Matz Ohlen - Director - Transformer test systems
Accurate information about the condition of insulation in a transformer is essential for effective asset management and accurate risk assessment. Information about moisture content limits the loading of the transformer accelerates the aging of the paper insulation, thus reducing transformer life. Dielectric frequency response measurement provides a cost- and time-effective method for determining transformer health, including the moisture content of the oil/paper insulation system.
Traditionally insulation condition is assessed by measuring tan delta/power factor at the mains frequency of 50 or 60 Hz. This approach is however of limited value. To see why, consider for example three power transformers that have the same tan delta/power factor at mains frequency. Further examination may, however, reveal that one of the transformers is wet with good oil and needs to be dried out, another is dry but has oil that needs reconditioning or regeneration and the third transformer is just in normal service-aged condition. The mains frequency tan delta test provides no way of distinguishing between these three different cases. Figure 1 clearly illustrates how misleading tan delta results can be.
In contrast, a dielectric frequency response (DFR/FDS) diagnostic test will provide accurate and conclusive information about the condition of the transformer insulation. Such data is essential in prioritising the maintenance of transformer fleets and making sure assets will reach and exceed their expected service lives. See Figure 1.
High moisture in the transformer insulation gives rise to a number of well-known issues, including:
- Reduced load capability – The load capability is limited because of the decreased bubbling inception temperature
- Reduced dielectric strength – The dielectric strength of the oil is decreased, along with the partial-discharge inception voltage
- Increased ageing – High temperatures and moisture dramatically accelerate ageing that lowers the mechanical strength of the cellulose insulation
For good life management of transformers, the moisture content of the insulation must be kept at the lowest possible level. Transformers are dried during the manufacturing process until measurements or standard practices yield moisture content in the cellulose-based insulation of less than 1%. After this initial drying, however, the moisture content of the insulation system will continually increase.
Figure 1 – Comparison of three transformers in very different conditions, but with the same tan delta.
There are several sources of water that drive this increase, including:
Residual moisture not removed during the factory dryout process
Moisture on the insulation surface during assembly and/or maintenance
Moisture ingress from the atmosphere (gasket leaks, breathing during load cycles, site erection process)
Ageing and decomposition of cellulose and oil
It is important to understand that the moisture in the transformer mainly resides in the solid insulation: typically 99% of the moisture in the transformer resides in the cellulose-based insulation. Knowing the moisture content is essential for transformer owners whether this information is used to verify that the transformer is dry as part of the commissioning process or to assess the condition of an older unit to determine if there is a need for preventive or reactive measures.
Assessing moisture content in transformer insulation based on oil sample tests is unreliable as the water migrates between the solid insulation and oil as the temperature and/or transformer load changes. An oil sample for moisture analysis has to be taken at relatively high temperature and when the transformer is in equilibrium. Unfortunately, this is a rare state for transformers, and the result is that most assessments based on oil sampling are unreliable. Experience has shown that this method tends to overestimate the amount of water in the insulation.
The DFR/FDS method is, in contrast, a very reliable method providing a high degree of accuracy in assessing the moisture content of the paper insulation, since it derives the moisture content in paper or pressboard from the measured dielectric properties.
The curve of dissipation factor (tan∂) plotted against frequency has the typical shape shown in Figure 2. With increasing temperature the curve shifts towards the higher frequencies. Moisture influences the low and the high frequency areas. The linear, middle section of the curve reflects oil conductivity. Insulation geometry determines the knee points, which are located to the left and right side of the steep gradient.
With DFR/FDR instruments like the Megger IDAX moisture determination is based on a comparison of the transformer’s measured response to a modelled dielectric response. The insulation model used is the internationally recognized X-Y model described in CIGRE TB 254 and 414. The insulation duct between the LV and HV windings of the transformer is reduced to a single capacitor where the dielectric materials consist of cellulose and oil.
The single capacitor can be analytically described. A matching algorithm compares the model data with the measured data and adjusts the dielectric response of the single capacitor until a best fit with the measured response from the transformer is achieved. From the modelled response, the moisture in the cellulose material and oil conductivity/dissipation factor is determined as well as temperature dependence of the dissipation factor. Only the insulation temperature (oil/winding temperature) needs to be entered as an input parameter.
Figure 2 – Dissipation factor plotted against frequency for a power transformer
Several international standards, guides and reports give guidelines for insulation assessment in terms of power frequency tan delta/power factor at 20°C, moisture content in solid insulation and conductivity/dissipation factor of oil. Insulation assessment using DFR/FDS techniques provides results and categorization for tan delta/power factor of the whole insulation system at 20°C, moisture in cellulose and oil conductivity/dissipation factor at 25°C, giving the user a detailed description of the insulation condition.
Guidelines for oil conductivity/dissipation factor at 25°C are given in IEEE C57.152-2013. For power frequency tan delta/power factor at 20°C the assessment limits are based on industry practice and the guidelines in CIGRE TB 445 and IEEE C57.152-2013.
In summary, DFR/FDS analysis of transformer insulation is a timesaving method based on the use of multi-frequency techniques to give the shortest possible measurement time. It offers temperature independent results and, when a wide frequency range and the latest modelling methods are used, it gives results that are accurate and reliable. Further, the DFR/FDS test method has been validated by more than ten years of practical experience, and is now included in international standards and guides for transformer diagnostics.