In the context of big data, machine learning plays an important role in many fields. With the increasing scale of power system and capacity of power grid, it becomes more and more complicated to accurately evaluate the transient stability of power system. In this paper, a power system transient stability assessment method based on XGBOOST algorithm is proposed. The XGBOOST algorithm is introduced to train the decision tree model and evaluate the transient stability of power system by converting the simulated power system operating data into the characteristic variables of power system. The results show that the training model of the algorithm can solve this kind of problem accurately and quickly.

Complex geographical conditions and bad weather have caused great difficults to inspect transmission line of power grids. This work aims to develop a micro multi-rotor Unmanned Aerial Vehicle (UAV) system for inspecting power grid. The proposed system integrates with an intelligent robot and mobile communication networks. The high-resolution aerial images of the inspection line can be obtained to improve the operational efficiency and safety. The time period from the front-end signal acquisition to terminal decoding and playback is 1.5s for condition-based maintenance. The system supports TDD-LTE, FDD-LTE, WCDMA, CDMA2000, and other network standards. It also supports the video input signals with the resolution of standard 576i, HD 720p, FHD 1080i, and FHD 1080p. It is compatible with a variety of transmission formats and codes, help to achieve timely, comprehensive and efficient high-resolution aerial image acquisition of patrol lines, and improve operational efficiency and safety, real-time monitoring and condition-based maintenance of power transmission lines.