To find an efficient method to reduce or to isolate the ground vibration due to seismic waves is an important topic of actual engineering. According to scholar's study, using trenches to isolate seismic waves is an economical and effective method. There are many suggestions and relations have been developed. Nevertheless, these empirical relations come mostly from statistics methods. In the present study, three kinds of artificial neural network approaches are suggested to be used to predict an effect of isolating vibration by using trenches. In addition, results obtained will be compared with this result of multiple variable regression method.
How to isolate the man-made vibrations has become an important issue in densely populated metropolitan areas and for structures housing sensitive equipment. In the various methods of seismic shakes isolation, by placing a suitable barriers such as in-filled or open trench between the source of waves and the area to be protected is an effective method of screening vibration. In the various methods of seismic shakes isolation, by placing a suitable barriers such as in-filled or open trench between the source of waves and the area to be protected is an effective method of screening vibration. Concerning the problem of vibration isolation, numerous researches have been conducted in the past. In early studies, the efforts of screening vibration have directed mainly toward the experimental work or analytical study. For instance, Woods (1968) performed an experimental investigation on the effectiveness of open trenches in reducing the amplitude of vertical ground vibrations. Pat and coworkers studied analytically the wave diffraction by spherical and parabolic obstacles. However, their solutions are confined to simply geometry and idealized conditions, as close-formed solutions cannot be easily gained for complex conditions.
