Automatic Microseismic Event Detection Using Deep Learning: a Classification is Detection Method
Z. Zhao, L. Gross, Y. Zhou, L. Li, H. Li and R. Yang
Event name: 81st EAGE Conference and Exhibition 2019
Session: Poster: Deep Learning and Data Analytics – Seismic Applications A
Publication date: 03 June 2019
Info: Extended abstract, PDF ( 1Mb )
Price: € 20
Automatic event detection is of vital importance for real-time microseismic or passive seismic monitoring in the oil and gas industry. With the recent advances in artificial intelligence and computing power, deep learning has now become a reliable tool for automatic microseismic event detection and other problems. We present a novel “classification is detection” strategy by leveraging convolutional neural networks (CNN) based deep learning method for automatic microseismic event detection in low signal-to-noise ration environment. The CNN model is trained using thousands of manually picked and labelled genuine event and noise segments. Validation on additional field data demonstrates the model’s capability in recognising microseismic event patterns and distinguishing them from various noises. The trained model can also be deployed in real-time monitoring scenario to provide automated real-time event detection capability.