烧结工况智能识别与质量在线预估系统的应用

Application of Intelligent Identification And Quality Online Prediction System for Sintering Conditions

由于烧结过程恶劣的生产环境、高昂的在线检测成本及较低的在线检测精度等原因,导致目前公司生产过程中的工况和质量只能通过离线检测和人工观察等方式获取,不能及时、有效地监控生产、调整工艺,难以提高质量。基于此,针对厂内烧结工况识别困难、识别结果主观性大以及烧结质量离线化验滞后等问题,通过数据及图像处理、深度学习及多信息融合等技术,实现了烧结生产过程多级工况在线判断,系统工况判断准确率达90%以上,质量在线预估误差在95±5%以内。

Due to the harsh production environment of sintering process,high cost of online detection and low accuracy of online detection and other reasons,the working conditions and quality of the company's production process can only be obtained through offline detection and manual observation,resulting in failure to timely and effectively monitor production,adjust technology and improve quality.In order to solve the problems such as the difficulty in the identification of sintering conditions in the plant and the lag in the offline testing of sintering quality by subjective large machine,the multi-stage condition online judgment of sintering production process was realized through data and image processing,deep learning technology and multi-information fusion technology.The accuracy rate of system condition judgment was more than 90%,and the quality online prediction error was less than 95±5%.