Struvite purity prediction by response surface methodology and chemical equilibrium modeling combination

The practical phosphorus recovery process lacks a useful tool to predict the purity of struvite production, identify the precipitation species and evaluate the influence of impurities on struvite formation. This study developed an integrated struvite purity prediction process based on response surface methodology (RSM) and Visual Minteq combination. The integrated process combined the advantages of the RSM and chemical equilibrium model to predict the purity during the recovery of struvite and identify the impurities in the recovered products. The second-order polynomial models were established by RSM to predict the nitrogen and phosphorus removal efficiency, and struvite purity in the presence of calcium. Visual Minteq was used to simulate precipitation process of impurity groups based on 528 different reaction conditions. Each reaction condition was verified by the RSM results. The precipitation group including struvite, CaHPO 4 , MgHPO 4 , brucite, and Ca 8 H 2 (PO 4 ) 6 ⋅ $\cdot$ 5H 2 O, with the smallest Z-score root mean square error weighted mean (-0.43), was most suitable for predicting the formation of struvite. The importance of the 5 possible precipitation combinations is ranked. The integrated process reduces the deviation of the prediction results from the reality and provides a useful tool to predict and optimize the quality of recovered products in the field of practical phosphorus recovery process.