(1) Due to water disinfection techniques and procedures
implemented in Sant Joan Despı´ WWP (Barcelona, Spain),
THM compounds are ubiquitously formed at relative large
amounts, especially in spring season, and with predominance
of brominated type of compounds over chloroform.
(2) Twenty-three different WWP variables were measured and
investigated for their possible correlation with measured
concentrations of the different THM compounds at the
exit of the plant. Among these WWP variables, water
temperature, carbon filters age, water turbidity after
carbon filters, as well as input wells water amount, UV
absorbance, total concentration of organic carbon (the two
later after carbon filters) and post-chlorination were found
to have significant correlations and influence formation of
THM compounds. Some of these variables, like carbon
filter age and input wells water can be modulated by
appropriate plant management to decrease the levels of
THM concentrations at the exit of the plant.
(3) Using these WWP variables, multivariate regression
techniques, such as MLR and PLS, allowed for a good
average THM concentration modeling and prediction.
Average relative errors ranged between 12% and 23% for
the modeling of THM concentrations in calibration
samples and between 16% and 29% in external validation
samples for all investigated trihalomethanes. Total
trihalomethane concentrations were predicted the best,
(1) Due to water disinfection techniques and proceduresimplemented in Sant Joan Despı´ WWP (Barcelona, Spain),THM compounds are ubiquitously formed at relative largeamounts, especially in spring season, and with predominanceof brominated type of compounds over chloroform.(2) Twenty-three different WWP variables were measured andinvestigated for their possible correlation with measuredconcentrations of the different THM compounds at theexit of the plant. Among these WWP variables, watertemperature, carbon filters age, water turbidity aftercarbon filters, as well as input wells water amount, UVabsorbance, total concentration of organic carbon (the twolater after carbon filters) and post-chlorination were foundto have significant correlations and influence formation ofTHM compounds. Some of these variables, like carbonfilter age and input wells water can be modulated byappropriate plant management to decrease the levels ofTHM concentrations at the exit of the plant.(3) Using these WWP variables, multivariate regressiontechniques, such as MLR and PLS, allowed for a goodaverage THM concentration modeling and prediction.Average relative errors ranged between 12% and 23% forthe modeling of THM concentrations in calibrationsamples and between 16% and 29% in external validationsamples for all investigated trihalomethanes. Totaltrihalomethane concentrations were predicted the best,
การแปล กรุณารอสักครู่..