The demand for cable diagnostics and monitoring techniques has increased significantly in recent decades. Various diagnostic tests such as partial discharge, dielectric loss, and elongation-at-break tests are available in real-world applications. Among the cable diagnostic methods, reflectometry can be used to detect the location of a fault according to the reflected signal at the impedance-discontinuity point. Because it is a nondestructive method and can be applied regardless of the cable type, reflectometry is commonly used for monitoring real-world applications. Time–frequency domain reflectometry, an improved type of reflectometry, is more accurate than conventional reflectometry and employs time–frequency localized signals that are robust to noise. To compensate for the errors occurring in the time–frequency domain reflectometry, several algorithms have been reviewed. In the past, because the signal of time-frequency domain reflectometry contains both time and frequency information, a high-specification signal generator and measurement sensors were required, which made practical application of the method difficult. However, with recent developments in instrumentation and sensing technology, reduction in the instrumentation size and cost facilitates the use of time–frequency domain reflectometry in various industrial applications, e.g., control and instrumentation cables in nuclear power plants, superconducting cables, and submarine cables. This work reviews the use of time-frequency domain reflectometry for cable diagnostics and monitoring in real-world applications. The purpose and results of the experiments on industrial applications are introduced and analyzed. This study also suggests directions for further research to make time–frequency domain reflectometry a diagnostic standard.

Fault diagnosis has been studied actively across the electrical industry to help maintain the stability of electrical equipment. Among these equipment, shielded cables, which are widely used in various industrial sectors, require careful and periodic diagnosis, owing to their poor installation environments and potential for creating huge economic losses. Reflectometry is a representative solution to locate the cable faults; however, conventional reflectometry techniques require prior knowledge about the cable under test, such as the reference wave velocity, total length of the cable, etc. Moreover, the degree of failure cannot be determined using conventional methods. In this paper, a novel reflectometry technique is proposed to locate and evaluate the faults in a cable, without requiring any prior knowledge. General regression neural network (GRNN) based on kernel density estimation are utilized with special feature extraction procedures. The proposed method is tested in an actual test bed with two types of emulated faults, and is found to estimate both the fault location and reflection coefficient successfully. It is expected that the proposed method can improve the stability of industrial equipment.

In recent years, as reliance on factory automation increases, real-time surveillance techniques for electrical systems have received substantial attention. In particular, the fault diagnosis of shielded cables has become crucial in the industrial sector due to their roles in interconnecting each electrical element. The time-frequency-domain reflectometry (TFDR), which is an advanced cable diagnostic technique, has been used to diagnose various types of shielded cable with high accuracy in fault location. However, in the case of reflected signals with a low signal-to-noise ratio (SNR) caused by any soft faults, the method faces ambiguities in interpreting the presence of failures and locating the faults. Thus, this article proposes an algorithm that simultaneously enhances the fault detection and localization performance of TFDR. In addition, the proposed method provides a statistical model-based threshold for fault detection. The performance of the proposed algorithm is tested via three experiments on actual shielded cables, and the efficacy of the proposed method is verified based on statistical analyses with theoretical discussion.

Load forecasting is a key issue for efficient real-time energy management in smart grids. To control the load using demand side management accurately, load forecasting should be predicted in the short term. With the advent of advanced measuring infrastructure, it is possible to measure energy consumption at sampling rates up to every 5 min and analyze the load profile of small-scale energy groups, such as individual buildings. This paper presents applications of deep learning using feature decomposition for improving the accuracy of load forecasting. The load profile is decomposed into a weekly load profile and then decomposed into intrinsic mode functions by variational mode decomposition to capture periodic features. Then, a long short-term memory network model is trained by three-dimensional input data with three-step regularization. Finally, the prediction results of all intrinsic mode functions are combined with advanced measuring infrastructure measured in the previous steps to determine an aggregated output for load forecasting. The results are validated by applications to real-world data from smart buildings, and the performance of the proposed approach is assessed by comparing the predicted results with those of conventional methods, nonlinear autoregressive networks with exogenous inputs, and long short-term memory network-based feature decomposition.

A defect occurring in cable joints is much more severe than a defect on cables due to its high incidence and an intensive electrical stress. In conventional reflectometry, it is hard to distinguish between a reflected signal from normal cable joints and that from faulty cable joints. This paper proposes a novel time–frequency domain reflectometry (TFDR) method based on an unsupervised neural network model combining long short-term memory (LSTM) and variational autoencoder (VAE) that can detect joint defects as well as cable defects. To verify the proposed method, a test bed is constructed with two failure scenarios; (1) defects on cables, and (2) defects in cable joints. In both scenarios, the proposed method successfully detects the failure using an anomaly score that conventional TFDR does not have. The proposed anomaly detection technique is expected to become the cornerstone of systems that can detect anomalies of earlier-stage of defects.

In this paper, new time-frequency-based anomaly detection methodology is proposed for the condition monitoring of high-temperature superconducting (HTS) cable systems. The time-frequency-based anomaly detection methodology includes two indices, which are obtained via cross Wigner-Ville distribution and scattering parameter. For the validation of the proposed methodology, the methodology is applied to an ac 22.9-kV/50-MVA HTS cable system connected to the real power grid network. Furthermore, the changes in dielectric properties of the HTS cable system during the cooling process and the current imbalance failure are monitored. The proposed anomaly detection indices are shown to be able to detect the heat and the current imbalance caused by the malfunction of the joint box. For the application of the two validated indices, new anomaly detection procedure for HTS cable systems is also presented.