| Abstract |
The digital wave has given birth to the arrival of smart power and smart cities, a large number of sensor systems deployed for traditional smart power sites have high fragmentation, large and scattered data collection, difficult system integration, and unsatisfactory application effects. A key technical solution for engineering brain based on artificial intelligence and joint learning in intelligent power plant power has been proposed. At the same time, under the premise of data not leaving islands, achieve a learning mode from the model side to multiple data sides, achieving the effect of data being available but not available. The experimental results indicate that: Although federated data governance technology increases model training time due to encrypted data transmission, with the increase of data silos, it can gradually approach data aggregation mode in testing accuracy, greatly improving the problem of data silos. The key technology solutions for the power of engineering brain based on artificial intelligence and federated learning have been demonstrated in practice. The use of digital twin technology to improve automation collection capabilities, the use of artificial intelligence technology to upgrade automation collection devices, and the use of federated learning technology to enhance the intelligence level of engineering power can effectively improve the overall level of intelligent power. In terms of engineering safety supervision, quality monitoring cost control and other aspects can provide reference for similar projects. © 2024 The Authors. |
