End of life estimation and optimisation of maintenance of HV switchgear and GIS substations

SUMMARY Efficiency is an important driving force for network operators in the field of operative asset management. Hence, condition and lifetime considerations as well as reflection of the effect of preventive maintenance are important issues with this respect. In this Paper new methods and tools are presented for support of the network operator in the decision making process to find an optimal balance between costs reduction and supply quality. Service experience and diagnostic measurements can provide the basis for this assessment. With this respect gas-insulated substations as well as conventional switchgear are subject to investigations. With a view to GIS the objective of this paper is to analyse the service experience gained during more than four decades with particular regard to dielectric failures and to assess the residual life based on mentioned analysis and on additional diagnostic measurements after nearly 40 years service time. The conclusions from the investigations are as follows: With respect to the insulation performance a service life of 50 years for GIS of the first and second generation is achievable, if some few measures for lifetime extension are introduced. The modern GIS generation seems to be more reliable as the first and second generation since certain deficiencies were overcome by design improvements, application of better material and advanced manufacturing technology. The results of the inquiry of CIGRE WG A3.06 show a similar tendency. With regard to conventional switchgear a condition based maintenance strategy is regarded as an optimal application in terms of overall costs, for planned maintenance measures and unplanned outages (repair). Enabling this strategy, a condition assessment has to be performed. By application of sophisticated methods like probabilistic data analysis the optimal maintenance can be obtained. Starting point is a general condition assessment model which is applicable for all assets. In the following, the asset condition is the degree of ability of each grid component to run the function or functions for which it is created without any major failures. The model considers the results of previous service periods combined with the damage occurrences of other assets of the same type. To predict future damage occurrences and to avoid that by adequate maintenance is the main aim in this content.