基于灰狼算法优化的多隐层径向基神经网络铅锌烧结返粉料水分预测  

Moisture prediction of lead zinc sintering recycled-fine-ore based on multi-hidden-layer radial basis function neural network optimized by gray wolf optimizer

作　　者：刘再涛 王震 贺建军[1] 沈玲 邵雄奇[3] 江新辉 LIU Zaitao;WANG Zhen;HE Jianjun;SHEN Ling;SHAO Xiongqi;JIANG Xinhui(School of Automation,Central South University,Changsha 410083,China;College of Information Science and Engineering,Hunan Normal University,Changsha 410081,China;Shaoguan Smelter,Shaoguan 512024,China)

机构地区：[1]中南大学自动化学院,湖南长沙410083 [2]湖南师范大学信息科学与工程学院,湖南长沙410081 [3]韶关冶炼厂,广东韶关512024

出　　处：《中南大学学报（自然科学版）》2023年第2期754-764,共11页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(61873282)。

摘　　要：返粉料冷却是铅锌烧结流程中烧结配料回收利用的一道流程,对冶炼厂生产成本、烧结效率、环境保护有着重要的影响。针对冷却圆筒返粉料水分预估困难的问题,以影响冷却圆筒返粉料水分的主要变量为输入,冷筒出料端返粉料的水分为输出,构建水分预测模型。首先,采用孤立森林算法对采集的生产数据进行预处理以去除异常值。然后,以改进的灰狼优化算法(IGWO)优化的多隐层径向基函数神经网络(MRBFNN)建立冷却圆筒返粉料水分预测模型(IGWO-MRBFNN),实现对物料水分的精准预测,对不同模型的预测效果进行对比。研究结果表明:IGWO-MRBFNN模型的预测精度最高,误差最小;该预测模型的建立对于提高冶炼厂的生产效率和烧结质量以及环境保护具有重要意义。Cooling the recycled fine ore(RFO) is a procedure of recycling sintering ingredients in the lead-zinc sintering process, which has an important influence on the production cost, sintering efficiency and environmental protection of smelters. To solve the problem of difficulty in estimating the moisture of the RFO in the cooling cylinder, the main variables affecting the moisture of the RFO in the cooling cylinder were used as input variables,and the moisture of the RFO at the discharge end of the cooling cylinder as the output to construct a moisture prediction model. Firstly, isolation forest was used to preprocess the collected production data to remove outliers.Then, the moisture prediction model, multi-hidden-layer radial basis function neural network optimized by improved grey wolf optimizer(IGWO-MRBFNN) was established, so that the accurate prediction of moisture content of RFO was realized, and the prediction effects of different models were compared. The results show that the IGWO-MRBFNN model has the highest prediction accuracy and the smallest error. The establishment of this prediction model is of great significance for improving the production efficiency, the quality of sintering and protecting the environment for smelters.

关 键 词：烧结返粉料 水分预测 孤立森林 RBF神经网络 灰狼算法 

分 类 号：TP278[自动化与计算机技术—检测技术与自动化装置]