<p>Seismic torsional response has always been a principal cause of structural failure in every major earthquake. There are numerous observations of damages caused by excessive torsional response in buildings, bridges, and lifeline structures. The torsion-induced failures have been especially catastrophic for multi-storey buildings because torsional response changes the uniform translational seismic floor displacements and causes concentration of demand in elements at the perimeter of the building. This often leads to failure of the over-loaded elements, which in turn initiates progressive collapse of the building. This study provides a conceptual explanation for the poor seismic performance of torsionally flexible asymmetric structures. These are buildings with a low level of torsional stiffness. Post-earthquake observations and also studies on single storey buildings have shown the vulnerability of these buildings to seismic damages. The study reported in this thesis extends the findings of previous research to multi-storey buildings and provides a theoretical foundation for understanding their seismic performance. Guidelines are developed and formulated to enable designers to identify torsionally flexible asymmetric buildings. This is of prime importance, as it is shown here that corrective measures taken by building codes in the form of statically applied torsional provisions are not effective for this type of structure. The future generation of codes for designing new structures and retrofitting existing structures will be performance based. A structure designed with such an approach has to meet a specific set of seismic performance criteria for a specific level of seismic hazard. Dynamic analyses of the building will be needed to assess the performance of the building at the many levels of seismic hazard. Preparation of input and interpretation of the large amount of output, resulting from an inelastic response analysis of multi-storey buildings is not practical, even for symmetric buildings. To overcome this, several simplified procedures based on inelastic static analyses are formulated in this thesis. Application of these procedures to some example multi-storey asymmetric buildings has shown that they are sufficiently simple and yet accurate for use in design offices. It is believed that a combination of these methods with sound engineering judgement will provide a practical and economical tool for the earthquake-resistant design profession to implement the performance-based design codes, currently being prepared by many countries, to protect life and property in urban centres in the event of an earthquake.</p> / Doctor of Philosophy (PhD)

<p>A computer model is developed for the solution of regional sanitary sewerage systems on a minimum cost basis. The regional system is considered to consist of a substantial number of alternative transportation (collection) routes, several potential treatment facilities and a receiving body of water which is able to accept the effluents from these facilities. The model, using a steady state analysis, facilitates the selection of a good collection network from the large number of possible networks, the selection of good depth-diameter combinations for each link in this final network and the selection of the size and number of treatment facilities which should be involved in the regional system. In addition, the required treatment efficiency of each facility will be determined so that a specified water quality goal may be satisfied.</p> <p>A major point to be considered in attempting to select a least cost system is the tradeoff in costs between constructing a few, large, centralized treatment facilities (thereby benefitting from the economies of scale which generally exist) and the extra cost associated with transporting the wastes an extra distance to these centralized plants. The possibility of higher treatment efficiencies being required at these centralized treatment facilities (due to a greater stress being exerted on the water quality of the receiving body of water) should also be considered. The piping costs are a function of the wasteflow. Fixed costs may be included in the cost function. The treatment plant costs are a function of both wasteflow and treatment efficiency. Fixed costs may also be included when determining the treatment plant costs.</p> <p>Linear programming and dynamic programming techniques are used as a basis for the computer model. Additional subroutines were requires to ensure that the technical constraints (for example, the minimum and maximum velocity constraints) were satisfied and to provide the missing components which are necessary to develop a complete computer package.</p> / Master of Engineering (ME)

<p>This thesis studies the behaviour of multi-storey wall-frame structures subjected to lateral and/or torsional loads. A planar structure with rotational and vertical restraints imposed between the wall and the frame is analysed. Parametric studies of a nominal structure are presented to examine the general effects of such restraints and structure height on wall-frame interaction. In the light of the understanding acquired from this two-dimensional analysis, a more general method of analysis for three-dimensional wall-frame structures is developed. For both methods, the variation of structural properties along the structure height is considered. A more flexible method of analysis for general three-dimensional building structures analogous to the stiffness matrix method is also presented. Applying this method to determine the structure stiffness matrix and using the lumped-mass approach, an attempt is made to study the free-vibration dynamic responses of the Banco de America Building both before and after damage was sustained by the building during the Managua earthquake of December 23, 1972. Lastly, conclusions regrading the behaviour and design of wall-frame structures are drawn, and the applicabilities and limitations of the three-dimensional analytical methods developed in this thesis discussed.</p> / Master of Engineering (ME)

