Water’s corrosion and scaling potential prediction using artificial neural networks and gene expression programming in several rural water distribution networks in Kermanshah Province, Iran
نویسندگان
1 Department of Soil Science, Razi University, Kermanshah, Iran
2 Department of Soil Science, Razi University, Kermanshah, Iran
3 Department of Water Science and Engineering, Razi University, Kermanshah, Iran
doi
10.22104/aet.2025.7194.1980چکیده
Water quality causes severe restrictions on the utilization of water resources. Corrosion and scaling are the most common problems in the operation and maintenance of water facilities. Corrosive indices are an indirect method of detecting and measuring water's tendency to corrosion and scaling. Water corrosion and scaling are complex phenomena that cannot be easily modeled. This study used meta-heuristics methods, such as artificial neural networks (ANN) and gene expression programming (GEP), to predict the water’s corrosion and scaling potential of the distribution network in some rural areas of Kermanshah Province. Equations were extracted to estimate water corrosion and scaling indices using linear regression and GEP. The results showed that ANN could reveal water corrosion and scaling indices with the highest correlation coefficient (0.95, 0.91, 0.96, 0.92, and 0.99) and the lowest percentage errors (0.20, 0.44, 0.40, 0.44, and 0.08) for the Langelier saturation index (LSI), Ryznar stability index (RSI), Puckorius scaling index (PSI), Aggressive index (AI), and Larson–Skold index (L-SI), respectively. Also, the linear and nonlinear relationships obtained by a high-precision GEP model (0.80 to 0.97) can estimate corrosion and scaling indices with lower cost and more accuracy by measuring the most influential physicochemical parameters.
