Self-Restoring and Non-Self-Restoring Insulation

Short answer

Self-restoring insulation fully recovers its properties after a disruptive discharge during a test; non-self-restoring insulation loses them or does not recover them completely. That difference defines how it is tested (many impulses with statistics, or few with no discharge at all) and the safety factor applied: IEC 60071-2 recommends 1.15 for internal insulation and 1.05 for external.

Short answer

Insulation is self-restoring if, after a disruptive discharge during a test, it fully recovers its properties within a short time. It is non-self-restoring if it loses them or does not recover them completely. IEC 60071-1 defines both in clauses 3.4 and 3.5. Air between electrodes and the surface of a clean insulator are the typical example of the first; paper-oil insulation, solid insulation and transformer windings, of the second.

Internal and external

The standard classifies insulation by where it is, as well as by how it behaves:

External insulation: air clearances and surfaces of solid insulation in contact with the atmosphere. It is exposed to pollution, humidity, wildlife and other site conditions. It may be protected from the weather (indoors) or not.
Internal insulation: internal distances of solid, liquid or gaseous insulation, protected from atmospheric and external conditions.

The two classifications overlap, but they are not equivalent. The standard notes that self-restoring insulation is generally, though not necessarily, external.

What changes in the test

Non-self-restoring insulation. A disruptive discharge degrades its properties, and even tests that produce no discharge can affect it. IEC 60071-2 (clause 5.3.2) explains that power-frequency overvoltage tests and impulse tests with polarity reversal can initiate treeing in polymeric insulation and generate gases in liquid or impregnated insulation. That is why it is tested with few applications at the standard level: three impulses per polarity, with no discharge at all.

Equipment that passes that test is considered, for coordination, to have an assumed withstand voltage equal to the applied voltage. Since the impulses are few and no failure is allowed, the result gives no statistical information about the actual withstand voltage.

Self-restoring insulation. A large number of impulses can be applied, limited only by the test requirements and not by the insulation, even with discharges. The advantage is statistical information. IEC 60060-1 standardizes methods that estimate the withstand voltage with 90 % probability, and IEC 60071-2 indicates as preferred the up-and-down method with seven impulses per group and at least eight groups. In IEC 60071-1 this is the statistical withstand voltage (withstand probability of 90 %); that of non-self-restoring insulation is the conventional assumed withstand voltage (probability of 100 %, no discharges).

Mixed insulation

Many pieces of equipment combine both types and cannot be tested in parts: bushings, instrument transformers and power transformers with higher-level bushings. The test must balance two opposite demands: few applications so as not to damage the non-self-restoring insulation, and enough to distinguish good self-restoring insulation from a defective one.

15-impulse procedure. IEC 60071-2 considers as an acceptable compromise applying 15 impulses and allowing up to two discharges in the self-restoring parts.
Adapted “two out of fifteen” procedure. To verify that there are no discharges in the non-self-restoring part, IEC 60071-1 (clause 6.3) requires, per polarity, a minimum of 15 impulses; that no discharge occurs in the non-self-restoring part, which is confirmed with five consecutive impulses without discharge after the last one; and no more than two discharges in total. The maximum possible is 25 impulses per polarity.
Transformers with higher-level bushings. If a discharge during the test would significantly damage the non-self-restoring part, the assembly is treated as non-self-restoring for the test.

The standard warns that, when non-self-restoring insulation is in parallel with self-restoring insulation, some procedures may apply voltages above the rated one and produce many discharges, and that this must be weighed when choosing the test. The full procedures are explained in impulse tests on instrument transformers and in type tests of switchgear.

Test selectivity

IEC 60071-2 (clause 5.3.4) describes the selectivity of a procedure as the difference between the actual withstand voltage levels that would give a pass probability of 5 % and of 95 %. The narrower that difference, the better the test distinguishes good insulation from deficient insulation. That is why procedures with few impulses, which protect the non-self-restoring part, are less selective than those with many impulses, and the fifteen-impulse one is the middle ground the guide considers acceptable for mixed insulation.

How it shows up in a specification

The type of insulation decides which impulse procedure the specification must request and what counts as a failure. For a transformer, a bushing or an instrument transformer, the product committee chooses among the standard’s procedures; for air and insulators, procedures with many impulses can be used. When a test report cites “fifteen impulses” or “two out of fifteen”, it is applying mixed-insulation logic.

Effect on insulation coordination

The type of insulation also enters the safety factor Ks. IEC 60071-2 (clause 4.3) explains it this way:

Internal insulation: ageing can be significant, and the recommended safety factor is 1.15.
External insulation: it is not assumed to be subject to ageing, except for organic materials exposed to the weather, and the recommended factor is 1.05. The atmospheric (altitude) correction is added.

These are values to use if the product committee does not specify others. For range II gas-insulated substations, the standard warns that larger factors may apply and that on-site tests should be considered.

To go deeper, see Insulation coordination: what it is and what it is for, Statistical method and insulation risk of failure, Safety factors and test conversion factors and Lightning and switching impulse: standard waveshapes.

When to call us

If you need to verify the dielectric test protocol of new equipment or repeat tests after transport or a repair, at Tiza Group we run field tests, in Colombia or elsewhere on request, and review the factory reports remotely. Learn about our testing and commissioning service and get in touch for a quote.

References

IEC 60071-1:2006, Insulation co-ordination - Part 1: Definitions, principles and rules: definitions 3.2 to 3.5 and 3.23; clause 6.3.
IEC 60071-2:1996, Insulation co-ordination - Part 2: Application guide: clauses 4.3 and 5.3.
IEC 60060-1, High-voltage test techniques (cited by the standard).

Frequently asked questions

What is the difference between self-restoring and non-self-restoring insulation?

Self-restoring insulation fully recovers its insulating properties shortly after a disruptive discharge during a test; non-self-restoring insulation loses them or does not recover them completely (IEC 60071-1, definitions 3.4 and 3.5). These definitions apply to discharges during a dielectric test: in service, a discharge can also partially damage self-restoring insulation.

Is external insulation always self-restoring?

Generally, but not necessarily. The standard notes that self-restoring insulation is usually external, without that being a rule. Air and insulator surfaces are the typical examples; internal insulation of paper, oil or solid material is non-self-restoring.

How is equipment with both kinds of insulation tested, such as a bushing or an instrument transformer?

With a compromise procedure. IEC 60071-2 (clause 5.3.4) states that 15 impulses with up to two discharges allowed in the self-restoring parts is an acceptable compromise. IEC 60071-1 (clause 6.3) adapts that procedure to verify that there are no discharges in the non-self-restoring part.