Dissolved gas analysis has long been the gold standard for transformer diagnostics. But even the best laboratory results are only as good as the sample behind them. Even worse, time delays often mean critical changes go unnoticed.

When KAMO Power detected acetylene in an 84 MVA transformer for the first time in its operational history, the utility faced a dilemma familiar to grid operators nationwide: How aggressively should the team respond? The fault gas had jumped from 0 to 1.2 ppm in a year, then climbed to 19 ppm over 6 months despite intensive manual sampling.

Each sample required pulling two technicians from regular duties, sending oil to external labs and waiting days for results. Costs mounted quickly, but without real-time data, maintenance managers lacked confidence in their next move. Should they drain the transformer for inspection? Plan for an outage? Keep sampling and hope the trend reverses?

Through this experience, the Oklahoma-based electric cooperative recognized that traditional dissolved gas analysis protocols, while proven over decades, increasingly struggle to meet the operational and financial demands of modern grid management. As transformer populations age, load demands intensify and O&M budgets tighten, utilities need monitoring approaches that deliver actionable insights without escalating costs or exposing personnel to unnecessary risk.

Transformer reliability = grid stability

For utilities and electric cooperatives, transformer reliability is synonymous with grid stability. Forming the critical junction between transmission and distribution, these high-value assets convert power and maintain voltage balance across increasingly complex networks. Yet as demand climbs and many transformers near or exceed their designed service lives, utilities face growing challenges in monitoring health, preventing faults and optimizing maintenance budgets.

When electrical or thermal faults occur within the oil, they generate gases — hydrogen, methane, acetylene and others — that can point to a larger problem. If concerning trends emerge, sampling frequency increases, sometimes dramatically. Yet for most utilities, traditional DGA still means manual sampling and laboratory testing, sometimes performed only once a year.

A low sampling frequency limits visibility into developing faults and inherently introduces delays between sample collection and results, potentially obscuring rapid changes in transformer condition. The process also incurs operational costs — sending technicians to remote sites, managing sample logistics and paying lab fees — all while exposing personnel to field risks.

The hidden costs of manual sampling

KAMO’s 2012 Waukesha transformer had logged clean DGA results since its installation. Annual samples from 2016 through 2021 showed no acetylene, a key indicator of arcing in oil. But the July 2022 sample revealed 1.2 ppm of acetylene, where none had been detected before.

Following lab recommendations, KAMO maintenance staff shifted to manual sampling every two weeks. Within a month, two additional samples showed acetylene climbing to 3.7 ppm. The electric cooperative consulted the manufacturer and ran comprehensive field tests: power factor, transformer turns ratio, winding resistance, insulation resistance and oil power factor evaluations. Staff exercised all DETC no-load taps, tested ratios and performed resistance testing on reactor core ground and main core grounds.

The tests returned no abnormalities, yet acetylene continued rising to 4.3 ppm.

The management of this sampling protocol remained burdensome and expensive. More troubling, a clear picture of the situation remained elusive to maintenance managers. Without confidence in the data story, they believed continued manual sampling was the only option.

KAMO’s service territory spans parts of four states, requiring two technicians to travel to substation sites for each sample. Those staff hours represented direct costs and opportunity costs from work left undone. Sample collection, shipping, lab analysis and results interpretation added administrative overhead. Perhaps most concerning, repeated site visits increased personnel exposure to potential safety hazards.

Consequently, KAMO sought a way to make its decisions with confidence without sending people into the field every few weeks.

The real-time revolution

The situation crystallized a broader industry trend: While foundational to transformer asset management, manual sampling protocols provide only periodic glimpses into asset health. Between sample measurements, operators essentially operate blindly, unable to correlate fault gas trends with loading conditions, temperature variations, power quality issues or other operational factors that might explain changes in transformer behavior.

In spring 2024, KAMO’s maintenance managers decided to deploy a mobile online multi-gas DGA monitor to collect real-time data during the transformer’s summer high-load operational periods. The Vaisala Optimus OPT100 monitor installation took just a few hours and began providing hourly dissolved gas measurements immediately.

The difference proved revelatory. Rather than waiting weeks between snapshots, operators could observe how acetylene levels responded to loading and thermal conditions in near real time. Patterns that remained hidden in periodic sampling data emerged when viewed continuously. The rate of change, visualized through trending slope data, allowed maintenance managers to evaluate fault severity with new precision.

The Vaisala sensor provided the maintenance team with the confidence to cut back on O&M manual sampling costs while improving and increasing the amount of transformer gassing data.

Summer 2024 data told a story that manual sampling had only hinted at: Despite higher loads during peak season, acetylene levels continued to decline, stabilizing around 6.5 ppm by season’s end. The transformer was not experiencing active, progressive faulting. Whatever had caused the initial acetylene generation had resolved, and the unit could safely remain in service.

Equally important, the OPT100’s real-time trending enabled KAMO’s asset managers to correlate gas behavior with temperature and load patterns, providing context that static lab results could not.

That insight carried enormous value. KAMO avoided planning for backup unit deployment, scheduling an outage during critical summer months and incurring inspection costs that could have run into six figures when factoring in labor, equipment, lost capacity and potential repair expenses if issues had been found.

