Least squares support vector machine electric shock current detection method based on parameter optimization

Abstract

The invention discloses a least squares support vector machine electric shock current detection method based on parameter optimization. First, the total leakage current and electric shock current waveform of an organism in the process of electric shock at the peak moment, zero-crossing moment and any moment of power voltage are acquired on a residual current action protection device electric shock physical test system platform through a fault recorder, and signal data of 800 sampling points of the cycle before electrical shock and the three cycles after electrical shock is taken as electric shock test sample data; and then, filtering preprocessing is performed on the electric shock test sample data, and the total leakage current of multiple sample sampling points after preprocessing is combined into a feature vector input least squares support vector machine. As a new tool for pattern identification, the least squares support vector machine has superiority unmatched by other identification tools. A new detection method is provided for identifying electric shock current of an organism from total leakage current.

Description

Technical field [0001] The invention belongs to the technical field of electric power detection, in particular to a least squares support vector machine electric shock current detection method based on parameter optimization. Background technique [0002] Residual current protection devices, as an important measure to effectively prevent grid leakage accidents (biological body electric shock accidents or equipment leakage accidents), have been widely used and promoted in rural low-voltage power grids in recent years. However, since the commonly used residual current protection devices are based on whether the effective value of the total leakage current detected in the voltage circuit is greater than a certain setting value to determine whether it is operating, misoperation and refusal of operation often occur in actual grid operation. The main reason is that the operation basis of the residual current protection device cannot truly identify the current drawn signal of the electr...

AI Technical Summary

Problems solved by technology

For this reason, some scholars have tried to use intelligent control technology to establish an electric shock current detection model, but the neural network has shortcomings such as easy to fall into local optimum, unstable training results, and difficult to determine the number of hidden layer units.

Images