Metal-Loss Defect Sets
The metal loss defect set utilized in this project is installed in two sections of pipe. An 80-foot (24.4 m) long section contains the 'characterization' defect set in five rows along the pipe. A 'detection' defect set arranged into three rows is contained on an additional 40-foot (12.2 m) length of pipe. The defects contain variations in factors that affect defect severity (depth and length), as well as variations in factors that affect the ability to measure depth and length (width, surface angle, plan roundness). These factors cover some but not all of the sources of variation in natural corrosion defects.
The defects are uniquely defined by five parameters. Three of the five geometry parameters are defined in an axial cross section through the defect. These parameters are the defect's equivalent length (L), depth (D), and surface angle (SA). The depth is taken as the maximum depth throughout the defect. The equivalent length is taken as its cross sectional area divided by the depth. |
The other two geometry parameters were included to generalize the design to three dimensions - width (W) and plan roundness (PR). Width is taken along the circumferential direction of of the pipe. Like length, width is reported as equivalent width; equivalent width is found by taking a circumferential cross-section through the defect and dividing by the maximum depth. The characterization defect set consists of 128 defects located on the outside surface of two pipe joints. The defects range in depth from 20 to 80 percent, length and width from 1 to 6 inches, surface angles of 22.5 to 90 degrees, and plan roundnesses that are elliptical to almost rectangular. The parameter variations are given in the following link: |
The defects are layed out in rows along the pipe lengths. Efforts were made to prevent two or more larger defects from being located near each other. The layout of all defects is given in the following link. Each defect's length, width, and plan roundness and surface angle are reproduced. Different colors are used to indicate percent depth, and surface angle is represented by varying shades of a color. The relative position of each row is indicated on the vertical axis.
The detection defect set contains 33 smaller defects that are primarily useful for evaluating threshold detection limits. Fifteen of these defects extend the range of defects considered in the characterization defect set, allowing development of more complete characterization and detection routines. Eighteen have depths less than 20% or lengths less than 1 inch (25.4 mm). Since primary objective here is the characterization of larger defects, these 18 defects are not discussed further here.
The range of parameter values for depth, equivalent length, and equivalent width were chosen for the following reasons. A wide range of depths (i.e. 20% to 80%) is covered. Lengths and widths smaller than one inch were excluded since the primary objective is defect characterization rather than detection. The maximum defect length and widths are six inches (152 mm) because the additional effects of length and width beyond six inches (152 mm) are minimal.