Why PD Mapping Isn’t Optional—It’s the First Real Test for Epoxy Resin Potential Transformers in GIS

You’ve just finished GIS bay assembly. The SF6 gas is at spec. Busbars are torqued. CTs and VTs are mounted—and yes, those epoxy resin potential transformers look clean, compact, and perfectly aligned. But here’s what most project managers don’t see until it’s too late: that flawless exterior hides a dielectric structure with zero self-healing capability. No oil to migrate, no gas to recombine, no buffer against micro-defects formed during casting, handling, or thermal cycling. That’s why energizing without partial discharge (PD) mapping isn’t just risky—it’s like signing off on a pressure vessel without hydrotesting.

At Jinshida Electric Power Technology Co., Ltd., we’ve supported GIS commissioning across over 30 substation projects in the past three years—mostly 110 kV to 500 kV, with increasing adoption of solid-dielectric instrument transformers. What stands out isn’t how often PD is found—but how consistently it’s missed when mapping is skipped or done as a checkbox test. One recent case in Guangdong involved a batch of 220 kV epoxy resin potential transformers supplied by a Tier-1 OEM. Visual inspection passed. Megger results were >10 GΩ. Yet PD mapping revealed localized discharges (>80 pC) at the HV electrode–resin interface—traced later to minor shrinkage voids from uneven post-cure cooling. Had they been energized, failure likely would have occurred within 6–18 months under transient overvoltage stress.

The Physics Behind the Risk

Unlike oil-immersed units, where PD activity may be partially suppressed or even masked by oil’s high permittivity and ion mobility, epoxy resin is rigid, homogeneous, and unforgiving. Its dielectric strength is excellent—if the material is continuous and bonded. But real-world manufacturing introduces variables: moisture ingress during demoulding, incomplete degassing before casting, thermal mismatch between conductor and resin during cooldown, or mechanical stress during transport that initiates micro-delamination. These defects don’t show up on insulation resistance tests or power factor measurements—they only manifest under AC stress, precisely where PD mapping excels.

And GIS environments compound the risk. The confined metal enclosure creates a resonant cavity. Any PD pulse couples efficiently into the GIS bus—potentially triggering false alarms in monitoring systems, accelerating aging of adjacent components, or—in worst cases—initiating flashover cascades if multiple weak points align spatially and temporally.

What “Mapping” Actually Means (and Why Generic PD Detection Falls Short)

“PD detection” isn’t enough. You need mapping: time-synchronized, phase-resolved, multi-sensor acquisition that correlates pulse amplitude, polarity, phase angle, and time-of-flight across at least three coupling points (typically on the transformer flange, GIS enclosure, and grounding strap). Only then can you distinguish surface discharge on the housing from internal voids—or separate noise from genuine insulation stress.

We’ve seen teams use handheld ultrasonic detectors and call it “done.” That catches loud, surface-level arcing—but misses the 20–50 pC discharges inside the resin body that grow exponentially under repeated voltage cycles. True mapping requires calibrated HFCTs or UHF sensors, synchronized to the supply waveform, with software capable of separating pulse families using statistical clustering. It’s not about finding *a* signal—it’s about answering: Where is it? Is it repeatable? Does it scale with voltage? Does it change after thermal soak?

When to Map—and When Not to Trust the Factory Report

Factory PD tests are valuable—but rarely replicate field conditions. Most manufacturers test at ambient temperature, with ideal grounding, and often at reduced voltage (e.g., 1.2 × U_m/√3 instead of full system voltage). Field installation introduces new stresses: mechanical compression from mounting bolts, thermal gradients from nearby bus heating, and grounding impedance variations across the GIS bay. A unit passing factory PD at 145 kV may emit measurable pulses at 168 kV once bolted to a live 220 kV GIS section.

So map after final installation, but before SF6 filling—or at minimum, before final pressurization. That way, if anomalies appear, you can isolate the unit without GIS bay depressurization. And always perform a second mapping after 24 hours of thermal stabilization. We’ve observed PD magnitude shifting by 30–40% between cold start and thermal equilibrium—especially near epoxy–metal interfaces.

Real-World Trade-Offs You’ll Face on Site

Yes, PD mapping takes time—typically 2–4 hours per transformer, depending on sensor setup and ambient noise. Yes, it requires trained personnel and calibrated gear. But compare that to the cost of a GIS bay shutdown mid-commissioning: crane mobilization, SF6 recovery/recharge, engineering review, and delay penalties. One client in Anhui avoided a 17-day delay by catching interfacial PD in a 500 kV VT—just two days before grid connection was scheduled.

Also worth noting: some project teams try to skip mapping by relying on extended AC withstand tests. That’s dangerous logic. Withstand tests verify gross integrity—not progressive degradation. A unit may hold 1.5×U_m for 60 seconds and still harbor sub-threshold voids that will nucleate PD under normal operating voltage over weeks.

Where Mobile Energy Storage Fits In

Commissioning windows are tight—and power availability isn’t always guaranteed onsite. That’s where flexible, on-demand backup matters. For remote GIS sites undergoing PD mapping and thermal soak testing, having an independent, trailer-mounted source eliminates dependency on temporary diesel gensets or unstable grid taps. Our 100kW/215kWh Mobile Trailer Energy Storage System has supported six GIS commissioning campaigns this year—providing stable, low-noise AC power for PD acquisition systems, control cabinets, and cooling fans during extended test sequences. Its LiFePO4 battery bank maintains voltage stability within ±1.2% under variable load, critical for sensitive measurement gear.

Final Call: Treat PD Mapping Like Commissioning Fuel

You wouldn’t pour fuel into a turbine without verifying lubrication integrity. Don’t energize epoxy resin potential transformers without validating their dielectric continuity under operational stress. PD mapping isn’t insurance—it’s diagnostics. It’s the only method that reveals what visual inspection, IR thermography, and basic electrical tests cannot: the true condition of the solid insulation where it matters most.

If your next GIS project includes epoxy resin potential transformers, build PD mapping into the critical path—not as an afterthought, but as the first functional verification step. Because reliability isn’t assumed. It’s measured.