Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2009. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3673. Adviser: Imad L. Al-Qadi. Includes bibliographical references (leaves 164-171) Available on microfilm from Pro Quest Information and Learning.

<p> Modernization and rehabilitation of existing school facilities is a key factor to increase the life expectancy of facilities and to benefit students by increasing educational standards. Since United States' last recession in 2009, many school districts have been confronted with difficulty in funding their projects. The existing process of decision making on distributing budget is subjective and depends on the judgment of decision makers. To avoid the errors involved in the existing approach, an optimization model is developed in this research study to optimize the project selection. This optimization system increases the horizontal and vertical equity in the educational system while it reflects the importance of different categories of facility projects and the various educational demands of students. Two powerful methods of Genetic Algorithm and Dynamic Programming is utilized to solve the optimization problem. Moreover, the proficiency of the developed model is shown in a case study. </p>

This thesis presents the results of two-dimensional, transient flow/chemical transport experiments. The porous medium consisted of glass beads mixed with resin pellets. The simulated chemical contaminant is a fluorescent dye, Rhodamine WT, which is adsorbed by the resin pellets. The concentration of the dye in solution is measured using a fluorometer. By varying the resin/glass beads weight ratio, porous media of various adsorptive capacities were created. Adsorption is the primary property that was studied. The reported results of these experiments include the densities and permeabilities of the porous media, the location of the two-dimensional phreatic surface, the downstream flowrate and the output concentration under transient conditions. These results are significant and contribute to the understanding of chemical transport and adsorption effects, and provide a methodology for laboratory simulation of pollutant transport in granular porous media.

The strengthening technique using advanced composite materials has been implemented worldwide into the retrofit of existing concrete structures for static and seismic loading. But, there is still a gap in knowledge regarding the effect of impact and explosion loading on concrete structures strengthened with composite materials. This situation is especially difficult since many serious problems exist in material modeling, load definition, and selecting effective approaches for addressing the behavior of complicated geometrical domains in a structure. A series of twenty-seven concrete beams were prepared and tested to investigate the behavior of concrete beams strengthened with composite materials. Twenty-five beams were tested with impact load induced by a steel cylinder drop weight; two beams were statically loaded as control specimens. Two of the twenty-seven beams were unretrofitted, one for impact test group and the other for static test. Two types composite materials were used in strengthening, Kevlar and carbon fabrics. Each type of fabric had two different weights. Five beams were reinforced with steel bars and the other beams without steel bars. Investigation during testing included monitoring reaction force at support of simply supported beam, history of strain vs. time in composite laminates, variation of beam stiffness and deflection, crack and failure modes and stress concentration at the cut-off end of composite laminates. The test results revealed that bonding composite laminates to concrete beam can significantly improve the performance of concrete structures to resist impact load, increase the cracking, flexural strength, and residual stiffness of beam and reduce number of cracks, crack width and maximum deflection. The residual stiffness of strengthened beam after first impact was two to three times of that of un-retrofitted beam; the maximum deflection decreased 30 to 40%. The gain depends on the type and weight of composite laminates. Comparing to the static test results, the ultimate deflection and crack width were smaller than those of beam under static load. But, the maximum reaction force was three to four times larger than that under static load. The residual stiffness of strengthened beam after first impact can be calculated using a regression equation. The impact force can be obtained with a semi-empirical equation, which is derived from Spring-Mass models and modified by test results. From flexural wave theory, an equation has been developed for predicting the deflection of beam caused by impact loading. Based on the test and analytical results, a retrofit technique and design guidelines have been proposed. To control failure modes of strengthened beam, the minimum and balanced ratio of composite plate reinforcement have been derived.

