Iron release in drinking water distribution systems (DWDSs) has attracted significant attention in recent years. There are many factors influencing iron release in DWDSs. In this study, we select pH, dissolved oxygen (DO), alkalinity (Alk), hardness, chloride (Cl-), sulfate (SO42-), temperature (T), and hydraulic retention time (HRT) as factors to investigate. Stepwise regression and partial correlation analysis are used as methods for establishing nonlinear power exponent iron release models and determining the relative importance of these factors by comparing standardized regression coefficients and partial correlation coefficients. The results show that the iron release models fit well. The important factors are SO42-, HRT, Cl-, and T, and the relative importance of the effect of the factors on iron release is SO42-≈HRT > Cl- > T, when measured by standardized regression coefficients; whereas, the important factors are HRT, SO42-, Alk, Cl-, and T, and the order of importance is HRT > SO42- > Alk > Cl- > T, when measured by partial correlation coefficients. Combining these two results, we conclude HRT, SO42-, and Cl- are the most important factors affecting iron release in pipes for the conditions in this study. The relative importance of T and Alk changes depending on the actual situation.
Keywords: drinking water distribution system; iron release; partial correlation coefficient; standardized regression coefficient; stepwise regression.