<p>Reinforced concrete beam column joints are critical members in frame structures they can be subjected to high shear forces under earthquake loading. As a consequence, they can experience high shear and bond slip deformations that contribute significantly to the story drift. Moreover, the joint capacity may be exceeded leading to a joint shear failure that can have a major impact on the overall stability of the entire structure. This condition is particularly pronounced in lightly reinforced concrete structures where the beam column joints are typically the weakest link in the lateral load resistant frame. There is a persistent need to develop an analytical model that accounts for their shear and bond slip deformations in order to predict realistically their response and assess their safety. A finite element based analytical model is developed in this thesis for the beam column connection region. The model overcomes the need of using refined meshes of simple elements by using high power elements in the critical regions of the joint panel and the plastic hinge zones in the beams and the columns. The proposed model takes into account the shear and bond slip deformations in the joint panel as well as flexural and shear deformations in the plastic hinge zones in the beams and the columns. Material non-linearities associated with the concrete and steel behaviour are taken into account. Bond slip relationship between the beam reinforcement and concrete in the joint panel is considered. The material models developed in this thesis are verified at the element level before the verification is made to the entire beam column connection model. The predicitions of the model are compared with experimental data for beam column subassemblies experiencing high shear and/or bond slip deformations. The success of the proposed model is demonstrated by the good correlation achieved with the experimental data. The model is then used in the analysis of a three story reinforced concrete frame structure designed without consideration of earthquake loads. The structure is analyzed using different joint detailing schemes using pushover and time history analyses to investigate the effect of the joint detailing on the response of the structure. It is concluded that the proposed beam column connection model can be used successfully for the dynamic analyses of a complete multistory structure.</p> / Doctor of Philosophy (PhD)

<p>The objective of this research was to develop and calibrate a dynamic mechanistic model for biological nutrient (nitrogen and phosphorus) removal activated sludge systems treating municipal wastewater. The IAWPRC (ASM1) model for non-polyP heterotrophic and autotrophic organisms (Henze et al., 1987a,b) and the Wentzel et al. (1989b) model for polyP organisms were merged to form a general activated sludge model. After a number of initial modifications the model was tested against literature data from laboratory-scale nitrification denitrification biological excess phosphorus removal (NDBEPR) systems. Based on the preliminary results, a number of specific areas were identified which required further study. These included (1) accounting for sludge production and oxygen utilization in BEPR systems; (2) denitrification behaviour in BEPR systems; and (3) other issues such as hydrolysis under unaerated conditions.</p> <p>The main body of this thesis is presented as a series of five papers. The first paper (Chapter 4) presents a study of COD and nitrogen balances in activated sludge systems. The results suggest that systems incorporating anaerobic zones exhibit low COD balances compared to aerobic and anoxic-aerobic systems. Possible mechanisms for this "loss" of COD are discussed, including the possibility that the COD loss is related to fermentation processes occurring under anaerobic conditions.</p> <p>The second paper (Chapter 5) presents a study of denitrification behaviour in BEPR activated sludge systems. Results of a review of microbiological studies and many continuous and batch reactor experimental studies indicate that a significant fraction of the polyP organisms can use nitrate as an electron acceptor in the absence of oxygen for oxidation of stored PHB and simultaneous uptake of phosphorus.</p> <p>The development of a general activated sludge model for biological nutrient removal activated sludge systems is discussed in the third paper (Chapter 6). Several modifications were made to both the ASM1 and Wentzel et al. (1989b) model components, based on the results of literature review and model simulations. A fermentation process has been included for the conversion of readily biodegradable COD to short chain fatty acids (assuming a loss of COD). Hydrolysis of enmeshed slowly biodegradable COD under anaerobic conditions has been incorporated, as well as anoxic growth of polyP organisms. These modifications and others are discussed in this paper. The matrix representation and a description of the model processes are also presented, as well as a brief outline of influent wastewater characterization.</p> <p>The application of the general model is demonstrated in the fourth paper (Chapter 7) for aerobic and anoxic-aerobic systems, as well as a number of nutrient removal (NDBEPR) systems for both steady state and dynamic conditions. Results of simulations show the model is capable of predicting sludge production and oxygen utilization for a range of system types and configurations, as well as tracking changes in a number of parameters including soluble phosphorus and nitrate concentrations.</p> <p>In the final paper (Chapter 8) the consequences of the COD loss assumptions incorporated in the model are demonstrated for a number of experimental anoxic-aerobic and anaerobic-anoxic-aerobic systems. Results of model simulations indicate that without the assumption of COD loss, predictions of oxygen consumption and volatile suspended solids production are significantly over-estimated for NDBEPR systems (and to a lesser extent anoxic-aerobic systems). These systems apparently consume less oxygen and produce less volatile solids than aerobic systems for the same amount of COD removal.</p> <p>In conclusion, the merits and weaknesses of the general model are discussed. An important feature of the model is that a single set of kinetic and stoichiometric parameters produced quite accurate predictions for the wide range of systems to which the model was applied (with the exception of the nitrifier growth rate - discussed in Chapter 6). This provides a degree of support for the model structure and integrity. Many aspects of NDBEPR modelling require further investigation, including: the COD loss phenomenon, the fermentation processes occurring under anaerobic (and possibly anoxic) conditions, the hydrolysis of slowly biodegradable colloidal and particulate organics (particularly under anoxic and anaerobic conditions), and the impact these aspects have on denitrification behaviour in NDBEPR systems.</p> / Doctor of Philosophy (PhD)

