Earthquake destructive power prediction method and device based on bidirectional gating circulation unit

Abstract

The invention relates to an earthquake destructive power prediction method based on a bidirectional gating circulation unit. The earthquake destructive power prediction method comprises the following steps: acquiring earthquake oscillation time history data from an open source earthquake oscillation database to obtain sample data and sample labels; respectively decomposing various data included in the seismic oscillation time history data into a plurality of intrinsic mode function components through an empirical mode decomposition algorithm; taking each intrinsic mode function component obtained by decomposition as an input feature of a BiGRU model, and performing training in combination with a sample label to obtain an EMD-BiGRU model; and predicting the earthquake destructive power through the model. According to the method, the BiGRU model is adopted as a basic model to predict the seismic destructive power level, and past and future seismic wave data can be associated with current seismic wave prediction, so that model prediction is more facilitated; the EMD and the BiGRU model are combined, the new model comprises local feature signals of different time scales of seismic oscillation data, and high prediction precision can be achieved.

Description

technical field [0001] The invention belongs to the technical field of earthquake destructive force prediction, and relates to an earthquake destructive force prediction method and device based on a bidirectional gated cycle unit. Background technique [0002] After the earthquake, it is very important to quickly and accurately obtain the earthquake damage situation in the disaster area to dispatch rescue forces in a timely and reasonable manner and reduce casualties and losses in the disaster-stricken area. The uncertainty of earthquake disasters and the high complexity of the environment in the disaster area make it difficult to accurately and timely predict the earthquake damage in the disaster area. In addition to seismic parameters, many factors such as fault development, structural environment, and epicentral distance may affect the final prediction results. Traditional machine learning methods, such as Support Vector Machine (SVM) and Decision Tree (Decision Tree), a...

AI Technical Summary

Problems solved by technology

One is the Fragility Analysis method, which simplifies the earthquake intensity index and the regional structure model, which is difficult to fully reflect the complex time-domain and frequency-domain characteristics of earthquake motion, and the prediction speed is fast but the accuracy is insufficient

One is the nonlinear time history analysis method (Nonlinear Time-history Analysis, NLTHA), which has good retention of ground motion parameters and building structures, and high prediction accuracy, but this type of method has a large amount of calculation and requires high performance Computers, the timeliness of forecasting is somewhat lacking

Images