基于一维卷积神经网络的微波加热钛精矿温度预测

Temperature Prediction of Titanium Concentrate by Microwave Heating Based on One Dimensional Convolutional Neural Network

以微波碳热还原低品位钛精矿工艺研究为背景,准确预测微波加热物料的温度对提高加热过程的安全性和可靠性具有重要意义。针对微波加热钛精矿过程进行温度预测,以微波输入功率、加热时间、初始温度三个因素作为神经网络的输入量,构建一维卷积神经网络预测模型,并将预测结果与热有效能传递模型和通用传热模型预测结果进行对比,三者预测结果的均方根误差和决定系数的值分别为1.4527、0.9927与6.0355、0.8734及6.5986、0.8486。结果表明,该模型能有效预测结果,为后续生产过程提供理论指导。

Based on the research of microwave carbothermal reduction of low-grade titanium concentrate,it is of great significance to accurately predict the temperature of microwave heating materials for improving the safety and reliability of the thermal process.For the process temperature prediction of microwave heating titanium concentrate via microwave input power,heating time,initial temperature,three factors as input of neural network,builds a one-dimensional convolutional neural network predictive model,the root mean square error and coefficient of determination of the predicted results were 1.4527,0.9927 and 6.0355,0.8734 and 6.5986,0.8486,respectively,compared with the predicted results of the thermal effective energy transfer model and the general transfer model.The results show that the model can effectively predict the results and provide theoretical guidance for the subsequent production process.