Development of online soft sensors and dynamic fundamental model-based process monitoring for complex sulfide ore flotation

摘要: Abstract Complex sulfide ores are difficult to process and often require multi-stage sequential flotation. Process outputs such as grade recovery in each stage affected by various sub-processes the system, it is crucial monitor performance order maximize production. In this work, we have proposed implemented a dynamic monitoring scheme using fundamental modeling an online soft sensor network for real-time measurements of recovery. Dynamic models lead zinc were developed represent rougher flotation lead-zinc ores. A was built measure support vector machine classification regression on multivariate image data. factorial design with feed particle size, collector dosage stage, variables used obtain diverse conditions validation. Successful validation at entire range demonstrates potential technique use control applications. Changes monitored stages state parameter estimates model structure. The framework can be extended other key process.