The insulation system in form-wound stator windings is exposed to electrical, mechanical and thermal stresses, in addition to factors such as humidity and contamination. These stresses combine to give rise to at least 15 different aging mechanisms, which are described in some detail in [1, 2]. Some of these aging processes are readily identifiable with a visual inspection of the winding. For example, loose windings in the slot (leading to PD), poorly made PD suppression coatings (either the ‘semicon” in the slot area or the silicon carbide grading coating just outside of the slot), and endwinding electrical tracking due to contamination, are easily visually identified. However, there are a number of aging processes that cannot be seen by this method. These processes include thermal aging of the turn and groundwall insulation due to operation at high temperature, delamination of the groundwall insulation due to load (thermal) cycling and PD due to voids or other manufacturing imperfections within the groundwall. The only definitive method to assess the degree of aging with these latter failure processes is to remove multiturn coils or Roebel bars from the stator core and dissect them.
In 2008, CEATI International, a consortium of electric power utilities that sponsors research into common problems, initiated a project to determine the state of aging of the stator winding insulation system in a wide variety of hydrogenerators. A number of utilities from North America, Europe and Australia sponsored the research. The project involved removing a few bars or coils from selected hydrogenerator stators, performing some electrical tests on these bars and coils, and then dissecting segments isolated from each of the coils or bars. Coils/bars were removed from 10 hydrogenerators from 9 utilities (3 from Europe and 7 in the USA). The bars and coils came from machines that were being rewound due to failure, old age or uprating. In cases where the coils/bars were not damaged during extraction from the generator, off-line electrical tests were performed and compared to the condition of the groundwall insulation, as determined by dissections. Only one of the windings had essentially unaged insulation. All the rest showed various degrees of thermal or thermo-mechanical aging. Only one of the 10 windings also showed severe aging due to surface partial discharge. The coin tap test was a good predictor of the degree of thermal and thermo-mechanical (load cycling) aging. The PD test was a good predictor of the coils/bars that had any of the three types of aging. This proposed contribution presents the results from the dissections, and discusses the causes of the aging that was found.
Stator winding, insulation, dissection, delamination, diagnostic tests