Thesis (B.A.(Hons.))--University of Adelaide, Dept. of History, 1972.

EU har planer for en storstilt utbygging av uregulerbar kraft, og etterspør regulerbar kraft som kan balansere forskjellen mellom kraftforbruk og kraftproduksjon. Norge har et fleksibelt vannkraftsystem, og kan bli en leverandør av balansekraft ved å bygge ut store pumpekraftverk. Urevatn og Vatndalsvatn i Setesdalen i Sør- Norge har potensiale for utbygging av pumpekraftverk. Det er særlig vannmagasinenes store vannvolum, korte innbyrdes avstand og høydeforskjellen mellom dem som gjør de interessante for en pumpekraftutbygging. Formålet med denne oppgaven er å skissere en mulig utforming av et pumpekraftverk mellom Urevatn og Vatndalsvatn. Hovedfokuset i oppgaven er tunnelsystemet og at pumpekraftverket som helhet skal representere kombinasjonen av optimal økonomi og gode tekniske løsninger. Det er gjort en rekke tekniske og markedsmessige vurderinger for å kunne optimalisere de nødvendige elementene av pumpekraftverket. Den installerte effekten er dimensjonert ved å analysere variasjonen i dansk vindkraftproduksjon over en 10- års periode. Den optimale installerte effekten er beregnet til å være 1600 MW slik at kraftverket får en pumpeevne på 549 m3/s. Ved design av vannveien er det i hovedsak benyttet beregningsmetoder av overslagsmessig karakter. Vannveien er 2630 meter lang og er designet med skrå tilløpstunnel og horisontal avløpstunnel. Det optimale tunneltverrsnittet er funnet til å være 360 m2. Aggregatet må dykkes 44 meter i forhold til undervannet for å unngå kavitasjon. De totale kostnadene for bygging av det designede pumpekraftverket er beregnet til å være 3762 MNOK. På bakgrunn av enkle kriterier for stabilitet av det dynamiske systemet, er det funnet av det er behov for et svingesystem. Et luftputekammer er derfor valgt for å bedre stabiliteten. Likevel er det konkludert med at det må gjøres frekvensanalyser av det dynamiske systemet for å kunne avgjøre behovet for et svingesystem. Det er konkludert med at et pumpekraftverk mellom Urevatn og Vatndalsvatn er meget godt egnet for å kunne levere balansekraft til Europa, der store mengder effekt kan bygges ut til en relativt lav kostnad.

