Strip tracking in hot strip mills

摘要

In the finishing mill, steel strip is rolled from thick slabs through pairs of rollers housed in a continuous train of seven roll stands. As the strip is rolled, unwanted lateral movement, known as strip tracking, can cause the strip to collide with the edge of the mill. Strip tracking control is currently a manual operation, relying on the skill of the operators. When tracking is observed, the stand tilt is adjusted asymmetrically, causing a camber in the strip, steering it towards the centreline. Tracking control can be automated if a reliable measurement of position is available. A vision-based system was developed to measure strip position. Cooling water, steam, high temperatures and electrical noise create a hazardous environment for electronic equipment and hamper image analysis. Hardware was specified to protect all equipment against the environment. A novel image analysis method combining predictive elements, filtering and Bezier curve fitting was created to allow measurements to be made with large amounts of cooling water obscuring the strip edges. The measurement system was designed to integrate with the existing mill systems, using the OPC protocol for communication. The system was created as a development system with only two cameras included, but allowed for additional cameras to be easily added and automatically detected. The results of the system showed that the image analysis techniques were effective, providing an estimated final resolution of 3.5mm/pixel, with measurements ±2mm within 60% confidence. Hardware performance provided good protection of the equipment against the environment but poor quality installation limited overall system performance. A computer model was developed to simulate tracking behaviour in the mill with non-linear variations of strip properties across the strip. The model was not completed to a satisfactory standard capable of producing useful results but the theories described could be developed further.