Thesis (MScEng)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: When a sailing ship passes a moored ship the moored ship experiences forces and moments. These forces and moments cause the moored ship to move. The resulting ship motions due to the passing ship can sometimes be more severe than the ship motions due to ocean waves and can cause serious accidents at moorings such as the failing of mooring lines or even the total break away of the ship from the berth. Since bulk carriers and tankers were traditionally the largest seafaring ships, passing ship studies have focused mainly on these vessels, but recently container ships have grown to a comparable size. In this study an existing numerical model “Passcat” is validated with physical model measurements for a Post Panamax container ship passing a Panamax bulk carrier. Other existing mathematical formulae are also evaluated by comparison with these model tests. In the physical model tests the passing speed (V), passing distance (G), depth draft ratio (d/D) and the presence of walls and channels were varied. It was found that the passing ship forces are proportional to the passing speed to the power of 2.32. This is slightly higher than the generally accepted quadratic relationship for passing ship induced forces. Similar relationships were found for the other variables. The numerical model results were compared to the physical model measurements by determining agreement ratios. A perfect agreement between the numerical and physical models would result in an agreement ratio of 1. Agreement ratio boundaries, wherein agreement would be regarded as good, were drawn between 0.7 and 1.3. The numerical model, Passcat, was found to under predict the passing ship forces. It was found that Passcat is valid for a wide range of sensitivities and remains within the agreement ratio limits as long as passing speed is limited to 10 knots (kt), depth draft ratio to more than 1.164, passing distance to less than four times the moored ship beam (Bm) for surge and sway estimation and passing distance to less than three times the moored ship beam for yaw estimations. These limits are true for no structures in the water. For structures in the water only the passing speed limits are different. When quay walls are present, the surge and sway forces will only provide acceptable answers at passing speeds below 9kt. When 9Bm or 12Bm channels are present, the sway force will only provide acceptable answers at passing speeds below 7kt. When a 6Bm channel is present, the yaw moments will only provide acceptable answers at passing speeds below 6kt. From the mathematical model evaluation study it was found that empirical or semi empirical methods can not provide answers with good agreement to the physical model when walls or channels are present. For the open water case, it is only the Flory method that can provide answers with good agreement to the physical model for surge, sway and yaw forces. The Flory method can provide answers with acceptable agreement within narrow boundaries of passing distance (1 to 2 times the beam of the moored ship), passing speed (4 kt to 14 kt) and depth draft ratio (less than 1.7). The numerical model, Passcat can be used with little effort to provide answers with better agreement to the physical model for a larger range of variables. / AFRIKAANSE OPSOMMING: Wanneer ’n skip verby ‘n vasgemeerde skip vaar, ondervind die vasgemeerde skip kragte en momente. Hierdie kragte induseer beweging van die vasgemeerde skip. Die beweging kan soms groter wees as die effek van wind of golwe. Indien die bewegings groot genoeg is kan dit van die vasmeer lyne van die skip laat breek, of al die lyne laat breek sodat die skip vry in die hawe ronddryf. Aangesien erts skepe en tenk skepe vir jare die grootste skepe in the wêreld was, het die meeste van die skip interaksie studies op daardie skepe gefokus. Die grootte van behouering skepe het egter in die onlangse tye gegroei om dimensies soortgelyk aan die van erts en tenk skepe te hê. In hierdie studie word ’n bestaande numeriese model “Passcat” gestaaf met fisiese model metings op ’n Post Panamax behoueringskip wat verby ‘n Panamax erts skip vaar. Bestaande wiskundige formules is ook getoets deur dit met dieselfde fisiese model metings te vergelyk. In die fisiese model studie is die spoed van die skip (V), tussenafstand (G), diepte diepgang verhouding (d/D) en die teenwoordigheid van kaai mure en kanale in die water getoets. Daar is gevind dat die kragte op die vasgemeerde skip direk eweredig is aan die spoed van die skip tot die mag 2.32. Dit is effens meer as die algemeen aanvaarde kwadratiese verhouding tussen vloeistof sleurkrag en vloeisnelheid asook tussen skip interaksie kragte en vaar snelheid. Soortgelyke verhoudings is vir al die veranderlikes bereken. Numeriese model resultate is vergelyk met die fisiese model om die verhouding van ooreenstemming te bepaal. ’n Perfekte ooreenstemming word voorgestel deur ’n verhouding van ooreenstemming van 1. Grense waarbinne die verhouding van ooreenstemming as goed beskou word is getrek tussen 0.7 en 1.3. Daar is gevind dat die numeriese model, Passcat, kragte oor die algemeen onderskat. Passcat is geldig vir 'n breë reeks van veranderlikes en sal geldig bly solank die skip spoed tot 10 knope, diepte diepgang verhouding tot meer as 1.164, tussenafstand tot minder as vier skipwydtes (Bm) vir 'surge' en 'sway' kragte en tot minder as drie skipwydtes vir 'yaw' momente beperk word. Hierdie grense is opgestel vir geen strukture in die water. Vir strukture in die water word slegs die skip spoed aangepas. Wanneer daar mure in die water is sal 'surge' en 'sway' slegs geskikte antwoorde gee as die skip spoed tot 9 knope beperk word. Vir 9Bm of 12Bm kanale sal geskikte antwoorde vir 'sway' kragte slegs voorkom met 'n skip spoed minder as 7 knope. Vir 6Bm kanale sal geskikte antwoorde vir 'yaw' momente slegs voorkom met 'n skip spoed van minder as 6 knope. Van die wiskundige model evaluasie studie is gevind dat empiriese of semi empiriese metodes nie resultate met goeie ooreenstemming tot the fisiese model metings kan gee, wanneer daar kaai of kanaal mure in die water is nie. Vir die oopwater geval is dit slegs die Flory metode wat antwoorde kan voorsien wat goed ooreenstem met die fisiese model vir 'surge', 'sway', en 'yaw' kragte. Die Flory metode voorsien hierdie resultate binne noue grense vir tussenafstand (1 tot 2 wydtes van die vasgemeerde skip), verbyvaar spoed (4 knope tot 14 knope) en diepte diepgang verhouding (minder as 1.7). Die numeriese model, Passcat, kan met min moeite antwoorde bereken wat beter ooreenstemming vir 'n groter reeks veranderlikes gee.

