2 Interpretation of measurements
Significance of temperature Any sudden changes in ambient temperature will increase
the measurement error since the temperature of the appara-
Most insulation measurements have to be interpreted based
tus will lag the ambient temperature.
on the temperature of the specimen. The dielectric losses
of most insulation increase with temperature; however, e.g. Dissipation factor-temperature characteristics, as well as
dry oil-impregnated paper and polyethylene of good quality dissipation factor measurements at a given temperature,
exhibit decrease of dielectric losses when temperature is may change with deterioration or damage of insulation.
raised moderately, e.g. from 20°C to 30°C. It is also known This suggests that any such change in temperature char-
that the effect of temperature depends on the aging status acteristics may be helpful in assessing deteriorated condi-
of the insulation. In many cases, insulations have failed due tions. As an example, bushings have typically a rather flat
to the cumulative effect of temperature, i.e., a rise in tem- temperature correction with only slightly elevated values
perature causes a rise in dielectric loss which in turn causes at high temperatures. Generally a bushing showing highly
a further rise in temperature, etc (thermal runaway). increased dissipation factor at elevated temperature should
be considered “at risk”.
It is important to determine the dissipation factor-tempera-
ture characteristics of the insulation under test. Otherwise, Be careful making measurements below the freezing point
all tests of the same specimen should be made, as nearly as of water. A crack in an insulator, for example, is easily
practicable, at the same temperature. detected if it contains a conducting film of water. When
the water freezes, it becomes non-conducting, and the
To compare the dissipation factor value of tests made on
defect may not be revealed by the measurement, because ice
the same or similar type apparatus at different temperatures,
has a volumetric resistivity approximately 100 times higher
it is necessary to convert the value to a reference tempera-
than that of water. Moisture in oil, or in oil-impregnated
ture base, usually 20°C (68°F). Examples of standard tables
solids, has been found to be detectable in dissipation factor
of multipliers for use in converting dissipation factors at
measurements at temperatures far below freezing, with no
test temperatures to dissipation factors at 20°C are found in
discontinuity in the measurements at the freezing point.
the Appendix A of this document.
Insulating surfaces exposed to ambient weather conditions
In reality, temperature correction for a specific compo-
may also be affected by temperature. The surface tempera-
nent is always individual and pending age/condition.
ture of the insulation specimen should be above and never
DELTA 4000 has a unique and patented feature for estimat-
below the ambient temperature to avoid the effects of
ing the individual temperature correction (ITC). By measur-
condensation on the exposed insulating surfaces.
ing dissipation factor over frequency and using mathemati-
cal formulas and models of insulation characteristics, the
correct temperature correction can be determined from
5 to 50°C measurement temperature to 20°C reference
temperature. The input data for the calculation is dissipa-
tion factor measured from 1 to 500 Hz and the method
is principally based on Arrhenius’ law, describing how the
insulation properties are changing over temperature.
κ = κ0·exp(-Wa/kT)
With activation energy Wa and Boltzmann constant k
The test temperature for apparatus such as spare bushings,
insulators, air or gas filled circuit breakers, and lightning
arresters is normally assumed to be the same as the ambient
temperature. For oil-filled circuit breakers and transform-
ers the test temperature is assumed to be the same as the
top oil temperature or winding temperature. For installed
bushings where the lower end is immersed in oil the test
temperature lies somewhere between the oil and air tem-
perature.
In practice, the test temperature is assumed to be the same
as the ambient temperature for bushings installed in oil-
filled circuit breakers and also for oil-filled transformers
that have been out of service for approximately 12 hours.
In transformers removed from service just prior to test,
the temperature of the oil normally exceeds the ambient
temperature. The bushing test temperature for this case can
be assumed to be the midpoint between the oil and ambient
temperatures.
12 DELTA 4000 ZM-AH02E
