Thesis (MSc)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: Rivers are one of the earth's major readily available sources of fresh water. Abstractions from rivers are however not without problems. Firstly, river ow is variable and to deliver a constant yield is difficult. Secondly, rivers transport sediment which will be included in the diverted ow. Sediment control at diversion works have been studied for many years and this study attempts to gain further knowledge on certain sediment control features of diversion works. Sediment control at diversion works and abstraction works is crucial to prolong the life of the mechanical components like pumps and turbines. A Commonly used diversion works design is one with a low weir and a graveltrap. The weir dams water for abstraction, which is of importance in South Africa with its variable rainfall and river ow. The study focused on the following design features of diversion works: (1) the intake angle, which is the angle at which the structure is pushed into the river, (2) the intake opening height above a datum, (3) the river ow range where sediment is sufficiently scoured from the graveltrap and (4) the efficiency and river ow range of sediment ushing through a sluice gate at the graveltrap. A Physical model study was conducted in the laboratory of the University of Stellenbosch, which consisted of designing the diversion works that were to be tested. The designs were based on guidelines from previous studies, case studies and hydraulic principles. The above mentioned features (1-4) were studied at three structures with prototype weir sizes of 2.5 m, 3.5 m and 4.5 m. The river was modelled as a straight rectangular channel with a loose bed surface, which was simulated with crushed peach pips. Sediment was also fed into the system with a conveyor belt feed system. Pumps were used to abstract water and sediment through the intake opening, during the diverted sediment tests. Flow was diverted at a specific ow rate for each structure. The diverted sediment was caught and weighed. Each structure was designed to divert sediment through one of three intake opening heights, to determine whether a higher intake opening sufficiently reduces the amount of diverted sediment. The self-scour efficiency at the graveltrap was determined with a sediment level survey in the graveltrap. From the survey a clearance ow was determined, which is the minimum river ow that clears the intake opening of sediment along its complete length. It was also determined what intake angle induces secondary ow which results in the lowest clearance ow. The sediment ushing through the sluice gate was evaluated by recording the time it takes a full graveltrap to be ushed clean at various river ow rates. The maximum river ow at which the graveltrap still ushes efficiently was determined for each structure. It was found that between the 300, 450 and 600 intake angle that were tested, the 60 0 angle yields the lowest diverted sediment ratio (DSR) over the range of structures as well as river ows tested. The tests yielded a river ow at each structure where the DSR is at minimum. During the self-scour tests of the graveltrap, it was determined that a 450 intake angle promotes better self-scour at the graveltrap. To promote both features, a 450 intake angle is suggested, as it reduces diverted sediment and has a lower risk of issues due to too large ow constriction. The intake opening height was evaluated with analysis of diverted load and concentration. The conclusions on the intake opening vary between structure sizes. In the case of the smallest structure, with a 2.5 m weir height, the improvement observed for intake openings higher than the first (lowest) were variable. In the case of the 3.5 m weir structure, the results showed three consecutive intake openings could be feasible. In the case of the 4.5 m weir structure, less improvement was observed between the highest two intakes. Flood frequency should determine whether an intake opening with top-of-inlet of 1.6 m or 3.3 m above the minimum operating level should be designed. It was observed during the sediment ushing tests that submergence of either the graveltrap wall and/or the downstream water level affects the ushing efficiency. y3/y2, which is the downstream ow depth over the contracted ow depth under the sluice gate of the graveltrap, was evaluated as an indicator of efficient ushing. The study found that a good guideline would be to ush during river ows where y3/y2 < 1, while also ensuring the ow over the graveltrap wall entrains the sediment in the graveltrap. A figure which plots the downstream ow depth over sluice gate opening size was developed to serve as an operational guideline to efficient sediment ushing. The figure shows zones of efficient and non-efficient ushing. Further, the observed sediment ushing and self-scour ranges at each structure are also represented graphically. The fact that there was designed for a specific river scenario and also the lack of varied model sediment size, limits the applicability of the findings and conclusions. / AFRIKAANSE OPSOMMING: Riviere is van die aarde se hoof, maklik beskikbare bronne van vars water. Onttrekking uit riviere is wel nie sonder probleme nie. Eerstens is rivier vloei wisselvallig en om 'n konstante lewering te handhaaf is moeilik. Tweedens, vervoer riviere sediment wat ingesluit sal wees in die uitgekeerde vloei. Sediment beheer by uitkeerwerke word al vir baie jare bestudeer en hierdie studie poog om verdere kennis te verkry oor sekere sediment beheer funksies van uitkeerwerke. Sediment beheer by uitkeerwerke en onttrekkingswerke is noodsaaklik om die lewensduur van meganiese komponente soos pompe en turbines te verleng. 