Evaluation of insulation condition based on the distribution of current and resistance as a function of measurement duration
The insulation resistance measurement involves measuring the current flowing through the insulation material at a test voltage suitable for the object. Based on Ohm's law the insulation resistance of the material from which the insulation is made is calculated. This parameter - thanks to the possibility of comparing it to the required values - is generally regarded as ultimately sufficient for assessing whether the condition of the insulation of an object is satisfactory or not.
However, attention must be paid to the phenomena that occur during the test (resulting from capacitance and absorption), especially for objects with significant capacitance (cables) or objects such as motors or transformers. In such cases the observation of changes in resistance and measurement current as a function of time may reveal dangerous weakening of the insulation, despite the fact that the obtained result of the insulation resistance measurement will be evaluated positively. The article will present the principles of insulation resistance measurement, the calculation of DAR and PI coefficients and examples of their practical application.
Introduction
Every insulating material is characterised by some ability to conduct electricity, as there is no ideal insulation that is completely non-conductive. This property is used to determine the resistance of insulating materials in electrical equipment. The inspection of the technical condition of electrical installations, as required by the building or electrical code and carried out in accordance with IEC 60364-6:2016 [1], must include the measurement of the insulation resistance of, among other things, the electrical conductors of which the installation is made. This is also how these measurements are commonly perceived in the electrical measurement community. This determines a certain stereotype regarding the measuring instruments used in this area. It is based on the assumption that any meter with the correct test voltage can be used for insulation resistance measurements. In principle, one can agree with this - but only when it comes to the aforementioned measurements of electrical installations.
The only problem you have to deal with is the preparation of the electrical installation to carry out the measurements. In operating electrical installations, this is a labour-intensive process, requiring the disconnection of voltage, loads and control devices (e.g. actuators), which is not always possible. However, you can avoid this inconvenience if you use the possibility of short-circuiting live conductors for the duration of the measurement in individual circuits, as allowed by IEC 60364-6:2016 p 61.3.3 [1].
The case is quite different when measuring more advanced measurements, e.g. of electric motors, transformers, MV cables or equipment requiring high test voltages. It is necessary in this case to look at the phenomena occurring during the measurement, as well as the meters that will allow the test to be carried out correctly.
Capacitive charging current, polarisation current, leakage current
Fig. 1. Overview graph of currents during insulation resistance measurement
Fig. 1. presents an illustrative diagram of the currents as a function of time. When the test voltage is applied to the circuit to be measured, the capacitance-induced current is the greatest and disappears after it is charged much faster than the absorption current, which also dissipates, which in turn is due to the material's ability to accumulate charge. Only after these phenomena have ceased can the insulation resistance result be read. It should be noted that the changes in the readings of the measuring instrument caused by these currents do not actually represent changes in the insulation resistance, but are merely the result of the measurement process.
Routine testing is actually limited to reading out the measured insulation resistance value once the result has stabilised. However, observing this process over time is a very important and useful diagnostic tool. This is because it is possible for insulation failure to occur in a way that will disqualify the tested object despite meeting the acceptable resistance criterion. Such cases often occur when specific, incomplete insulation damage occurs during the overhaul of, for example, a motor. This applies not only to strictly rewinding motors, but also to mechanical repairs, e.g. replacement of bearings (fitting a motor deckle can lead to damage to the winding face insulation). The correct use of the information provided by the test process will protect us from making an erroneous judgement about the full efficiency of the inspected object.
DAR and PI coefficients
When measuring good undamaged and non-moisturised insulation, it is to be expected that the readings of the measuring instrument will increase in time, as a result of the change in total current shown in Fig. 2 Based on the time intervals, the coefficients specific to the measurement process are calculated: DAR (dielectric absorption ratio) and PI (polarisation index).
An increase in the meter reading will, of course, cause the aforementioned coefficients to take values greater than 1, obviously for good insulation.
The dielectric absorption ratio (DAR) is the ratio of the measurement result in MΩ within 1 minute to the result in MΩ after 30 seconds. When the measured leakage current stabilises within 1 minute, the test is usually no longer used to determine the PI factor, as the ratio of the measured values will be about 1. The PI coefficient, as the formula implies, is respectively the result in MΩ after 10 minutes to the result in MΩ after 1 minute.
The DAR and PI values commonly given in the literature and by test equipment manufacturers for assessing insulation condition are given in Table 1.
Table 1. Values of DAR and PI coefficients
| Item | DAR | PI | Condition of insulation |
|---|---|---|---|
| 1. | <1,2 | 1 - 2 | Unclear, questionable |
| 2. | 1,2 – 1,6 | 2 - 4 | Good |
| 3. | >1,6 | >4 | Excellent |
The coefficients should not take on values less than 1, because the insulation resistance readings must not decrease during the measurement.
It should also be noted that the changes in currents at high insulation resistance values can be very small, of the order of a few or several nA. Therefore, the IEEE 43-2013 standard [2] states the following:
“When the insulation resistance reading Riso obtained after applying voltage for 1 minute is greater than 5000 MΩ, based on the magnitude of the applied DC voltage, the total measured current (I total) may be in the sub-microampere range. At this level of required sensitivity of the measuring instrument, small changes in supply voltage, ambient humidity, test connections and other unrelated components can significantly affect the total current measured during the 1 min -10 min interval required for the PI test. Because of these phenomena, when Riso is higher than 5000 MΩ, PI may or may not reflect the correct insulation condition and is therefore not recommended as an assessment tool.”
