Value from free text: An active learning approach

This is a presentation from WESC 2019 mini-symposium 8.1b on Artificial intelligence and big data. This presentation gives an overview of the O&M insights and value contained in free text maintenance records and highlights some of the issues that prevent operators from exploiting this value analytically. A text mining methodology is described to help operators overcome these issues and save up to 90% analysis time. The focus of the research was on involving real wind farm data analysts in the software prototyping process and create a solution that is both analytically effective as well as user-oriented and accessible for an operator. (Please note that a whiteboard was used to illustrate part of the methodology, which is not covered in the slide pack.)