Author(s): Ignnacio Andréss-Doménech; M. Estherr Gómez-Marrtín; Josep R. Medina; Juan B. Marrco
Linked Author(s):
Keywords: Combined sewer ooverflow; COD load; Multivariaate model; Neurral network; Semmi-arid climate
Abstract: Estimation of environment. assess their CSO from on modelled and (T) ,rainfall d receiving wa event magnit First, an ana basis, an art layer. Both m model best fi and M is also COD load (M lead to very and are not because of th f pollution load. According to impact on the e of the main d analysed du uration (D) ,p ter body were tude and I to lytical multiva tificial pruned models highligh it shows a qui o deduced fro M) with a 10%large anteced already influe he huge amou ds from comb o EU directive e water bodie sewer trunks uring the perio eak rainfall int e obtained. T erosive proce ariate regress neural netwo ht the same co ite linear relat om the NN mo%relative mea dent dry perio enced by T. C unt of pollutant ined sewer ov es, pollutant lo es. This study in Valencia (S od 2008-2012tensity (I) ,rain is related to sses (wash-o sive model is ork (NN) was ounterintuitive ionship betwe odel, which el an squared err ods. Accumula onsequently, ts accumulate verflows (CSO oads must be focuses on e pain) to asse2 (quantity an nfall volume (pollutant acc off) .In this pa adjusted con trained to es e result in the s een R (or V) a iminates the T ror on test da ated pollutants the higher ra ed in the syste O) is a major is estimated in stimating the ess impacts on d quality data R) ,runoff volu cumulation in aper, two diffe sidering relev stimate M, dep studied case: and the COD l T, D and I inp ta. Semi-arid s in the catch ainfall or runof em and mobilis ssue for mitiga a frequency-m chemical oxy n the waterfro a) .For each e ume (V) and C the catchmen erent models a vant explanato pending on in M does not de oads. This str puts, and only conditions of ment have re ff volumes are sed during eac ation of impac magnitude an ygen demand nt. 42 events vent, anteced COD load (M) nt (build-up) ,R re analysed tory variables put variables epend on T. T rong depende considers R the Valencia eached their m e, the higher ch event. cts on the wate alysis to bette (COD) load were recorded dent dry perio spilled into th R and V to th to estimate M. On the sam s with a hidde The multivariat nce between to estimate th rainfall regim maximum rate pollutant load er er in d, od he e M. me en te R e me es ds
Year: 2015