Prediction of casing damage: A data-driven, machine learning approach

Abstract

Casing damage is the result of a number of factors in the long process of oilfield development, so it must be correctly judged and repaired in time to ensure the normal production of the oil fields. With the development of data science, it has always been an imperative problem remained to be solved. In this paper, we adopt a data-driven and the machine learning approach to casing damage forecasts. Firstly, from the fields of geology, engineering and development, a lot of history data is collected and processed. Then, based on these dynamic and static data samples, the random forest algorithm is used to create the casing damage prediction model. Finally, after the model is tested in two fault blocks, the results indicate that accuracy rates are 91% and 75%, which proves the validity and performance of the mode.