A new empirical method based on piecewise linear model to predict static Poisson's ratio via well logs

Abstract

Fracturing often occurs in exploration and development of unconventional oil and gas. Hence, rock mechanical parameters, like Poisson's ratio are very important. However, they are very difficult to predict. In order to solve this problem, the low permeability reservoirs of Chang 8 Member, Yanchang Formation, Triassic, in western Ordos Basin, China, have been selected as the study area. Firstly, 18 typical core plunger samples from 12 wells were selected and the rock mechanical parameters were measured and analyzed, including the helium porosity, the static Young's modulus, the static Poisson's ratio and the compressional and shear wave velocity. Secondly, after analyzing the experimental data, an abnormal phenomenon was found, showing that when the porosity is higher than 7%, the reduction in speed of static Poisson's ratio with the increase of porosity is greater than that one found when the porosity is lower than 7%. Thirdly, from the perspective of rock deformation mechanisms, the origin of the above abnormal phenomenon is proposed, explaining that the deformation is influenced by grains, pores, clay and their contact relationship. When porosity is lower than 7%, it is mainly determined by pores and clay, while it is influenced by not only the two former factors, but also the connected grains when porosity is higher than 7%. Fourthly, based on the above origin, a new empirical method based on piecewise linear model with the boundary of porosity of 7% is proposed in order to predict the static Poisson's ratio through the well logs. Their correlation coefficient squares are 0.92 and 0.70 and their root mean square errors (RMSE) are 0.02 and 0.05 for the modeling and testing data respectively, which are much better than the values occurring in the linear model. In the field study and another study area, the predicted static Poisson's ratio and brittleness indices via piecewise linear models well match with the experimental results, exhibiting a better effect of the application. This study provides a new idea and feasible solution for the static Poisson's ratio prediction and is of critical importance for well design and prediction of formation fracturing.