Gas-liquid Jet Pump Crystallizer in Phosphorus Recycling Technology – neural Network Model

AbstractThe struvite (MgNH4PO46H2O) continuous reaction crystallization process from phosphorus mineral fertilizers industry wastewater containing 0.445 mass % of phosphate(V) ions and typical impurities was investigated. Original gas–liquid jet pump DTM MSMPR (Draft Tube Magma, Mixed Suspension Mixed Product Removal) crystallizer fed with compressed air generating ascending suspension flow in a mixing chamber was applied. Influence of pH (8.5 – 10), mean residence time of suspension in a crystallizer τ (900 – 3600 s) and magnesium ions excess (0 – 20%) on population density distribution of product crystals was identified. Artificial neural network (ANN) was used for creation of numerical, fully functional model of the process