Abstract:

Objectives With the growing integration of renewable energy, the complexity of new power system has significantly increased. The power industry requires the integration of large-scale multi-source data, more complex analysis and decision-making processes, and more intelligent approaches to enhance system flexibility and adaptability. Large models represented by large language models have attracted significant attention due to their robust natural language processing capabilities and reasoning abilities across various complex tasks. Based on this, the study reviews implementation technologies for applying large language models in the power industry and summarizes relevant achievements to inform future applications. Methods Firstly, key technologies for implementing large language models are introduced, including prompt engineering, retrieval-augmented generation, model fine-tuning, and the development of intelligent agents. These technologies enhance the accuracy and practicality of large language models in real-world applications and broaden their range of use cases. Secondly, the study outlines the research progress of large language models in areas such as power knowledge services, assisted decision-making, equipment fault diagnosis, and power system prediction. Lastly, the challenges in applying large language models in the power industry are analyzed. Conclusions The application of large models in the power industry is currently focused on the cenarios based on large language models, which are relatively mature. In contrast, applications involving multimodal models, time-series models, and large-small model collaboration remain in the exploratory and rapidly evolving stage.

Key words: power industry, new power system, artificial intelligence (AI), large language model, knowledge services, assisted decision-making, fault diagnosis, prediction