Publications

The Wanda Complex field was first developed in 1971 and produces from the sandstone reservoir of Talangakar Formation. Over time, oil production in the Wanda Complex Field experienced sand problems which had a negative impact on production wells and surface facilities. Parameters are needed that can be used to predict when a sand problem will occur so that engineers can mitigate the problem, one of which is the cementation factor. In practice, determining the cementation factor value requires core measurements which require large costs and time. Besides that, reservoir pressure as dynamic data can also be used to determine when a well stops producing due to sand problems. Currently, a feasible approach to establish correlations between cementation factor and reservoir pressure to predict the occurrence of sand problems has not been developed.

The application of machine learning methodology is used to predict cementation factor within a certain depth range of a well, by utilizing electrical log data. Various machine learning models will be used to determine the most appropriate model in predicting cementation factor throughout the well interval. In addition, this research aims to identify reservoir dynamics data that shows a strong correlation with predicted cementation factor.

Analysis of the developed models revealed that ensemble or decision tree-based models showed good performance in predicting cementation factor. The observed correlation between cementation factor and reservoir pressure explains that a lower cementation factor requires a greater reservoir pressure to hold the bonds between rock grains. The observed linear relationship between cementation factor and reservoir pressure suggests that this correlation can be used to predict when sand problems occur and carry out related mitigation.