentations, Conferences, and Papers from the Archive
TRACC papers at the 90th Annual Transportation Board Meeting
The research and development performed by the TRACC staff has resulted in several technical papers that were accepted for the presentation at the 90th Annual Transportation Research Board Meeting, Washington, D.C., January, 2011:
- Seismic and Traffic Load Modeling on Cable Stayed Bridge
- Modeling of Soil-Structure Interaction in Presence of Large Deformations in Soil
- Comparison of ECE-R66 and FMVSS 220 Test for a Selected Paratransit Bus
- Congestion Pricing, Heterogeneous Users, and Travel Time Reliability: Multicriterion Dynamic User Equilibrium Model and Efficient Implementation for Large-Scale Networks
- Integrated Nested-Logit Mode Choice and Dynamic Network Microassignment Model Platform to Support Congestion and Pricing Studies: New York Metropolitan Case
- Robust Optimization Approach for System-Optimal Dynamic Traffic Assignment with Demand Uncertainty
- Time-Dependent Origin-Destination Demand Estimation: Challenges and Methods for Large-Scale Networks with Multiple Vehicle Classes
Seismic and Traffic Load Modeling on Cable Stayed Bridge
by Cezary Bojanowski and Ronald F. Kulak
Abstract
Long span cable stayed bridges are usually characterized by numerous, closely spaced low natural frequencies below 1 Hz. For a particular combination of cable and deck frequencies, large amplitude cable vibrations due to parametric excitation may occur even for considerably low traffic, wind or seismic loading. The mechanism of parametric excitation is well described theoretically for simple academic cases. However, due to the complexity of the parametric resonance phenomenon, analysis of bridges for this purpose is only possible by using explicit Finite Element (FE) simulations.
This paper presents a FE analysis of the Bill Emerson Memorial Bridge subject to seismic and traffic loading and their influence on vibrations of the cables. The development of the FE model of the bridge for use with the LS-DYNA® code is described. The model was validated by comparison of calculated natural frequencies with experimentally determined values.
The bridge was subjected to seismic loading registered during the Northridge earthquake. This simulation revealed large oscillations of the cables that were induced by the tower lateral vibrations. However, the vibrations did not lead to excessive stresses in cables.
A simplified model representing the AASHTO HS 20-44 truck was also developed. The truck model was used to simulate traffic loading on the bridge. The interaction between the truck and the bridge deck was defined through a special LS-DYNA algorithm - RAIL. In this paper, cases with one and twenty trucks on the bridge were examined. No excessive cable vibrations due to traffic loading were observed.
Simulation of traffic loading on the Bill Emerson Memorial Bridge
via the LS-DYNA® code (Livermore Software Technology Corporation)
Modeling of Soil-Structure Interaction in Presence of Large Deformations in Soil
by Ronald F. Kulak and Cezary Bojanowski
Abstract
In current design practice, soil-structure interaction analysis often assumes linear elastic properties of the soil and uses small displacement theory. However, there are numerous problems which require a more advanced approach. One such problem is the analysis of bridge pier stability under scour conditions where complex interactions occur between the bridge piers with footings and the surrounding soil. This type of problem requires special solution algorithms that account for soil-structure interaction and appropriate constitutive models for soil.
This paper presents comparison of four numerical approaches to modeling soil-structure interaction in the presence of large soil deformations. The commercial code LS-DYNA®/MPP was used to investigate the Lagrangian, Element Free Galerkin (EFG), Smoothed Particle Hydrodynamic (SPH) and Multi Material Arbitrary Lagrangian-Eulerian (MM-ALE) algorithms. To establish the accuracy and computational efficiency of each method, simulations were performed for the in-situ experiment of a steel loading pad penetrating into silty clay sand. The efficiency of the methods was assessed in terms of preprocessing complexity, robustness and computational cost. Solution accuracy was assessed by comparison to experimental results.
The results show that three of the four formulations can produce reasonable predictions at very large loading pad penetrations. However, the most reliable and efficient approach turned out to be the SPH method. This method was further used to investigate failure conditions of the pier of the Oat Ditch Bridge on I-15 in California during the August 2003 flood. The numerical simulations show that scour of the riverbed must have triggered excessive movement of the footing that led to failure of the bridge.
Deformations and vertical stresses in the soil via the LS-DYNA® simulation (Livermore Software Technology Corporation)
Comparison of ECE-R66 and FMVSS 220 Test for a Selected Paratransit Bus
by Bronislaw Gepner, Christopher Rawl, Leskaw Kwasniewski, Cezary Bojanowski and Jerry Wekezer
Abstract
The main objective of this paper is to compare the performance of a selected paratransit bus subjected to two roof integrity tests, dynamic rollover test according to UN-ECE Regulation 66 (ECE-R66), and the quasi-static symmetric roof loading according the standard FMVSS 220. The investigation is carried out primarily using a numerical study accompanied by experimental tests on components and by a full scale roll over test. The first part of the research consisted of the FE model development, verification, experimental validation, final check using full scale rollover test, and calibration. The verification was primarily done through an energy balance check. The hierarchical validation involved three levels of experiments: material characterization, static and dynamic component testing and determination of the center of gravity. Next, the verified and validated model was used for prediction of the full scale test conducted following ECE-R66 procedures but with an increased drop height of 909 mm. The FE model was calibrated through a numerical parametric study and comparison with the full scale experiment. The calibrated model was used in the next stage which consisted of two numerical tests: dynamic rollover according to ECE-R66, and quasi-static roof loading following requirements of FMVSS 220. Results presented in the paper illustrate the fact that the final crashworthiness assessment of the paratransit buses depend on the selected evaluation codes. In addition to these discrepancies, the paper underscores consequences of using quasi-static testing methods for making a judgment on possible performance behavior of paratransit buses during violent, dynamic, rollover accidents.
