基于卷积神经网络的带钢质量预测  

Strip Quality Prediction Based on Convolutional Neural Network

作　　者：林刚[1] 王庆强[1] 宋泽平 杨英华[2] LIN Gang;WANG Qingqiang;SONG Zeping;YANG Yinghua(Ningbo Baoxin Stainless Steel Co.,Ltd.,Ningbo 315807,China;College of Information Science and Engineering,Northeastern University,Shenyang 110004,China)

机构地区：[1]宁波宝新不锈钢有限公司,浙江宁波315807 [2]东北大学信息科学与工程学院,辽宁沈阳110004

出　　处：《仪表技术》2020年第5期5-7,40,共4页Instrumentation Technology

摘　　要：带钢产品组织性能是产品质量参数的主要组成部分,参数的准确预测对生产线实现精准质量控制及质量稳定性具有重要意义。采用卷积神经网络,建立带钢产品质量与其过程参数之间的预测模型;利用该模型实现对带钢产品的延伸率、屈服强度、抗拉强度、硬度等质量指标的预测,从而可以及时准确地在线估计带钢质量。通过对某钢厂实际生产过程带钢质量的预报,预报的产品质量和现场实测值有较好的一致性,证明了所提出方法的可行性和有效性。The micro-structure and properties of strip products are the main components of parameters on the product quality.At the same time,the accurate prediction of the parameters is of great significance to the quality control and quality stability of the production line.In this paper,a convolutional neural network is used to establish a prediction model between the quality of the strip products and its process parameters.By using this model,the quality indexes such as elongation,yield strength,tensile strength and hardness of strip products can be predicted,so that the strip quality can be estimated timely and accurately.Through the prediction of strip quality in a steel plant,the predicted product quality is consistent with the measured value,which proves the feasibility and validity of the proposed method.

关 键 词：带钢 质量预测 卷积神经网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TG335.5[自动化与计算机技术—计算机科学与技术]