Transient stability evaluation method for Bayesian optimization LightGBM

Abstract

The invention discloses a transient stability evaluation method for Bayesian optimization LightGBM. The method comprises the following steps: obtaining a transient stability data set of a power system; training the Bayesian optimized LightGBM by using the data in the transient stability data set to obtain an optimal parameter of the LightGBM, and then obtaining the trained LightGBM; obtaining thedata used for evaluating the transient stability of the power system on line after the power system has a fault, preprocessing the obtained data, and inputting the preprocessed data into the trained LightGBM to obtain a power system transient stability evaluation result after the power system has the fault. According to the method, the transient stability state under multiple complex uncertain factor 'combinatorial number explosion' can be quickly and accurately evaluated, and the online evaluation of the transient stability of the power system is facilitated.

Description

technical field [0001] The invention relates to a method for evaluating transient stability of a power system, in particular to a method for evaluating transient stability of a Bayesian optimized LightGBM. Background technique [0002] With the large-scale grid integration of wind power, photovoltaic and other renewable energy sources, their volatility, randomness, and low inertia characteristics make the power system face great challenges in transient stability assessment and control. As we all know, the early stages of power system accidents are often accompanied by transient faults. Once the dispatcher cannot make appropriate decisions on transient faults and intervene in time, the transient stability of the system may be destroyed, and even develop into a subsequent Cascading failures, which will lead to mass blackouts. Due to the rapid development of transient accidents and the short response time, it is difficult to make correct judgments and decisions in a very short...

AI Technical Summary

Problems solved by technology

However, decision trees are prone to overfitting; when dealing with large-scale power system data, the ability of extreme learning machines to extract input features is limited; support vector machines are easily sensitive to noise and have weak processing capabilities for large data; random forests Can handle high-dimensional data samples, but the amount of calculation is large; the classification performance of XGBoost is weaker than that of LightGBM; deep belief network and convolutional neural network have better feature extraction ability and generalization ability, and can automatically determine the best parameters of the network The problem is still not effectively resolved

Images