Summary The oil and gas industry is awash with sub‐surface data, which is used to characterize the rock and fluid properties beneath the seabed. This drives commercial decision making and exploration, but the industry relies upon highly manual workflows when processing data. A question is whether this can be improved using machine learning, complementing the activities of petrophysicists searching for hydrocarbons. In this paper, we present work using supervised learning with the aim of decreasing the petrophysical interpretation time down from over 7 days to 7 minutes. We describe the use of mathematical models that have been trained using raw well log data, to complete each of the four stages of a petrophysical interpretation workflow, in addition to initial data cleaning. We explore how the predictions from these models compare against the interpretations of human petrophysicists, and numerous options and techniques that were used to optimize the models. The result of this work is the ability, for the first time, to use machine learning for the entire petrophysical workflow.

No abstract

Our study applied a geophysical well log analysis, rock physics diagnostics and rock physics modelling to an exploration well log data from a shale gas exploration area in the Sichuan Basin of South China. The study established an unconsolidated model (80% quartz plus 20% clay in the shale gas formation) transform between the acoustic and elastic impedance on the one hand and lithology, porosity, water saturation, clay content, quartz content, and TOC content on the other hand. Through our geophysical well log analysis, we calculated mineral volumes using best available data, total and effective porosity, water saturation, and bulk density and V S prediction where it was missing. For rock physics modeling, the shale gas formation matrix substitution (Clay, Quarzt and TOC) and porosity modeling were performed in this exploration well. Crossplots are also used to analyse the elastic properties of the shale gas formation including V P velocity vs density, Acoustic Impedance (AI) vs total porosity ( T ), AI Ratio (PR), and V P vs V S . The results were quality controlled by core sample laboratory analysis data. To understand seismic effect as a result of rock physics modeling, ray traced synthetic modelling will be applied. The Ray-traced synthetics will be generated for the in situ and modeled scenarios for future AVA analysis. These transforms will be upscaled and applied to acoustic and elastic impedance inversion volumes to map lithology, porosity, and TOC distribution in the shale gas exploration area.

No abstract

No abstract