Research Overview
Resiliency refers to the ability of a community to overcome the disruptions and return to a normal state following a hazardous event (e.g., hurricane, earthquake, or tsunami). There is a strong consensus among researchers that an essential element of resiliency is the preparation for and conduct of the rapid and efficient assessment of the post-event situation that includes the assessment of buildings, critical infrastructures, utilities, communication systems as well as the economy and the overall wellbeing of the community. The overarching goal of my research is to enhance post-disaster resilience by developing methods for state-of-the-art monitoring of infrastructures, reconnaissance, and recovery of the disaster-stricken community using the major progress made in artificial intelligence algorithms, computational simulation, advanced materials, sensing systems, communication technologies, and science-based understanding of natural hazards. Some of my on-going and past projects are listed below.
Community sentiment analysis for identifying social vulnerability following earthquakes
In this project, we will develop a framework to conduct sentiment analysis and assess the earthquake's impact on a community. The framework will include methods to collect social-media data and evaluate its usability, i.e., community representation. It will further include NLP capabilities for sentiment analysis. We will utilize the framework to monitor the progression of sentiments of a community before, during, and after an earthquake. We will then study its correlation to observed damage data identified through reconnaissance efforts usually conducted by EERI, GEER, or StEER Network. We will further explore the impact of the recovery process in the communities by computing correlations between observed sentiments and retrofit efforts. We will also conduct topic analysis to find specific issues the communities in need are discussing.
Printed strain sensor arrays for large-area structural health monitoring
This project aims to develop and test printed and flexible strain sensor and wireless communication electronics into a conformal sensor patch that will improve post-earthquake large-area structural health monitoring (SHM). Large area SHM is essential for large scale structures such as bridges, dams, etc. Printed electronics can be fabricated in roll-to-roll processes on flexible substrates, making them feasible and ideal for large area SHM. The sensor patch will be able to monitor strain during and following an earthquake and transmit the acquired data to a Bluetooth-enabled device.
Human-Machine Collaboration Framework for Bridge Health Monitoring
Structural health monitoring (SHM) of the bridges is becoming essential as 50% of California’s bridges have exceeded their design life. This research project aims at combining the strengths of ML tools with that of domain expertise of structural engineer’s to develop a bridge SHM framework. The framework will utilize recorded earthquake response from undamaged bridge and develop a novelty model for individual bridge. At the same time, a simplified analytical model will be developed to simulate damage information. The developed and verified framework will be implemented on selected CSMIP instrumented bridge structures. When developed successfully the framework is expected to automate the process of damage detection and aid in efficient post-earthquake response.
Damage Detection of CSMIP Instrumented Buildings using CAV: A Machine Learning Approach
This research project aims at developing a damage detection algorithm using acceleration data of CSMIP instrumented structures. The developed method will utilize cumulative absolute velocity (CAV) of sensed data as a damage feature and apply a statistical pattern recognition (SPR) approach in order to identify existence as well as location and extent of the damage. When developed successfully, this method will not only ensure better structural integrity and life safety but also mitigate economic losses associated with major seismic events.
Damage Detection of CSMIP Instrumented Buildings using CAV: A Machine Learning Approach
This research project aims at developing a damage detection algorithm using acceleration data of CSMIP instrumented structures. The developed method will utilize cumulative absolute velocity (CAV) of sensed data as a damage feature and apply a statistical pattern recognition (SPR) approach in order to identify existence as well as location and extent of the damage. When developed successfully, this method will not only ensure better structural integrity and life safety but also mitigate economic losses associated with major seismic events.
Localized Damage Detection using Cumulative Absolute Velocity
In this study, a new local damage detection methodology using Cumulative Absolute Velocity (CAV) is developed. Application of the CAV as a local damage indicator is evaluated by analyzing acceleration response during a shaking table test and two instrumented buildings damaged during an earthquake. The presented method enables identifying the onset and location of damage to ensure structural integrity and resiliency.
Response of a Tall Building of a Long Distance Earthquake
Developed a finite element model of a 49-storey building and carried out time-history analysis to asses the performance of that building under Northridge (near fault,strong) and Chichi (Longdistance,weak) earthquakes. The model was verified with CSMIP records which showed good correlation
Ground Motion Modelling Projects
-NGA-West2: worked as a part of the team that developed attenuation models for active tectonic regions.
-NGA-East: worked as a part of the team that developed attenuation models for Central and Eastern North America.
-NGA-Sub: worked as a part of the team that developed attenuation models for subduction regions.
-Investigation of Ground Motions Recorded during the 2014 South Napa Earthquake
-Multivariate Prediction of Moment Magnitude for Small-to-Moderate Magnitude Earthquakes in Iran
