An adaptive, standalone, and scalable solar combined heat and power system with AI-enhanced universal energy management

To address the operational flexibility challenges of small-scale distributed solar combined heat and power plants under unstable solar input and fluctuating user demand, this study proposes an artificial intelligence-driven dynamic energy management approach integrated with knowledge graph technology for autonomous peak shaving. Conventional control strategies suffer from inadequate information exchange between upstream generation and downstream consumption modules, limiting their ability to handle system complexity. To solve this, first a simplified heat hub was developed with manual intervention to establish basic system information connectivity, which achieved desired flexibility but was constrained by low responsiveness and accuracy. Leveraging operational data from this preliminary system, a knowledge graph was constructed which effectively captured inherent information flow patterns and revealed key system behavioral insights. Guided by these insights, an enhanced heat hub was designed and optimized, significantly improving information transmission efficiency and peak-shaving accuracy. The results demonstrate that the knowledge graph-enabled approach overcomes critical energy management challenges in distributed solar systems by bridging the information gap between upstream and downstream subsystems. This study not only verifies the technical feasibility of integrating knowledge graph with solar combined heat and power systems for improved operational performance but also showcases the promising application prospects of artificial intelligence technologies in advancing the flexibility and reliability of distributed energy systems, providing valuable guidance for the development of efficient and stable solar energy utilization solutions.