A 10.8 km long cable part, consisting of 12 cables, is installed in the meshed and OHL-based Dutch 380 kV grid. At this operating voltage the cable part will produce a huge amount of reactive power which needs to be compensated by shunt reactors. These shunt reactors, together with other (inductively behaving) components, can exhibit resonances with the (capacitive) cable. To be able to predict the sensitivity for resonate behavior, system simulations are required. There is however a need for practical guidelines concerning the level of detail the model for the surrounding network needs to be built to obtain accurate results. Much effort can be saved in obtaining detailed data for an extensive but proper model, while on the other hand a simpler and less detailed model might not describe the network behavior properly. In this study we test different simplification options, which can reduce in the future the time to construct a simulation model. The reduced models are compared with the reference model, the complete Dutch 380 kV grid, and judged based on four indicators which describe the level of accuracy. As a results, guidelines on how extensive the network has to be simulated and in which detail the model has to be build are derived. Three modeling areas are distinguished, with a preferable transmission line model, requirement of accurate earth resistivity and phase order information.
|Date of Award||31 Aug 2013|
|Supervisor||Peter A.A.F. Wouters (Supervisor 1) & J.P.W. Jong, de (External coach)|