BP神经网络预测软锰矿与黄铁矿干法制取工业硫酸锰最佳条件的研究  

BP Artificial Neural Network Calculating Optimum Conditions for Dry Preparation of Manganese Sulfate by Roasting Pyrolusite and Pyrite

作　　者：张红萍[1] 唐爱东[2] 彭宏[3] 

机构地区：[1]邵阳学院,湖南邵阳422000 [2]中南大学化学化工学院,湖南长沙410083 [3]中南大学资源加工与生物工程学院,湖南长沙410083

出　　处：《矿冶工程》2004年第5期57-60,共4页Mining and Metallurgical Engineering

摘　　要：采用软锰矿和黄铁矿焙烧制取工业硫酸锰进行了实验研究 ,通过热力学计算表明温度高于 5 0℃时反应可自发进行 ,并由人工神经网络预测得出了最佳工艺条件并进行实验验证。当浸出条件为 :焙烧时间 1.2h ,焙烧温度 5 40℃ ,MnO2 /FeS2 (纯物质重量比 ) =1.0 5 5时 ,Mn在稀硫酸锰溶液 (pH >3 .8)中 ,常温浸取 2 4h的浸取率达到 85 .74%。该工艺具有能源消耗少 ,生产成本低 ,锰的回收率高等优点。利用BP神经网络进行计算误差小 ,可以预测最优工艺条件。The process of preparing manganese sulfate by roasting pyrolusite and pyrite is studied. The thermodynamic calculation shows that the reaction can proceed automatically at the temperature over 50 ℃. The optimum technical conditions are calculated by BP artificial neural network and verified by experiments. The leaching rate of Mn can reach to 85.74% in the case that the ores are roasted at MnO 2/FeS 2 (mass ratio) of 1.055 and 540 ℃ for 1.2 hours and leached in dilute manganese sulfate solution(pH>3.8) at room temperature for 24 hrs. The technology has advantages such as low consumption of power, low cost of production and high recovery of manganese. Using BP artificial neural network can lead to less calculation deviation and calculate the optimum technical conditions.

关 键 词：BP神经网络 软锰矿 黄铁矿 焙烧 浸取 

分 类 号：TF111[冶金工程—冶金物理化学]