28 November 2023 in IEEE Transactions on Power Delivery
Partial discharge (PD) is indicative of insulation defects in power cable joints. Cross-bonding cable joints have a break in the joint shield that interrupts the screen of the jointed cables. It is possible for radio frequency (RF) signals of PD in the joints to emit from the shield break. This paper presents the use of off-the-shelf RF sensors at the shield break for non-intrusive PD detection. Measuring results from a 35 kV XLPE cable joint sample with artificial defects suggest that PD-induced electric and magnetic fields can be detected simultaneously with this approach. Typical pulses measured are limited to approximately 500 ns in duration, and the FFT spectrum of the pulses is mainly distributed at frequencies below 200 MHz. Average pulse magnitude ranging from about 6–22 mV can be measured between 10–15 kV. Increasing the measuring distance can cause the measured pulse magnitude to decease exponentially. Additionally, numerical simulations show that reducing the length of the shield break can cause the pulse magnitude recorded at the shield break to decrease gradually. The presented approach can potentially be used for cost-effective and non-intrusive detection of PD in cross-bonding cable joints.
8 November 2022 in IEEE Transactions on Power Delivery
Inappropriate manufacturing of cable joints can cause defects in the joint insulation. During operation, partial discharge (PD) may occur at these defects. The development of PD can cause accelerated aging even breakdown of the insulation, resulting in faults in the electric grid. In this paper, for radio frequency (RF) detection of PD in cable joints, a finite-difference time-domain (FDTD) model of a 35 kV XLPE cable joint is established, an excitation source is defined in the model to represent PD of different defects, and then propagation characteristics of RF electromagnetic signals induced by PD in cable joints are studied through the established model. Transfer functions (TF) of RF PD signals obtained with the FDTD model show that both magnitude and shape of the TF are closely related to the location and direction of the excitation source in the model. Two peaks in the magnitude, locating at 380 MHz and 790 MHz, respectively are present in the obtained TF regardless of the excitation source settings. Moreover, the TF magnitude for the excitation source defined in the radial direction of the cable joint model is found to be generally more than one order larger than that of the same source defined in the axial direction. The effectiveness of the FDTD model in representing high frequency properties of the cable joint has been checked against both time domain reflectometry (TDR) and S 21 measurements on a cable joint made in laboratory. The TF results from this work are expected to be helpful for applications of radiometric PD detection to cable joints, e.g., enabling improvement of the detection sensitivity through optimizing frequency responses of RF sensors.