For further information please contact: Ronald F. Kulak or Cezary Bojanowski
Congestion Pricing, Heterogeneous Users, and Travel Time Reliability: Multicriterion Dynamic User Equilibrium Model and Efficient Implementation for Large-Scale Networks
by Lan Jiang, Hani Mahmassani and Kuilin Zhang
Abstract
With increasing interest in road pricing strategies to alleviate congestion and improve network performance, this study develops a multi-criterion dynamic user equilibrium (MDUE) traffic assignment model which explicitly considers heterogeneous users that seek to minimize three essential decision attributes: travel time, out-of-pocket cost, and travel time reliability in the underlying path choice framework. The value of time (VOT) in this study is treated as a continuous random variable distributed across the population of trip-makers, and travel time reliability is estimated at a path level. Furthermore, the proposed MDUE model is capable of considering different vehicle classes, namely low occupancy vehicles (LOV) and high occupancy vehicles (HOV); thus, with greater realism in trip-makers' path choice behavior, the MDUE model is applicable to analyze a variety of road pricing scenarios. To address the practical application aspects for large-scale congested networks, novel implementation techniques are proposed to overcome computational obstacles to real world application. An application to the New York metropolitan regional network is demonstrated, and a set of numerical experiments are conducted on this network to explore the convergence behavior, solution quality and required computational time of the MDUE algorithm.
For further information please contact: Kuilin Zhang
Integrated Nested-Logit Mode Choice and Dynamic Network Microassignment Model Platform to Support Congestion and Pricing Studies: New York Metropolitan Case
by Kuilin Zhang, Hani Mahmassani and Peter Vovsha
Abstract
This paper presents an integrated nested logit mode choice and dynamic network simulation-assignment model platform to support congestion and pricing studies. The integrated model, coupling a nested logit mode choice model with the algorithmic procedures developed for large-scale dynamic network assignment applications, is demonstrated for the New York metropolitan region "best practice" network using actual data. Specifically, the proposed model simultaneously addresses two important problems on multidimensional networks: a time-dependent stochastic user equilibrium mode choice problem and a multi-class multi-criterion dynamic user equilibrium route choice problem. This study reformulates these two equilibrium assignment problems as gap function based nonlinear optimization problems. Then a projected gradient based descent direction method is proposed to solve these nonlinear optimization problems. This solution algorithm is embedded in a simulation-assignment algorithmic framework that implements individual choices of both travelers to modes and vehicles to routes. This micro-assignment approach is shown to provide the integrated model with rich behavioral modeling capabilities which enable realistic modeling of heterogeneous users' mode and route choices in response to a variety of traffic management measures, such as dynamic pricing, as well as congestion induced unreliability.
For further information please contact: Kuilin Zhang
Robust Optimization Approach for System-Optimal Dynamic Traffic Assignment with Demand Uncertainty
by Chung-Cheng Lu, Jiuh-Biing Sheu and Kuilin Zhang
Abstract
This study deals with the path-based system optimal dynamic traffic assignment (SODTA) problem with uncertain demands. We consider that the probability distribution of uncertain demands is unknown, so classical stochastic programming techniques are not applicable. Instead, uncertain demands are assumed to be bounded by a prescribed uncertainty set, and a robust optimization approach is adopted to address this problem. The objective is to minimize the total network travel time under the worst-case scenario defined by a pre-determined demand uncertainty set. We formulate the robust counterpart optimization problem of SODTA for a general uncertainty set, and show that solving the robust SODTA (RSODTA) problem is not more difficult than solving the deterministic SODTA problem for some specific types of uncertainty set, such as convex hull, box, and ellipsoidal. Moreover, a column generation-based algorithmic framework that embeds a scaled gradient projection algorithm is proposed to solve the SODTA problem. Numerical experiments were conducted to demonstrate the effectiveness of the algorithm and to examine the impact of different types of demand uncertainty set on solution quality.
For further information please contact: Kuilin Zhang
Time-Dependent Origin-Destination Demand Estimation: Challenges and Methods for Large-Scale Networks with Multiple Vehicle Classes
by Ismail Omer Verbas, Hani S. Mahmassani and Kuilin Zhang
Abstract
This paper proposes a modified bi-level optimization algorithm to estimate the time-dependent origin-destination trip matrices for large-scale networks with multiple vehicle classes. Methodologies to overcome the challenges due to the scale of the problem are presented. The upper-level problem, which is a bound-constrained quadratic problem, has an enormous number of variables and parameters for a network with around 68,000 links, 28,000 nodes and 3,700 zones. Techniques to reduce the number of variables and parameters are described in this study along with an approach to reduce the time and memory requirement of the lower-level problem. Furthermore, the basic approach, which was previously only applied to a single vehicle class, is extended and adapted in this study to estimate matrices for single-occupancy and high-occupancy vehicles jointly. Two different solution packages, MINOS and KNITRO are tested for the upper-level problem. The solution package KNITRO is run with an option to use an interior point/conjugate gradient algorithm, which is very well suited for large-scale nonlinear problems. The modified bi-level algorithm is applied to estimate the time-dependent demand patterns for the New York Regional Network.
For further information please contact: Kuilin Zhang