How Do You Draw a Representative Transformer Oil Sample?
- Knowledge Center
- Articles
- Oil Sampling
- How Do You Draw a Representative Transformer Oil Sample?
This article is Part 2 of a three-part series on transformer oil sampling. Part 1 addressed the equipment data, nameplate information, and operational conditions that must be documented before submitting a sample for analysis. Here we will address the field practices that determine whether an oil sample is truly representative of internal equipment conditions — covering safety, valve preparation, flushing, ambient considerations, and sample handling. In Part 3, we will cover how laboratory results are interpreted to diagnose transformer health.
A Good Sample Requires More Than Method Compliance
Accurate oil analysis depends on two inputs of equal importance: complete equipment documentation and a representative sample. Part 1 covered the former. This article addresses the latter.
Most transformer oil sampling in the U.S. is performed in accordance with ASTM D923 (Standard Practices for Sampling Electrical Insulating Liquids). IEC 60475 and IEC 60567 are the applicable international standards. These methods are well established and will not be reviewed here. What they do not fully address are the field-level details that most often compromise sample quality — preparation, contamination control, ambient conditions, and handling. Those are the focus of this article.
Safety and Pre-Sample Preparation
Safety planning should precede every sampling activity. Treat the worst-case scenario as the baseline, not the exception.
Before beginning, obtain all required permits — lockout/tagout and any site-specific authorizations — and confirm you have properly rated tools and appropriate personal protective equipment. NFPA 70E provides guidance on PPE selection for electrical work environments.
Conduct a site survey before approaching the equipment. Check for electrical hazards, trip hazards, and signs of wildlife. Rodents, skunks, and other small animals are commonly found in and around transformer enclosures. Bees and snakes are also worth checking for before opening access panels or working near ground-level drains.
Prepare your supplies before starting: bring extra sample bottles with airtight lids, ground glass syringes, and sufficient labels and markers. Label each sample container as it is filled — not after the fact — to eliminate any risk of mix-up when pulling multiple samples in the field.
Confirming Positive Pressure Before Sampling
Never attempt to draw a sample from a unit that is under negative pressure (vacuum). Doing so risks introducing an air bubble into the equipment, which can cause dielectric failure.
Verify that the unit is under positive pressure using the procedure in ASTM D923 or an equivalent proven method. Do not rely solely on the pressure gauge — gauges can be plugged or non-functional. If the unit is found to be under vacuum, reestablish positive pressure using Type III nitrogen per ASTM D1933 before proceeding.
Valve Cleaning and Oil Flushing
Clean the Drain Valve Exterior and Interior
Contaminants on the outside of the drain valve are a common and easily overlooked source of sample contamination. Clean the exterior of the valve thoroughly before opening it. If accessible, clean the interior as well.
Flush Before You Fill
Contaminants — including water, metal debris, and paper fiber particulates — tend to settle near the drain valve over time and are not representative of bulk oil conditions. A minimum of two liters (approximately two quarts) of oil should be flushed through the valve into a waste container before drawing the actual sample. In some cases, particularly with LTCs and OCBs, two to four liters may be required to fully clear settled contamination.
A practical field technique: use a spare sample bottle to collect flush oil, fill and discard it several times, and use the final flush collection — just before discarding — to measure the oil sample temperature. That temperature reading is required by the laboratory to calculate percent moisture saturation and moisture by dry weight for oil-filled transformers.
Operational and Ambient Conditions at Sampling
Load History Affects Moisture Results
Sampling during periods of relatively stable load and temperature produces the most representative results. A unit that has been operating under heavy load for an extended period will show elevated moisture levels in the oil — not because the insulation system is wet, but because heat-driven moisture migration from the paper into the oil has not yet reached equilibrium. If that unit is sampled before moisture redistributes back into the solid insulation, the result may indicate a wet condition that does not reflect the true steady-state moisture content of the transformer.
Weather and Environmental Contamination
Adverse ambient conditions can compromise sample integrity. Wind can introduce airborne dust into an open sample container. High ambient temperatures increase the risk of perspiration contamination. Rain, snow, and high humidity present additional exposure risks. When conditions are poor, take extra precautions to shield the sample container during filling, and minimize the time the lid is off.
Sample Handling and Shelf Life
Samples should be submitted for testing as soon as practical after collection. For oil quality and dissolved gas analysis, samples approaching 45 to 60 days of age are nearing the end of their useful shelf life. Properly sealed samples in the correct containers can still yield accurate results near that threshold, but as samples approach 60 days, resampling should be considered.
Furans analysis and PCB testing are exceptions — those analyses can be performed on samples up to one year old, provided samples are stored under controlled conditions, shielded from heat and light.
Summary:
- ASTM D923, IEC 60475, and IEC 60567 are the established sampling standards; this article addresses the field-level practices those methods do not fully cover.
- Safety and site preparation must precede sampling — obtain all required permits, use properly rated PPE per NFPA 70E, and survey the area for electrical hazards and wildlife before beginning.
- Positive pressure must be confirmed before drawing a sample; sampling under vacuum risks introducing air into the equipment and causing dielectric failure. Use ASTM D1933 Type III nitrogen to restore pressure if needed.
- Valve cleaning and flushing are critical contamination controls — flush a minimum of two liters before collecting the sample, and use the final flush to measure oil temperature for moisture calculations.
- Load history influences moisture results — sampling too soon after a period of heavy load can produce artificially elevated moisture saturation values that do not reflect steady-state insulation condition.
- Ambient conditions including wind, high temperature, rain, and humidity can contaminate samples; take precautions to protect the open container during filling.
- Sample shelf life for oil quality and DGA is 45–60 days; furans and PCB samples are stable for up to one year under controlled storage.
