Machine‐Learning‐Guided Structural Design of Interlayer‐Confined PEI‐Based Composites for Superior Dielectric Energy Storage

Structural design plays a pivotal role in enhancing the performance of composite materials for electrostatic energy storage, particularly in pulsed power systems operating under extreme conditions. Herein, guided by a sequential multi‐target machine learning (ML) framework, a series of polymer‐based nanocomposites with an interlayer‐confined architecture are fabricated, featuring two‐dimensional in‐plane network formed by interlaced Na 0.5 Bi 0.5 TiO 3 (NBT) nanofibers embedded into the polyetherimide (PEI) matrix. This strategy leverages NBT nanofibers to tailor the spatial electric field distribution via their high dielectric constant, while the fillers simultaneously suppress chain mobility and thereby reduce the matrix diffusivity to reinforce mechanical strength, collectively enhancing breakdown strength and polarization, and thus contributing to excellent electrostatic energy storage capacity. The resulting polymer composite showcases impressive energy storage performance (energy storage density U e = 11.08 J/cm 3 , energy storage efficiency η = 95.46%) and excellent high‐temperature performance ( U e = 5.32 J/cm 3 with η > 90.00%, 150°C), which are far superior to those of most advanced polymer‐based dielectrics reported recently. This work establishes an ML‐guided structural design strategy for next‐generation multilayer polymer dielectrics toward superior energy storage.