Construction and Application of a Power Industry Knowledge Graph

In the era of big data for the power sector, AI, cloud computing, and IoT technologies are widely applied, and knowledge‑graph technology emerges as a new driver for industry innovation.

The presentation is organized into four parts: (1) building a power‑industry knowledge system, (2) constructing the power knowledge graph, (3) application scenarios, and (4) practical deployment and insights.

Knowledge modeling follows a top‑down approach, defining data schemas from high‑level concepts down to detailed entities, and combines manual expert input with automated computer extraction to ensure quality and authority.

The resulting knowledge graph contains over 200,000 entities and 300,000 relationships, covering five major dimensions (discipline‑based and four scenario‑based subsystems) and supports functions such as personal achievement aggregation, professional learning, job training, generic knowledge services, and power‑production decision support.

Deployments include the public “China Power Encyclopedia” portal and internal systems for State Grid utilities, demonstrating the value of high‑quality domain data, scenario‑driven graph construction, and close collaboration between publishing and technical teams.