New Techniques for Intelligent Offset Data Analysis Using Artificial Intelligence

For well construction, planning is one of the most important phases in the delivery of a well. To select the right engineering solutions, it is important to analyze offset data. It is common for engineers to search, index, and catalog offset data. Today, as it has been historically for the industry, this is mostly a manual task. In the past few years, many advances in the way data have been collected, stored, and sorted have made the data easier to access, but actually determining which data are similar enough to use as offset data is still done manually. In simple cases, engineers draw a circle around the coordinates of the well to be drilled and any data within that circle would be considered offset data. This method considers only the geospatial similarity of the wells, and it is used because of time and technology constraints. In our work, we developed a method for a data-driven solution to this problem of identifying offset data. Data are critical for producing digital products and services. More than 2 years of research have led to the implementation of a powerful tool that allows users to search all global data to discover similar attributes for oil and gas data. In addition, the method includes the ability for users to define importance or allocate priorities to various features in the data. The mentioned advances in collecting, storing, and sorting data have been a major enabler in the evolution of this technology. This work introduces a developed framework that aids in identifying offset data by implementing a dynamic data-driven solution to analyze and display the similarity of the offset well data in a new way.