In response to the growing demands for digitalization and intelligent capabilities in new power systems

this paper reviews the evolution of digital-intelligent development in power equipment. The construction goals of power grid and digital-intelligent technology characteristic of power equipment at different developmental stages are analyzed

and the corresponding technical requirements and current status are summarized. Based on the technical characteristics of large language model (LLM) agent at the current stage

a five-layer application framework for the intelligent operation and maintenance of power equipment is proposed. Potential applications of the proposed approach in natural language interaction systems

digital twin systems

and intelligent robotics are further explored. Finally

possible economic

reliability

feasibility

and efficiency challenges in applying LLM agent to power equipment operation and maintenance are identified

providing a reference for advancing digital-intelligent development in power equipment.