Abstract: t-SNE is an embedding method that the data science community has widely used. It helps two main tasks: to display results by coloring items according to the item class or feature value; and for forensic, giving a first overview of the dataset distribution. Two interesting characteristics of t-SNE are the structure preservation property and the answer to the crowding problem, where all neighbors in high dimensional space cannot be represented correctly in low dimensional space. t-SNE preserves the local neighborhood, and similar items are nicely spaced by adjusting to the local density. These two characteristics produce a meaningful representation, where the cluster area is proportional to its size in number, and relationships between clusters are materialized by closeness on the embedding. This algorithm is non-parametric. The transformation from a high to low dimensional space is described but not learned. Two initializations of the algorithm would lead to two different embedding. In a forensic approach, analysts would like to compare two or more datasets using their embedding. A naive approach would be to embed all datasets together. However, this process is costly as the complexity of t-SNE is quadratic, and would be infeasible for too many datasets. Another approach would be to learn a parametric model over an embedding built with a subset of data. While this approach is highly scalable, points could be mapped at the same exact position, making them indistinguishable. This type of model would be unable to adapt to new outliers nor concept drift. This paper presents a methodology to reuse an embedding to create a new one, where cluster positions are preserved. The optimization process minimizes two costs, one relative to the embedding shape and the second relative to the support embedding’ match. The embedding with the support process can be repeated more than once, with the newly obtained embedding. The successive embedding can be used to study the impact of one variable over the dataset distribution or monitor changes over time. This method has the same complexity as t-SNE per embedding, and memory requirements are only doubled. For a dataset of n elements sorted and split into k subsets, the total embedding complexity would be reduced from O(n2) to O(n2/k), and the memory requirement from n2 to 2(n/k)2 which enables computation on recent laptops. The method showed promising results on a real-world dataset, allowing to observe the birth, evolution and death of clusters. The proposed approach facilitates identifying significant trends and changes, which empowers the monitoring high dimensional datasets’ dynamics.
Abstract: The configuration of shear walls in plan of building will affect the seismic design of structure. The position of these walls will change the stiffness of each floor in the structure, the diaphragm center of mass displacement, and the drift of floor. Structural engineers preferred to distribute the walls in buildings to make the center of mass almost close enough to the center of rigidity, but to make this condition satisfied, they have many choices: construct the walls on the perimeter, or use intermediate walls, or use walls as core. In this paper and by using ETABS, each case is studied and compared to other cases according to three parameters: lateral stiffness, diaphragm displacement, and drift. It is found that the core walls are the best choice for the position of the walls in the buildings to resist earthquake loads.
Abstract: R.C.C. buildings with dual structural system
consisting of shear walls (or braces) and moment resisting frames
have been widely used to resist lateral forces during earthquakes. The
dual systems are designed to resist the total design lateral force in
proportion to their lateral stiffness. The response of combination of
braces and shear walls has not yet been studied. The combination
may prove to be more effective to resist lateral forces during
earthquakes. This concept has been applied to regular R.C.C.
buildings provided with shear walls, braces and their combinations.
Abstract: This study estimates the seismic demands of tall
buildings with central symmetric setbacks by using nonlinear time
history analysis. Three setback structures, all 60-story high with
setback in three levels, are used for evaluation. The effects of
irregularities occurred by setback are evaluated by determination of
global-drift, story-displacement and story drift. Story-displacement is
modified by roof displacement and first story displacement and story
drift is modified by global drift. All results are calculated at the
center of mass and in x and y direction. Also the absolute values of
these quantities are determined. The results show that increasing of
vertical irregularities increases the global drift of the structure and
enlarges the deformations in the height of the structure. It is also
observed that the effects of geometry irregularity in the seismic
deformations of setback structures are higher than those of mass
irregularity.
Abstract: This paper presents the retrofitting of beam-column joint using CFRP (Carbon Fiber Reinforced Polymer) and steel plate. This specimen was tested until failure up to 1.0% drift. This joint suffered severe damages and diagonal cracks at upper crack at upper column before retrofitted. CFRP were wrapped at corbel, bottom and top of the column. Steel plates with bonding were attached to the two beams and the jointing system. This retrofitted specimen is tested again under lateral cyclic loading up 1.75% drift. Visual observations show that the cracks started at joint when 0.5% drift applied at top of column. Damage of retrofitted beam-column joint occurred inside the CFRP and it cannot be seen from outside. Analysis of elastic stiffness, lateral strength, ductility, hysteresis loops and equivalent viscous damping shows that these values are higher than before retrofitting. Therefore, it is recommended to use this type of retrofitting method for beam-column joint with corbel which suffers severe damage after the earthquake.
