Sonel PQM analyzers case study #16: Burnt transformer

Poor power quality can cause a variety of negative effects. From relatively mild ones, like periodic flickering of lights, to very serious ones, like damage to electrical equipment. However, one should always bear in mind that poor power quality is not necessarily the only cause of problems. Other causes can be technical problems, construction problems, design errors, etc.

Description of the identified problem

A 1250 kVA block transformer at the CHP plant was damaged. After it had been replaced with a new one, the incident recurred after several months. The transformer manufacturer refused another replacement under the guarantee, accusing the user of poor operating conditions of the transformer, caused by too high levels of harmonics in the current. Was this the actual reason for the failure of the two transformers?

Measurement tools used

• Sonel PQM-710 analyser
• Set of F-3A clamps (In=3000 A)
• Sonel Analysis software

The transformer was fed by two generators, G-1 and G-2. To diagnose the problem, the measurements were averaged every 10 seconds. The following measurements were taken over a period of one week:

• in the G-1 generator circuit (800 kW) before the generator switch,
• in the G-2 generator circuit (400 kW) before the generator switch, 
• on the 0.4 kV buses of the 15/0.4 kV transformer.

1. Measurements in the G-1 generator circuit (800 kW)

Table 1. Weekly averaged measurements

Phase	Voltage [V]	Current [A]	THD U [%]	THD I [%]	K- factor
L1	235.0	1040	1.22	3.22	1.04
L2	234.3	995	1.14	3.03	1.03
L3	233.8	920	1.07	2.50	1.03

                 
             

       Figure 1 Current and voltage waveforms of the G-1 generator

 

 

2. Measurements in the G-2 generator circuit (400 kW)

Table 2. Weekly averaged measurements

Phase	Voltage [V]	Current [A]	THD U [%]	THD I [%]	K- factor
L1	236.3	618	1.70	11.3	1.11
L2	234.2	583	1.38	10.0	1.14
L3	234.4	542	1.31	8.8	1.17

 

 

Figure 2 Current and voltage waveforms of the G-2 generator

 

PRELIMINARY CONCLUSIONS:

1. The weekly averaged values of the total harmonic distortion factor (THD I) indicate a greater harmonic distortion in the current in the G-2 generator (of 8.8...11.3%), but this does not appear to be the cause of the transformer overheating.
2. One can note an asymmetrical loading of the generators, in particular the G-1 generator.

3. Measurements on a 1250 kVA transformer

 

Table 3. Operation of the G-1 and G-2 generators at 100% of rated power (weekly averaged measurements)

Phase	Voltage [V]	Current [A]	THD U [%]	THD I [%]	K- factor
L1	237.1	1625	1.04	3.22	1.02
L2	235.4	1578	1.05	3.21	1.02
L3	236.1	1410	1.06	3.26	1.02
N	-	90	-	-	-

                

Table 4. Operation of the G-1 and G-2 generators at 90% of rated power. With power limitation, the level of zero order harmonics remained practically unchanged, 90 A current in the neutral conductor (weekly averaged measurements)

Phase	Voltage [V]	Current [A]	THD U [%]	THD I [%]	K- factor
L1	235.8	1506	1.23	5.34	1.09
L2	235.8	1458	1.22	5.53	1.09
L3	235.8	1315	1.24	5.84	1.10
N	-	90	-	-	-

Figure 3 Transformer current and voltage waveforms for 100% generator output

 

FINAL CONCLUSIONS:

1. The proportion of zero-order harmonics at 90 A in the total current of 1,400...1,600 A, is approximately 5.5% and was not the cause of the transformer .
2. The additional thermal losses due to harmonics (K-factor = 1.1 at 90% of rated power) were not a significant influence with the losses estimated by the manufacturer at 12 kW.
3. Testing with the PQM analyser allowed the influence of current harmonics to be ruled out as the cause of the failure. Therefore, an additional inspection of the installation was performed and, among other things, the cooling air circulation system was examined. The transformer manufacturer required an air flow of 6,900 m3/h, while in reality the average was 2,100 m3/h. The immediate cause of the damage to the transformers was inadequate cooling due to faulty ventilation.
4. The above case makes it clear that one should always remember about a comprehensive analysis of the problem at hand, involving a power quality analysis as the first step, followed by an analysis of further risk factors.

 

Author:
Marcin Szkudniewski