A Correlation Feature Selection Method Adapting to Various Control Operation Scenarios of Power Grid

Abstract

The invention relates to a power system feature selection technology, in particular to a correlation feature selection method adapted to various control and operation scenarios of a power grid. First, from the perspective of the historical actual data of the power system, standardization and preprocessing are carried out to make it a direct Include the time series data of the analysis calculation and the classifier, and then study which feature variables may indicate the occurrence of power system scenarios, or which feature variables change with the appearance of typical scenarios, use this information to construct feature candidate sets, and then use gray The correlation degree analysis algorithm quantitatively calculates the correlation degree values ​​between the features and the target typical scenes, and sorts them. Then, a reasonable threshold is set to obtain the final feature subset, and the classifier accuracy is used to evaluate the feature subset. The method is beneficial for dispatchers to have better instructions when performing related operations, while reducing the data processing time of the power system and improving the efficiency of data analysis.

Description

technical field [0001] The invention belongs to the technical field of power system feature selection, and in particular relates to a correlation feature selection method adapted to various control and operation scenarios of a power grid. Background technique [0002] In recent years, with the integration and close interconnection of large power grids, the integration of sending and receiving terminals has become prominent. The access of high-penetration renewable energy, the security, stability and economic operation of the power system have also ushered in great challenges. A high proportion of power electronic equipment is put into the power grid, which makes the dynamic stability of the system more complicated. When these new energy sources and random loads are connected to the power grid on a large scale, their randomness and intermittency will greatly increase the difficulty of power grid planning and dispatching, and may also cause serious grid failures, heavy loads ...

AI Technical Summary

Problems solved by technology

If the data dimension significantly exceeds the number of training samples, it is likely to lead to overfitting and other phenomena.

And the current dispatching is passively waiting for the arrival of power grid accidents, and there is still room for improvement in the early warning and prevention of predictable risks such as natural disasters; there is no effective monitoring means for unit power fluctuations, relying on power plant reports; there is no monitoring means for power grid sudden public opinion events, relying on Lower-level scheduling reports; technical support system information is numerous and scattered, lack of centralized monitoring functions, etc.

[0003] The existing dispatching mode lacks characteristic analysis of typical scenarios of the power system, and cannot find out the relationship between typical scenarios and their associated features. The actual analysis of power system data is often not well targeted. The type judgment has the disadvantages of low accuracy, high time consumption, and large amount of data to be processed.

If the operating scenario of the power system cannot be predicted and judged well, dispatchers will not be able to make accurate operation decisions, which will cause great harm to the safe and stable operation of the power system

Images