Why Transformer Bushings Fail, and How to Stop It Happening

A bushing is a small part of a transformer and a large share of its failure statistics. It carries full line voltage through an earthed tank wall, so it lives permanently under electrical stress, thermal cycling and weather. This page sets out the failure mechanisms that actually cause bushings to fail in service, the condition-monitoring signs that precede each one, and the design and manufacturing decisions that remove them. Indisol is a Portuguese manufacturer of cast epoxy bushings, and where our own practice is relevant we say so plainly.

1. Why bushings fail more often than the transformer

The bushing is the one component that must be a conductor, an insulator, a pressure seal and a structural support at the same time. Its insulation sits in the narrow radial gap between a live central conductor and an earthed flange, which is the highest field region in the whole assembly. It is bolted to a tank that heats and cools with load, so the joint between organic and metallic materials is worked mechanically every day of its life. And unlike the windings, it is exposed: rain, salt, dust, UV and ice all act directly on it. None of that is a defect. It is simply the duty, and it explains why bushing problems appear early in the life of a transformer and why the mechanisms below repeat across every manufacturer and every technology.

2. The four failure mechanisms

Field failures cluster into four groups. They are not independent: moisture accelerates partial discharge, discharge accelerates thermal ageing, and thermal cycling opens the path that lets moisture in.

3. How each one announces itself

Bushings rarely fail without warning. The warning is just quiet, and it is electrical rather than visual. If you are monitoring a fleet, these are the indicators worth trending.

• Capacitance and tan delta (power factor). A rise in tan delta at the same temperature is the single most useful indicator of moisture and general insulation deterioration. A change in measured capacitance points to a shorted or open grading layer.
• Partial discharge measurement. Detects voids and interface problems long before they erode a path. This is why we test for partial discharge in production and not only at type approval.
• Insulation resistance. Cheap, quick, and sensitive to gross moisture ingress and surface contamination, though not to small internal defects.
• Infrared thermography. Finds a hot connection at the top terminal or a resistive joint inside, both of which drive local thermal ageing.
• Dissolved gas analysis, for oil-filled designs. Rising acetylene or hydrogen from a bushing compartment indicates arcing or discharge inside it.
• Visual inspection. Oil weeping at the flange, carbonised tracks on the surface, chalking, chipped sheds and a corroded or loose earth connection are all worth acting on.

Two practical notes. First, a single measurement means little; the value is in the trend for the same bushing measured under similar conditions. Second, cast epoxy has an advantage here that is easy to miss: it is a dry, sealed-for-life system, so there is no oil level to lose, no paper to dry out, and no reason for the insulation condition to depend on a maintenance regime.

4. Design: what has to be right before anything is cast

Every decision below exists to remove one of the four mechanisms, and all of them are taken at the drawing stage, where they are still free to change.

Field control in the radial gap

The insulation thickness, the shape of the conductor and the geometry of the earthed flange region set the local field. Sharp edges and abrupt transitions at the flange are where partial discharge starts, so the transition is rounded and, where the voltage requires it, graded.

Creepage distance and shed profile

External flashover is a function of creepage distance, shed profile and the pollution class of the site. A bushing specified for a clean indoor cubicle and installed at a coastal substation will fail on the surface long before its internal insulation is exhausted, which is why we ask about the installation environment and not only the voltage.

The metal-to-resin interface

This is where thermo-mechanical ageing does its work. Differential expansion between the metal insert and the resin has to be absorbed without opening a gap, so insert geometry, surface preparation and the resin system are chosen together. Our bushings are cast with the metal parts encapsulated, and we do the metallization in-house so the surface the resin bonds to is under our control.

Mechanical duty and mounting

Cantilever load from the connected conductor, torque applied when the terminal is tightened and short-circuit forces all pass through the same part. Mounting method, flange dimensions and the tightening arrangement are engineered for the load case rather than borrowed from a similar rating.

5. Resin: why we formulate our own

We do not buy a general-purpose epoxy and cast it. We formulate our own resin systems, which is what lets us match the material to the duty rather than the other way round. Two families cover most of what a bushing needs, and the choice between them is usually decided by where the part is installed.

The properties that matter for the failure mechanisms above are dielectric strength through life rather than on day one, resistance to thermal shock, low water absorption and non-tracking surface behaviour. There is more detail on the material itself in our reference page on epoxy for electrical insulation.

6. Manufacturing and testing controls

A void a fraction of a millimetre across is enough to start partial discharge, and no amount of good design survives a casting process that leaves one. Our process is vertically integrated for that reason: resin formulation, tooling, metallization, casting and testing are all under one roof, from prototype through to series production.

• Vacuum casting and APG. Resin is degassed and cast under vacuum, and automatic pressure gelation controls the gelation front so the part fills and cures without entrapped air or shrinkage voids.
• 29 clamping machines. Capacity from small parts up to pieces over 50 kg and 1 m, so the tooling suits the part instead of the part being redesigned to suit the tooling.
• In-house metallization. The bonding surface of every insert is prepared by us, and that interface is what governs thermo-mechanical life.
• 100% electrical testing. Every unit is tested, not a sample. Our laboratory tests to 200 kV at industrial frequency and to 200 kV for partial discharge, and also measures contact resistance and performs high-voltage and current metrology.
• X-ray inspection. Used where internal integrity has to be demonstrated rather than inferred, including parts destined for SF6 equipment.
• Type testing. Design qualification catches the problems that belong to the design, before the design reaches your transformer.

We are certified to ISO 9001 and ISO 14001. The approvals and test reports are listed on our certificates page, and the process itself is described in more detail under technology.

7. Epoxy or porcelain?

Porcelain has an excellent record and remains the right answer in some applications. Where cast epoxy changes the failure picture is mainly in brittleness and in weight.

• Thermal shock. Epoxy does not crack from thermal shock the way porcelain can, which removes one of the classic paths for moisture to get in.
• Weight. A lighter part is easier to handle and install, and imposes less on the mounting.
• Surface behaviour. Non-tracking and weather resistant, and available with silicone overmoulding for polluted or coastal sites.
• Electrical behaviour. Low partial discharge, high compression strength and good resistance to arcing.
• Failure mode. A cast part does not shatter, which matters for the equipment and the people next to it.

The comparison is set out at greater length in our comparison of epoxy and porcelain bushings, and the range of types is on the bushings product page.

8. Common questions

What is the most common cause of transformer bushing failure?

Moisture ingress, usually through a degraded seal or a crack, and usually revealed by a rising tan delta before it becomes a flashover. Partial discharge from casting voids or a poor metal-to-resin interface is the second, and the one that is entirely preventable at the manufacturing stage.

How long should a bushing last?

A correctly specified bushing should last the life of the transformer. Where they fall short it is normally a mismatch between the specification and the site: creepage distance chosen for a cleaner environment than the real one, or a mechanical load case that was underestimated.

Can a failing bushing be detected before it fails?

In almost every case, yes. Tan delta and capacitance trending, partial discharge measurement, thermography and, for oil designs, dissolved gas analysis will all show a deteriorating bushing well ahead of breakdown.

Does cast epoxy need maintenance?

It is a dry, sealed system with no oil to check or replace. In polluted environments the external surface should be inspected and, where necessary, cleaned, and the earth connection should be checked.

What do you need from us to quote a bushing?

Rated voltage and current, insulation levels, indoor or outdoor and the pollution class, mounting arrangement and flange interface, mechanical loads, the drawing or the part being replaced, and quantities. A sketch and the ratings are enough to start.