Research on Strip Steel Water Beam Mark Recognition Method Based on Data Statistical Characteristics

Aiming at the periodic temperature anomaly (water beam marks) caused by the cooling effect of walking beam heating furnace water beams during hot-rolled strip steel production, this study proposes an automatic recognition method based on temperature distribution feature analysis. By real-time acquisition of rough rolling exit temperature and rolling speed data, a temperature distribution matrix along the strip length direction is constructed. Combined with extremum point detection and interference elimination strategies, precise extraction of water beam mark features is achieved. Innovatively, a dynamic extremum merging algorithm is proposed to eliminate local interference through distance threshold determination. A pseudo water beam mark temperature difference screening mechanism is designed to retain significant temperature difference features based on actual water beam quantities. Finally, a scoring model is established to quantify water beam mark severity. Industrial experiments demonstrate that this method enables online identification of water beam marks, with scoring results showing 92% consistency with manual detection, providing data support for heating process optimization. Compared with traditional manual sampling, this method achieves fully automated production-line monitoring, reducing response time to under 30 seconds, significantly enhancing timeliness and reliability in strip quality control. It holds critical engineering value for reducing rolling defects induced by water beam marks.