STTR Phase I: An Inference Engine for an Intelligent Imaging System for Detecting and Eliminating Hot Rolled Seams

这项小型企业技术转移研究（STTR）计划将开发一套智能成像系统，作为流程设计和控制的智能顾问。将检测系统与推理机相结合，为成像系统提供智能。该推理引擎将基于预测多道次热轧过程中焊缝形成的软件，预测缺陷频率的概率方法以及设计优化和在线过程控制的框架。该项目将开发预测有限元算法，并为其提供处理材料、工艺和设计不确定性的能力。一种创新的基于模型的检测与控制系统，可以将检测与测量系统与基于故障发生概率的推理机相结合。这可以提供一种工具来确定缺陷的严重性和频率，并允许进行更改以纠正过程以消除缺陷。该推理引擎还可用于对不受控制的噪声因素不敏感的鲁棒滚动规程的离线设计。该技术更广泛的（商业）影响将是能够识别控制参数，以检测接缝中的表面缺陷。识别缺陷的能力将为轧钢工业提供提高生产率和改善表面质量的产品的能力。

This Small Business Technology Transfer Research (STTR) Project will develop an imaging system with intelligence so that it can be used as an intelligent advisor for process design and control. Integrating the inspection system with an inference engine provides the intelligence in the imaging system. This inference engine will be based on software for predicting the formation of seams in the multi-pass hot rolling process, a probabilistic approach for predicting defect frequency and a framework for design optimization and on-line process control. The project will develop the predictive finite element algorithm and to provide it the capability of handling material, process and design uncertainties. An innovative model based inspection and control system, could integrate inspection and measurement systems with an inference engine based on probability of on fault generation. This could provide a tool to determine the severity and frequency of defects and permit changes to be made to correct the process for defect elimination. This inference engine can also be used for off-line design of robust rolling schedules that are insensitive to uncontrolled noise factors. The broader (commercial) impacts of this technology will be the ability to identify control parameters to detect surface defects in seams. The ability to identify defects will provide the steel rolling industry to have the ability to increase productivity and have products with improved surface qualities.