Thesis (MScEng (Civil Engineering))--University of Stellenbosch, 2002. / 198 leaves double sided printed, preliminary pages i-xx and numberd pages 1-1-12-6.Includes bibliography. List of tables, figures and appendices and acronyms. Scanned with a HP Scanjet 8250 Scanner to pdf format (OCR). / ENGLISH ABSTRACT: In this study the numerical short wave model SWAN is evaluated for application to a selected coastal region in South Africa. The aim of this study was to evaluate the degree of accuracy with which SWAN can simulate prototype nearshore wave spectra and wave parameters (e.g. wave height, mean wave direction and mean wave period) for an Algoa Bay field case. Algoa Bay represents a typical deep, sheltered embayment on the South African south coast, which is exposed to high-energy swell. Sensitivity analyses on various wave-related processes were also done, with the aim of establishing the dominant physical processes and appropriate model setup for the Algoa Bay field case. With the dominant wave-related processes and appropriate model setup for the Algoa Bay field case established, selected final runs were performed to determine the degree of accuracy with which SWAN can simulate prototype conditions, by comparing its results with available field recordings. This study comprises a review of the SWAN evaluation work conducted to date by others, an overview of South African coastal conditions, and numerical model simulations. The model simulations, which represent the main focus of this study, were conducted for a selection of available offshore wave conditions (at 85 m water depth) observed during the Algoa Bay field case and were compared to available nearshore observations (at 17 m water depth). Environmental conditions of waves, wind and currents were included in these simulations. The study focuses on model application and sensitivity analysis, rather than model development, and includes evaluation of all relevant processes, without focussing on any specific model aspect. The results of this study show that SWAN simulations correlated well with observations at the nearshore station in Algoa Bay, both in wave spectral shape and its associated parameters. Dominant processes identified for the field case were depth-induced refraction, bottom friction and directional spreading. This finding agrees with those of previous evaluations of SWAN and previous modelling experience by others. It is shown that high-energy swell is relatively more sensitive to the choices of model setup than wind sea. Based on the simulation results of high-energy swell, it is concluded that the calculation of depth-induced refraction in SWAN seem to contain a degree of inaccuracy. It is also concluded that the findings of this study could be used as a guideline to SWAN modelling studies along the South African south coast. / AFRIKAANSE OPSOMMING: In hierdie studie word die toepassingsmoontlikhede van die numeriese kortgolf model SWAN vir 'n geselekteerde gedeelde van die Suid-Afrikaanse kuslyn beoordeel. Die doel van hierdie studie is om die vlak van akkuraatheid waarmee SWAN prototipe golfspektra en golfparameters (bv. golfhoogte, gemiddelde golfrigting en gemiddelde golfperiode) in die vlakwater kan simuleer te beoordeel, vir 'n Algoabaai gevallestudie. Algoabaai verteenwoordig 'n tipiese diep, beskermde baai aan die Suid-Afrikaanse kuslyn, wat blootgestel is aan hoe-energie deining. Sensitiwiteitstoetse is ook uitgevoer vir verskillende golfprosesse, met die doel om die dominante fisiese prosesse en gepaste modelopstelling vir die Algoabaai gevallestudie te vind. Nadat die dominante golfprosesse geidentifiseer is, en die toepaslike modelopstelling gevind is, is finale simulasies uitgevoer vir geselekteerde gevalle om die mate van akkuraatheid te bepaal waarmee SWAN prototipe kondisies kan simuleer, deur simulasie resultate met beskikbare veldmetings te vergelyk. Hierdie studie bestaan uit 'n samevatting van die evaluasiewerk verrig op SWAN deur andere, 'n samevatting van golf-, wind- en stroomtoestande aan die Suid-Afrikaanse kus en numeriese modelsimulasies. Die modelsimulasies, wat die hooffokus van hierdie studie is, is uitgevoer vir 'n seleksie van beskikbare diepsee golftoestande (in 85 m waterdiepte) uit die Algoabaai gevallestudie en is vergelyk met beskikbare vlakwater metings (in 17 m waterdiepte). Omgewingstoestande van golwe, wind en seestrome is ingesluit in hierdie simulasies. Die studie fokus op modeltoepassing en sensitiwiteits-analise, eerder as modelontwikkeling, en behels die beoordeeling van alle toepaslike modelprosesse, sonder om te fokus op enige spesifieke model aspek. Die resultate van hierdie studie toon aan dat die SWAN simulasies goed korrileer met vlakwater meetings in Algoabaai, vir beide golfspektraalvorm en verwante golfparameters. Bodemrefraksie, bodemwrywing en rigtingsspreiding is geidentifiseer as dominante modelprosesse. Hierdie resultaat kom ooreen met bevindings van vroeere beoordeling van SWAN en modelleer-ervaring deur andere. Dit word aangetoon dat hoe-energie deining relatief meer sensitief is vir modelopstelling as wind-see. Gebasseer op resultate van simulasie met hoe-energie deining, word die gevolgtrekking gemaak dat die berekening van bodemrefraksie in SWAN 'n mate van onakkuraatheid toon. Die gevolgtrekking word ook gemaak dat die resultate van hierdie studie as riglyn gebruik kan word vir modelleerwerk met SWAN aan die Suid-Afrikaanse suidkus.