The brittleness index (BI) is a key indicator for hydraulic fracture in shale reservoir development. At present, it is mainly estimated by elastic properties or mineral contents. But it remains controversial that which estimation is suitable for shale reservoir. Therefore, we propose a BI estimation workflow guided by triaxial fracture experiments. Firstly, according to a good BI indicator from triaxial fracture experiments, we choose a more suitable estimation method between elastic properties and mineral contents. Then, we use the selected BI as output, velocities and density as input to establish a nonlinear model with Relevance Vector Machine (RVM) algorithm. And that model is applied in logging BI estimation. The application of this workflow in a gas field from southwest China shows that triaxial fracture experiments can effectively screen the more suitable BI estimation method, and the BI estimated by mineral contents is selected. Compared to Multiple Linear Regression, RVM has a higher correlation coefficient and higher prediction accuracy. Above all, the BI estimation workflow for shale reservoir in this paper is reliable in logging BI estimation, as well as seismic data, can be an important guidance for field production of unconventional shale reservoirs.

This content is only available via PDF.
You can access this article if you purchase or spend a download.