Influence of liquid gases on insulation of high voltage motor coils
Stanisław Azarewicz, Dominika Gaworska, Jarosłąw Koniarek, Bogumił Węgliński Politechnika Wrocławska, Instytut Maszyn, Napędów i Pomiarów Elektrycznych, ul. Smoluchowskiego 19, 50-372 Wrocław
Quarterly No. 2, 2006 pages 8-13
DOI:
keywords: insulating materials, kriogenic temperatures, techno-climatic test
abstract The paper presents the results of techno-climatic and electric research of the chosen insulating materials (Table 1) and of coils applied for a motor designed to work in LNG environment. Due to high work safety restrictions, the tests were taken in liquid nitrogen environment (instead of in LNG) which ensures cryogenic temperature, and concurrently in LPG environment, which has similar chemical composition to LNG (as far as the presence of compounds and not their percentage Is concerned). The examined materials and coils underwent fixed thermal stress cycles. The run of one thermal stress cycle is presented in Figure 1. The samples were visually tested and the condition of their structure was photographically registered. The structure tests were registered prior to thermal stress cycles and following a certain number of tests. For the sake of insignificant changes in the examined structures after following stress cycles, photographic registration was done before and then after 1st, 10th, 20th, 60th, and 120th thermal stress cycle. No mechanical changes were noticed for most tested materials. The samples did not split, crack or chip; the thermal stress cycles, however, resulted in the change of the color of the insulation. This does not result from the change of their parameters but from multiple change of their environment: liquid nitrogen and air. Additionally, the placing of the samples in temperature of 393 K causes the resin to displace and the partially melted resin to accumulate locally. This phenomenon is particularly visible in Ergofol EW-2 and Ergofol W-2 samples (Fig. 2). Ergofol W-2/E and Ergofol W-2 samples are made from identical materials but they have different resins. In case of the former - epoxy-, while in the later one - polyurethane resin was used. In Ergofol W-2/E sample accumulation effect is unnoticeable, while in Ergofol W-2 - it is significant. It suggests that the selection of suitable resin for the insulating materials is very important. Unfortunately for Ergofol NKN and double sided NEN FpPuPpa materials, subsequent cycles resulted in unfavorable changes of their structure (Fig. 3). Following the 20th temperature stress cycle, the delamination of the laminations is visible, and air bubbles appear, decreasing the mechanical properties of these samples (Fig. 3d). Both Ergofol NKN and double sided NEN FpPuPpa contain Nomex polyamide paper which, following a subsequent number of cycles, undergoes further separation from the other layers in the tested samples. The changes presented did not cause the decrease of electrical strength of the insulating systems working in liquefied gasses but it can influence the mechanical strength significantly. During the test in LPG environment in Epoksterm 5, Remikaflex and Conductofol samples already after 1st cycle, delamination of subsequent layers of the tested materials structure is visible (Fig. 4). Tests in LPG point to the fact that each insulation utilizing epoxy resin will delaminate, if it had not been properly cured. Also, the fragments of coils with two different insulating structures were tested: 1) winding wire with polyester foil insulation, 2) winding wire with polyamide-imide resin (Thermex 200G2 1Daglas). The tests of cured fragments of coils shows that polyester foil is characterized by weak adhesion to copper bar and the next insulating layer (glass fiber braid) what causes its displacing within the coil (Fig. 6). In case of coil No. 2 (with enamel polyamide-imide resin insulation) this phenomenon is not visible (Fig. 7). Within the frame of electrical strength tests, each of the tested insulation samples underwent break-down strength tests, which consisted of a 10 kV voltage being applied to a single insulation layer in blade-plate system of reversed blade polarity. The electrical strength of the tested insulating materials following thermal stress cycles did not drop (resistivities of the examined samples exceed 1012 Ω). When the fragment of coil No. 1 is concerned, immersing winding in LN2 causes drop of circuit resistance that is probably caused by displacement of winding wire. In case of coil No. 2 the decrease of resistance is not observed.