Commissioning a 1500 kVA Dry-Type Transformer With Two Anomalies
You are commissioning a 1500 kVA dry-type transformer, 13,800 V delta primary to 480Y/277 V secondary, installed as a separately derived system feeding a 4000 A switchboard. The secondary is carried in five sets of 500 kcmil copper per phase. The installing contractor bonded the transformer neutral to ground in the switchboard with a 3/0 copper jumper. Field data: turns-ratio test on the nominal tap gave 28.79, 28.76, and 28.11 against a nameplate ratio of 28.75. Insulation resistance at 5 kV on the primary was 12 GΩ after 60 seconds; at 1 kV on the secondary it was 480 MΩ. Both were taken at 24 °C ambient and 45% RH. There is no prior baseline for this unit. The plant wants the transformer energized tonight to support a startup.
1. Interpret the three turns-ratio readings. Do you energize? What do you check next?
Read the worked answer
Two phases are correct to within 0.2%; phase C is out by more than 2%. Do not energize.
- Nominal ratio = 13,800/480 = 28.75
- A: (28.79 − 28.75)/28.75 = +0.14%
- B: (28.76 − 28.75)/28.75 = +0.03%
- C: (28.11 − 28.75)/28.75 = −2.23%
Typical acceptance is within about 0.5% of nameplate on the marked tap. Phase C is roughly four times outside that, and the fact that two phases agree closely makes a wrong nameplate or a global instrument error unlikely. Ranked causes:
- Tap link on the wrong position on phase C - check the physical link before anything else.
- Shorted turn in the C winding - a shorted turn shows low winding resistance and low exciting current on that phase.
- Test lead or connection error on C - clean the studs and repeat before drawing conclusions.
Next tests if it repeats: ratio at every tap, winding resistance on all three phases, and single-phase exciting current. Energizing tonight is not justified until the cause is identified and corrected.
2. The 480 MΩ secondary reading passes any table minimum you are likely to find. Why is that not the end of the analysis, and what would you require before energizing?
Read the worked answer
A single absolute reading is the weakest possible basis for an insulation judgment.
- 480 MΩ is very likely above any published minimum for a 600 V class winding, so a pass/fail call on the number alone clears it. The problem is the 25:1 disparity against the 12 GΩ primary measured at the same temperature and humidity. Winding geometry on a dry-type unit does not explain a gap that large; moisture, surface contamination, or a damaged barrier does.
What to require before energizing:
- Repeat the secondary test after cleaning and drying the bushings and using a guard terminal to eliminate surface leakage.
- Run a polarization index (10 min / 1 min). A PI of roughly 2.0 or better indicates a stable, dry insulation system; a flat or falling PI indicates moisture or contamination that a 60-second value can hide.
- Apply temperature correction - IR falls sharply as temperature rises, so compare only same-temperature readings.
- With no prior baseline, use the PI and the phase-to-phase comparison as the trend. If the PI is acceptable and the reading is stable after cleaning, energize with a follow-up reading scheduled.
The defensible answer is neither a blind accept nor a blind reject: the single IR number is not sufficient evidence in either direction.
3. Evaluate the 3/0 copper system bonding jumper. What size is required?
Read the worked answer
The 3/0 jumper is undersized.
For a system bonding jumper where the supply conductors are larger than 1100 kcmil copper, the required size is 12.5% of the area of the largest phase conductor, rounded up to the next standard size.
- Largest phase conductor total = 5 × 500 kcmil = 2500 kcmil
- 12.5% × 2500 = 312.5 kcmil
- Next standard size up = 350 kcmil copper
A 3/0 is about 168 kcmil - roughly half the required area. It must be replaced with a 350 kcmil copper jumper (or the equivalent in the permitted alternate material).
Also verify while you are in the switchboard: this is a separately derived system, so the neutral-to-ground bond must exist at exactly one point - the source or the first disconnecting means, per the design - and must not be duplicated at the transformer and again at the switchboard. A double bond puts the bonding jumper and the neutral in parallel and creates a parallel neutral path through the grounding system.