Reservoir Induced Seismcity

Abstract

Reservoir Induced Seismcity In most cases of reservoir-induced seismicity, seismicity follows the impoundment, large lake-level changes, or filling at a later time above the highest water level achieved until then. We classify this as initial seismicity. This ‘‘initial seismicity’’ is ascribable to the coupled poroelastic response of the reservoir to initial filling or water level changes. It is characterized by an increase in seismicity above preimpoundment levels, large event(s), general stabilization and (usually) a lack of seismicity beneath the deepest part of the reservoir, widespread seismicity on the periphery, migrating outwards in one or more directions. With time, there is a decrease in both the number and magnitudes of earthquakes, with the seismicity returning to preimpoundment levels.

Conclusion

In short, the results from three different predictive methods are quite same with each other, especially distinction between mathematics statistics model and artificial neural network model, which is also conform to past macroscopical analogical prediction, it is proved that the neural network method to predict the seismic tendency is of certain significance. The results are obtained in so harsh condition, which show the preferable selection of induced earthquake factors, to certain degree, it also reflects the basic characteristics of induced earthquake mechanism.