论文部分内容阅读
应用人工神经网络技术开发了0Cr18Ni9不锈钢碳含量在线预报模型。模型经过生产现场连续58炉在线验证,预报碳质量分数与实际碳质量分数误差在±0.015%之内的命中率达到89.66%。
The on-line prediction model of carbon content in 0Cr18Ni9 stainless steel was developed by using artificial neural network technology. The model was verified on-line continuously by 58 furnaces at the production site, and the predicted hit rate was 89.66% within ± 0.015% of the predicted carbon mass fraction and actual carbon mass fraction.