Abstract: The effect of different combinations of response
feedback on the performance of active control system on nonlinear
frames has been studied in this paper. To this end different feedback
combinations including displacement, velocity, acceleration and full
response feedback have been utilized in controlling the response of
an eight story bilinear hysteretic frame which has been subjected to a
white noise excitation and controlled by eight actuators which could
fully control the frame. For active control of nonlinear frame
Newmark nonlinear instantaneous optimal control algorithm has been
used which a diagonal matrix has been selected for weighting
matrices in performance index. For optimal design of active control
system while the objective has been to reduce the maximum drift to
below the yielding level, Distributed Genetic Algorithm (DGA) has
been used to determine the proper set of weighting matrices. The
criteria to assess the effect of each combination of response feedback
have been the minimum required control force to reduce the
maximum drift to below the yielding drift. The results of numerical
simulation show that the performance of active control system is
dependent on the type of response feedback where the velocity
feedback is more effective in designing optimal control system in
comparison with displacement and acceleration feedback. Also using
full feedback of response in controller design leads to minimum
control force amongst other combinations. Also the distributed
genetic algorithm shows acceptable convergence speed in solving the
optimization problem of designing active control systems.
Abstract: Conventional concentrically-braced frame (CBF)
systems have limited drift capacity before brace buckling and related
damage leads to deterioration in strength and stiffness. Self-centering
concentrically-braced frame (SC-CBF) systems have been developed
to increase drift capacity prior to initiation of damage and minimize
residual drift. SC-CBFs differ from conventional CBFs in that the
SC-CBF columns are designed to uplift from the foundation at a
specified level of lateral loading, initiating a rigid-body rotation
(rocking) of the frame. Vertically-aligned post-tensioning bars resist
uplift and provide a restoring force to return the SC-CBF columns to
the foundation (self-centering the system). This paper presents a
parametric study of different prototype buildings using SC-CBFs.
The bay widths of the SC-CBFs have been varied in these buildings
to study different geometries. Nonlinear numerical analyses of the
different SC-CBFs are presented to illustrate the effect of frame
geometry on the behavior and dynamic response of the SC-CBF
system.
Abstract: A method to predict the column size for displacement based design of reinforced concrete frame buildings with higher target inter storey drift is reported here. The column depth derived from empirical relation as a function of given beam section, target inter-story drift, building plan features and common displacement based design parameters is used. Regarding the high drift requirement, a minimum column-beam moment capacity ratio is maintained during capacity design. The method is used in designing four, eight and twelve story frame buildings with displacement based design for three percent target inter storey drift. Non linear time history analysis of the designed buildings are performed under five artificial ground motions to show that the columns are found elastic enough to avoid column sway mechanism assuring that for the design the column size can be used with or without minor changes.
Abstract: Analytical seismic response of multi-story building
supported on base isolation system is investigated under real
earthquake motion. The superstructure is idealized as a shear type
flexible building with lateral degree-of-freedom at each floor. The
force-deformation behaviour of the isolation system is modelled by
the bi-linear behaviour which can be effectively used to model all
isolation systems in practice. The governing equations of motion of
the isolated structural system are derived. The response of the system
is obtained numerically by step-by-method under three real recorded
earthquake motions and pulse motions associated in the near-fault
earthquake motion. The variation of the top floor acceleration, interstory
drift, base shear and bearing displacement of the isolated
building is studied under different initial stiffness of the bi-linear
isolation system. It was observed that the high initial stiffness of the
isolation system excites higher modes in base-isolated structure and
generate floor accelerations and story drift. Such behaviour of the
base-isolated building especially supported on sliding type of
isolation systems can be detrimental to sensitive equipment installed
in the building. On the other hand, the bearing displacement and base
shear found to reduce marginally with the increase of the initial
stiffness of the initial stiffness of the isolation system. Further, the
above behaviour of the base-isolated building was observed for
different parameters of the bearing (i.e. post-yield stiffness and
characteristic strength) and earthquake motions (i.e. real time history
as well as pulse type motion).