Performance seismic oscillation risk analysis method based on three-layer data set neural network

Abstract

The invention relates to an anti-seismic technology analysis method, in particular to a performance seismic oscillation risk analysis method based on a three-layer data set neural network. The methodcomprises the following steps of: (1) performing seismic oscillation data acquisition and data noise removal; (2) performing data characteristic parameter extraction and initialization processing; (3)generating a training set, an interval set and a test set ; (4) training a multi-layer neural network based on the training set; (5) training a neural network output value based on the interval set,and calculating a mean value and a standard deviation of the relative errors of the output value; (6) training the neural network based on the test set to determine an output value, and calculating amagnitude interval based on interval confidence; (7) performing probabilistic earthquake risk analysis, and determining the annual transcendental probability and recurrence period of a performance earthquake; and (8) based on performance seismic oscillation and consistency probability, determining the magnitude and epicenter distance of performance seismic oscillation. And a new neural network training method is adopted to predict a seismic oscillation attenuation relationship, so that the universality and toughness of the attenuation relationship are improved.

Description

technical field [0001] The invention relates to an analysis method of anti-seismic technology, in particular to an analysis method of earthquake risk based on a neural network. Background technique [0002] Earthquake "time, space and intensity" are the three core issues of seismic hazard analysis technology. In 1968, Cornell proposed the probabilistic seismic hazard method. The magnitude-frequency function relationship reflects the distribution requirements of earthquake magnitude intensity; the Poisson distribution reflects the probability of earthquake occurrence over time; Spatial distribution needs. In order to correct the approximation of the assumption of a uniform distribution of potential source areas, the 1990 "Seismic Intensity Zoning Map of China" proposed a comprehensive probability seismic hazard method, which divided the potential source areas into secondary structures. The 2015 "Seismic Motion Parameter Zoning Map of China" included The potential source are...

AI Technical Summary

Problems solved by technology

At present, the ground motion attenuation relationship is restricted by the form of expression. There are only three parameters of ground motion (including intensity), magnitude, and epicentral distance. It is difficult to fully reflect the complexity of earthquakes, and it is only suitable for regional ground motion attenuation simulation.

Images