<p>The scope of this thesis is to present a framework for the modelling of two-component, liquid/solid mixtures using the finite element method. The presentation is applicable to a wide range of two-component phenomena, however, special attention is paid to the liquefaction of sandy soils which is of particular concern to the civil engineer. In the past, much of the focus of research has been placed on the creation of models capable of capturing the mechanism of excess pore pressure development leading to liquefaction. Comparatively little attention has been placed on modelling the post-liquefaction event. This thesis provides a framework that allows the civil engineer to model soil deposits, predicting the onset of liquefaction and simulating the events that follow, namely post-liquefaction flow and the reconsolidation of the soil. The mixtures considered in this thesis are treated as two viscous fluids having momentum exchange between them via hydrodynamic drag. Two sets of Navier-Stokes equations are used to model the two-component mixture. The thesis presents the derivation of these equations along with the closure equations required to model the hydrodynamic drag and shear stress constitutive behaviour. A simple, novel approach to modelling the compaction/dilation behaviour of granular materials under the action of shear strain is also presented. It is shown that a recasting of the equations with anew set of variables is helpful in solving the equations via standard Galerkin finite element methods. An outgrowth of this recasting is also presented, that allows, for one-dimensional problems, to reduce the variable set from four unknowns to one, greatly simplifying the solution process and computation effort. Finally, several applications of the model are presented in order to validate the model and to demonstrate the wide range for which the model may be used.</p> / Doctor of Philosophy (PhD)

<p>This thesis presents a comprehensive approach to numerical modelling of human bone. Bone has been shown, in general, to be a heterogeneous material with orthotropic symmetry. The geometric arrangement of its porous micro-structure can be detected by high resolution tomographic imaging and then characterized by a 'fabric tensor.' This tensorial measure of material fabric can be correlated with mechanical properties and subsequently employed within numerical analyses of bone. In this work, the fabric tensor is incorporated into an elastic constitutive framework and a novel failure criterion for bone is proposed, which is seen as an important contribution to the numerical analysis of bone within the finite element (FE) methodology. The identification of fabric from micro-computed tomography (micro-CT) images of representative bone samples is achieved by a unified computational framework, described by a language independent pseudo-code. As a contribution to the constitutive representation of bone material, a new measure of fabric is defined and then identified using synthetic data of simple geometric shapes and micro-CT scans of human trabecular bone. The current potential for improvement in FE modelling of the mechanical behaviour of bone is illustrated through a discussion of bone fracture. The numerical analysis is an extension of the results presented in Pietruszczak et al. (1997, 1999) wherein a high resolution geometric model with heterogeneous distribution of orientation-dependent mechanical properties was employed. The use of the material model within a FE analysis is illustrated by a FE analysis pertaining to the prediction of fracture within a femur, under the simulated conditions of a fall to the hip. In particular, the distribution of damage within a femur is assessed under two porosity distributions, simulating a healthy and an aged bone.</p> / Doctor of Philosophy (PhD)

