Detecting the start of roll damage

. The Predictive Repeating Defect Analyser (PRDA) developed by ABB in Finland takes images that are longer than 10m and process them on a PC. The algorithm used goes through the images one by one, processing even the weakest anomalies, and determines whether something similar is turning up at different positions along the length. If there are at least three instances of an anomaly that appear at regular intervals, a cyclic effect is reported. The method has been applied to the hot rolling mill at the Outokumpu Tornio Works in Finland. An ABB Surface Inspection System (SIS) is located before the down coiler on the mill. Before the SIS, there is a Steckel mill and three mill stands. The diameters of all four working rolls, as well as the reductions achieved by each working roll are input into the SIS. It is then able to tell the operator when a working roll is starting to produce cyclic marking on the strip – and which stand is the most probable source of the defect. All detected defects are classified and the information shown on the operator’s PC. The operator can view the image and decide whether the defect will affect the end product quality. More information from Email harri.melin@fi.abb.com