'n Algemeen toegepaste uitkeerwerke ontwerp is een met n lae keerwal en gruisvangkanaal. Die keerwal dam water op, wat nodig kan wees om die lewering te handhaaf, veral met Suid-Afrika se wisselvallige reënval en rivier vloei. Die studie het gefokus op die volgende ontwerp funksies van uitkeerwerke: (1) die inlaathoek, wat die hoek is waarteen die struktuur in die rivier ingedruk is, (2) die inlaatopening hoogte bo 'n datum, (3) die rivier vloei reeks waar sediment voldoende uitgeskuur word uit die gruisvangkanaal uit en (4) die effektiwiteit en rivier vloei reeks van 'n sediment spoel aksie deur 'n sluishek in die gruisvangkanaal. 'n Fisiese model studie was onderneem in die laboratorium van die Universiteit van Stellenbosch, wat bestaan het uit die ontwerp van die uitkeerwerke wat getoets sou word. Die ontwerp is gebasseer op riglyne van vorige studies, gevallestudies en hirouliese beginsels. Die bogenoemde funksies (1-4) was bestudeer by drie strukture met prototipe keerwal hoogtes van 2.5 m, 3.5 m en 4.5 m. Die rivier was gemodelleer as 'n reguit, reghoekige kanaal met 'n los bed oppervlakte, wat gesimuleer is met fyngemaakte perske pitte. Sediment was ook in die sisteem ingevoer met 'n vervoerband voer sisteem. Pompe was gebruik om water en sediment te onttrek deur die inlaatopening tydens die uitgekeerde sediment toetse. Vloei was uitgekeer teen 'n spesifieke vloeitempo vir elke struktuur. Die uitgekeerde sediment was gevang en geweeg. Elke struktuur was ontwerp om sediment uit te keer deur een van drie inlaatopening hoogtes, om te bepaal of 'n hoër inlaatopening hoogte die hoeveelheid uitgekeerde sediment voldoende verminder. Die self-uitskuur effektiwiteit van die gruisvangkanaal was bepaal deur 'n sediment vlak opmeting in die gruisvangkanaal. Vanaf die opmeting was 'n skoonmaak vloei bepaal, wat die minimum rivier vloei is wat die inlaatopening skoon maak van sediment oor die totale lengte. Dit was ook bepaal watter inlaathoek veroorsaak sekondêre vloei wat die laagste skoonmaak vloei oplewer. Die sediment spoel aksie deur die sluishek was geëvalueer deur die tyd wat dit neem om 'n vol gruisvangkanaal skoon te spoel, teen verskeie rivier vloeitempos te bepaal. 'n Maksimum rivier vloei waarteen die guisvangkanaal steeds effektiewelik skoon spoel was bepaal vir elke struktuur. Dit was bevind dat tussen die 300, 450 en 600 inlaathoeke wat getoets is, lewer die 600 hoek die laagste uitgekeerde sediment verhouding (USV) oor die reeks van strukture, asook rivier vloeitempos wat getoets is. Die toetse het 'n rivier vloei opgelewer by elke struktuur, waar USV 'n minimum was. Gedurende die self-uitskuur toetse was dit bepaal dat 'n 450 inlaathoek beter uitskuur in die gruisvangkanaal bevorder. Om beide funksies te bevorder word 'n 450 inlaathoek voorgestel, omdat dit ook uitgekeerde sediment verminder en 'n laer risiko van probleme as gevolg van te groot vloei vernouing het. Die inlaatopening hoogte was geëvalueer met analise van die uitgekeerde sediment lading en konsentrasie. Die gevolgtrekkings oor die inlaatopening hoogte varieer tussen struktuur groottes. In die geval van die kleinste struktuur, met 'n 2.5 m keerwal hoogte, was die verbetering wat waargeneem was by inlaatopeninge hoër as die eerste (laagste) inlaat, wisselvallig. In die geval van 'n 3.5 m keerwal struktuur het die resultate getoon dat drie opeenvolgende inlaatopeninge kan uitvoerbaar wees. In die geval van 'n 4.5 m keerwal struktuur was minder verbetering waargeneem tussen die hoogste twee inlate. Vloed frekwensie moet bepaal of 'n inlaatopening hoogte met 'n bokant-van-inlaat vlak van 1.6 m of 3.3 m bo minimun bedryfvlak moet ontwerp word. Dit was waargeneem dat gedurende die sediment spoel toetse dat versuiping van die gruisvangkanaal muur en/of die stroomaf watervlak die spoel effektiwiteit beïnvloed. y3/y2, wat die stroomaf vloeidiepte oor die vernoude vloeidiepte onder die sluishek van die gruisvangkanaal is, was geëvalueer as 'n indikator van effektiewe spoel aksie. Die studie het bevind dat 'n goeie riglyn sal wees om te spoel tydens rivier vloeie waar y3/y2 < 1 is, terwyl dit ook verseker moet word dat vloei oor die gruisvang kanaal sediment meevoer in die gruisvangkanaal. 'n Figuur wat die stroomaf vloeidiepte teenoor die sluisopening grootte plot was ontwikkel om te dien as 'n bedryfsriglyn tot effektiewe spoel aksie. Die figuur toon zones van effektiewe en nie-effektiewe spoel aksie aan. Verder is die waargeneemde sediment spoel aksie en self-uitskuur reekse van elke struktuur ook grafies voorgestel. Die feit dat daar ontwerp is vir 'n spesifieke rivier scenario asook die gebrek aan variëerende model sediment grootte, beperk die toepasbaarheid van die bevindings en gevolgtrekkings.