One of the main issues of wastewater treatment is the silica deposition in steam turbines. Evaporation of silica with the steam in adequate concentration is one of the main sources of scale formation in steam turbines. In this study, the authors introduce the utilization of a genetic‐based approach—gene expression programming (GEP)—for solubility prognostication of the silica in superheated steam of boilers with respect to water silica content and pressure. The result of GEP mathematical approach is a new algebraic formula to achieve our goals. Developed model predicts the silica solubility in the range of 0.8–22.1 MPa and 1–500 mg/kg for pressure and boiler water silica content, respectively. The results show that the constructed model outperforms pre‐existing method and provides acceptable predictions with average absolute relative deviation (AARD%) of 9.46% and determination coefficient (R2) of 0.98. Outcomes of the performed sensitivity analysis reveal that the effect of pressure is more pronounced than the impact of the boiler water silica content on the estimation of the silica solubility. Moreover, an outlier analysis based on Leverage approach was performed confirming the validity of the used database in this study. Finally, it can be concluded that the recommended tool in this study can be a suitable candidate for prognosticating the silica scaling in industries dealing with steam. © 2018 American Institute of Chemical Engineers Environ Prog, 38:e13089, 2019
Read full abstract