Use of artificial intelligence for phased array data analysis


In the past 20 years, several steps have been taken to better the quality of non-destructive testing. The introduction of portable phased-array instruments led to recorded ultrasonic data, which in turn led to a better control of the quality of the inspections. The introduction of online training led to a standardization of theoretical teaching, which in turn led to better trained inspector and analysts. Still, problems remain in terms of defect detection, identification and sizing. Data analysis still fully relies on the skills of a human analyst even if the volume of data is exploding. Unfortunately, while there is an explosion in the volume of data to analyze, the number of qualified data analysts is not increasing accordingly. The result is shorter time to assess the data and longer working hours for the data analysts. New tools must be provided to data analysts, so they can perform their task more efficiently and accurately. In this paper we are going to present how artificial intelligence is used to analyse non-destructive testing data.