In industrial systems, cables are used to connect various pieces of equipment. Diagnosing soft faults in cables is a consistent step to ensure the system's stable operation. The bottleneck with existing diagnostic methods is locating and evaluating multiple faults in a cable with low error. In this article, a novel fault diagnosis method is proposed to locate multiple faults and estimate their reflection coefficients; the latter is used to assess the level of the fault. First, the reflection coefficient spectrum (RCS) (the data to be processed) is modeled, the relationship with faults is elaborated, and cable model parameters are estimated. Second, a matched filter matrix is constructed to analyze the RCS to locate faults. Third, the compensation model based on the cable transfer function is constructed to compensate for the attenuation of the signal and then estimate the reflection coefficients. Finally, sidelobe analysis is performed to improve detection accuracy, and multiple reflection analysis is performed to elaborate on the complex reflection phenomenon between multiple faults and filter out their interference in detecting real faults. Experimental results show that the proposed method locates multiple faults, estimates reflection coefficients with low error, and exhibits good robustness in the presence of noise.
25 February 2025 in IEEE Transactions on Instrumentation and Measurement
Fault diagnosis serves as a consistent approach to ensuring cable system stability, with high-frequency transmission characteristic parameters (HFTCPs) enhancing sensitivity, accuracy, and fault assessment capabilities. This article introduces a method for measuring HFTCPs using reflection coefficient spectrum (RCS) and optimized multivariate variational mode decomposition (MVMD), applicable to various shielded cables. Firstly, we model the RCS under impedance mismatch and elaborate on the error sources of traditional methods. Secondly, we calculate the signal's energy spectral density to determine the initial center time and mode number, and dynamically adjust the bandwidth balance parameter based on the SNR of RCS components. Subsequently, we utilize the optimized MVMD to decompose the RCS into mode banks, shift the modes' spectrum to the baseband using single-sideband modulation, and perform multiscale decomposition to eliminate noise mixed into the modes. Lastly, we design a homemade transmission line to interconnect the large cable and measuring instrument, correct the measuring plane using OSL calibration, and remove frequency dependence using the least squares method. The results show that the proposed method has high decomposition accuracy and execution efficiency, good noise robustness and large improvement in the accuracy and bandwidth of the acquired HFTCPs.
27 January 2026 in IEEE Transactions on Instrumentation and Measurement
Soft faults in coaxial cables, such as insulation aging and micro-cracks, can evolve into severe failures if not detected timely. Existing reflectometry-based techniques are capable of promising fault localization but lack the ability and accuracy in evaluating fault severity. In this study, we propose a method that combines Electromagnetic Time Reversal (EMTR) and Stepped-Frequency Waveform Reflectometry (SFWR) to enhance both fault location and evaluation. The proposed technique utilizes stepped-frequency signals for time reversal, enabling precise reconstruction of reflected waveforms in both time and frequency domains. A signal fitting and amplitude compensation algorithm is further applied to estimate reflection coefficients, allowing quantification of fault degree. Experimental validation on coaxial cables demonstrates that the method performs well both in fault location and reflection coefficient estimation. Moreover, the approach can simultaneously detect and assess multiple faults with an accuracy exceeding conventional methods. Compared to traditional SFWR and BIS, the proposed method significantly improves diagnostic accuracy and provides a more precise fault evaluation. The study also investigates the performance of the proposed method under noisy conditions, where the results demonstrate its favorable applicability, making it suitable for predictive maintenance in industrial cable networks.
1 January 2026 in IEEE Transactions on Instrumentation and Measurement