| dc.description.abstract | The recently proposed Interferometric Neighboring Fracture (INF) localization method places unique and demanding constraints on relative travel time measurement accuracy and precision, while sampling a function of relative travel times between pairs of microseismic events as measured along a linear array. Conducting two synthetic trials, I analyze the relationship between event-receiver geometry and relative travel time measurement error and its effect on the feasibility of INF localization. The results indicate that even for typical hydraulic fracturing monitoring geometries, measurement error can exceed the feasible error limits of INF localization.
In order to mitigate this error, I propose a new relative travel time measurement technique, Modified Adaptive Steering (MAS), along with a unique preprocessing methodology, Progressive Template Extraction (PTE). Analyzing synthetic data sets with varying SNR ratios, and a field recorded microseismic data set, I compare the performance of PTE-preprocessed MAS to conventional cross-correlation (CXC). Results of both synthetic and field recorded data analysis indicate that PTE enhanced MAS outperforms CXC as a general lag measurement technique, reducing average lag error by as much as 1.25 ms at SNRs below 10. With respect to the unique constraints of the INF method, PTE-MAS produces as many as 4.2 times as many usable samplings of the relative travel time function, while reducing error in stationary position and lag by up to 15 m and 2.5 ms, respectively. | en |
