Stability index construction method and device based on machine learning

Abstract

The invention discloses a method and device for constructing a stability index based on machine learning. The method includes: obtaining training samples of the stability index, the training samples including stable samples and unstable samples; Slack variables are introduced into the constraint conditions of machine model SVM; the slack variables in the constraints of stable samples and unstable samples are removed respectively to obtain the aggressive support vector machine model ASVM and the conservative support vector machine model CSVM; according to the boundary distance between ASVM and CSVM The instability evaluation index DD is obtained by the difference; DD is transformed according to ASVM and CSVM, and combined with the Ridge regression algorithm for fitting to construct a stability index. This method combines the advantages of SVM with the Ridge regression algorithm to obtain the stability index to represent the stability margin of the system, which is used for auxiliary decision-making of stability judgment, thereby improving the decision-making ability and precision of emergency control.

Description

technical field [0001] The invention relates to the technical field of power system network source coordination management, in particular to a machine learning-based stability index construction method and device. Background technique [0002] In recent years, the scale of the power system has been increasing, and the characteristics of the power grid have become more and more complex; a large number of new energy sources have greatly increased the uncertainty and volatility of the power system's operating status; the aging and damage of existing equipment have brought hidden dangers of system failure. . The above factors make the safe and stable operation of the power grid more and more difficult, and power outages occur from time to time. [0003] An important factor leading to these blackouts is that at the beginning of the accident, the control center of the power grid cannot quickly predict the transient stability of the system after the fault and take effective contro...

AI Technical Summary

Problems solved by technology

These two evaluation methods are completed based on simulation scans of typical faults, which cannot fully take into account various situations in actual operation, which may affect the accuracy of evaluation results

In this case, in order to ensure the conservatism of the application, the stability margin of the system operation is large, resulting in the waste of power equipment

Images