Machine Learning Confidently Predicts Subsurface Attributes for Unconventional Basins

Synthetic well log curves fill data gaps in the subsurface for five standard log types resulting in full quad-combo-equivalent curve coverage at a basin scale.

Wireline logs are a fundamental aspect of subsurface property characterization. However, economic constraints often limit data acquisition from specific logs or depth intervals, resulting in incomplete information from the well’s surface to its base in many areas. An alternative that addresses the lack of data is to create synthetic curves. With the advent of data science and the availability of digitized well data, machine learning (ML) algorithms can be used to predict missing logs or log intervals. TGS’ Analytics Ready LAS (ARLAS) leverages both the vast amount of well log data in the TGS library as well as data management infrastructure. TGS algorithms provide curve predictions for five standard curves, including the confidence intervals of each log which can be used for automated interpretation such as facies classification or basin stratigraphy.

Read the full article at Well Data Insights - Well Intel.

Click here to learn more about ARLAS and request sample data.