|Title||An investigation of Network Splitting for Fault Level Reduction|
|Publication Type||Tyndall Working Paper|
|Series||Tyndall Centre Working Papers|
|Tyndall Consortium Institution|| |
|Secondary Title||Tyndall Centre Working Paper 25|
|Authors||Wu, X., J. Mutale, N. Jenkins, and G. Strbac|
|Year of Publication||2003|
The UK government has set targets that by the year 2010, 10% of the electrical energy consumed in the UK will be provided from renewable sources and 10GWe capacity of CHP plants will be installed by the same date. Achievement of these very ambitious targets will most likely entail the connection of a very large number of distributed generation to existing distribution networks. Connection of distributed generation on the scale that is envisaged will require practical solutions to a number of identified technical, commercial and regulatory challenges. One of the technical challenges is fault level management as connection of distributed generation often results in increased fault levels beyond the capacity of existing switchgear, especially in urban areas. In this paper, results of investigations carried out to assess the effectiveness of the options available for fault level management as well as, where necessary, their impact on voltage profiles and system stability are presented and discussed. Five main methods for fault level reduction have been reviewed and discussed, namely current limiting reactor, Is-limiter, superconducting fault current limiter, solid-state fault current limiter and network splitting. Network splitting was found to have the greatest potential for fault level reduction in the short term as it is relatively inexpensive and furthermore it has high reliability and flexibility. Modelling and simulations for various network-splitting scenarios performed using PASCAD/EMTDC confirm that fault level can be reduced significantly by network splitting. Furthermore, the quality of power supply to the customer and the transient stability of the DG can be improved by using fast-closing switchgear or a static transfer switch in the bus-section. The analysis shows that the fault current is sensitive to closure of the normally open point (NOP) on the ring effectively connecting the cable network in parallel with the bus-section breaker or the reactor.