Search

The ever increasing level of penetration of Distributed Generation (DG) in power distribution networks is not without its challenges for network planners and operators. Some of these challenges are in the areas of voltage regulation, increase of network fault levels and the disturbance to the network protection settings. Distributed generation can be beneficial to both electricity consumers and if the integration is properly engineered the energy utility. Thus, the need for tools considering these challenges for the optimal placement and sizing of DG units cannot be over emphasized. This dissertation focuses on the application of a soft computing technique based on a stochastic optimisation algorithm (Particle Swarm Optimisation or PSO) for the integration of DG in a power distribution network. The proposed algorithm takes into consideration the inherent nature of the control variables that comprise the search space in the optimal DG sizing/location optimisation problem, without compromising the network operational constraints. The developments of the proposed Multi-Search PSO algorithm (MSPSO) is described, and the algorithm is tested using a standard, benchmarking 69-bus radial distribution network. MSPSO results and performance are compared with that of a conventional PSO algorithm (and other analytical and stochastic methods). Both single-objective (minimising network power loss) and multi-objective (considering nodal voltages as part of the cost function) optimisation studies were conducted. When compared with previously published studies, the proposed MSPSO algorithm produces more realistic results since it accounts for the discrete sizes of commercially available DG units. The new MSPSO algorithm was also found to be the most computationally efficient, substantially reducing the search space and hence the computational cost of the algorithm compared with other methods, without loss of quality in the obtained solutions. As well as the size and location of DG units, these studies considered the operation of the generators to provide ancillary voltage support to the network (i.e. with the generators operating over a realistic range of lagging power factors, injecting reactive power into the network). The algorithm was also employed to optimise the integration of induction generation based DG into the network, considering network short-circuit current ratings and line loading constraints. A new method for computing the reactive power requirement of the Abstract V induction generator (based on the machine equivalent circuit) was developed and interfaced with the MSPSO to solve the optimization problem, including the generator shunt compensation capacitors. Finally, the MSPSO was implemented to carry out a DG integration problem for a real distribution network and the results validated using a commercial power system analysis tool (ERACS).

Steel tubular members are widely used in offshore structures and in infrastructure entities such as oil platforms and bridges. The tubular members are exposed to repetitive loading from sea waves, wind and traffic. Repetitive loading causes deterioration of the structural members through crack initiation and propagation which is called ‘fatigue’. Fatigue strength has always taken most attention in the design of steel tubular members. Due to geometric discontinuity at the connected steel tubular members, high stress concentration exists at the joint. The location of high stresses is usually the location of fatigue crack initiation. In the last century, a large amount of research was conducted on empty (non-filled) steel tubular joints which resulted in the development of various fatigue design guidelines. At the beginning of the current century, researchers have effectively used concrete filled steel tubes to reduce stress concentration at the joint. With regard to the fatigue strength of welded steel tubular joints, there is still lack of understanding and further research is required. In this study, a comprehensive study, which covers experimental and numerical analysis, on the fatigue performance of CHS-to-CFCHS T- and K-joints based on the hot spot stress method has been conducted. In the experimental program, three CHS-to-CFCHS T-joints have been tested experimentally under axial tension, axial compression, in-plane bending and out-of-plane bending in the brace. SCF measurements in the three specimens under the three loading formats have been performed and compared with those for empty T-joints. Additionally, a CHS-to-CFCHS K-joint specimen was tested under balanced axial loading. SCF measurements around the brace-chord intersection have been performed. The experimental program was followed by an extensive numerical analysis study. A large number of finite element CHS-to-CFCHS T- and K-joints models were developed. The effect of the joint geometric parameters on the maximum SCF in CHS-to-CFCHS T- and K-joints has been studied and compared with those for empty T- and K-joints. Parametric equations for predicting the maximum SCF in CHS-to-CFCHS T-joints under the three basic loading formats and in CHS-to-CFCHS K-joints under balanced axial loading have been derived.