Yi Li

The advancement of power systems imposes increasingly stringent performance requirements on insulating materials, making the development of high‐performance insulating materials a critical challenge in the field of electrical engineering. Traditional material development relies heavily on empirical trial‐and‐error approaches, which are time‐consuming and cost‐intensive and thus cannot keep pace with the rapid iteration of modern power systems. Machine learning (ML) offers a powerful data‐driven paradigm to overcome these limitations. By extracting complex structure–property relationships from experimental and computational datasets, ML enables efficient performance prediction, formulation optimisation and inverse molecular design. This review first systematically presents the fundamental workflow of ML‐assisted materials research and summarises representative ML algorithms and their applicability to specific material science problems. Subsequently, it reviews key advances in ML applications for gaseous, liquid and solid insulating materials, highlighting the value of ML in supporting insulating materials research. Finally, it discusses major challenges in ML‐assisted insulating materials research and proposes potential solutions. By outlining the workflow of ML‐assisted materials research, summarising landmark advancements and exploring critical challenges in ML‐assisted insulating material research, this review aims to provide guidance for researchers seeking to leverage ML to elucidate structure–property relationships and accelerate the discovery of novel high‐performance insulating materials.

Precise control over the stacking of two-dimensional (2D) metal-organic framework (MOF) nanosheets enables the fabrication of MOF films or membranes with adjustable interlayer spacing and/or pore size, offering great potential for numerous applications, yet remains highly challenging. We herein propose a postsynthetic modification strategy to functionalize solution-processable NUS-8 nanosheets with photo-responsive monodentate carboxylic acids undergoing a “trans-cis” conformational change upon irradiation or with non-responsive counterparts. This yields azobenzene-functionalized NUS-8 (designated as NUS-8-Azobenzene) and/or tetra-phenylethylene-functionalized NUS-8 (NUS-8-TPE). The 2D nature and excellent solution processability of resultant NUS-8-Azobenzene and NUS-8-TPE facilitate the fabrication of large-area, highly homogeneous films/membranes with (00l) preferential orientation. Responsive interlayer spacing was observed in the NUS-8-Azobenzene stacked membranes, confirmed by multiple characterizations. Subtle variations in interlayer spacing of NUS-8-Azobenzene stacked membranes under irradiation enable a certain control over the ubiquitous trade-off between membrane selectivity and permeability. This study presents a strategy for fabricating 2D MOF membranes with scalable production and new functionalities, even for rigid and non-responsive MOFs, holding significant potential for practical applications in separation.

The liquid–solid triboelectric nanogenerator (LS‐TENG) can efficiently convert low‐frequency, unstable ocean wave, droplet energy into electricity. Additionally, it holds the potential to function as a self‐powered sensor for liquid physicochemical composition analysis. Herein, we summarize recent advances on the principles, method, and application of LS‐TENG. Firstly, we introduced the fundamental principles of liquid–solid contact electrification and the electrical double layer. Second, we highlighted the strategies for enhancing the output of LS‐TENG in terms of material selection, modification, and structural optimization. These strategies encompass the regulation of solid triboelectric materials, optimization of liquid polarity and ion concentration, and innovative designs. Furthermore, we showcased the latest applications of LS‐TENG in areas such as blue energy harvesting, contact electro‐catalysis, smart interconnection, and physical/chemical/biological signal sensing. Finally, the challenges faced by LS‐TENG, including understanding its mechanisms, material selection, structural innovation, and the expansion of application scenarios are proposed. This work anticipates future trends‐including the development of high‐efficiency energy management, hybrid energy systems with higher energy density, deep interdisciplinary fusion with artificial intelligence, microfluidics, biomedicine, and the establishment of unified international standards and testing protocols, thereby providing targeted and actionable insights for advancing the design optimization, functional expansion, and industrial commercialization of high‐performance LS‐TENG.

