Jørgen Tolstrup Sørensen

Sensitivity of identified transfer functions in transformer diagnosis

The performance of a new transformer diagnosis method is tested experimentally. The method is called model-based diagnosis because the diagnosis is based on an identified model of the component, i.e., the transfer function. A single test series is carried out in order to evaluate the sensitivity of the method in transformer diagnosis. The basic purpose of this sensitivity test series is to determine how close the identified transformer model is connected to the physical condition of the modeled transformer. Based on the results of the sensitivity test series it is concluded that the detection of failures and aging phenomena in transformers by transfer functions is dependent on detectable changes in relatively few characteristic transformer parameters, i.e., basically lumped winding capacitances, total losses and core reluctance.

An improved model for the calculation of the electrical onset in gas insulated medium inhomogeneous gaps

Most hardware for use in HV systems requires a thorough geometric design concerning the inception of breakdown (partial discharges, corona). Work shows an interest in such corona onset design optimization criteria. Corona onset optimization is based on numerical E-field calculations and different methods for determination of onset/no onset. Practical HV equipment conductor surfaces will normally form nonuniform E-fields possessing a breakdown behaviour determined by both the surface electric field and the rate of decay of the field. Besides that, among other issues, one has to take into account the type of gas, the temperature and the pressure. Therefore it seems important to apply an accurate onset criteria in corona onset design. This work presents the first step of a validated onset criteria useful for practical electrode geometries. This is accomplished for atmospheric air on the basis of the well-known Townsend approach for nonuniform fields with modifications concerning the gas parameters and validation against IEC 52 sphere-sphere gap breakdown voltages. It was found that the onset/breakdown voltage for medium nonuniform fields can be calculated by means of the same criteria for varying electrode curvatures and degrees of field utilization with an accuracy better than 4%.

Curve-fitting method for transfer function determination

A curve-fitting method for determination of transfer functions by frequency sweep analysis is established. The method uses linearization of the modeled signals by use of a Taylors series and minimization of the error between the estimated and the measured values by the least-squares method. It is found that the transfer function determined by this method is more certain and stable than it would be if the transfer function were determined by detecting maximum amplitudes and zero-crossings of the input/output signals even if these signals had been averaged and filtered. Some errors are detected in the phase angles at very low currents, but the errors are within +or-30 degrees.<<ETX>>