The planning and management of coastal and river structures such as breakwaters, groynes, jetties, bridges and tidal inlets require accurate predictions of equilibrium morphologies. Generally these types of situations are modelled numerically using process - based models, where wave, current and sediment transport modules are applied over a number of time - steps until a steady - state morphology is obtained. Two alternative methods have been developed and applied in this thesis, based on self - organisation and entropy approaches. The self - organisation - based method utilises a cellular automata model, where local rules produce a global stable pattern through positive and negative feedback. The entropy - based method is able to predict equilibrium morphologies directly. It compares different randomly generated morphologies using an objective function and optimisation, instead of moving to an equilibrium morphology through intermediate states. This avoids some potential problems associated with traditional models such as error propagation and reliance on accurate initial conditions. The models developed in this thesis have been applied to a number of case studies. It was found that the cellular automata model obtained a higher Brier Skill Score than a comparable process - based model when predicting the equilibrium morphology associated with a channel obstruction. The entropy - based method was able to predict a realistic erosional channel in a coastal lagoon, similar to field observations at the Murray River Mouth in South Australia. It had difficulties predicting the deposition pattern due to the bias of the objective function towards erosional environments. The entropy - based method outperformed a conventional model prediction of the equilibrium erosional channel associated with a laboratory - sized lagoon, but similar problems were observed with its deposition predictive ability. The modelling methods developed in this thesis are a first step into the use of non - traditional, entropy - and self - organisation - based models for the prediction of complex equilibrium morphologies. They have made use of non - conventional models in order to explore different objective function formulations or self - organisation rules and the sensitivity of these, and have compared the models to laboratory results. The work documented in this dissertation shows that it is possible to use self - organisation - and entropy - based modelling methods to predict stable, equilibrium morphologies in coastal and river environments. / Thesis (Ph.D.)--School of Civil and Environmental Engineering, 2006.

Not available / text

published_or_final_version / Civil Engineering / Master / Master of Philosophy

Thesis (M.S.)--University of Hawaii at Manoa, 2006. / Includes bibliographical references (leaves 36-38). / x, 38 leaves, bound ill., map 29 cm

No description available.

Box-girder bridges supported by single reinforced concrete (RC) columns are expected to sustain seismic shocks with minor structural damages in seismically active regions where transportation is substantially required for rescuing and evacuating tasks. Such viaducts are vulnerable to damage when they are subjected to strong ground motions and acceleration pulse records, especially when responding in a flexural mode or having relatively low core confinement. Using a nonlinear dynamic solver that applies the fibre element method, global and local damage curves are computed based on the dissipated energy under hysteretic curves and based on constitutive curves, respectively. The RC bridge with seismic isolation bearing is used as an alternative system to control the damage, and modelled using linkage elements between the substructure and super structure. It was found that seismic isolation can be controlled to dissipate partial seismic energy so that the RC column gains the least possible minor damage. Using a MatLab program, a fibre element nonlinear model was built using a simplified iterative process and simplified constitutive relations. The number of fibres and elements under the dynamic loading was found to be affecting the final results of the analysis. Using crack growth modelling based on fracture mechanics, the combined discrete element/finite element explicit-Elfen code was applied to investigate the crack growth in 3D dynamically loaded RC columns. Despite its excessive computational cost and time, this code provides reliable information about local damage in the RC column core. Earthquake records with the pulse acceleration phenomenon have a severe damage potential on the structure. The difference in damage intensities was detected by crack growth modelling for the same problem using different loading rates. Critically stressed zones can be investigated independently by using the relative response technique, in which responses from the numerically analysed structure are re-used as applied loads onto a small-scale crack model for the critical member. Two general conclusions can be obtained; bridges with single RC columns designed by the demand/capacity criterion could suffer severe damage and possible collapse when subjected to strong ground motions. Secondly; hysteresis-based methods provide a global damage evaluation based on strength and ductility only regardless of the damage growth inside the concrete core and the buckling of bars, which could lead to progressive collapse.

No description available.

Load Factor Design (LFD), adopted by AASHTO in the mid-1970, is currently used for bridge superstructure design. However, the AASHTO specifications do not have any LFD provisions for foundations. In this study, a LFD format for the design of shallow foundations for bridges is developed. Design equations for reliability analysis are formulated. Uncertainties in design parameters for ultimate and serviceability limit states are evaluated. A random field model is employed to investigate the combined inherent spatial variability and systematic error for serviceability limit state. Advanced first order second moment method is then used to compute reliability indices inherent in the current AASHTO specifications. Reliability indices for ultimate and serviceability limit states with different safety factors and dead to live load ratios are investigated. Reliability indices for ultimate limit state are found to be in the range of 2.3 to 3.4, for safety factors between 2 and 3. This is shown to be in good agreement with Meyerhof's conclusion (1970). Reliability indices for serviceability limit state are found to be in the range of 0.43 to 1.40, for ratios of allowable to actual settlement between 1.0 to 2.0. This appears to be in good agreement with what may be expected. Performance factors are then determined using target reliability indices selected on the basis of existing risk levels. / Master of Science

Master of Science

The collapse of the I-35W Bridge in Minneapolis, MN on August 1, 2007 brought into question the design and inspection of gusset plates in steel truss bridges. The experimental tests performed in this research study the sway-buckling strength and behavior of large-scale steel gusset plates in an isolated truss connection. Parameters studied include plate thickness,combination member loading, initial out-of-plane imperfection, diagonal compression member out-of-plane flexural stiffness, corrosion, and alternative retrofits to increase lateral stiffness. The flexural stiffness of the diagonal compression member and retrofit designs were unique to the testing program. The variables monitored during testing include gusset plate surface stresses and strains, member axial strains, out-of-plane displacement of the gusset plate free edge, and buckling capacity. The results were compared with previously established design models for predicting buckling capacity of gusset plates which include the Whitmore effective width, the Modified-Thornton method, and the FHWA Load Rating Guidelines. A parametric finite element model was developed to determine the lateral stiffness of the gusset plate connection and the additional stiffness provided by the alternative retrofit options. The results showed interaction between the diagonal compression member and gusset plate occurs, which affects sway-buckling capacity. Combination of member loads showed evidence of detrimental effects on sway-buckling capacity. Corrosion of the gusset plates along the top edge of the bottom chord did not lead to significant reduction in sway-buckling capacity. The two retrofit designs showed increases in both lateral stiffness and buckling capacity as well as economic benefits over traditional retrofit methods. Comparison of the results to the current design guidelines showed that the current methods are conservative and do not accurately represent the true behavior of gusset plate connections. The research concludes with two proposed models for future use in design and retrofit of gusset plates. The first is a member-gusset plate interaction model based on a stepped column analogy that takes into account the effects of member flexural stiffness and gusset plate stiffness. The second is a general design guideline developed for retrofit of gusset plate connections dominated by sway-buckling behavior which uses a stiffness based approach to increase the capacity of gusset plate connections. / Graduation date: 2012