融合处理速度和加速度记录的地震检测模型及其在新丰江水库的应用
摘要
With the construction of the "National Seismic Intensity Rapid Reporting and Early Warning" project, acceleration records data will be increasingly applied in earthquake science research.However, most current earthquake detection models use velocity records for training, which results in poor detection performance for acceleration records.This study utilized seismic records from the Guangdong Earthquake Network to train the PhaseNet_GD model for detecting velocity records and the PhaseNet_ITS model for detecting acceleration records.Based on this, a new intelligent earthquake data processing system was developed by combining the GaMMA, phase association method, and the HYPOSAT, earthquake location method.This system was used to process the 2023 M L 4.8 earthquake sequence in XinfengjiangReservoir, Heyuan, and detected events 3.8 times more 基金项目：