To process the difficulty of equipment operation and maintenance supervision caused by the large number and wide distribution of smart grid power equipment, an innovative architecture integrated IBA is proposed, which combines big data and AI technology in different dimensions of the architecture to achieve hierarchical intelligence as needed. We provided the design process and key modules of the power grid system based on the IBA architecture: firstly, we introduce human-machine data under the architecture and propose the concept of hierarchical cognition. Then, real-time data collection is performed through the Internet of Things, and further mining and analysis of the data is achieved using cloud platforms. By Spark clustering based on k-means to partition state data, we establish corresponding Bayesian causal network models, and achieve data-driven modeling of smart grid equipment monitoring and operation. Finally, an application case of artificial intelligence is used to demonstrate the feasibility and effectiveness of the system. The experimental results demonstrate that the IBA fusion technology architecture has expanded the application scope of the IoTs in smart grids, and improved the intelligent level of smart grid construction and management.