With the introduction of the ubiquitous power Internet of Things, power system analysis based on data mining will play an increasingly important role in the construction of the energy internet. The boundary problem of critical transient stability is one of the core issues in evaluating the transient stability of power systems. At present, in the transient stability assessment, it is often impossible to give an accurate judgment on the critical state of the grid, so that there is a fuzzy area in the safety assessment. From the perspective of data analysis, the related research on the feature extraction of critical transient stability boundary of power system was carried out. Based on the transient stability single machine model and energy function, the boundary phenomenon of critical transient stability of power system was characterized. Then, based on the BPA automatic simulation program, the transient stable data sample generation problem was solved. Finally, the selection of the input characteristics of the "source-network" and the correlation analysis of the input features were completed, and the verification was completed by the IEEE-39 node system.
