Our objective in earthquake engineering research is to improve the state of knowledge, through fundamental and applied research, for the purpose of aiding decision makers in reducing seismic hazards. In this context, decision makers are all individuals and agencies that affect the planning and design/construct process, such as planning or regulatory agencies, owners, investors and insurers, and engineers who have to provide protection against seismic hazards through earthquake resistant design.

We view earthquake engineering as a multi-phased process that ranges from the description of earthquake sources, to characterization of site effects and structural response, and to description of measures of seismic protection. These basic phases in turn bring about the following components of research that we are presently engaged in: occurrence modeling, geophysical modeling, ground motion modeling, stochastic and nonlinear dynamic analysis, and design and experimentation. These components pertain to the individual phases but also, and perhaps more importantly, to aspects that incorporate some or all of the phases of earthquake engineering. An example of the latter is seismic risk analysis, which requires knowledge and developments in source, attenuation, and ground motion modeling, as well as in response evaluation and protective measures.

Our research activities in earthquake engineering consist of individual projects, initiated and supervised by one or several faculty members, but performed under the common umbrella of the John A. Blume Earthquake Engineering Center. The following summary of research activities focuses on those aspects in which Stanford's contributions to
earthquake hazard mitigation have been and are expected to remain most effective.

Seismic Hazard and Risk Analysis

For over thirty years, research has been conducted by researchers at the John A. Blume Earthquake Engineering Center in the general field of seismic hazard and risk analysis. Early work focused mainly on modeling sources, occurrence and attenuation. In recent years, considerable efforts are placed into introducing mechanistic models to occurrence and attenuation phenomena. Stanford University researchers were responsible for developing probabilistic hazard analysis methodologies, using Poisson models and Bayesian models. Over the years, time and space dependent models have been introduced to represent the fault rupture mechanics and the stress accumulation and release cycles of large earthquakes. Most recently advanced computational tools, such as geographic information systems (GIS) and database management systems (DBMS), have been utilized to capture, analyze, integrate and display the tectonic, seismological, geological and engineering information needed in seismic hazard assessment.

Stanford researchers have worked with various countries in Central America, North Africa, Asia and Europe to develop seismic hazard maps and structural design criteria. Furthermore, our faculty and graduate students have significantly contributed in the development of models and methods for earthquake vulnerability and risk assessment. Early work was based on empirical damage assessment models using damage data from past earthquakes. More recently, analytical models for damage and structural vulnerability assessment have been formulated that are based on nonlinear structural response simulation. A key question currently being addressed is the assessment of losses resulting from structural damage. Damage and vulnerability models are developed for individual structures within the context of performance based engineering. More generic vulnerability models are formulated for application over large regions to many different types of structures. These risk assessment tools have been implemented and utilized by the practicing engineering community in individual building seismic risk assessment as well as by government agencies, insurance/reinsurance and financial institutions for large portfolio or building inventory analysis.

Over the past decade researchers in our department are also working on seismic risk assessment models for transportation systems. These models utilize GIS and transportation network analysis tools to estimate the losses from damage to components of the system as well as those due to traffic time delays or inaccessibility of particular locations. Tools for emergency response and resource allocation following disasters are key features currently under development. Significant component of this research is supported through the Pacific Earthquake Engineering Research Center (PEER).

Ground Motion Modeling

Prediction of strong ground motion has been and continues to be a major research area in earthquake engineering. The topics in this research area that receive specific attention at Stanford include (1) simulation of ground motion models for seismic hazard analysis, (2) stochastic- physical rupture process models for ground motion prediction, (3) prediction of ground motion for engineering applications, and (4) study of the nonstationary characteristics of simulated and recorded ground motions for nonlinear analysis of structures. Various geophysical models are being considered for the purposes of simulating strong ground motion. Recorded motions from recent earthquakes are being studied for their characteristics and damage potential. Recent seismological studies have focused on the understanding and characterization of strong ground motion in the near-field. The effect of near-field motions on structures has been observed from past earthquake events to be particularly important; however, systematic studies of these effects have not been conducted and is the focus of current research.

Damage Potential of Ground Motions

Experience in past earthquakes has shown that the engineering profession has not yet succeeded in defining ground motion parameters that correlate well with observed damage. From an engineering perspective we are seeking representations of the seismic "demand" that can be used, through convolution with the structural "capacity", to assess structural reliability. Thus, both demand and capacity need to be evaluated, the latter with due regard to structural characteristics and cumulative damage effects that depend on strong motion duration. If this can be achieved, seismic risk analysis can be based on reliability concepts, and design parameters can be derived that are consistent with the damage potential of the ground motions.

Research is in progress in which seismic hazard analysis, input and response characterization, structural reliability, and design are treated as interrelated subjects through a consistent and coordinated approach. The following are the major components of this research: development of damage models for structural response; characterization of ground motions based on damage potential; reliability evaluation; seismic risk analysis; and development of design parameters.

Design and Experimentation

Considerable effort is devoted to design research that can be implemented directly in engineering practice. This work is concerned with methods to evaluate and improve the behavior of new and existing structures in severe earthquakes. Important topics, on which research is being carried out, include:

• Development of a deformation based seismic design methodology;
• Dynamic stability considerations and P-delta effects;
• Evaluation of the effects of stiffness and strength irregularities in plan and elevation;
• Cumulative damage modeling;
• Retrofit measures for existing structures
• Exploration of new materials and new structural systems for earthquake resistance.

Our research facilities include a laboratory with equipment for static and dynamic testing of structural materials, components and system models. Past research has been devoted to the development and applications of techniques for testing of small-scale reinforced concrete and steel structures. Shaking table experiments have also been performed on scale models of adobe houses to develop
simple and effective seismic strengthening techniques for housing in developing countries.

More recently, the structural testing is focusing on research to validate computational models to predict dynamic nonlinear response of structures and for developing health-monitoring technologies. This includes, for example, shaking table tests to examine structural collapse phenomena as affected by the complex interactions of degrading structural response and random earthquake input motions. Shake table testing is also an important component of the research to develop more robust wireless strong motion sensors. Other projects involve quasi-static testing of structural components and materials to evaluate fiber-optic sensors and to investigate the effect of localized failure mechanisms on structural
performance.

Data Base Management and Knowledge-Based Systems

In Stanford's Center for Integrated Facility Engineering (CIFE), civil engineering and computer science faculty cooperate on applications of computer science techniques to civil engineering problems. Several faculty members are now utilizing these techniques successfully in earthquake engineering research.

Data base management is a powerful tool for storage, retrieval, and evaluation of ground motion and structural damage data. Knowledge-based systems are being employed for problems that do not submit to traditional analytical descriptions and methods. Much of our knowledge is encapsulated in collections of linguistic statements of relationships. Such characterizations, while not processable by traditional methods, can be coded in a knowledge-based expert system that utilizes a series of artificial intelligence tools. Initial applications of such systems in building performance, site hazard, and seismic risk assessment have indicated that such methods offer substantial opportunities to organize judgmental knowledge, particularly those that require the aggregation of expert opinions. Examples of areas in earthquake hazards reduction suitable for such applications are: aggregation of expert opinions to complement incomplete earthquake catalogs; earthquake and ground motion prediction models; formulation of observational knowledge from damaged regions; collateral hazard analysis (landslide, flood, fire); utilization of geographic information systems (GIS); damage evaluation for various classes of buildings and facilities.