Seismic performance of steel frames with "post-Northridge" bolted web, welded flange, beam-column connection with adequate ductility (BWWF-AD) and "pre-Northridge" bolted web, welded flange, beam-column connection (BWWF) is comprehensively evaluated using a nonlinear finite element method (FEM). The material and geometric nonlinearities, major sources of energy dissipation, and the flexibility of the connections are considered. The Richard Model is employed to analytically describe the moment-relative rotation (M-theta) curves of BWWF-AD connections using limited available data of full-scale, ATC-24 laboratory experiments. From the limited experimental data, an analytical method is developed to predict and to extrapolate the M-theta curves of BWWF-AD connection with any sizes of beams and columns. This method is used to extend the test results of an individual BWWF-AD connection to a frame with several or more BWWF-AD connections. The inelastic seismic response of two SAC Steel Project benchmark steel moment-resisting frames (SMRF) with BWWF and BWWF-AD connections modeled as partially restrained (PR) and as fully restrained (FR) are calculated and compared. Furthermore, the effects of BWWF and BWWF-AD connections on the frames' maximum top lateral and interstory displacement, maximum connection rotation, drift, and the frames' base shear forces are evaluated and compared. Moreover, the calculated drift and base shear forces of the frames are compared with the 2000 Edition of International Building Code (IBC 2000). This study suggests that the Richard Model captures the experimental M-theta curves of BWWF-AD connections accurately. BWWF-AD connections have significantly larger elastic stiffness than other types of PR connections. Their elastic stiffness is similar to that of bolted, double web angle with top and seat connection, but their ductility is much larger. In most cases, the presence of BWWF-AD connections improved the seismic performance of the steel frames. The proposed method analytically confirmed the beneficial effects of BWWF-AD connections observed in full-scale testing.

The contents of this dissertation present five submitted journal papers corresponding to experiments and prediction analysis of the bonding characteristics of fiber-reinforced-polymers (FRP) to concrete and steel. The papers are submitted to the American Society of Civil Engineers (ASCE) Journal of Structural Engineering and the American Concrete Institute (ACI) Journal of Composites for Construction. The first paper provides the background on the importance of resin systems in composites, particularly, for retrofitting and rehabilitating of existing structures. The paper provides a general understanding of the structure and important characteristics of epoxies for civil engineers involved in research and the application of FRP in construction. This paper also presents an overview of studies conducted on the durability of epoxy bonded joints in moist environments. The second and third paper provide the results of a comprehensive study on evaluation of bond strength of fiber-reinforced-plastics (FRP) to concrete and steel in simulated aggressive environments. The severity of bond deterioration varies in different environments and is quantified in this study. Fracture toughness is evaluated at the initiation of cracking and during the propagation of cracking. The environments consist of high temperature (120°F), acidic (pH = 3), alkaline (pH = 12), seawater (pH = 8.3), and high humidity <100%. The experiments consist of testing in shear 728 unidirectional carbon FRP coupons bonded to concrete and steel blocks. The lap shear tests evaluate the bond strength and fracture toughness of the FRP/substrate interface as related to different exposure times in different environments. The last two papers focus on accelerated degradation testing techniques, and discuss the lifetime prediction on the bond of carbon (FRP) to concrete and steel. Fick's law of diffusion is used to model the moisture penetrating through the FRP and into the bonded joint. A reliability analysis is developed to predict the fracture toughness at any relative humidity and temperature, based on the 100% relative humidity conditions at ambient temperature. The results in this study generally show an excellent durability for the FRP bonded to either concrete or steel, even in the most severe environments.

Effective and accurate algorithms are developed to evaluate the reliability of frame and shear wall structural system subjected to both static and dynamic loadings. The basic deterministic finite element algorithm is based on the assumed stress-based finite element method in which the tangent stiffness can be expressed in explicit form and fewer elements are required to realistically capture the structural behavior. These features are desirable for developing an efficient reliability analysis algorithm for both static and dynamic cases. The presence of shear walls is represented by plate elements. The stiffness matrix for the combined system is then developed. To verify the accuracy of the deterministic algorithm, a 2-bay 2-story building consisting of five similar frames is considered. Only one frame is assumed to have shear walls. The responses of the frame with shear walls subjected to static and dynamic loadings are evaluated. The responses of the same structural system are also evaluated using a commercially available computer program. The results match very well, implying that the deterministic algorithm developed in this study is accurate. The deterministic algorithm is then extended to consider the uncertainty in the random variables. For the static case, a stochastic finite element-based approach consisting of the reliability approach, the first-order reliability analysis procedure and the finite element method is proposed. For the dynamic case, a hybrid approach consisting of the response surface method, the finite element method, the first-order reliability method and the linear iterative scheme is used. The unique feature of this algorithm is that the earthquake loading can be applied in the time domain. The material and cross-sectional properties, the damping and the magnification factors of earthquake time histories are considered to be random variables in this study. The reliability of a frame without and with shear walls is evaluated for the strength and serviceability performance functions. The results are verified using the Monte Carlo simulation technique.

