Optimal Demagnetization of Transformer After Winding Resistance Measurements

Sammanfattning: In this thesis, transformer demagnetization in connection to winding resistance tests is investigated. In transformer testing, winding resistance measurement (WRM) is an important test used to detect winding or connection faults. It is common practice to measure the resistance by injecting a DC current into the windings of a transformer. The method magnetizes the core, which leads to the transformer core remaining magnetized even after the test. A magnetized core can cause various problems, such as inrush currents that are several times higher than the rated current. Furthermore, residual magnetism may influence other AC diagnostic tests, such as the SFRA (Sweep Frequency Response Analysis) test and the excitation current test. Therefore, it is recommended to demagnetize transformers after WRM. The core-type and winding configuration may impact the demagnetization process. This makes the different demagnetization approaches of interest, so that the most efficient method can be used. This thesis presents demagnetization experiments on several power transformers with different core-types and winding configurations. The transformers were magnetized with WRM and then a TRAX (transformer and substation test system) was used to demagnetize the transformers with different approaches. Numerous measurements were performed on a three-limb and a five-limb power transformer. The effectiveness of the demagnetization methods was measured by using the excitation current in magnetized and demagnetized state as a comparative tool. It is shown that the demagnetization approach for the three-limb core had little impact on the magnetization result. The excitation currents at demagnetized state for the different combinations was roughly equal. The five-limb transformer was demagnetized on the middle limb, and then demagnetized on the outer limbs to see if the demagnetization gave better results. For the five-limb transformer, the measurements showed that the demagnetization approach impacted the results. Different Yy configurations were tested, and the result showed that the remanence was removed after one demagnetization with a phase to phase approach. However, if a Yy configuration was demagnetized with phase to neutral, then the transformer had to be demagnetized twice to remove the remanence. The result for the Yd configuration showed that it was not improved by a phase to phase approach. The phase that was magnetized did in general have a slightly larger excitation current than the other phases, even after demagnetization.