Abstract
Faced with an ageing transformer population, Norwegian authorities, utilities and research organizations together with international manufacturers of apparatus and materials have focused research on thermal ageing and end-oflife assessment of transformer insulation. Based on a suggested possible life-extension of 40 years until insulation system end-of-life the value a postponed reinvestment is about 30 % of the present value of the transformer park. This requires focus on maintenance and improved thermal monitoring. The paper will describe the Norwegian situation with age distribution of transformers, typical designs and loading patterns and failure statistics. The main results from more than 10 years of laboratory ageing studies of cellulose insulation will be summarized, with focus on ageing kinetic formulas and parameters from a 5 year long ageing experiment. The results from this experiment suggested that ageing could be described by either oxidation or acid assisted hydrolysis. This is supported by more dedicated investigations on these two mechanisms. In scrapping investigations cellulose insulation material is sampled and its ageing condition measured. This gives information about how cellulose ages in service. These results will be used to calibrate the ageing kinetic formulas derived from laboratory experiments, and for improving correlation between ageing markers and experienced ageing. The assumption is that it is the transformers weakest spot - the part of the windings where paper is aged to the lowest degree of polymerisation (DP) - that determines the end-of-life. Variation of DP with local temperatures and conditions makes it difficult to locate this spot. One therefore applies a statistical approach where first the hot-spot region is identified and described. Then based on experienced variations in DP over the winding, a certain number of samples are collected. Based on analyses of these results mean values and low percentile values are estimated. Materials are collected from service aged units and partitioning curves for ageing markers are developed using an in-house developed test rig. Findings from scrapping investigations and analysis of operating conditions and designs indicate that even if the age of the population is high, one may still postpone new investment for many years. This is partly due to low loading; in the grid because of load sharing due to redundancy and in the power stations due to loading below nominal load because of best-point operation of generators. Also high load situation will occur during wintertime where temperatures of cooling air and water is low. It also helps that a high percentage of the transformers is made with thermally upgraded paper. Obstacles for establish a road-path for end of life calculation based on ageing kinetic formulae's are incomplete and poor records on historical loading and temperature profiles. Furthermore there are often uncertainties on thermal design parameters like hot spot factors. To control the situation the Norwegian Regulator has introduced a mandatory reporting scheme of some simple condition parameters. These have been developed in collaboration with a Norwegian "User Group for Power and Industrial Transformers". The research has shown the importance of good design and specifications. Now sealed transformers and use of thermally upgraded paper has become dominating. It has also resulted in an increased focus on documentation of thermal design and testing of new transformers. © 2014 CIGRE. All rights reserved.