Cable defect studying and judging method based on outlier detection algorithm

• Main Authors: REN SHUQING, YUE KUN, ZHANG SICONG, LING TIANHAN, GONG BOREN, SUN FUFENG, YANG ZHOU, LI ZHAN, HAO ZEQI, HE YULONG, WANG JUNPENG, WANG BO, CHEN XIAOXU, LUO KAIXIN, SHAO QIANG, ZHOU BAOZHU, ZHANG JINHUA, LIU YUHENG, DU ZHIMIN
• Format: Patent
• Language: Chinese; English
• Published: 11.08.2020
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; HANDLING RECORD CARRIERS; MEASURING; MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS; TESTING

The invention discloses a cable defect studying and judging method based on an outlier detection algorithm. The cable defect studying and judging method comprises the following steps: 1, acquiring andpreprocessing data; 2, analyzing the data by using a DDPOS algorithm, if the outlier factor of the outlier is greater than 1.2, determining the outlier as a suspicious outlier, otherwise, determiningthe outlier as a normal point; 3, setting a threshold value, and if the data of the suspicious outliers is greater than the threshold value, determining the suspicious outliers as fault points; otherwise, continuing to perform outlier detection by using the DDPOS algorithm again, if the outlier factor of the suspicious outlier is greater than 1.05, determining the suspicious outlier as a suspicious fault point, and otherwise, determining the suspicious outlier as a normal point; and 4, fault prediction: comparing and matching the determined data of the fault point and the suspicious fault point with historical data in