A fuzzy neural network prediction model of the principal motions of earthquakes based on preliminary tremors

A technique to predict principal motions of earthquakes using preliminary tremors, has been developed. Taking advantage of the time lag between them, we can take suitable countermeasures against the principal motions that affect urban structures; e.g. an escape from dangerous zones, stopping elevators and gas supply, and activating AMD (active mass damper) systems. A structured neural network is used to construct a peak ground acceleration prediction model, where inputs are fuzzified shaking direction data, and power spectrum and maximum acceleration of preliminary tremors. The proposed model has been improved by handling some earthquakes in Ibaraki-ken south-west zone that least fit the model as exceptions. Mean square error of the improved model is reduced to one third of the statistical model.