Analysis of Inspection and Approval for Partial Discharge Testing on Several Sides of the Transformer

Variable-potential multi-terminal measurement is primarily used to determine whether the location of partial discharge lies between the winding and ground or within the winding itself. This is because, under these two wiring configurations, the UU14 two-phase support with one phase remaining unchanged persists. However, the induced voltage multiplication factor Et is significantly lower in the latter case than in the former; therefore, if the partial-discharge data before and after the test are essentially consistent, it is more likely that the partial discharge originates at the end of the winding. Conversely, if there is a substantial difference between the pre- and post-test partial-discharge data, it is more probable that the partial discharge occurs inside the winding—particularly between turns, between饼s, or between layers.

Comparing cases (a) and (b), it can be seen that although the induced voltage multiplier decreases significantly when two phases are supported while one phase is open, the potential at each point of the U1-phase winding either remains unchanged or increases to some extent.

Therefore, the generalization that “if partial discharge remains unchanged regardless of whether support is provided, then partial discharge must be present at the line end” is somewhat inappropriate. Typically, for transformers, aside from the end regions, the insulation structure and clearances are designed with a certain margin; hence, problems are more likely to occur at the ends.

The relationship between the windings and the potential under the two wiring configurations deserves attention. The reason for using two-phase support while omitting single-phase support is as follows: Taking phase U1 as an example, when single-phase excitation is applied, phases V1 and W1 are not short-circuited. Due to the differing lengths of the magnetic paths, the voltages across V1 and W1 are not equal, resulting in UV1 > UW1. If only one phase were supported, the voltage at terminal U1 would be either higher or lower than the original value, thereby reducing the comparability of the test results. In contrast, when two-phase support is employed and phases V1 and W1 are short-circuited, UV1 is forced to equal UW1, ensuring that the magnetic flux is evenly distributed between the two magnetic paths; thus, UV1 = UW1 = 0.15UU1. This allows the test voltage to be calculated, thereby guaranteeing the comparability of the results before and after the test.

During a partial discharge test on phase V1 of a three-phase transformer, the measured partial discharge quantity for phase V1 was 2000 pC, while that for phase W1 was 3000 pC. At this time, UV1 = 0.15 Um, U1/11732, UW1 = 0.125 Um, U1/11732, and UV12W1 = 0.175 Um, U1/11732. Based on analysis, it was suspected that the partial discharge source was located between phases V1 and W1; therefore, the wiring configuration shown was adopted, resulting in UV12W1 = 0, UV1 = UW1 = 0.18 Um, U1/11732. Subsequent measurements showed that the partial discharge quantities for both phases V1 and W1 were less than 100 pC. The test confirmed that the partial discharge source is not related to the grounding connection between the V1 and W1 terminals, but rather to the potential difference between phases V1 and W1. Upon inspection of the V1–W1 interphase region, metallic microparticles were found adhering to the shielding screen of phase W1; after their removal, the fault disappeared.

When a certain three-winding transformer, with low-voltage support for medium-voltage operation, was subjected to a partial-discharge test on phase V1, the discharge quantity on the medium-voltage side, phase V2, was measured at 2,700 pC, along with the potential relationships at various measurement points. Under normal wiring conditions, the potential relationships were as follows; however, after switching to a configuration in which two medium-voltage phases are supported while one is left floating (vector relationship), the discharge quantity on phase V2 still reached 2,800 pC. At this point, two potential sources of concern emerged: first, a possible issue with the V2 terminal; second, an increased potential difference between V2 and V3, which also appeared highly suspicious. To investigate further, the low-voltage side was used to support the medium-voltage side, while maintaining the original potential difference between V2 and V3—specifically, by grounding the V3 terminal and connecting the W3 terminal to the neutral point—thus establishing the corresponding vector relationship. Under these conditions, the measured discharge quantity on phase V2 dropped to 250 pC. Subsequent inspection revealed that the issue stemmed from the medium-voltage bushing; after replacing the bushing, the test results were found to be compliant.

In conclusion, the principle of multi-terminal potential variation measurement is simple and convenient for testing, playing a significant role in fault detection. During testing, the key to effectively applying the method of using two high-voltage phases and one low-voltage phase for potential support lies in: a) there are often multiple ways to change the potential at a particular terminal; the choice should be based on the internal insulation structure of the test specimen, selecting the approach with the least adverse side effects. b) Any changes in other related potentials caused by voltage application or alterations in wiring configuration must be carefully evaluated; hasty conclusions should be avoided. c) Whenever possible, verification or elimination should be carried out using different methods. d) There are many methods for varying potential, and one should not be rigidly confined to a single approach; the degree of flexibility in their application depends on practical experience.