A novel diagnostic algorithm for AC series arcing based on correlation analysis of high-frequency component of wavelet. Issue 1 (3rd January 2017)

• Record Type: Journal Article
• Authors: Wang, Yangkun; Zhang, Feng; Zhang, Shiwen; Yang, Guang
• Is Part Of: Compel. Volume 36:Issue 1(2017)
• Page Start: 271
• Page End: 288
• Publication Date: 2017-01-03
• Subjects: Correlation coefficient -- Diagnostic algorithm -- Low voltage arc fault -- Wavelet transform; Electrical engineering -- Data Processing -- Periodicals; Electrical engineering -- Mathematics -- Periodicals; Electrical engineering -- Periodicals; Electronics -- Data Processing -- Periodicals; Electronics -- Mathematics -- Periodicals; 621.3
• DOI: 10.1108/COMPEL-08-2015-0282
• Languages: English
• ISSNs: 0332-1649

Abstract

Abstract : Purpose: A multi-load available, response reliable and product-friendly method is in urgent need to diagnose the signs of incipient arcing. This paper aims to propose a novel algorithm that originates the application of correlativity analysis of wavelet high-frequency component in state discrimination and further in arcing detection. Design/methodology/approach: The proposed method calculates the correlation coefficient between the extraction by wavelet transform of arcing series current and that of normal, compares it with a predefined threshold and outputs a trip signal when eight qualified arcing half cycles within a period of 0.5 s are detected. Findings: Typical appliances are selected in laboratory for arc detection to test the method which carries on independently of impedance type. The algorithm could be optimized to identify arcing for different kinds of loads, including resistive, inductive, capacitive and switching power supply loads, with a same correlation coefficient threshold. Practical implications: The arithmetic operations of the method are addition and multiplication, which contribute to efficient data computation and transmission for micro-processor to undertake. The reference optimal sampling rate recommended for the algorithm helps to reduce the processed data volume and shows its promising prospect for portable product development. Originality/value: This proposed correlativity analysis of wavelet transform component algorithm could classify the tested signal into two categories, which benefits the discrimination of normal and fault states in condition monitoring. Laboratory tests prove that it works effectively in arc detection for the commonly used impedance types of loads and needs no offline self-learning or training of samples.