STTR Phase II: An Inference Engine for an Intelligent Imaging System for Detecting and Eliminating Hot Rolled Surface Defects

小企业技术转移研究（STTR）第二阶段项目将为智能成像系统开发一个推理引擎，该系统可以检测和消除热轧操作中的表面缺陷。这些缺陷约占钢材废品的50%。 申报产品是一个自动系统，可生成适当的纠正措施以消除缺陷。建议进一步开发推理机，并在选定的工业案例上进行验证。 这项研究的潜在价值是每年为美国钢铁行业减少超过20万吨钢铁的材料浪费，或1.2亿美元的生产力。 预计它还将在北美带来效益，每年节省1.14 Tetra W-hr的能源，减少94，000吨的碳当量排放。 其他好处包括减少用水量和更有效的下游工艺。 该项目具有很强的教育意义，公司与学术界密切合作，并为学生实习提供便利。

The Small Business Technology Transfer Research (STTR) Phase II project will develop an inference engine for an intelligent imaging system that can detect and eliminate surface defects in hot rolling operations. These defects account for roughly 50% of steel rejects. The proposed product is an automatic system that generates appropriate corrective actions for defect elimination. It is proposed to further develop the inference engine and validate it on selected industrial cases. The potential value of the research is to reduce material waste by over 200,000 tons of steel, or $120 million in productivity, per year for the US steel industry. It is also expected to deliver benefits in North America with energy savings of 1.14 Tetra W-hr and reduced carbon-equivalent emission of 94,000 tons per year. Other benefits include reduced water usage and more efficient downstream processes. The project carries strong educational implication, with the company working closely with academia and facilitating student interns.