SummaryRock Bolts, improved design and possibilitiesMaster thesis NTNU 2012Student : Capucine Thomas-LepineSupervisor : Leif LiaKey words : rock foundation, small concrete dam, rock mass classification, rock joints, shear strength of rock discontinuities, fully grouted passive rock bolts designMasters Thesis : “Rock bolts, improved design and possibilities” is a continuation from the Masters Thesis NTNU 2011 “Rock bolts in dams, expected capacity” by Lars Kristian Neby. Internationally, dam engineering focuses mainly on pre-stressed anchors in rehabilitation and improvement of stability of large dams, which is undergoing constant research in North America. Passive rock bolts are used in small concrete dam foundations to ensure sufficient stability against overturning moment from ice loads. This concerns the majority of dams in Norway, over 98% of whose electricity comes from hydropower developed over the last 100 years and still developing. Design is ruled by regulation from NVE (Norwegian water resources and energy directorate) published first in the 1980s, and regularly revised until the retroactive “Retningslinjer for betongdammer” in 2005. This design method for passive rock bolts is conservative with regards to rock capacity, as it is worldwide. The model, developed in the early age of rock bolt development in 1977 by Littlejohn and Bruce, considers the rock resistance as equivalent to the weight of the cone of rock around the bolt. Rock engineering has improved since, often with regards to underground engineering, which is not necessarily transposable to dam engineering. The inherent uncertainty in rock mass characterization slowed development of new design method for passive rock bolts. This is however of great interest in the Norwegian hydropower industry, and for applications to other civil engineering structural foundations. This thesis is meant to develop knowledge of qualitative and quantitative rock mechanisms in passive rock bolts in order to improve their design.The work is composed of three parts, as follows :A study on rock mass capacity and mechanisms in dam foundations, comes first. An empirical and quantitative estimation of rock mass strength with regards to recognized classification Q, RMR or GSI is proposed, based on Wyllie (1992) and results from Lars K. Neby. Full scale tests are then performed to the assess validity of empirical relationships developed in the first part between rock mass quality and rock bolt capacity (maximal tension load in pull out tests). 50 steel bolts with diameter 25mm and mortar grouted length 0.4m were pulled out with logging of strength and deformation in a rock quarry presenting various degrees of rock quality (RMR 40 to 80), representative of expected conditions for dam foundations. Rock quality was assessed for each bolt by laboratory testing (intact material properties) and rock mass characterization on site supplemented by core drilling or video inside hammer drilled holes. This program of tests was an improvement on the protocol developed and performed by Lars K. Neby on 18 bolts, whose results give the first relevant clues on parameters such as limit range of quality of rock (RMR>40), length of bolt (0.4m), maximal capacity (more than 20 tons, when conservative calculations assessed 0.2 tons). The results of testing confirm the relationships developed in a more statistical approach.Conclusions from these two parts lead to a proposition for a new design model, with a higher resistance contribution from the rock. The three modes of failure (rock, steel, grouting) are considered for ensuring resistance of a maximal stress of 180MPa (to avoid important deformations in the structure). Factors of safety and the range of validity in the rock mass condition for the proposed design are also considered. The thesis concludes with propositions for further works in order to :- Further extend the domain of validity of the proposed design.- Review methods of control of installed passive rock bolts.- Document and improve knowledge of load transfer mechanisms in passive rock bolts in dams.

Bakgrunnen for oppgaven er et ønske om å redusere lekkasjetapet på vannledningsnettet i Narvik by. Reduksjon av lekkasjetapet kan gjøres ved utskiftning/rehabilitering av ledninger, lokalisering og reparasjon av lekkasjer og trykkreduksjon.Målet med prosjektet er å gi et grunnlag for videre arbeid med reduksjon av lekkasjer, ved hjelp av fin-lokalisering, trykkreduksjon og generell planlegging av hvilke ledningsstrekk som skal saneres i de kommende år.I rapporten gjennomgås vannbalanse for lekkasjesonene på ledningsnettet i Narvik, lekkasjetap med ulike indikatorer presenteres og lekkasjesonene rangeres. Det utredes muligheter for reduksjon av lekkasjetapet ved å redusere trykket der det er praktisk gjennomførbart. Videre sammenlignes lekkasjetapet i Narvik med andre kommuner og forslag til organisering av lekkasjesøk fremstilles.Utarbeidelsen av vannbalansen er gjort på grunnlag av foreliggende data om målt forbruk på kommunalt ledningsnett, innrapporterte målerverdier fra større abonnenter og folkeregisterdata. Det er i enkelte tilfeller manglende data fra målere på ledningsnettet som har måttet blitt erstattet av erfaringsdata som gir visse muligheter for feil. Resultatene viser et totalt vanntap på ca. 55 % for Narvik. Etter sonebalanse er satt opp og lekkasjesonene rangert er sone 7, 12, 10 og 4 rangert som verst. Disse tre sonene står for ca. halvparten av det totale lekkasjetapet. Resultatene fra simuleringene ved en trykkreduksjon i sone 12 gir en lekkasjereduksjon på ca. 13 % i denne sonen. Ut fra erfaringer fra andre kommuner og størrelsen på forsyningsområdet i Narvik anbefales det å benytte ca. et halvt årsverk til lekkasjelytting.

"This thesis proposes to investigate the theory posed by Comer Vann Howard in 'The Strange Career of Jim Crow.' Woodward claims that complete physical segregation of Negroes was not legally established in the Southern states until the turn of the century. He further contends the period from Reconstruction until the 1890s was an era when Negroes participated in many activities with whites. This work investigates Woodward's theory in its applicability to Texas between 1874 and 1898. The study begins with redemption, which came to Texas in 1874 with the election of the first Democratic governor since the Civil War. The concluding year of 1896 was chosen because the last Negro to serve in the Texas Legislature ended his term that year."-- leaf [i].