In summary, for motors with zero or low absorption current, where the total leakage current stabilises within 1 minute, PI values are close to or equal to 1. In this case, PI is not an appropriate assessment tool. This often happens in rotating devices with random winding.
When determining the coefficients, it is not necessary to make a temperature correction because both DAR and PI are ratios, so if the test runs over the entire range under constant conditions, this is understandable. It is recommended that motors with low insulation resistance values should not be measured with higher test voltages than recommended. Since it is often possible to find devices for which there is no requirement for insulation resistance values in Polish standards. In this situation, the ANSI/NETA MTS-2011 standard can be used (Table 2) [3].
Table 2. Test voltage and required minimum insulation resistance values depending on the nominal voltage of the object under test
| Nominal Rating of Equipment (Volts) | Minimum Test Voltage (DC) | Recommended Minimum Insulation Resistance (Megohms) |
|---|---|---|
| 250 | 500 | 25 |
| 600 | 1,000 | 100 |
| 1,000 | 1,000 | 100 |
| 2,500 | 1,000 | 500 |
| 5,000 | 2,500 | 1,000 |
| 8,000 | 2,500 | 2,000 |
| 15,000 | 2,500 | 5,000 |
| 25,000 | 5,000 | 20,000 |
| 34,500 and above | 15,000 | 100,000 |
Practical example
In order to test the insulation resistance and DAR and PI coefficients, one motor overhaul company provided a randomly selected motor stator after rewinding. A Sonel MIC-10k1 meter with the function of creating graphs with current and voltage waveforms as a function of time was used for the measurements.
Fig. 2. Measurement result on the IMI meter and the tested 40 kW / 400 V motor stator
The 40kW/400V motor stator under test was rewound and refurbished. After heating and cooling, it was subjected to measurements. An IMI meter indicated an overrange at a test voltage of 500 V (>300 MΩ). Measurements made with a Sonel MIC-10k1 meter with a higher measuring range indicated an insulation resistance value of 430 MΩ. However, the DAR coefficient reached a value of 1.2. This created reasonable doubt about the insulation condition despite the result being many times above the required threshold (5 MΩ or ANSI NETA 25 MΩ). As the stator was fitted with new porcelain junction insulators and was also heated and dried, the influence of possible moisture was excluded. The current and resistance waveform during the measurement is interesting under these circumstances.
Fig. 3. Current and resistance graph. Visible anomalies
In about the fortieth second there is a sharp increase in the insulation resistance, after which the resistance decreases almost immediately and the process repeats itself. For reasons as above, interfering external factors were excluded. The problem therefore had to be sought in the insulation of the stator stacked windings themselves.
A thorough visual inspection did not establish the cause. This was only made possible by the successive removal of the windings. The problem was one pressboard spacer, which was cracked. When placing the windings in the grooves of the stator, the winding wire got between the pre-span and the groove and rested directly in the groove, touching the stator stripped of its pressboard insulation. The winding wire is enamelled, so the indications obtained were the result of current flow at the point of contact between the winding wire and the stator. Running the motor in this condition would certainly have damaged it. The stator was rewound again.
Fig. 4. Current and resistance graph after rewinding. Correct graph
As shown on the graph in Fig. 4, re-winding correctly allowed the insulation resistance to reach 12 GΩ (previously 430 MΩ) with a correct, constant increase in resistance during the measurement, which was also shown by the DAR coefficient = 2.
In summary, the example presented demonstrates the effectiveness of the method, with the DAR coefficient and the shape of the graph used to detect insulation damage. The company where the tests were carried out, after diagnosing this case, introduced full test and measurement procedures for the overhauled engines, changed the way the pressboard was trimmed, and equipped the workshop with appropriate measuring equipment. Since then, a similar incident has not occurred again at this company. On the other hand, breaches of insulation were detected several times after mechanical overhauls (e.g. replacement of bearings).
Fig. 5. Damage to the insulation sheath as a result of trimming with a clevis
Thus, given the consequences of downtime, which, although small in number, can after all be avoided, a full test for more than just insulation resistance alone seems particularly justified. Relative to the scale of damage that can be caused to an object with a latent defect, the cost of purchasing a suitable meter is small and certainly acceptable.
The information that has been included in this article is merely to signal the subject in order to broaden the perspective and to indicate to avoid routine when testing the insulation resistance of electrical equipment. This knowledge should encourage the exploration of the causes of phenomena occurring during measurements and their practical use in assessing the insulation condition of the objects under test. Annex 1 contains a sample motor test protocol.
Bibliography
[1] Standard IEC 60364-6:2016 Low voltage electrical installations - Part 6: Verification
[2] IEEE 43-2013 IEEE Recommended Practice for Testing Insulation Resistance of Electric Machinery
[3] ANSI/NETA MTS-2011 Standard for Maintenance Testing Specifications for Electrical Power Equipment and Systems
SONEL S.A. materials and author's own materials
Roman Domański
SONEL S.A.
