Life Prediction Study And Maintenance Strategy Of Top Blowing Smelting System

Top blowing smelting system is a melting equipment which always used in metallurgical industry.This system is widely used in fire smelting,such as metal copper,lead,zinc,etal.This system has a lot of advantages,for example: compact structure,easy to control and operate,high productivity,low pollution,high adaptability,etal.In 2002,a smelting company introduced first top blowing smelting equipment in China for copper metallurgy.In the actual metallurgical production process,the system had frequent failures,and it's a great problem in prediction.So there are lots of influences on normal production efficiency and economic benefits.Therefore,in this article,nearly four years' production data of a smelting company is used,and based on predecessors' reliability analysis of top blowing system and life prediction study of core component spray gun,feature fusion and life prediction of key components,furnace lining and flue,are made.Then maintenance strategy of top blowing smelting system is put forward.In this article,these contents are researched:(1)Feature fusion of furnace lining and flue in top blowing smelting systemBy analyzing collected data in nearly four years from a smelting company,feature fusion of key components in system is made by manifold learning method.Feature fusion of furnace lining and flue in top blowing system is made by principal component analysis,isometric feature mapping and local linear embedding.Results got are compared and analyzed.It's found after analyzing results,effect of feature fusion based on local linear embedding is the best.This makes a pave for later life prediction study.(2)Life prediction analysis of furnace lining and flue in top blowing smelting systemLife prediction analysis of key component furnace lining and flue in top blowing smelting system is made by models support vector machines,sensor,B-P neural networks,etal.Results of life prediction are compared and analyzed.It's found that effect of life prediction based on support vector machines is the best.Results of feature fusion and life prediction analysis are sum up to prepare for later research.(3)Optimization of preventive maintenance strategy in top blowing smelting systemBased on feature fusion and life prediction analysis of key components,furnace lining and flue,in top blowing smelting system in past two parts and previous research,model of preventive maintenance strategy is optimized by quadratic interpolation.The aim of saving maintenance costs and reducing maintenance frequency can be achieved.