Publications

Activities during drilling and workover operations are reported regularly by field personnel in the Drilling Reports. Drilling reports contain a lot of important information that can be used to identify critical events or issues that has happened on a well. But extracting information from drilling reports is quite challenging. The large number of reports (activities from thousands of wells) and its unstructured format (free text with regularly found typos) are the main issues. Aligning with Digital Innovation and Acceleration program, PT. Chevron Pacific Indonesia conducted study to use Machine Learning Based Data Mining to extract critical information from drilling reports to identify Wellbore Integrity issues such as casing leak and casing parted. The study is conducted in Mahandini field which has the largest number of well in PT. Chevron Pacific Indonesia (around 10,000 wells). This Machine Learning approach eliminates the manual labor efforts to run through all the drilling reports one by one to identify any Wellbore Integrity Issue.