In this dissertation, generalized additive neural networks (GANNs) and multilayer perceptrons (MLPs) are studied and compared as prediction techniques. MLPs are the most widely used type of artificial neural network (ANN), but are considered black boxes with regard to interpretability. There is currently no simple a priori method to determine the number of hidden neurons in each of the hidden layers of ANNs. Guidelines exist that are either heuristic or based on simulations that are derived from limited experiments. A modified version of the neural network construction with cross–validation samples (N2C2S) algorithm is therefore implemented and utilized to construct good MLP models. This algorithm enables the comparison with GANN models. GANNs are a relatively new type of ANN, based on the generalized additive model. The architecture of a GANN is less complex compared to MLPs and results can be interpreted with a graphical method, called the partial residual plot. A GANN consists of an input layer where each of the input nodes has its own MLP with one hidden layer. Originally, GANNs were constructed by interpreting partial residual plots. This method is time consuming and subjective, which may lead to the creation of suboptimal models. Consequently, an automated construction algorithm for GANNs was created and implemented in the SAS R statistical language. This system was called AutoGANN and is used to create good GANN models. A number of experiments are conducted on five publicly available data sets to gain insight into the similarities and differences between GANN and MLP models. The data sets include regression and classification tasks. In–sample model selection with the SBC model selection criterion and out–of–sample model selection with the average validation error as model selection criterion are performed. The models created are compared in terms of predictive accuracy, model complexity, comprehensibility, ease of construction and utility. The results show that the choice of model is highly dependent on the problem, as no single model always outperforms the other in terms of predictive accuracy. GANNs may be suggested for problems where interpretability of the results is important. The time taken to construct good MLP models by the modified N2C2S algorithm may be shorter than the time to build good GANN models by the automated construction algorithm / Thesis (M.Sc. (Computer Science))--North-West University, Potchefstroom Campus, 2011.

In this dissertation, generalized additive neural networks (GANNs) and multilayer perceptrons (MLPs) are studied and compared as prediction techniques. MLPs are the most widely used type of artificial neural network (ANN), but are considered black boxes with regard to interpretability. There is currently no simple a priori method to determine the number of hidden neurons in each of the hidden layers of ANNs. Guidelines exist that are either heuristic or based on simulations that are derived from limited experiments. A modified version of the neural network construction with cross–validation samples (N2C2S) algorithm is therefore implemented and utilized to construct good MLP models. This algorithm enables the comparison with GANN models. GANNs are a relatively new type of ANN, based on the generalized additive model. The architecture of a GANN is less complex compared to MLPs and results can be interpreted with a graphical method, called the partial residual plot. A GANN consists of an input layer where each of the input nodes has its own MLP with one hidden layer. Originally, GANNs were constructed by interpreting partial residual plots. This method is time consuming and subjective, which may lead to the creation of suboptimal models. Consequently, an automated construction algorithm for GANNs was created and implemented in the SAS R statistical language. This system was called AutoGANN and is used to create good GANN models. A number of experiments are conducted on five publicly available data sets to gain insight into the similarities and differences between GANN and MLP models. The data sets include regression and classification tasks. In–sample model selection with the SBC model selection criterion and out–of–sample model selection with the average validation error as model selection criterion are performed. The models created are compared in terms of predictive accuracy, model complexity, comprehensibility, ease of construction and utility. The results show that the choice of model is highly dependent on the problem, as no single model always outperforms the other in terms of predictive accuracy. GANNs may be suggested for problems where interpretability of the results is important. The time taken to construct good MLP models by the modified N2C2S algorithm may be shorter than the time to build good GANN models by the automated construction algorithm / Thesis (M.Sc. (Computer Science))--North-West University, Potchefstroom Campus, 2011.