Affiliations: Electronic Information Engineering, College of Information Science and Engineering, Hunan First Normal University, Changsha, China
Correspondence:
[*]
Corresponding author. Di Zhao, Electronic Information Engineering, College of Information Science and Engineering, Hunan First Normal University, 410000, Changsha, China. E-mail: [email protected].
Abstract: Due to poor noise elimination effect and low detection accuracy in traditional methods, a method to detect DC power disturbance data for charging piles based on linear algebra was presented. Firstly, a filter was constructed based on the mathematical morphology theory of linear algebra to preprocess the disturbed waveform, so as to filter out the random noise and impulse noise in signal. Secondly, the filtered waveform was analyzed according to the change rule of grid. And then, a simple and fast singularity detection criterion was proposed to detect the disturbance accurately and quickly and thus to locate the time. According to the location result, the clustering algorithm based on GK was used to identify various types of disturbance data, so as to make countermeasures. Finally, voltage sag, voltage swell, harmonic wave and the combined disturbances were used to verify the proposed method. The experimental results show that the maximum denoising result of this method is 37.3, and the maximum detection accuracy of disturbance data is 0.952, which is much higher than the traditional method. The results show that the signal-to-noise ratio is larger and the detection effect is better based on the proposed method, so that the Intersection-over-Union (IoU) is improved and the detection accuracy is better. In conclusion, the overall detection performance is better.
Keywords: Mathematical morphology, GK clustering algorithm, signal recognition, SNR