<p>This research uses Geographic Information System (GIS) databases to manipulate the actual travel routes used by urban commuter cyclists to address issues related to bicycle route choice and bicycle safely.</p> <p>Chapter 2 compares previously collected routes from Guelph to the shortest path routes. Most commuters divert little (0.4 km on average) and use major road routes. Although the cyclists tend to avoid grades, grade-separated railway crossings and high activity areas, they do not avoid high speed traffic or bridges. The cyclists use traffic signals especially for crossing major roadways and for turning. High-quality off-road paths are used infrequently, the lower quality ones even less.</p> <p>Chapter 3 uses a logit model for route choice. The choice set determination is critical logic that checks alternative routes for duplication of attribute combinations is used. A multinomial logit model that exhibited independence from irrelevant alternatives is estimated. The model identifies route attributes similar to those described above. In addition, personal variables (age, gender, winter cycling) were found to be significant.</p> <p>Chapter 4 describes the survey methodology used to collect two larger commuter bicycle route datasets in Toronto and Ottawa. Questionnaires containing a map for route collection as well as questions relating to cycling patterns and accident history were placed on cross-bars of parked bicycles. The return rate was 47% of the 6043 questionnaires.</p> <p>The overall rates per 100,000 commuter kilometres of collisions, falls, injuries and major injuries were found to be 3.26, 9.51, 7.60 and 1.10 respectively (Chapter 5). These are approximately 10 times comparable rates for automobiles. The GIS was used to determine the distance in the regular route that was undertaken on-road, off-road or on sidewalks in the Ottawa study area. Collision rates on different infrastructure were not statistically different. However, injury and fall rates were highest on sidewalks (4 times that of roads) followed by off-road paths (1.6 times that of roads).</p> / Doctor of Philosophy (PhD)

<p>There are many existing buildings which have been designed according to earlier codes. In these codes, either design for seismic loads was not a requirement or design was for lower levels of seismic forces. One of the major challenges that faces structural engineers is to determine the seismic capacity of an existing building and to rehabilitate these buildings to upgrade their seismic capacity if needed. At present, there are no guidelines available for the rehabilitation of existing structures and for the design of different retrofitting techniques. One of the most effective ways of rehabilitation is the addition of reinforced concrete (RC) walls. Until now, there is no robust analytical model available to study the effect of adding RC walls on the building behaviour up to failure. To determine the building seismic capacity and to study the effect of adding RC walls, the analytical model used should be capable of representing all the deficiencies in existing buildings. Moreover, it should be able to represent all possible failure modes such as beam-column joint shear failure, cumulative concrete crushing and bond slip failure. The objective of the present research is to evaluate the behaviour of buildings retrofitted by the addition of reinforced concrete structural walls. This was done by the development of a suitable analytical model that fully describes the behaviour of existing buildings before and after rehabilitation using RC structural walls. This model was verified using test results on specimens representing existing structures and found to give results that are closely correlated to the experimental results. The model was able to detect the failure mechanisms observed in the experiments. This model was used in the analysis of an existing three-storey building to determine its seismic capacity. The effect of ignoring the beam-column joint shear deformations, bond slip and concrete crushing was studied. Ignoring any of these parameters can change the building failure mechanism which could lead to misleading results. After that, a suitable rehabilitation technique using RC walls was chosen and the building behaviour after rehabilitation was studied.</p> / Doctor of Philosophy (PhD)

<p>Curbside collection of municipal solid waste is an expensive and increasingly complex operation. Municipalities across North America have been expanding their waste collection fleets steadily over the past decade due to population growth, the introduction of new collection services such as curbside recycling, and a decline in the number of active landfill sites. Despite the increasing cost and complexity of municipal solid waste collection systems, many collection programs are designed and operated without a clear understanding of the parameters responsible for those costs or the relationships between those parameters. Existing models of municipal waste collection operations often deal only with average system performance, ignoring large variations in important parameters such as the quantity of waste set out for collection or the percentage of households participating in a collection program. This research develops two different analytic models of municipal solid waste collection that explicitly address the variability of municipal solid waste collection operations. The first model is based on probability theory and vehicle dynamics, while the second model is based on queuing theory. Despite different starting assumptions, both models provide similar results and both models agree well with Monte Carlo computer simulation results. Both models are easier to use than computer simulations of the waste collection process, can be applied to any municipal waste collection operation, and can be coded on spreadsheets. The potential utility of the developed models has been demonstrated by application to a number of practical municipal solid waste collection problems. The models are not used to optimize systems of collection vehicles in this research, although they are used to generate improved strategies for the specific problems presented. However, either of the two models could be further incorporated into large scale optimization models for complete waste management systems. The models are of interest primarily to solid waste management practitioners. It is anticipated that they would make use of the models both to design new collection systems and to improve existing collection operations. Application of the models to local design and operational problems should result in more efficient and less costly waste collection operations. Specifically, these models can be used to minimize the size of the collection fleet and the amount of time required to collect municipal solid wastes, resulting in lower capital and operating costs, lower fuel consumption, and reduced air emissions from collection vehicles.</p> / Doctor of Philosophy (PhD)