Scientists have developed a new technology that significantly improves the accuracy of predicting the properties of oil reservoirs. This will allow more accurate assessment of field reserves and optimization of production.
Predicting the properties of oil reservoirs is a key stage in assessing the potential of fields. It involves creating 3D models based on geophysical data and determining rock characteristics: porosity, density and permeability. However, due to the heterogeneity of the structure of oil-bearing formations, traditional methods such as core analysis, electrical and acoustic logging do not always provide accurate results.
Specialists from the Perm National Research Polytechnic University (PNRPU) proposed a solution – machine learning technology to improve forecasting. They analyzed data from 238 wells from six fields, comparing them with laboratory core studies. Based on these data, a model was created that clarifies the geological structure of the deposits and recalculates reserves.
The results showed that oil reserves were higher than previously expected. This is due to the increase in average reservoir porosity in the new model. In addition, new promising areas were discovered, which made it possible to adjust the production plan.
The new technology has proven its effectiveness even in complex geology. Its advantages are high accuracy of forecasts and process automation. In the future, such solutions may become a standard in the study of oil fields, combining big data and modern analysis methods.
Source: geonews.ru
