| Summary: | The prediction of mass oscillation behaviour is an important part of the hydraulic design of a hydro-electric power scheme containing one or more surge chambers. Currently design practice is to employ computational models. However, the use of hydraulic models in the design of large or complex surge systems provides a useful means of corroborating results from computational studies. This thesis is concerned with the development of techniques for improving the accuracy of both methods. The first part of the thesis deals with the formulation of equations and techniques required for the development of a computer program capable of analysing a wide range of chamber configurations. The majority of these techniques are employed in a computational model, the salient features of which are discussed in some depth. The second section is concerned with the derivation of scaling relationships between prototype schemes and their corresponding hydraulic models. This is supplemented by a discussion of some practical aspects in the design and construction of models. The study culminated in the construction of a hydraulic model of the 1800 MW Dinorwig scheme. Advances in microcomputer technology have enabled them to be used, at reasonably low cost, in customised monitoring/control systems which improve the quality of hydraulic model tests. A number of techniques were developed which were incorporated in a program package for a 32 kilobyte microcomputer for use with the Dinorwig model. Further techniques are discussed in which monitored data is susbjected to further computational analysis. This includes a technique for combining results from hydraulic and computational models to produce a more accurate prediction. These techniques were used in a study of the Dinorwig scheme which revealed that as a result of an amendment to the chamber design, the original specification for the limits of surge amplitude are not satisfied.
|
|---|
