Transformer dew point measurement and insulation moisture estimation part 2

Author: Tom Melle


As discussed previously, elevated levels of moisture in the solid insulation of power transformers have a detrimental effect on the insulation’s integrity and will accelerate the degradation of the paper insulation. High moisture content will also reduce the dielectric strength in oil and can lead to possible bubble formation during overload conditions.


Prior to final oil filling, the dew point measurement technique is widely used to estimate moisture content of dry insulation inside of power transformers. Experience strongly suggests that the technique provides an accurate and reliable method of moisture estimation in a gas space. When used with proper understanding, it can be a very useful tool.


Background on dew point measurements


The dew point inside a transformer can be related to the partial pressure of water vapor at the surface of the insulation material. This is called the vapor pressure and it is measured in thousandths of a millimeter of mercury (micrometers of Hg.).


The vapor pressure and the temperature of the insulation material determine the moisture content, in percent by weight, at the surface of the insulation material. Therefore, it is critical that all partial pressures of gases in the gas space and in the insulation have come to equilibrium.


Estimation of insulation moisture content


The purpose of the following measurement “best practice” is to establish a consistent process for calculating the transformer insulation moisture content from dew point measurements. Equipment and conditions can vary as well as acceptance criteria. It is best to check with the equipment manufacturer for advice on acceptance criteria.


A good moisture content estimation should foremost have an approved procedure, as well as a high quality, calibrated meter to measure the dew point of the transformer insulation system.


Instruments and tools required:

  • Dew point meter

  • Stainless steel or PTFE (Teflon) sampling line

  • Sample line adaptors (e.g. ¼” NPT fittings)

  • Flow gauge and/or pressure regulator

  • Moisture equilibrium charts or an approved data entry program/spreadsheet.

Best measurment practices


The unit atmosphere must be allowed to stabilize in order to obtain an accurate sample and dew point reading.


After a unit has been pressurized with dry air or nitrogen, the unit must sit for 6-12 hours to allow temperature and moisture stabilization. The time and date of final pressurization is to be recorded.


In the field the best time for the insulation and tank temperature to be in equilibrium is in the early morning hours. Taking a dew point reading at this time will provide the best opportunity for accurate results.


When ambient air is below zero (Celsius), taking dew point measurements involves special considerations and a reliable transformer service group should be consulted.


Recommended steps

  1. Sampling will typically be performed at the ¼” sample valve on the pressure/vacuum gauge or at the sample valve provided in the inert gas system box (if present). In the instance that a sampling valve is not provided, a temporary ¼” sampler can be installed on the UPPER oil fill valve (the highest possible location) in order to collect the sample.

  2. After the proper sit time, open the valve at the sample point to be tested and allow enough air to escape to clear the valve and line of any dirt, oil or moisture condensate. If the line is clear, close the valve after one minute.

  3. A calibrated dew point meter should always be used to perform the dew point measurement. The meter should be left in the same environment as the unit being tested for one to two hours to reach ambient temperature before use. Inspect the instrument condition prior to use. There should be a “zero” or calibrate process for the dew point meter.

  4. Prepare the sampling line (stainless steel or PTFE only). Tygon, plastic, rubber, or nylon tubing should not be used under any circumstances due to moisture adsorption tendencies. The best tubing material is stainless steel. Even the best nonmetallic tubing can absorb and desorb water vapor, creating a lag in measurement response. Teflon tape or PTFE sealant can be used to keep the connections tight. Connect one end of the sample line to the transformer sampling valve and repeat the valve clearing process (open the valve to purge the sampling line). After purging the sampling line, connect the other end to the dew point meter sensor assembly.

  5. Adjust the sample valve or regulator to give a gentle flow through the instrument. It is suggested to reduce the line pressure to below five PSI. Adjusting the flow rate of the sampling valve may slightly improve the dewpoint reading. This is acceptable. A flow rate of less than 20scfh (10L/m) is ideal. Any excessive air flow rate will create a local pressure drop at the sensor. Because dewpoint temperature is pressure sensitive, the excessive flow rate will create an error in the measurement. (A resulting line pressure of two PSI or less is preferred, otherwise a pressure correction should be made when calculating the final moisture content value).

  6. Allow the sampled air to flow for a minimum of five minutes in order to purge the sensor assembly. Once purging and/or calibration is complete, the instrument reading will change and tend toward stabilizing. Stabilization usually takes several minutes. It is recommended to allow the meter to become stable (analog needle stops moving or digital readout in the “tenths” position not fluctuating). Ideally you should wait an additional 30 seconds before recording the dew point reading in case of a false stabilization. (Note: the drier the gas space, the longer it will take for stabilization to occur).

  7. If the reading continues to fluctuate after five minutes then either the sample line was not purged properly or the sample is becoming drier during the test. In either case the test should be continued until there is no doubt that the reading is accurate.

  8. In addition to recording the dew point measurement, record the unit temperature from either of the following: a fiber optic sensor, a laser temperature gauge, the Liquid Temperature Indicator, or the ambient temperature (in order of preference). Also record the unit pressure from the pressure/vacuum gauge on the transformer.

  9. If necessary, a second (confirmation) sample can be taken after the unit is allowed to sit for another 6-12 hours. Follow the previous purging and sampling procedure and record the dew point, unit temperature and unit pressure. If the unit temperature is changing during the sampling period, an average of multiple samples is widely considered to be more accurate than just one sample.

  10. When satisfied with the consistency of the results, shut off the sample valve at the tank and disconnect the sampling hose. Shut the meter down and return the instrument to storage.

Acceptable criteria


After the dew point temperature (degrees C), unit pressure reading, and unit temperature has been obtained, the unit moisture content can be calculated (estimated). The moisture content calculation is to be accomplished using the “Piper” charts from C57.93-2007 Annex B - Moisture Equilibrium Charts. The moisture content calculation can also be performed using computer programs or spreadsheets that duplicate the chart results.


Generally 0.5% is the maximum acceptable moisture content of the insulation material of a power transformer prior to leaving the factory. The moisture level may increase slightly over time due to stabilization or atmospheric changes.

Transformers with dew point levels indicating a percent moisture that is significantly higher than 0.5% should be allowed to sit for an additional 6-12 hours and the dew point test retaken. If the percent moisture indication has increased, allow the transformer to sit another 6-12 hours and repeat the dew point test. Continue to repeat this cycle until stabilization is reached or a field vacuum/hot oil process is deemed necessary.


Variables to consider when taking a dew point measurement


Equilibrium - Equilibrium conditions must be reached. This requires a minimum of 6-12 hours of sit time following a change (temperature, pressure, etc.) to the gas space.


Insulation temperature(s) - Insulation temperature must be known accurately. Varying conditions can create a temperature delta inside the transformer (inner windings hotter than outer, top of tank hotter than bottom of tank, etc.). Insulation temperature is by far the most critical variable other than the dew point measurement itself.


Tank pressure – the tank pressure directly affects the moisture content calculation and must be accurate.


Sensor pressure – the dew point sensor pressure affects the dew point measurement. High flow rates may create a “pressure penalty” and negatively affect the moisture content calculation.


Type of meter - The measuring equipment shall be properly calibrated and be in good working order. Different meters have varying degrees of error and range. Ensure the meter matches the desired dew point measurement range.


Tdew/Tfrost - Some meters have settings for a Frost Point rather than a Dew Point. When comparing measurements, always ensure you are using the same measurement scale.


Sampling line – Proper choice of sampling line is critical to an accurate dew point measurements.