Scientia Iranica
Volume:18 Issue: 2, 2011

A matrix free unstructured Galerkin Finite Volume Method (GFVM) is adopted for solving plane-stress two dimensional Cauchy equilibrium equations. The algorithm is developed based on the Galerkin method, for the solution of structural problems on unstructured linear triangular element meshes. The developed shape function free Galerkin Finite Volume solver computes stresses and displacements of solid mechanic problems via some iteration. The performance of the introduced algorithm on coarse unstructured meshes is assessed by comparison with computed results of a plane-stress case (with uniformly distributed load on one of its elliptic boundaries and two straight sliding support boundaries), for which an analytical solution is available. The results of the introduced method are presented in terms of stress and strain contours, and the sensitivity of the GFVM solver to mesh coarseness, as well as to the utilized gradual load imposing parameter (which affects the convergence behavior of the model), is assessed. Furthermore, the accuracy of the present matrix free GFVM is compared to the previous matrix manipulation based solution methods.

The Endurance Time (ET) method is a time history-based dynamic analysis procedure which uses special intensifying acceleration functions for evaluation of the seismic response of structures. One of the potential applications of the ET method is in the three-dimensional analysis of buildings under multi-directional excitations. In this paper, considering horizontal components of excitation, an algorithm for the multi-component analysis of building structures by the ET method is proposed, and results of the ET method for various steel moment frames with 1 to 7 stories are compared with results from time history analysis with real earthquakes. Results show that based on recommendations of structural codes for bi-directional time history analysis, which requires applying horizontal components of earthquakes simultaneously, the ET method can be used to predict the seismic response of structures with appropriate approximation.

In this paper, a new constitutive law for concrete is proposed to predict the non-linear behavior of reinforced concrete members. The proposed model is intended to provide improvements on modeling the non-linear hysteretic behavior of concrete structures in the finite element coding. The independent damage and fracture parameters in compression and in tension have been introduced to the constitutive law of concrete degradation, due to increasing loads. In the sample of non-linear monotonic compressive loading, the law has been derived from a basic mathematical model obtained previously. In addition, in the sample of monotonic tension loading, the same law is adopted based on the latter basic mathematical model. The main novelty of the proposed Uniaxial Constitutive Law (UCL) lies in the fact that all the required input data can be obtained through conventional monotonic and cyclic compression and tension tests.

As part of human resource management policies and practices, construction firms need to define competency requirements for project staff, and recruit the necessary team for completion of project assignments. Traditionally, potential candidates are interviewed and the most qualified are selected. Precise computing models, which could take various candidate competencies into consideration and then pinpoint the most qualified person with a high degree of accuracy, would be beneficial. This paper presents a fuzzy adaptive decision making model for selection of different types of competent personnel. For this purpose, human resources are classified into four types of main personnel: Project Manager, Engineer, Technician, and Laborer. Then the competency criteria model of each main personnel is developed. Decision making is performed in two stages: a fuzzy Analytic Hierarchy Process (AHP) for evaluating the competency criteria, and an Adaptive Neuro-Fuzzy Inference System (ANFIS) for establishing competency IF-THEN rules of the fuzzy inference system. Finally, a hybrid learning algorithm is used to train the system. The proposed model integrates a fuzzy logic qualitative approach and neural network adaptive capabilities to evaluate and rank construction personnel based on their competency. Results from this system in personnel staffing show the high capability of the model in making a high quality personnel selection.

This paper addresses a numerical algorithm for nonlinear analysis of frames, using the unit displacement method, in generating a reduced stiffness matrix of the structure. This algorithm can properly be used in nonlinear static analysis or in the incremental response spectral method. Here, the instantaneous reduced stiffness matrix of the structure is calculated, considering its linear behavior at the latest state, by performing a set of numerical tests on the whole structure. Each numerical test consists of imposing prescribed displacement fields on the lateral displacement of stories and calculating the reactions of the structure. The solution procedure of each test is based on the division of degrees of freedom into three parts: (1) predefined lateral displacement of joints, (2) vertical displacement of joints, considered as linear degrees of freedom, and (3) rotation of joints, regarded as nonlinear degrees of freedom. The stiffness matrices are generated distinctly for all mentioned parts. Therefore, the linear stiffness matrix is inverted once at the beginning of the analysis. The suggested method is not limited to any special case or physical assumptions. This model has good accuracy in representing structural responses and modal properties, confirmed by different numerical examples. Regarding computational cost, the proposed algorithm is more efficient in comparison with the conventional method.

Traffic assignment is a step of travel demand estimation. Given a trip origin–destination demand matrix, this step determines traffic flow in each link, according to assumptions based on the behavior of drivers. Conventional assignment algorithms, which are mostly based on the Wardrop first principle of user equilibrium, assume that all drivers choose the shortest path to the destination, based on the same travel time computed by travel time functions. However, in reality, driver perception of travel time varies for a specific route. This paper presents a traffic assignment algorithm which assumes that driver perception of travel time affects route choices. Fuzzy set theory is used to define travel time perceived by drivers. A fuzzy equilibrium is suggested for the prediction of network flows. Next, a Fuzzy Incremental Traffic Assignment algorithm (FITA) is developed to utilize route Perceived Travel Time (PTT) for reaching the suggested fuzzy equilibrium. The FITA is used for a real network traffic assignment in Mashhad, which is large city in Iran. Traffic flow that is estimated by a FITA and by conventional algorithms is compared to real observed volumes, which indicate that a FITA is more accurate than conventional traffic flow estimation algorithms.

This study presents an energy-based approach to the performance-based optimization of steel moment resisting frames at the so-called operational, immediate occupancy, life safety and collapse prevention performance levels. Two objective criteria are identified for a performance-based seismic design: minimizing structural cost (interpreted as weight) is one; the other concerns minimizing earthquake damage with respect to the maximum hysteretic energy capacity of the structure. That is, the overall objective for the design of a building framework is to have minimum weight and maximum energy dissipation capacity.