Heat Is the Silent Killer of High Frequency Transformers
High frequency transformers (HFTs) are the core components of power electronic transformers (PETs) and solid state transformers (SSTs). Unlike conventional 50 Hz power transformers, HFTs operate at frequencies ranging from several kilohertz to hundreds of kilohertz.
Higher frequency brings smaller size – but it also brings a cost.
The move to high frequencies leads to a sharp increase in both core losses and winding copper losses. Data shows that HFT losses can account for as much as 35% of total SST losses. If heat cannot be extracted in time, it can cause core demagnetisation and insulation degradation. Conventional epoxy potting compounds have a thermal conductivity of less than 0.5 W/(m·K), making it difficult to meet the increasing power density demands of HFTs.
Epoxy resin is the mainstream insulating material for HFTs – good insulation performance, easy to process, and low cost. But its weakness is equally critical: poor thermal conductivity. Under the combined effects of high frequency voltage and elevated temperature, heat accumulates and creates a “electro thermal coupling effect” that accelerates insulation degradation, compromising both transformer life and reliability.
One Interface Design That Solves Two Problems at Once
In June 2026, a research team from Nanjing University of Aeronautics and Astronautics, State Grid Shandong Electric Power Company, Peking University, and other institutions published a study in the journal Advanced Science.
The team’s core idea was: not to change the main composition of the epoxy resin, but to modify the thermally conductive filler being added.
They chose boron nitride (BN) as the thermally conductive filler – BN has good thermal conductivity and excellent electrical insulation. But BN has a natural drawback: its surface is chemically inert, making it difficult to disperse uniformly in epoxy and resulting in poor compatibility.
The team designed a two step surface modification strategy:
Step 1: KH550 coupling agent grafting. KH550 is a silane coupling agent – one end “grabs” onto the BN surface, the other “reaches out” to the epoxy resin, building a “bridge” between the BN and the epoxy.
Step 2: G POSS grafting. G POSS is a cage structured nanoparticle. After grafting onto the KH550 modified BN surface, it further optimises the interface structure.
The result of the two step modification is that the BN@G POSS nanofiller can both disperse uniformly in the epoxy and form a strong interfacial bond with the epoxy matrix – thermal conduction pathways are kept clear, while interface defects are effectively suppressed.
Three Numbers That Show the Scale of the Breakthrough
Thermal conductivity up 492.5%. The optimised formulation contains 7 wt.% G POSS and 20 vol% BN@G POSS nanofiller. Compared with conventional BN/EP composites, thermal conductivity increased by 492.5%.
Partial discharge inception voltage up 266.5 V. Under a 100 kHz electric field, the material’s resistance to electro thermal degradation was significantly enhanced. Partial discharge inception voltage increased by 266.5 V, while total discharge amplitude dropped by 256.6 V.
Viscosity down to 3,493 mPa·s. High filler loading typically causes a sharp rise in viscosity and poor processability. The new formulation has lower viscosity than conventional BN/EP composites – meaning it can be produced in batches using existing casting equipment and processes, without redesigning the production line.
Why This Breakthrough Matters
The insulation materials for HFTs have long faced an “impossible triangle”: high thermal conductivity, high insulation performance, and good processability – improving one often comes at the expense of another. Adding thermally conductive filler may reduce insulation performance; adding too much causes viscosity to spike, making potting impossible.
The value of this research lies in pushing all three goals forward at the same time through interface design. The KH550/G POSS two step modification strategy essentially gives the BN nanofiller a “multi functional coat” – helping it disperse uniformly in the epoxy (thermal conductivity), suppressing interface defects (insulation), while maintaining acceptable viscosity (processability).
At the critical industry inflection point where solid state transformers are moving from prototypes to mass production, breakthroughs in epoxy insulating materials are one of the key technical bottlenecks. As the power density of HFTs continues to climb, heat dissipation and insulation issues will only become more pressing. This research offers a scalable solution – not by replacing the material system entirely, but by optimising the interface within the existing system.
Conventional epoxy has a thermal conductivity of less than 0.5 W/(m·K). Through interface engineering, this study increased thermal conductivity by nearly five times, while raising partial discharge inception voltage by 266 V. The twin challenges of “heat dissipation” and “insulation” for HFTs are being overcome – simultaneously.
Post time: Aug-04-2026