Earthquake early warning system using real-time signal processing

An earthquake warning system has been developed to provide a time series profile from which vital parameters such as time until strong shaking begins, intensity of shaking, and duration of shaking, can be derived. Interaction of different types of ground motion and changes in the elastic properties of geological media throughout the propagation path result in a highly nonlinear function. We use neural networks to model these nonlinearities and develop learning techniques for analysis of temporal precursors occurring in the emerging seismic signal. The warning system is designed to analyze the first-arrival from the three components of an earthquake signal and provide a profile of impending ground motion, in as little as 0.3 sec after first ground motion is felt. For each new data sample, at a rate of 25 samples per second, the complete profile of the earthquake is updated. The profile consists of a magnitude-related estimate as well as an estimate of the envelope of the complete signal. The envelope provides estimates of damage parameters, such as time until peak ground acceleration (PGA) and duration. The system is trained using previous seismogram data. It has been implemented in hardware using silicon accelerometers and a standard microprocessor. The proposed warning units can be used for site-specific applications, distributed networks, or to enhance existing distributed networks. By producing accurate, and informative warnings, the system has the potential to significantly minimize the hazards of catastrophic ground motion.