The effect of different weight functions on calibrating corrosion measurements using KNN technique

Pipeline corrosion scan devices have different accuracies which may affect the reliability of corrosion measurements used to investigate the pipeline's integrity assessment. In this paper K-Nearest Neighbor (KNN) interpolation technique was used to enhance the accuracy of corrosion measurements collected by two corrosion scan devices. Considering that KNN is a weighted interpolator, the use of the proper weight function will raise the chance of representing the original points correctly. The effect of using various weight functions within (KNN) were tested to determine the most appropriate interpolator for calibrating corrosion metrics. The comparison has relied on the effect of the weight function on the neighbor points, the position of the interpolator, and the error between the original and the enhanced metrics using the interpolation technique. (KNN) interpolation showed the potential to enhance corrosion measurements collected by several scan devices with different accuracies. This enhancement will be used to improve the integrity assessment report that depends on the disturbed corrosion metrics of oil and gas pipelines, to decide whether the pipeline is fit for service or needs certain maintenance. © 2016 Author(s).