A transformer manufacturer tries to minimize construction cost by using solid conductors for tap changer connections. Lack of conductor flexibility appears to be leading cause of problem.
A three-year-old, oil-filled, 5MVA, 21kV/4.16kV power transformer feeding a refinery fails unexpectedly. Because the facility has no adequate back up, the refinery experiences extensive production losses. The owner repairs the failed unit and places it back in service. Shortly thereafter, the transformer fails again. The result: substantial production loss.
The refinery turns the matter over to its legal department for possible litigation. The law firm representing the refinery contacts our organization and requests an investigation to find the cause of the transformer failures. A careful examination reveals that a small cost-saving change in the transformer’s construction is the most likely cause of failure.
Investigation details. We direct the refinery to remove the transformer cover and drain the oil prior to our arrival. This helps speed up our inspection process. (The owner had already removed the failed unit from service and placed it in storage.)
We conduct extensive interviews with facility engineers and operating personnel, but they indicate nothing unusual prior to the transformer failure. There were no lightning storms in the area at anytime preceding the failure, which tends to eliminate lightning impulse as a possible cause. However, we can’t easily rule out switching transients, since electric utilities routinely switch feeders and capacitor banks on their system. This switching is known to produce voltage transients.
Investigation determines coils are okay. Our examination of the complete assembly reveals that the transformer coils were not involved in the failure. The transformer core also seems untouched by this incident. However, the tap changer—supported by the core clamping structure—is severely damaged by heat. Portions of the tap changer insulation supports are broken, possibly due to mechanical stress caused by high temperature.
To evaluate the transformer components, we arrange to have the non-load tap changer removed from the transformer tank along with the phase leads for further investigation. We then test the transformer coils to see if any of the turns are shorted. The tests indicate the coils are not shorted, and therefore they did not cause the problem. The insulation resistance of the coils is also tested. The HV to LV, HV to ground, and LV to ground insulation resistances are high enough that we consider them satisfactory. This indicates the transformer core and the coils are probably in good condition. Finally, we focus our investigation on the tap changer itself.
Tap changer design is the problem. Tap changers can fail for a number of reasons. Improper contact pressure at the tap contacts can cause hot spots and result in thermal degradation of the switch. Overvoltages, due to switching or lightning impulses, produce voltage stresses at the tap connections. This can cause insulation breakdown. These tap changers were designed for tap change under no load, which is common. Using tap changers of this type to change taps under load produces excessive contact wear due to arcing, resulting in premature failure. However, the investigation reveals the damaged tap changer was not operated under load conditions.
The tap changer consists of three decks, one for each phase and six posts per deck for connection of the transformer tap leads. A cam driven shorting link is used to change the tap setting by operating the tap changer handle. The tap changer posts are mounted on insulating disks supported by a heavier phenolic board. Areas surrounding two of the posts are severely burnt. One of the posts has come loose from the supporting insulation disk. There is also evidence of arcing between Deck 1 (closest to the operating handle) and Deck 2 of the switch.
The examination reveals probable deficiencies in the tap changer design and construction. Normally, to reduce mechanical stress on phase leads and the tap changer mechanism, flexible stranded conductors are used to make the electrical connection between the tap changer and the transformer coil. The transformer manufacturer in this case provided rectangular copper magnet wires for tap changer connections instead of flexible stranded wires. The magnet wires are fairly rigid, heavy and impose considerable mechanical stress on the tap changer components. In addition, several drilled holes are in the insulating disk, which mechanically weakened it.
During a typical day, a transformer can go through temperature cycling due to load and ambient temperature changes. This introduces expansion and contraction of the tap changer components, which further increases the mechanical stress on the unit. The combination of the physical stress produced by the rigid tap leads and the expansion and contraction of the components due to temperature variation can cause the insulating disk to crack. Prior to our involvement, the transformer manufacturer claimed that the transformer failure was caused by an electrical surge.
Pinpointing the cause. Attaching flexible stranded tap leads to a transformer coil is typically a time consuming process. Magnet wires are heavier, and they more easily attach to transformer windings by brazing. Providing magnet wires for tap connections is viable if the leads are mounted on posts supported on rigid insulating material that is not subject to movement. A tap changer, however, is a mechanical device subject to movement.
We determine that using magnet wires for tap leads is a misapplication. We don’t know if the transformer manufacturer changed its method of tap lead attachment in transformers manufactured subsequent to this incident, or if this method was only used on this repaired unit. Fortunately, we have no further reported failures at this facility since the investigation. The transformer manufacturer settles this case out of court.