CE SEMINARS
The NetQuakes Project
Thursday, May 8, 2008
The USGS has funded the development of a new generation of inexpensive and robust strong-motion recorders that are designed to be deployed in urban areas. The recorders incorporate MEMS sensors, wireless communication, NTP and GPS timing, expanded data storage, secure communications, remote management capabilities, and low power hardware within a Linux-based operating system. The goal of the project is to greatly increase the density of strong motion instruments near active faults without increasing telemetry costs or maintenance efforts. The talk will review the motivation for the project, instrument capabilities, results of instrument testing at the Albuquerque Seismological Laboratory, and strategies for deploying the instruments.
Steel Moment-Resisting Frame Responses in Simulated Strong Ground Motions: or How I Learned to Stop Worrying and Love the Big One
Thursday, May 1, 2008
This talk presents the response of steel moment-resisting frame buildings in simulated strong ground motions. I collected 37 scenario and hypothetical earthquake simulations of crustal earthquakes in California. These ground motions are applied to nonlinear finite element models of four types of steel moment frame buildings: six- or twenty-stories with either a stiffer, higher- strength design or a more flexible, lower-strength design. I also consider the presence of fracture-prone welds in each design. Since these buildings experience large deformations in strong ground motions, the building states of interest are collapse, total loss, and if repairable, the peak inter-story drift. This talk maps these building responses on the simulation domains which cover the San Francisco and Los Angeles regions. The building responses can also be understood as functions of ground motion intensity measures, such as pseudo-spectral acceleration (PSA), peak ground displacement (PGD), and peak ground velocity (PGV). This talk develops building response prediction equations to describe probabilistically the state of a steel moment frame given a ground motion. The presence of fracture-prone welds increases the probability of collapse by a factor of 2--8. The probability of collapse of the more flexible design is 1--4 times that of the stiffer design. Six-story buildings are slightly less likely to collapse than twenty-story buildings assuming sound welds, but twenty-story buildings are 2--4 times more likely to collapse than six-story buildings if both have fracture- prone welds. A vector intensity measure of PGD and PGV predicts collapse better than PSA. Models based on the vector of PGD and PGV predict total loss equally well as models using PSA. PSA alone best predicts the peak inter-story drift, assuming that the building is repairable. As “rules of thumb,” twenty-story steel moment frames with sound welds collapse in ground motions with PGD greater than 1 m and PGV greater than 2 m/s, and they are a total loss for PGD greater than 0.6 m and PGV greater than 1 m/s.
An Ounce of Prevention: Probabilistic Loss Estimation for Performance-Based Earthquake Engineering
Judith Mitrani-Reiser, John Hopkins University
Wednesday, February 20, 2008
A variety of sourcesi may be considered in the loss estimation of a building in a seismic-prone region as part of the performance-based earthquake engineering design methodology. The main contributors to building seismic loss result from the human injury and life loss during a seismic event, the downtime of a building after a seismic event, and the cost to repair the damaged structure. A methodology was developed to estimate the direct economic losses due to repair costs as well as two types of indirect economic losses, those produced by building downtime and by human fatalities. A procedure for a virtual inspection is used to assess the safety of buildings, based on current damage assessment guidelines. Additionally, a model is established to estimate human fatalities caused by the partial and global collapse of buildings, using probabilities of fatality based on relevant empirical data and the results of the virtual inspection process. A simplified methodology will be presented for estimating building downtime after seismic events, including mobilization delays before construction begins and the elapsed time needed to repair damaged building components. The losses due to downtime and human fatalities are then added to the building repair cost in order to estimate the total building loss, which is used to perform a benefit-cost analysis of several designs of a reinforced-concrete moment-frame office building used in a previous benchmark study. This work is the most faithful attempt to estimating the main decision variables (termed the 3 D’s: dollars, deaths and downtime), proposed by the Pacific Earthquake Engineering Research (PEER) Center and the ATC-58 Project for performance assessment of structures.
A Plasticity Model to Predict the Effects of Confinement on Concrete
Julie Wolf, California Institute of Technology
Thursday, February 7, 2008
A plasticity model to predict the behavior of confined concrete is defined.
The model is designed to implicitly account for the increase in strength and ductility due to confining a concrete member. The concrete model is implemented into a finite element (FE) model. By implicitly including the change in the strength and ductility in the material model, the confining material can be explicitly included in the FE model. Any confining material can be considered, and the effects on the concrete of failure in the confinement material can be modeled. Test data from a wide variety of different concretes utilizing different confinement methods are used to estimate the model parameters. This allows the FE model to capture the generalized behavior of concrete under multiaxial loading. The FE model is used to predict the results of tests on reinforced concrete members confined by steel hoops and fiber reinforced polymer (FRP) jackets. Loading includes pure axial load and axial load-moment combinations. Variability in the test data made the model predictions difficult to compare but, overall, the FE model is able to capture the effects of confinement on concrete. Finally, the FE model is used to compare the performance of steel hoop to FRP confined sections, and of square to circular cross sections. As expected, circular sections are better able to engage the confining material, leading to higher strengths. However, higher strains are seen in the confining material for the circular sections. This leads to failure at lower axial strain levels in the case of the FRP confined sections. Significant differences are seen in the behavior of FRP confined members and steel hoop confined members. Failure in the FRP members is always determined by rupture in the composite jacket. As a result, the FRP members continue to take load up to failure. In contrast, the steel hoop confined sections exhibit extensive strain softening before failure. This comparison illustrates the usefulness of the concrete model as a tool for designers. The concrete model provides a flexible and powerful method to predict the performance of confined concrete.
