Automatic Event-Detection and Time-Picking Algorithms for Downhole Microseismic Data Processing

Microseismic event detection and time picking are critical for accurate microseismic locations. Both processes are often done automatically, considering the large volume of recorded microseismic data. Many event detection algorithms require a static or dynamic threshold criterion. In this paper, we propose a dynamic threshold criterion and evaluate its performance in comparison with static thresholds using STA/LTA and peak eigenvalue (PEV) methods. The comparison is based on a subset of microseismic data acquired during two different hydraulic-fracturing treatments in western Canada. We also compare the performance of several time-picking algorithms, STA/LTA, modified energy ratio (MER), modified Coppens’s method (MCM) and Akaike information criterion (AIC). In addition, we propose a hybrid time picking algorithm, joint energy ratio (JER), which combines peak eigenvalue ratio (PER) with STA/LTA. We find that the dynamic threshold approach yields detection results between the intermediate and low static threshold parameters, reduces false noisy detections and is capable of identifying weak signals. We also find that JER, MER and STA/LTA all perform equally well with good S/N data. However, JER provides more stable results with mixed S/N data. We recommend that a manual QC step should follow the automatic time picking to verify time picking accuracy.