Building a hybrid monitoring strategy

Armed with confidence from continuous monitoring data, KAMO strategically pivoted. The mobile multi-gas monitor confirmed the transformer’s condition had stabilized around 6.5 ppm acetylene. Rather than maintain that level of monitoring indefinitely, the co-op installed a Vaisala MHT410 moisture, hydrogen and temperature transmitter for ongoing surveillance.

This approach illustrates a cost-effective transformer asset management alternative. Different situations demand different monitoring intensities. When fault gases appear or trends raise concerns, multi-gas continuous monitoring provides the necessary detailed trending data for confident decision-making. Once situations stabilize, single-gas monitoring focused on key fault indicators offers early warning at a lower cost.

Hydrogen serves as a “smoke detector” gas because it typically appears early when faults begin developing. By monitoring hydrogen continuously, KAMO would receive advance warning if the previous acetylene issue resurged or if new fault conditions emerged. That assurance allowed maintenance managers to reduce manual lab sampling from monthly to twice yearly, seriously cutting both direct costs and staff exposure to risks in the field.

The operational math proved compelling. KAMO estimates it saved tens of thousands of dollars in O&M and testing costs on this single fault case, a figure that accounts for avoided lab fees, reduced technician travel time and the administrative overhead of managing intensive sampling protocols. The co-op also avoided potentially unnecessary repair expenses and capacity loss from an outage that continuous monitoring revealed was unwarranted.

Project stakeholders noted the savings could have multiplied if the mobile monitor had been deployed sooner in the investigation cycle. That lesson has shaped KAMO’s thinking about future transformer health management strategies.

Technical success factors

Several technical factors enabled KAMO’s successful monitoring strategy shift. The online monitors require no consumables and minimal scheduled maintenance, contrasting sharply with some competing technologies that demand regular calibration gas or sensor replacements. Plus, installation speed matters when assessing developing fault conditions. The two-hour deployment timeframe allowed KAMO to quickly collect continuous data without extensive outages or complex integration work.

Remote data accessibility also simplified the monitoring program. Rather than requiring site visits to download data or check readings, maintenance managers could access measurement information from office locations, reviewing trends and setting alerts without field trips. As utility workforces face skilled labor shortages and competing demands on technical staff time, remote monitoring becomes increasingly valuable.

Hourly measurements throughout the summer load period created a detailed picture of how the transformer responded to varying conditions. The monitoring systems’ granularity enabled correlation analysis between fault gas levels, loading patterns and thermal conditions that would be impossible with periodic manual sampling.

For utilities evaluating monitoring technology options, these operational characteristics deserve weight alongside accuracy specifications and measurement ranges. A highly precise monitor that requires frequent maintenance visits, consumes expensive calibration gases or demands complex data retrieval may deliver less practical value than a lower-maintenance system providing adequate precision for decision-making.

Scaling monitoring across the fleet

KAMO’s experience with the Waukesha transformer prompted broader strategic thinking about its fleetwide monitoring approaches. The electric co-op is now investigating how to integrate fixed DGA monitors and portable units into standard maintenance and asset health operations across its transformer population.

The KAMO case’s evolution reflects a maturing understanding of monitoring economics. Not every transformer justifies permanent multi-gas monitoring, particularly newer units with clean DGA histories and lower criticality to grid operations. But having mobile multi-gas monitors available for deployment when concerning trends emerge provides flexibility that pure manual sampling cannot match.

Capital for monitoring infrastructure is limited, but so is tolerance for unexpected transformer failures. By mixing monitoring technologies and deployment models, operators can enhance visibility into asset health without budget-breaking capital expenditures.

Data-driven resilience

The pressures driving KAMO’s monitoring strategy evolution affect utilities and electric co-ops nationwide, regardless of size or business model.

The transformer population is aging across the industry. Many units installed during major grid expansion periods in the 1960s through 1980s are reaching or exceeding design life. These aging assets face increasing failure risk just as load growth from data centers and electrification stresses grid infrastructure, creating heightened urgency around asset health visibility.

And with regulatory and stakeholder expectations around grid reliability rising, customers and regulators increasingly expect utilities to prevent failures rather than simply respond to them.

With the electric utility industry standing at an inflection point in transformer asset management, online monitoring technologies have matured to offer practical alternatives that can enhance visibility while controlling costs.

KAMO Electric Cooperative’s experience demonstrates that the path forward is not about abandoning proven manual sampling protocols but rather about deploying multiple monitoring tools strategically. Mobile multi-gas monitors for intensive investigation periods, permanent single-gas systems for ongoing surveillance and periodic manual sampling for baseline monitoring can work together in a hybrid approach that optimizes resource allocation.

For an industry facing mounting pressures on multiple fronts, KAMO’s journey from reactive manual sampling to proactive hybrid monitoring offers a roadmap. And the destination is a flexible, scalable approach that deploys the right monitoring intensity in the right situations. In that balance lies the future of transformer asset health management: better data, lower costs, improved safety and the confidence to make critical decisions that keep the grid running reliably.