基于蚁群算法-BP神经网络的漏钢预报模型研究  

Breakout Prediction Model Based on ACO-BP Neural Network

作　　者：生万宝 张本国[1] 吴迪 张瑞忠 Sheng Wanbao;Zhang Benguo;Wu Di;Zhang Ruizhong(School of Mechanical Engineering,Yancheng Institute of Technology;Process Research Institute,Hegang Group Steel Research Institute)

机构地区：[1]盐城工学院机械工程学院,江苏盐城224051 [2]河钢集团钢研总院工艺研究所

出　　处：《特种铸造及有色合金》2022年第11期1366-1369,共4页Special Casting & Nonferrous Alloys

基　　金：江苏省基础研究计划资助项目(BK20150429);盐城工学院研究生科研(实践)创新计划资助项目(SJCX21-XY005)。

摘　　要：针对传统BP神经网络在漏钢预报过程中识别速度慢、无法精准预测等问题,采用蚁群算法(ACO)对随机选取的权值阈值进行寻优,详细介绍了ACO算法的优化步骤,利用MATLAB软件建立了神经网络模型,并将优化后的模型应用到漏钢预报中。将现场采集的数据进行预处理,再输入到神经网络模型中进行训练和测试。结果表明,ACO-BP漏钢预报模型的识别精度明显高于传统BP漏钢预报模型,漏钢预报率可达96.77%,报出率达100%,不仅加快了网络模型的运行速度,也保证了模型的全局搜索能力及鲁棒性,具有良好的应用前景。Aiming at the problems of slow recognition and inability to achieve accurate prediction during the process of traditional BP neural network for breakout prediction, ACO algorithm was utilized to optimize the randomly selected weight threshold, and the optimization steps of ACO algorithm was introduced in detail. The neural network model was established based on the MATLAB software, and the optimized model was applied to breakout prediction. The historical data collected on the site was preprocessed, which was then input to the neural network model for training and testing. The result indicates that the recognition accuracy of ACO-BP breakout prediction model is significantly higher than that of traditional ones, where the forecast rate and reporting rate can reach 96.77% and 100%, respectively. The model not only speeds up the operation of the network model, but also ensures the global search ability and robustness of the model, which has desirable application prospects.

关 键 词：连铸 BP神经网络 ACO优化算法 漏钢预报 

分 类 号：F777.1[经济管理—产业经济] TP311[自动化与计算机技术—计算机软件与理论] TG249.7[自动化与计算机技术—计算机科学与技术]