Heat run test of Power Transformers - Algorithms for calculation of the winding oil temperature rise from measurements

摘要

Heat run tests are part of the type test of power transformers. As a first step the transformer is feeded with an AC current corresponding with the sum of no-load and load losses. After achieving a thermally stable condition the oil temperature at the top of the transformer is measured. Then, the transformer is feeded with its rated current during one hour. After finishing this period the average winding temperature is determined. Therefore, the windings are feeded with a DC current as quick as possible. Core Saturation shall be reached as soon as possible so that all transient conditions are finished. Then, the DC voltage drop along the windings and the current through the windings are measured during an intervall of e. g. 20 minutes. The average temperature of the copper can be determined from the winding resistance which is calculated from the DC voltage and current measurements. According to theory the winding temperature decayes exponentially. The application of linear and non-linear approximation algorithms allows to fit the measured resistance values with an exponential function. The average winding temperature at the instant of switching off the rated AC current can be determined from the exponential approximation of the measured resistance values. With the top oil temperature and the average winding temaperature the most essential temperatures in the transformer are calculated. In this contribution the characteristics of linear and non-linear algorithms for approximation of meaured value to an exponential function and their applicability in the described optimization problem are discussed.