Objective With the continuous development of power transmission systems from west to east, the advantages of ultrahigh voltage (UHV) direct current (DC) transmission have become increasingly evident. The optical fiber link used in the UHV DC control and protection system may fail, exposing the system to risks. Existing methods for inspecting the performance of the optical fiber link in such systems are deficient. Only a part of the optical fiber can be used to test the attenuation using a light source and an optical power meter, and the quality of the optical fiber link requires manual evaluation. Faults along the optical fiber link and decreasing transmission performance are difficult to determine during operation and maintenance, thus representing a security risk to the UHV DC control and protection system. Because the use of a light source and an optical power meter to detect link faults in the control and protection system requires cooperation on the two ends of the link, it is difficult to determine the fault location, which is not conducive to failure cause analysis and rectification. Considering the high security requirements of UHV DC transmission, phase-sensitive optical time-domain reflectometry (Φ-OTDR) is employed to detect and locate weak faults in the optical fiber link of the control and protection system.
Methods The proposed Φ-OTDR system prepared under laboratory settings comprised an optical transmission module, optical fiber interferometer, sensing optical fiber, optical receiver module, simulated weak fault source, signal processing module, etc. The fault status and environmental information of the key components of the UHV DC control and protection system are simulated, and information from time-domain backscatter phase modulation at the bottom of the Φ-OTDR system is obtained and analyzed experimentally. The information entropy algorithms of segmentation, multiplication, and integration are introduced to determine the failure status of the key components of the control and protection system of the simulated UHV DC transmission project and environmental information. Furthermore, the location accuracy of weak faults is evaluated.
Results and Discussions In the experimental system, different frequencies and weak displacement vibrations are simulated. First, disturbances are applied to collect 20 raw datapoints at 1.8 km of the optical fiber link. During the acquisition of the backscatter of the optical fiber link, the difference between the fault signal and background noise is not clear and the signal-to-noise ratio (SNR) is low (Fig. 3). Then, the information entropy algorithm is used to conduct a preliminary signal analysis. The segmented information entropy is proposed after identifying the shortcomings of the conventional information entropy algorithm. The amplitude of the fault signal is divided into k segments in the dynamic range, and the information entropy of the k-segment fault signal is calculated and the sum is used to represent the information entropy of the signal (Fig. 5). Finally, the information entropy location using segmentation and multiplication segmentation is proposed and our study shows that it is more suitable for the system. The improvement in the SNR using the signal processing of the information entropy location considering segmentation and multiplication segmentation is evaluated (Fig. 8). The optimal number of segments for the application of the system is determined (Fig. 9).
Conclusions Herein, two algorithms for information entropy location using segmentation and multiplication segmentation are proposed and the results of processing multiple types of mixed signals are compared and analyzed. The information entropy algorithm obtained by multiplying integral segments can considerably improve the SNR of the system when two segments are considered. The location accuracy of weak fault can reach ±1.5 and ±2.0 m, respectively, for 100 and 40 sampling periods, respectively.
