Evaluation of alternatives for hydraulic analysis of sanitary sewer systems

Van Heerden, George Adrian (2014-04)

Thesis (MScEng)--Stellenbosch University, 2014.

Thesis

ENGLISH ABSTRACT: This research project focuses on sanitary sewer systems. When performing an analysis of a sewer drainage system with known constraints, an appropriate model needs to be chosen depending on the objectives of the analysis. Uncertainties are also present in the analysis of sewer drainage systems. The uncertainties and the errors in hydraulic models need to be understood and considered. The required level of accuracy and the type of hydraulic problem that needs to be solved may alter the complexity of the hydraulic model used to solve a drainage system. The wide variety of available simulation models further complicates model selection. With various models available, selecting the most appropriate model for a particular drainage system simulation is important. The various models for sewer drainage system analysis can be categorised in different ways. For example, it is possible to categorise models according to their purpose, which could be evaluation, design or planning. Evaluation models are mainly used to test whether existing systems or planned systems are adequate and require the highest hydraulic detail. Design models are used to determine the size of conduits within a drainage system and require moderate levels of hydraulic detail. Planning models are primarily used for strategic planning and decision making for urban or regional drainage systems and require the least amount of hydraulic detail. An understanding of the available models is required in order to choose the most suitable simulation model for the desired purpose. Some models are derived from the Saint-Venant equations of flow. The most detailed models are typically referred to as fully dynamic wave models and utilise all the components of the Saint-Venant flow equations. By removing terms from the Saint-Venant equations a kinematic wave model can be created. Some less complex models ignore basic principles of hydraulics in order to make assumptions that simplify the process of simulating flows. In this thesis three different models were compared: a detailed model using fully dynamic flow equations, a simplified model using kinematic wave equations and a basic model using contributor hydrograph routing equations. For the drainage system analysis SWMM-EXTRAN was used as the fully dynamic wave model, SWMM-TRANSPORT was used as the kinematic wave model and SEWSAN was used as the contributor hydrograph model. Two drainage systems situated in South Africa were used as case studies and are referred to as Drainage System A and Drainage System B in this thesis. The actual flow rate was recorded at two points with flow loggers, one in each of the two systems. The flow rate was continually recorded at 1 hour intervals for the period 1 July 2010 to 9 July 2010 in Drainage System A as well as in Drainage System B. The same input parameters were used for each model allowing the modelled flow rates to be compared to the measured flow rates. The models provided peak flow results that were within 2% of the measured peak flow rates and the modelled mean flows were within 8.5% of the measured mean flows in most situations. However, when rapidly varied flows occurred the kinematic wave and contributor hydrograph models returned conservative results as they were unable to account for hydraulic effects such as acceleration. The effect of acceleration became most pronounced up and downstream of drop structures and sections where the slope changed considerably. The kinematic wave and contributor hydrograph models were therefore unable to accurately simulate surcharge conditions. The results suggest that the fully dynamic wave model can be used in all scenarios. The kinematic wave model can be used for a design analysis if no hydraulic structures occur in the system. The contributor hydrograph model should not be used for an evaluation analysis, but can be used for a design analysis if a relatively high level of confidence in the parameter set exists and no areas of rapidly varying flow or hydraulic structures exist within the system.

