Why Do New Transformers Get Short DGA Retest Intervals?

This article covers retest recommendations for newly installed transformers. Here we will address why new units with very low dissolved gas levels still receive 1- or 3-month retest recommendations, and what standards and best practices inform those intervals.

New transformers sometimes return DGA results so clean that total dissolved gases register at just 1 ppm — well within any normal operating threshold. Yet these same units often carry a 1- or 3-month retest recommendation. For asset managers and reliability engineers, this can appear counterintuitive: if there’s no fault signature, why the accelerated interval?

The short answer is that low gas concentrations on a new unit aren’t a clean bill of health — they’re a baseline. The retest schedule is there to monitor for emerging issues during the most failure-prone window of a transformer’s life.

What a Short-Interval Recommendation Actually Means

When a new unit receives a 1- or 3-month retest recommendation on a DGA report, it is often accompanied by a comment such as: “The analysis of this sample shows only minor amounts of combustible gas. This baseline indicates normal operation.” The recommendation is not driven by concern over current gas levels — it’s driven by the unit’s age and installation status.

The dissolved gas results themselves may show no fault indicators, no trending combustibles, and no ratios of concern. The accelerated interval is a separate, proactive decision based on where the unit is in its service life.

The Reliability Logic Behind Early Testing

The rationale for close monitoring of new equipment follows a well-established reliability principle: early-life failures are disproportionately likely to occur shortly after energization. The logic is sequential — if a unit survives the first minute, it will likely survive the first hour; if it survives the first hour, it will likely survive the first day, and so on. This reflects the infant mortality portion of the classic bathtub curve in equipment reliability.

Manufacturing defects, improper assembly, insulation weaknesses, and installation errors tend to manifest under initial load and thermal stress. A unit that passes factory testing but has a latent defect may not reveal that defect until it has been energized and loaded in the field. DGA is one of the most sensitive diagnostic tools for detecting developing faults before they progress to failure — but only if sampling intervals are tight enough to catch early gas generation.

Industry Standards That Support Early Sampling

IEEE C57.104-2019

Unlike IEEE C57.104-2008, the 2019 revision abandons generic, fixed sampling guidelines in favor of a data-driven, percentile-based framework. Instead, it requires establishing a statistical baseline (Status 1) using actual numbers from the unit. This new framework bases all subsequent monitoring frequencies strictly on the 90th and 95th percentiles of gas accumulation and the actual rate of gas generation (volume per day).¹ This reflects the same early-life concern: the risk profile of a transformer is highest immediately following energization, and sampling frequency should be calibrated to that risk.

In the 2019 revision, there are only three condition levels:

These conditions are determined using percentile-based statistical thresholds. The 90th percentile and 95th percentile gas concentrations define the boundary for DGA status 1 to 3.

IEEE C57.93-2019

IEEE C57.93-2019 provides a rigorous DGA sampling timeline, adding mandatory tests at 1 week, 2 weeks, and 1 month to establish baselines under the updated IEEE C57.104-2019 standard.² This updated schedule ensures better Rate of Change (RoC) tracking for detecting early thermal or electrical faults, creating a structured monitoring window during the critical early service period, allowing trending data to be established before extending to routine annual intervals.

SDMyers‘ Recommended Intervals by Unit Type

SDMyers applies a risk-tiered approach to DGA retest intervals on new units, segmented by transformer size, class, and criticality:

• Small to medium units (those that would otherwise receive a standard 1-year recommendation): retest at 3 months to establish baseline dissolved gas data
• Large, high-value, or high-maintenance units — including furnace transformers and generator step-ups (GSUs): retest at 1 month, given the elevated financial and operational consequences of an undetected fault

The 3-month interval for standard units aligns with the IEEE C57.93-2019 first milestone. The 1-month interval for critical assets reflects the heightened cost of failure and the greater complexity of those units, which may be more susceptible to manufacturing or installation anomalies.

Why Early Detection Is Cost-Critical

Catching defects while under warranty

The primary strategic value of tight early-interval DGA is that manufacturing defects can be identified while the unit is still under the manufacturer’s warranty. Defects detected early — before a fault progresses to failure — are far less expensive to address than post-failure remediation, and warranty coverage may offset repair or replacement costs entirely.

Once a transformer fails in service, options are often limited. The defect may not be correctable without full unit replacement, and the window for warranty remedies may have closed.

Consequential damage risk

In-service failures carry costs well beyond the transformer itself. Consequential damages can include:

• Equipment damage to connected or adjacent infrastructure
• Facility damage from fire, arc flash, or released dielectric fluid
• Production loss and associated revenue impact
• Contractual non-performance and associated liability
• Loss of goodwill with customers or regulators

Early DGA detection — before a fault reaches failure — eliminates most of these downstream risk vectors.

Warranty Compliance and Testing Programs

It is worth noting that failure to maintain manufacturer-recommended testing intervals can invalidate warranty coverage. Asset managers who do not document adherence to prescribed DGA schedules may find warranty claims denied on the basis of non-compliance.

As a general rule, the “newly installed” period is considered to extend until the first routine samples are drawn under a standard interval schedule. Manufacturer and warranty requirements should be treated as the minimum acceptable standard in any transformer monitoring program — not a ceiling, but a floor.

Summary:

• Low dissolved gas levels on a new unit are a baseline, not a pass. Short retest intervals reflect the unit’s age and installation status, not a concern with current gas concentrations.
• Early-life failures follow a well-established reliability pattern. Manufacturing defects and installation anomalies are most likely to manifest under initial energization and load.
• IEEE standards support early, frequent sampling. C57.104-2019 recommends daily or weekly sampling post-startup; C57.93-2019 establishes specific milestones at 1 week, 2 weeks, and 1 month.
• SDMyers applies a risk-tiered approach: 3-month retests for small-to-medium units; 1-month retests for large, critical, or high-value assets such as furnace transformers and GSUs.
• Catching defects early is financially significant. Warranty coverage, reduced remediation costs, and avoidance of consequential damages all depend on identifying developing faults before in-service failure.
• Warranty compliance requires documented testing. Failure to follow manufacturer-recommended intervals may invalidate warranty claims. Treat those requirements as the minimum threshold in any DGA program.

References:

¹ IEEE Guide for the Interpretation of Gases Generated in Mineral Oil-Immersed Transformers, IEEE Standard C57.104-2019, 2019.

² Guide for Installation and Maintenance of Liquid-Immersed Power Transformers, IEEE Standard C57.93-2019, 2019.