Application of B-Delta and UrEDAS on Seismometer Sensor Data to Model the Uncertainty in Time-Critical Detection of Earthquakes Affecting Turkish High Speed Railways

Since some sections of the Ankara-Istanbul High Speed Railway are very close to the North Anatolian Fault Zone and intersects the fault at two regions, it is important to detect the upcoming destructive earthquake before it hits to the target. Among few earthquake recognition methods, two effective methods were identified for investigation on the Turkish earthquake events. Both methods use the first 3 seconds of the vertical component of acceleration signal after the arrival of the P-wave to provide estimates of magnitude and epicentral distance. One method estimates the epicentral distance of earthquake events, and the other method estimates the magnitude of earthquake. The two methods then take the estimated value to the pre-defined empirical model between magnitude, amplitude, and epicentral distance of the same signal to estimate the other value. This shows that each method requires a valid relationship between magnitude, amplitude, and epicentral distance. Establishment of such empirical models, one per measurement point would require a relatively high number of data series, meaning many earthquake events. To avoid such far-reaching process, this research decided to apply both methods simultaneously on the first 3 seconds of the vertical component of acceleration signal after the arrival of the P-wave. The outcome showed a feasible resolution in terms of the detection time.