We explore the connection between microseismicity and stress flow through a machine learning approach that allows for clustering of major “events” in the stress field. By using an extension of Kostrov summation of moment tensor to obtain the strain tensors, in the context of a spatial-temporal clustering methodology, and making the assumption that the stress axes are aligned with the strain axes, we can image how stress evolves though space and time during a hydraulic fracture completion. Furthermore, we can use these estimates of stress/strain state using a machine learning approach to identify where major transition are occurring in the stress field.
Check out our expanded abstract published by SEG (Technical Program 2018) or contact us to get a copy.