Untersuchungen der inneren und äußeren Festigkeit von Epoxidharzisoliersystemen (English shop)

Preview

Table of Contents, Datei (31 KB)
Extract, Datei (130 KB)

Internal and external strength are components of the dielectric strength of an epoxy resin insulation system. The evaluation of both components is carried out with the aid of various measurement methods. A meaningful evaluation is only possible after the separation of disturbing influences that may possibly falsify the results. Influencing parameters can alter the external and internal strength, whereby the internal strength is generally to be seen in connection with internal discharges (partial discharges) and the external strength in connection with flashovers and (corona).

Influencing parameters of the internal strength are mechanical, electrical or thermal stresses. Mechanical stress is only to a limited extent an influencing parameter of the internal strength in epoxy resin insulation systems, since a firm bond between insulation and electrodes can be achieved. Electrical stress, by contrast, is the strongest influencing parameter of the internal strength, since the conduction losses, ionisation losses and polarisation losses in the form of partial discharges or heating of the insulating material can cause damage up to complete breakdown. Thermal stress occurs as an influencing parameter of the internal strength in connection with electrical stress and/or with chemical reactions, since when the thermal limits are exceeded a change in the insulation system and an electrical breakdown take place. Chemical reactions such as depolymerisation, random chain scission or scission of end groups can additionally be stimulated by thermal and/or electrical stress.

An essential influencing parameter of the external strength is the length of the creepage path, whereby an increase of the creepage path can be achieved through the choice of the contours and the shape. Air pressure also has an influence on the external strength in connection with the breakdown behaviour in gases. Temperature contributes to influencing the external strength through convection, condensation or diffusion in connection with humidity and foreign substances, whereby humidity in particular has an effect under DC voltage stress. Roughness and air flow are further influencing parameters of the external strength. The meeting of three different substances (electrode, insulation, ambient gas) represents a critical situation in the angle with respect to the discharge behaviour, whereby the assignment of the materials to one another represents the essential influencing variable.

The external strength can essentially be described by the breakdown mechanisms in gases, whereby the accumulation of charge carriers in the form of avalanches between the electrodes influences the breakdown as a function of the electrode spacing, gas pressure and polarity of the applied voltage. Depending on the external conditions, avalanche formation can take place in one or several generations. Other mechanisms are responsible for the internal strength, since breakdowns in solids stimulate, depending on the magnitude and duration of the applied voltage, mechanisms such as conduction losses, ionisation losses and heat losses, which can then lead to breakdown.

The measurement of the insulation resistance and of the dissipation factor serves to evaluate the losses in the insulating material. However, these measurements are not sufficiently suitable for the diagnosis of the insulating material in the initial stage of damage, since on account of their sensitivity only a rough evaluation of the insulation condition is possible. The partial discharge (PD) measurement and the measurement of the polarisation and depolarisation current are more sensitive and better suited for detecting microscopic damage, so that PD measurement was applied as the diagnostic method for damage in this work. Under certain circumstances, PD measurement also permits a location of the damage site in the insulating body.

Epoxy resin is fundamentally produced on the basis of the cross-linking of an epoxide ring with catalyst or hardener in reactions of homo- or hetero-polymerisation. The widespread use of epoxy resin on the basis of bisphenol for electrical engineering purposes is explicable because of the good dielectric properties. However, the good properties of epoxy resin based on bisphenol are impaired by weathering, so that the use of cycloaliphatic epoxy resins is necessary for applications in outdoor installations, since these exhibit a better resistance to weathering, though with the limitation of sensitivity to ultraviolet light (UV). Epoxy resin is therefore used in combination with fillers such as quartz powder for the production of post insulators for indoor applications.

In combination with mica, epoxy resin is used for the production of stator bars for rotating high-voltage machines. Further fillers in epoxy resin are used in order to achieve a certain tracking resistance for external insulation systems.

The test specimens for determining the internal strength in the homogeneous and inhomogeneous field were immersed in an ester liquid in order to avoid external discharge and to use the liquid as a heat transfer medium. These test specimens were electrically aged under 50 kV/cm in the inhomogeneous field arrangement or 100 kV/cm in the homogeneous field arrangement. In practice, maximum operating field strengths of 30–50 kV/cm are customary. The ageing of the test specimens for the investigation of the internal strength was divided into stress periods and rest periods. Cavities occur in random distribution in the interior of the test specimens. Through changes in the conductivity or in the gas pressure, the ignitions of the cavities are altered, so that a different PD behaviour was to be expected. Through the stressing of the test specimens, changes in the conductivity and the gas pressure were generated. The rest periods in the stressing of the test specimens can likewise produce changes in the gas pressure and the conductivity, so that here too a change in the PD behaviour was to be expected. The PD measurement for the diagnosis of these test specimens was carried out with a wideband PD measuring device at voltages from 0 kV to 6 kV for 160 s with a minimum sensitivity for the apparent charge of 10 pC. Because of the random distribution of the cavities, the PD measurements do not always permit a detection of the PD source. With the separation of the stress and rest periods, a decrease of the PD inception voltage and PD extinction voltage and an increase of the PD pulse rate could be recognised during the rest periods in test specimens with inhomogeneous field under electrical stress. The cause of this behaviour is a decrease of the gas pressure in the cavities of the test specimens during the stress interruption. With images of the interface between electrode and insulation of the test specimens with homogeneous field under electrical stress, the development and the carbonisation of the breakdown channel, the amorphous structures of the epoxy resin and particles from the electrodes could be recognised, whereby the breakdown channels run in several paths parallel to the electric field. The number of cavities made an evaluation of the PD measurement difficult in test specimens with homogeneous field under electrical stress.

