Missing Log Measurements
During the development of the field, log measurements can be missing due to many reasons, including economical and technical.
Extra Measurements are Costly
Predictive reservoir models require a constant set of input well logs from individual wells. In addition to missing logs, some logs should be eliminated from input due to erroneous data. Extra log measurements are costly and sometimes impossible to perform.
Wells Subdivided into Clusters
Using the Self-Organizing Maps (SOMs) approach, wells are subdivided into clusters. A minimum set of well logs is identified for the application of this approach. Any other, not measured well logs, are predicted by using similarities to other wells and their logs within established clusters.
Added Value Gained
Fulfilling the missing logs with synthetic ones improves accuracy and predictive power of reservoir models.
- Canadian CHOPS operator
- Major European operator