Synthetic Casing Corrosion Log Prediction Using Geospatial Analysis – A Digital Twin Concept

Downhole casing corrosion monitoring is a key element in production engineering as it ensures the integrity and safety of assets, maximizes the life and serviceability of a well, and contributes to a successful HSE management programs. Consequently, wells are frequently logged for corrosion and metal loss anomalies to monitor casing integrity. This study explores a method using geospatial analytical techniques to develop synthetic corrosion logs to optimize OPEX, supplement missing logs, and avoid production deferment and downtimes. The proposed method generates full synthetic corrosion logs using geospatial analysis based on available logs, then it maps metal loss defects across the entire field. The spatial mapping builds a 3D map based on depth using computational geometry and computer-aided engineering. Hundreds of thousands of data points from hundreds of logs, represented by (1) depth, (2) casing specifications, (3) cement properties, and (4) metal loss severity, have been fed into the framework to develop a variogram model using Kriging interpolation. By developing the variogram model, a map is generated at each depth point with casing metal loss ratio, and hence a full synthetic corrosion log is built. The data set of available corrosion logs was split into two parts; 70% for training the model and the remining 30 % for testing. Then a cross-verification check was done as well. The developed geospatial analytical model achieved an overall confidence level of 95% of all predicted logs generated using the geospatial analysis. Another scenario was initially studied that incorporates depth, metal loss percentages, and well age as the only input data points. However, this study yielded a lower accuracy level of only 90%. This percentage increased to 95% when incorporating formation characteristics, casing and cement properties into the model. The developed model enabled effective optimization of 1000 corrosion logs requirement through the generation of a full field metal loss severity map. The cost avoidance can be estimated to reach up to tens of millions of dollars due to the ability of predicting metal loss for critical wells without actual operation costs. On top of assuring well integrity, the developed method promotes health and safety of assets and personnel as it minimizes physical exposure of corrosive gases such as H2S.