Modeling ultrasonic transducers is an important aspect of research in nondestructive evaluation and testing. In most nondestructive evaluation applications, the ultrasonic transducers are traditionally modeled as: (1) point sources generating spherical wave fronts (2) line sources generating cylindrical wave fronts, or (3) planar surfaces generating plane wave fronts. In reality, the transducer front face has finite dimensions; it is neither point source nor planar source because the ultrasound that emits from a piezoelectric transducer does not originate from a point or an infinite plane, but instead originates from the finite surface of the piezoelectric element with flat or curved front face. Analytically modeling the fields radiated by ultrasonic transducers is a very difficult task because of the large number of possible transducer types, sizes and configurations that are used in practice. In this study, a semi analytical technique the Distributed Point Source Method (DPSM) is adapted to model ultrasonic transducers. The DPSM discretizes the transducer surface into a finite number of elemental surfaces. As a result, the complexity associated with the discretization of the three-dimensional problem geometry as done in the finite element technique is reduced. In the DPSM technique, the fundamental governing equations for elastic wave propagation in a fluid and in a solid are solved. For this reason, the DPSM technique is called a semi-analytical technique. In this research, computer codes for computing the ultrasonic field in a three dimensional inhomogeneous medium in front of a transducer of finite dimension have been written in MatLab. Two different cases are considered in this study, nonhomogeneous fluid and fluid-solid interface. Both normal and inclined incidence cases are investigated. This investigation shows that DPSM is an efficient technique for modeling ultrasonic transducers in nonhomogeneous media.

A novel nondestructive defect detection procedure is proposed for the health assessment of existing structures. The procedure can also be used to evaluate the health of structures just after natural events like strong earthquakes or high winds, or after a man-made event like explosion. The procedure is essentially a time domain system identification technique. The two most important features of the method are that it does not need excitation information to identify a structural system and it can identify defects at the local element level using noise-contaminated limited response information, i.e., responses need not be available at all dynamic degrees of freedoms. Any structure that can be represented by finite elements can be identified by this method. The method identifies defects by tracking the changes in the stiffness of the structural elements in the finite element representation. Once a defective element is identified, it can locate the defect spot more accurately within the defective element. The method is denoted as the Generalized Iterative Least-Square Extended Kalman Filter with Unknown Input (GILS-EKF-UI). Since the Kalman filter-based algorithm requires information on excitation force and initial state vector, the GILS-EKF-UI technique consists of two stages. In the first stage, based on the response information, a sub-structure model is developed that satisfying all the requirements for the generalized ILS-UI method. The unknown excitation force f(t) and all the elements in the substructure are identified. The identified stiffness and damping coefficients provide information on the initial values of the state vectors. In the second stage, the EKF-WGI method is used to identify the whole structure since all the information required to implement the EKF-WGI method is now available from Stage 1. With the help of numerous examples, the GILS-EKF-UI method was verified for relatively small and large structures in the presence of noise in the responses. It can detect relatively small defects. The study also helps to define the minimum response information required to implement the Kalman filter-based algorithm. The method is very accurate, robust, and economical and has a potential to be a non-destructive defect evaluation technique.

The damage suffered by steel structures during recent strong earthquakes forced the profession to reevaluate issues related to the seismic design of steel structures. The evaluation of the maximum inelastic deformation of a structure subjected to a strong motion earthquake is a critical part of this process. A ductility parameter can also be used to calculate the maximum inelastic deformation of a structure. It is pointed that there is no unanimity on the definition of ductility, although it is constantly used in the profession. In this research several definitions of story ductility for MDOF systems are studied and the most appropriate one is identified. Definitions for local and global ductility are proposed. The presence of PR and composite connections on the structural response is also addressed in this study. Conventional analysis and design of steel frames structures is based on the assumptions that beam-to-column connection are either fully restrained (FR) or perfectly pinned (PP) connections. However, almost all steel connections used in practice are essentially partially restrained (PR) connections with different rigidities. The effect of PR and composite connections on the nonlinear seismic response of steel frames is evaluated. For this purpose, first the structural responses in terms of maximum interstory displacements and maximum top lateral displacements of three steel frames are calculated considering all of the frame connections to be of FR-type. Then the structural responses are evaluated for the frames with PR connections and finally for the frames with composite connections. Responses are compared for the three different cases. The recommendations to consider the effect of the vertical component on the structural response of two major seismic design guidelines for buildings are also studied. The first one is the National Earthquake Hazard Reduction Program (NEHRP) Recommended Provisions for Seismic Regulations for New Building (1994) and the second one is the Mexico City Seismic Code. Specifically, the effect of the vertical component on the structural responses of steel frames is evaluated first analytically and then according to the NEHRP Provisions and the Mexican Code. Finally, the analytical results are compared with the codes' recommendations.