Ground Motions and Intensity Measures as a Link Between Seismology and Engineering
Jack Baker, Stanford University
Wednesday, October 24, 2007
Ground motions, and measures of ground motion intensity, serve as the link between seismology and engineering for the purpose of earthquake risk assessment and performance-based earthquake engineering. Seismology is used to quantify rates of earthquake activity and associated ground shaking at a site, while engineering is used to analyze response of structures to ground shaking input. There is a potential for erroneous conclusions to be drawn if this important interface is not properly addressed. One approach for this problem is to use recorded ground motions and quantify them using ground motion intensity measures. Recent developments in this field will be presented, and current initiatives and problems will be discussed. Recent past work includes development of multi-parameter (i.e., vector-valued) descriptions of earthquake intensity, and advanced intensity measures beyond parameters such as elastic spectral acceleration. Implications for selection of recorded ground motions and validation of simulated ground motions will also be discussed.
Stochastic System Design and Applications to Stochastically Robust Structural Control
Alexandros A Taflanidis, Duke University
Thursday, October 11, 2007
In engineering applications the knowledge about a planned system is never complete. Often a probabilistic quantification of the uncertainty arising from this missing information is warranted in order to efficiently incorporate our partial knowledge about the system and its environment into their respective models. In this stochastic framework the design objective is related to the expected value of a system performance measure, such as life-cycle cost or reliability. This system design process is called stochastic system design.
This seminar addresses general stochastic system design problems. Topics related to efficient system modeling, performance evaluation and design optimization are addressed. Stochastic simulation is suggested for evaluation of the overall performance of the system. In this setting, the response of the system model can be evaluated using computer simulation, rather than approximated analytically, to allow for explicit consideration of (a) nonlinearities of the models adopted for the system and excitation, and (b) complex performance metrics. This approach involves, though, significant computational cost for the associated optimization. A novel two-stage framework is discussed for this optimization. The first stage implements a new algorithm (Stochastic Subset Optimization) for efficiently exploring the sensitivity of the objective function and converging to a small candidate set for the optimal design variables. The second stage adopts some established stochastic optimization algorithm for pinpointing the optimal solution within that set.
Applications of this general methodology to control problems as well as general structural design problems are presented. Two specific examples are discussed. The first example considers the design of passive and active tuned mass dampers for improvement of the reliability of an offshore platform in a random sea environment. The second example discusses the retrofitting of a four-story structure with viscous dampers. The lifetime cost is adopted as design objective and a comprehensive methodology is presented for estimating the cost associated with future earthquake damages.
Optimal Control of Distributed Energy Harvesting Networks
Jeffrey Scruggs, Duke University
Monday, September 24, 2007
In many areas of vibration engineering, there is a growing interest in the use of transduction to couple the mechanical dynamics of a vibrating system with controllable electronic networks. One interesting application of this concept is called energy harvesting, which pertains to systems which are required to operate in energy-autonomy over the duration of a decades-long service life. In structural health monitoring applications employing embedded wireless sensing, for example, it is highly desirable for these sensors to somehow scavenge their power from their immediate surroundings, rather than requiring external power or long-term battery storage. One promising technology for such energy-harvesting systems entails the use of resonant piezoelectric devices, which harvest energy from ambient vibration. The fabrication of such devices constitutes a well-understood technology. However, for systems driven by broadband stochastic excitation, the determination of the optimal control of the electronic network was only recently shown to be a tractable problem in general.
This presentation illustrates that such energy harvesting control problems can be reduced to standard H2 optimal control problems, which may be solved to determine the optimal routing of power in the electronic network. This result is applied to a flexible cantilever beam, with a bimorph piezoelectric transducer bonded along its length, and subject to broadband stochastic vibration. It is shown that by subdividing the transducer into sections along the beam length, and controlling each of these sections through separate electronics, the overall efficiency of power absorption can be increased. For the case of transducer voltage feedback, the impact of measurement noise on the optimal conversion efficiency, and optimal beam configuration, is investigated. Additionally, the theory is extended to accommodate the impact of nonlinear dissipation in the electronic drive circuitry, due to semiconductor conduction losses, and it is shown that these losses result in a critical level of excitation below which energy conversion is infeasible. Finally, multi-objective optimization techniques are used to maximize generated power, subject to constraints on the dynamic responses of transducer voltages and beam stresses.