A thermal stress was carried out at 80 °C, which is to be regarded as the limit of the glass transition temperature for this epoxy resin system. The separation of the PD measurement according to rest periods and stress periods yields clear results for the thermal stressing of test specimens with homogeneous and inhomogeneous field, since the decrease of the PD inception voltage and PD extinction voltage always occurred in connection with a rest period, whereas the increase of the PD inception voltage and PD extinction voltage was always associated with a stress period. The scatter of the PD pulse rate is to be explained by the change in the geometry of the cavities as a consequence of the thermal stress.

An electrothermal stress was carried out with 50 kV/cm for the inhomogeneous field and 100 kV/cm for the homogeneous field at a temperature of 80 °C. The analysis of the electrothermal stress on test specimens with homogeneous and inhomogeneous field yields no clear findings on the basis of the PD inception voltage, PD extinction voltage and PD pulse rate. For test specimens with inhomogeneous field, a dominating effect of the thermal stress can be recognised. Images of the interface between electrode and insulation of test specimens with homogeneous field show the breakdown channel in the form of a hole. In addition, a white substance was found in the breakdown channel which, by its apparently crystalline structure, points to a salt-like substance. The PD inception voltage of test specimens with homogeneous field shows a stabilised range for periods with stress, which points to the formation of new PD channels. The PD pulse rate shows an approximately exponential decrease of the values for periods with stress, which is connected with the formation of larger cavities or of cavities which remain ineffective during the stress period.

Numerous investigations of the ageing behaviour at 50 Hz have been carried out by many authors. An increase of the frequency by a factor of 10 should cause more rapid damage. Therefore, stator bars were stressed for ageing with an alternating voltage at 500 Hz. The voltage was generated with a voltage-source inverter and subsequent filtering. The ageing voltage was 10 kV, which corresponds to the rated voltage for these stator bars. The electrothermal stressing of the stator bars is likewise carried out with 10 kV from the same voltage source at a maximum temperature of 60 °C, which corresponds to the operating temperature for these stator bars. The ageing of the stator bars at 500 Hz showed a clear damage of the outer corona protection for the electrical and electrothermal stress, whereby this damage does not occur under stress at 50 Hz. The PD measurement for the diagnosis of these test specimens was likewise carried out with a wideband PD measuring device at voltages from 0 kV to 10 kV for 160 s with a minimum sensitivity for the apparent charge of 100 pC. The PD inception voltage and PD extinction voltage show, for an electrical stress, a strong increase after a long rest period, while in the case of the PD pulse rate a clear drop after a long rest period can be recognised, which points to changes in the cavities. Under thermal stress, the PD parameters occur in complex form. In individual cases, extremely high PD pulse rates are linked with long rest periods and short stress periods. The electrothermal stress likewise shows a complex behaviour which cannot be explained by a direct superposition of the electrical and thermal stresses. The PD pulse rate shows a clear increase with the stress duration. After a long rest period, however, the pulse rate drops sharply.

The test specimens for investigations of the external strength were stressed with DC voltage, since numerous investigations with alternating voltage were already available. The roughness of these test specimens was varied by producing contours perpendicularly, in parallel and circularly with emery paper P400, P100 and P40. In addition, an air flow was employed for investigations of the external strength, from an air compressor and a nozzle with air velocities from 0.5 m/s to 40 m/s. In the investigations it was established that fillers cause different traces on the surface of the test specimens in the determination of the external strength. In test specimens with filler type M, tree-shaped traces occurred, while for test specimens with filler type K carbonised areas appeared. This behaviour points to an influence of the thermal conductivity through geometry and material of the filler. Consecutive flashovers on a test specimen cause a reduction of the flashover voltage. After a storage period this reduction is no longer present. A cleaning of the surface, however, influences the behaviour after storage only slightly, which points to a slow decay of the charge.

The surface roughness has a positive effect on the flashover voltage, provided the roughness did not reach certain limits in machining with emery paper P40. When a limiting value is exceeded, no increase of the flashover voltage is to be observed, since the breakdown channel no longer forms along the surface. Circular contours of the roughness show higher flashover voltages than contours of the roughness running perpendicular or parallel to the field direction. However, here too a maximum roughness can be recognised.

An air flow causes a removal of the charge carriers from the surface and therefore an increase of the flashover voltage. The increase of the flashover voltage with the air flow is independent of the polarity of the applied DC voltage. The fillers show differences in the flashover voltage. Likewise, the transition between turbulent and laminar flow as well as the surface condition of the material have an influence on the flashover voltage. With an air flow, the flashover takes place between the electrodes above the material surface and is thus no longer influenced by the material.