The advancement of the Industrial Internet of Things and digital‐intelligent machinery presents considerable demands in detecting impurities in industrial equipment lubricants or oils. Herein, a Tubular oil–solid triboelectric nanogenerator (TOS‐TENG) is designed as a self‐powered sensor for impurity detection in insulating oil flow. Specifically, an oleophobic triboelectric sensing layer with superior electron trapping capability is created based on Nano silica and 1H,1H,2H,2H‐perfluorodecyltriethoxysilane. Additionally, a periodic start‐stop oil flow control method that facilitates the transition between flowing and stationary insulating oil is proposed, thereby generating periodic output signals from the TOS‐TENG. Furthermore, a correlation between the output characteristics of the TOS‐TENG and impurity properties (type, content) is established through a deep learning algorithm based on convolutional neural networks, enabling the identification of common impurities such as moisture, metal, and fiber particles in oil. As a proof of concept, an industrial‐grade TOS‐TENG‐based device for in situ online monitoring of impurities in oil‐immersed transformers is designed. This work offers novel insights into the compositional analysis of continuous fluids and provides a novel strategy for oil condition sensing in smart industrial equipment.

Perfluoroisobutyronitrile (C ₄ F ₇ N) is recognised as the next‐generation dielectric medium for eco‐friendly gas‐insulated equipment (GIE). However, there is gas metal incompatibility in strong electromagnetic environments and environmentally friendly insulating gas atmospheres that endangers the operational safety and service life of GIE. Here, a solvent‐induced grain boundary reconstruction method was proposed to prepare corrosion‐resistant copper rapidly. Accordingly, the partial discharge‐induced gas–metal interaction was simulated and the corrosion resistance properties of copper were evaluated. Additionally, the gas–solid interaction mechanism was investigated and the improvement mechanism of the proposed method on the stability of the gas–solid interface was discussed. Related results provide a reference for the development and application reliability of C ₄ F ₇ N‐based eco‐friendly GIE.

Perfluoroisobutyronitrile (C₄F₇N), an eco-friendly insulating gas, has been extensively utilized in diverse gas-insulated equipment (GIE) to replace the most potent greenhouse gas SF₆. Nonetheless, given the low toxicity of C₄F₇N, the potential leakage risk at a high-pressure operational environment cannot be overlooked, making the development of highly sensitive gas sensors for leak detection essential. Herein, Co₃O₄ was selected as a specific gas sensing material for C₄F₇N detection based on the density functional theory screening. The Co₃O₄-based gas sensor was prepared, and its response performance was systematically evaluated. The interaction mechanism between Co₃O₄ and C₄F₇N was also examined by the oxygen vacancy theory. The sensor demonstrates a low detection limit of 0.15 ppm at its optimal operating temperature of 300 °C, exhibiting exceptional selectivity and repeatability, with a stability response time of 55.5 s and recovery time of 1478 s. The sensor's application potential under actual working settings was validated using a simulated leakage test. Relevant results provide guidance for the leakage monitoring of the C₄F₇N-based GIE.

Hybrid energy harvesters offer promising solutions for powering distributed sensors. However, achieving optimal synergy among multiple energy sources to attain superior performance remains challenging. Herein, we proposed an ultrahigh performance electric-field and vibration hybrid energy harvester, leveraging the induced charge excitation strategy (ICE-EVH). Induced charges generated by the electric-field energy harvester are pumped into a contact-separation triboelectric nanogenerator (TENG) to provide charge excitation. Subsequently, the mechanical motion of TENG-electrodes boosts the electric potential of induced charges and the output energy. This hybrid paradigm enables effective synergy for harvesting electric-field and vibration energy with highly-efficient hybrid power management and ultrahigh energy density. The peak and average power densities reach 5.15 MW m⁻² and 1.02 W m⁻², respectively, significantly surpassing the combined output of individual harvesters and demonstrating a “1 + 1 > 2” hybrid performance. Finally, an all-in-one ICE-EVH prototype was developed, which successfully powered a wireless camera. This work provides new insights for designing high-performance hybrid nanogenerators with broad application potential.