Mechanize feature-to-feature matching system utilizing repeated inspection data

Abstract

The advances of computational methods and tools can greatly support other areas in doing tasks from the most tedious or repetitive to the most complex. In this paper, these advances were manipulated in civil structures maintenance specifically in pipeline corrosion assessment. This paper describes mechanize method developed to automatically detect and quantitY important parameters for future prediction of corrosion growth using In-line inspection (Ill) data. The focal process in this system includes data conversion, data filtering, parameter tolerance or sizing configuration, matching, and data trimming. A sensitivity analysis using linear regression method was used to correlates defects from one inspection to the next. Issues and advantage gain from this mechanize system is threefold~ Firstly, timeliness (manual matching procedure consumed a great deal of time). Secondly, accuracy and consistencies in data sampling (current implementation, different researcher obtain a different number of sample even though the method used in matching were the same), and finally, reduction of data matching error (manual matching was prone to human error due to the masses of inspection data to be match. Furthermore this method is impractical when facing the large amount of real inspection data).