AFRIKAANSE OPSOMMING: Hierdie navorsing projek is gefokus op riool dreineringsisteme. Wanneer ʼn analise van ʼn riool dreineringsisteem met bekende beperkinge onderneem word, moet ʼn geskikte model gekies word afhangende van die doelwitte van die analise. Onbekendes is ook teenwoordig in die analise van riool dreineringsisteme. Dit word belangrik dat die onsekerhede en die foute in hidroliese modelle moet verstaan en oorweeg word. Die verwagte vlak van akkuraatheid en die tipe hidroliese probleem wat opgelos moet word mag die ingewikkeldheid van die hidroliese probleem, wat gebruik word om ʼn rioolsisteem op te los, verander. Die wye verskeidenheid van beskikbare simulasie modelle bemoeilik verder die keuse van ʼn model. Met etlike modelle beskikbaar vir seleksie, is die mees geskikte model vir ʼn spesifieke dreineringsisteem simulasie belangrik. Die verskeie modelle vir riool dreineringsisteem analise kan op verskillende wyses gekategoriseer word. Byvoorbeeld, dit is moontlik om modelle te kategoriseer volgens hulle doel, wat evaluasie, ontwerp en beplanning kan wees. Evaluasiemodelle word hoofsaaklik gebruik om te toets of huidige of beplande sisteme voldoende is en of hulle die hoogs moontlike hidroliese besonderhede benodig. Ontwerpmodelle word gebruik om die grootte van ʼn leipyp binne ʼn rioolsisteem te bepaal en benodig matige vlakke van hidroliese besonderhede. Beplanningsmodelle word hoofsaaklik gebruik vir strategiese beplanning en besluitneming vir stedelike en landelike rioolsisteme en benodig die laagste vlak van hidroliese data. ʼn Begrip van die beskikbare modelle is nodig om ʼn keuse te maak rakende die mees geskikte simulasie model vir die verlangde doelwit. Sommige modelle is afkomstig van die Saint-Venant vergelykings van vloei. Die mees gedetailleerde modelle word tipies na verwys as die volledige dinamiese golf modelle en benut alle komponente van die Saint-Venant vloei vergelykings. Deur die verwydering van terme van die Saint-Venant vergelykings kan ʼn kinematiese golf model daargestel word. Sommige minder gekompliseerde modelle ignoreer die basiese beginsels van hidrologie om aannames te maak wat die proses van golf simulering vereenvoudig. In hierdie tesis is drie verskillende modelle vergelyk; ʼn gedetailleerde model wat volledige dinamiese vloeivergelykings gebruik; ʼn vereenvoudigde model wat kinematiese golfvergelykings gebruik en ʼn basiese model wat bydraende hidroliese versending vergelykings. Vir die dreineringsisteem analise was SWMM-EXTRAN gebruik as die volledige dinamiese golfmodel, SWMM-TRANSPORT was gebruik as die kinetiese golfmodel en SEWSAN was gebruik as die bydraende hidroliese model. Twee dreineringsisteme in Suid-Afrika was gebruik as gevallestudies en word na verwys as Dreineringsisteem A en Dreineringsisteem B. Die werklike vloeikoers was aangeteken by twee punte met vloeimeters, een in elk van die sisteme. Die vloeikoers was deurlopend opgeteken met 1 uur tussenposes vir die periode 1 Julie 2010 tot 9 Julie 2010 in Dreinering Sisteem A sowel as Dreinering Sisteem B. Dieselfde inset parameters was gebruik vir elke model wat dit moontlik gemaak het dat die gemoduleerde vloeikoerse met die gemete vloeikoerse vergelyk kon word. Die modelle het spits vloeiresultate voorsien wat binne 2% van die gemete spits vloeikoerse was en, in die meeste situasies, dat die gemoduleerde gemiddelde vloei binne 8.5% van die gemete gemiddelde vloei was. Wanneer vinnig varierende vloei voorgekom het, die kinetiese golf and bydraende hidrograaf modelle konserwatiewe resultate gelewer het, aangesien hulle nie in staat was om hidroliese effekte soos versnelling te verklaar nie. Die effek van versnelling was op sy duidelikste stroomopwaarts en stroomafwaarts onder valstrukture en by gedeeltes waar die helling aansienlik verander het. Die kinetiese golf en bydraer hidrograaf modelle was gevolglik nie in staat om oorladingsomstandighede akkuraat te simuleer nie. Die resultate wys dat die volledige dinamiese vloeimiddel gebruik kan word in alle omstandighede. Die kinematiese vloeimiddel kan gebruik word vir ʼn ontwerp analise indien geen hidroliese struktuur in die sisteem voorkom nie. Die bydraer hidrograaf model behoort nie gebruik te word vir ʼn evaluerings analise nie, maar kan gebruik word vir ʼn ontwerp analise indien ʼn relatiewe hoë vlak van vertroue in die parameter stel bestaan en geen area van vinnig veranderende vloei of hidroliese strukture binne die sisteem bestaan nie.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/86534
This item appears in the following collections: