IAHR, founded in 1935, is a worldwide independent member-based organisation of engineers and water specialists working in fields related to the hydro-environmental sciences and their practical application. Activities range from river and maritime hydraulics to water resources development and eco-hydraulics, through to ice engineering, hydroinformatics, and hydraulic machinery.
Log On
About IAHRDirectoryCommitteesMy IAHRNews & JournalseLibraryeShopEventsJoin IAHRWorld CongressDonate
spacer.gif
spacer.gif eLibrary
spacer.gif eLibrary
You are here : eLibrary : IAHR World Congress Proceedings : 36th Congress - The Hague (2015) ALL CONTENT : Hydro-environment : Esstimating cod load s in combiined seweer overfloows with mmultivari ate and neurral netwoork model...
Esstimating cod load s in combiined seweer overfloows with mmultivari ate and neurral netwoork modells under ssemi-arid rrainfall cconditionss
Author : IGNNACIO ANDR¨¦SS-DOM¨¦NECH (1), M. ESTHERR G¨®MEZ-MARRT¨ªN (2), JOSEP R. MEDINA(3) && JUAN B. MARRCO (4)
ABSTRACT

Estimation off pollution loadds from comb ined sewer ovverflows (CSOO) is a major isssue for mitigaation of impaccts on the wateer
environment.. According too EU directivees, pollutant looads must be estimated in a frequency-mmagnitude analysis to betteer
assess their impact on thee water bodiees. This study focuses on eestimating the chemical oxyygen demand (COD) load in
CSO from onne of the main sewer trunks in Valencia (SSpain) to asseess impacts onn the waterfro nt. 42 events were recordedd,
modelled andd analysed duuring the periood 2008-20122 (quantity an d quality dataa). For each eevent, anteceddent dry periood
(T), rainfall d uration (D), p eak rainfall inttensity (I), rainnfall volume ( R), runoff voluume (V) and CCOD load (M) spilled into thhe
receiving wa ater body weree obtained. T is related to pollutant acccumulation in the catchmennt (build-up), RR and V to thhe
event magnittude and I to erosive proceesses (wash-ooff). In this paaper, two diffeerent models aare analysed to estimate MM.
First, an anaalytical multivaariate regresssive model is adjusted connsidering relevvant explanatotory variables . On the samme
basis, an arttificial pruned neural netwoork (NN) was trained to esstimate M, deppending on innput variabless with a hiddeen
layer. Both mmodels highlighht the same coounterintuitivee result in the sstudied case: M does not deepend on T. TThe multivariatte
model best fiit shows a quiite linear relat ionship betweeen R (or V) aand the COD l oads. This strrong depende nce between R
and M is alsoo deduced froom the NN moodel, which el iminates the TT, D and I inpputs, and only considers R to estimate thhe
COD load (MM) with a 10%% relative meaan squared errror on test daata. Semi-arid conditions of f the Valencia rainfall regimme
lead to very large anteceddent dry perioods. Accumulaated pollutantss in the catchhment have reeached their mmaximum ratees
and are not already influeenced by T. CConsequently, the higher raainfall or runofff volumes aree, the higher pollutant loadds
because of thhe huge amouunt of pollutantts accumulateed in the systeem and mobilissed during eacch event.
File Size : 808,781 bytes
File Type : Adobe Acrobat Document
Chapter : IAHR World Congress Proceedings
Category : 36th Congress - The Hague (2015) ALL CONTENT
Article : Hydro-environment
Date Published : 20/08/2015
Download Now