Search

The Phase-Resolved-Partial-Discharge pattern (PRPD) is a conventional technique used for the evaluation of partial discharges (PD) phenomena in High-Voltage-Alternating-Current (HVAC) systems. This map is constructed by plotting the peak of each detected pulses as a function of the phase angle of the supply voltage. Therefore it is obvious that this technique cannot be used for the analysis of data from PD mesaurement under different supply voltage condition (DC). The aim of this paper is to evaluate the application of the Time-Frequency map (TF map) for the analysis of a dataset obtained from PD measurement under DC voltage. A density-based clustering algorithm was also used to gain more insight from the collected data. The results show that, with this approach, it's possible to perform a noise rejection and identify PD pulses.

Rapid growth of HVDC cable line installations is expected in the coming years. Several research groups have investigated dielectric phenomena that occur in the insulation of HVDC cables and accessories that accelerate aging and reduce the reliability of this technology over time. Unlike in HVAC cables, under HVDC stress, the distribution of the electric field inside the insulation layer depends not only on the applied voltage but also on the temperature distribution and on other factors related to manufacturing such as the presence of impurities. Reaching local and transient electric field peaks can enhance the onset of Partial Discharges (PDs) which can lead to premature aging of the insulation. As demonstrated by several studies, the achievement of electric field peaks that can exceed the inception level for partial discharges is more likely in joints. This is due both to the greater probability that a defect is incorporated in two different dielectric layers, and to the greater thickness which determines a higher temperature drop. In this paper, the occurrence of PDs inception conditions in a void inside an HVDC joint during a Transient Over Voltage (TOV) event is investigated by means of FEM simulations. This type of event can be considered as one of the most severe from the point of view of the occurrence of partial discharges as the effect of the rapid variation of applied voltage is combined with the previously established electric field distribution due to the space charge accumulation.

DC cables are supposed to withstand a number of operations involving supply-voltage peak-value time variations, such as energization, switch off and polarity inversion. It is known that voltage-supply time variations cause electric field transients in cable insulation that can last for a long time before steady-state conditions are achieved, and during this time the electric field distribution is not only driven by electrical conductivity, as in steady state, but also by permittivity, as in AC. This paper shows that highly energetic degradation phenomena, that is, partial discharges, PD, can occur during voltage transients that a DC cable has to withstand during life. This results in a cumulative damage which can cause premature insulation breakdown, at times well shorter than those expected based on a cable design which is PD free under steady-state conditions. Quantification of the damage caused by PD, based on aging and residual life models, is proposed in the paper, and it is shown that damage can be significant enough to pose a threat for cable life when operation transient rate is e.g., a few per day, and that the faster the slew rate of the supply voltage variation, the highest the cumulative damage due to PD.

Medium voltage (MV)/high voltage (HV) cable lines are high-value assets and require substantial costs for their installation or replacement. For this reason, the capability to assess the condition of cable lengths and accessories on site is of great importance. Cross-linked polyethylene (XLPE) insulation is extensively used for MV, HV, and extra-high voltage (EHV) class cables [1] thanks to its excellent voltage-endurance and thermomechanical properties. In particular, XLPE-insulated cables feature substantially lower losses, easier manufacturing and jointing procedures, better environmental compatibility, and higher operating temperature compared with impregnated paper cables, which lead to lower costs in procurement and operation [2], [3].

The expression of the dimensional effect proposed by Enrico Occhini in the 70's is still commonly used to transform cable insulation breakdown and life data, obtained through laboratory accelerated testing on cable models to those relevant to full-size cables. However, the Occhini model does not take into account the dependence of dielectric strength on insulation thickness. Indeed, the balance between the insulation losses generated in the volume and the dissipated energy which occurs through cable insulation surface affect dielectric strength and this can have an impact in cable characterization and testing. This paper presents a modification of the statistical expression of dimensional effect, through the introduction of the dependence of dielectric strength on insulation thickness due to thermal instability. The modelling study carried out in the paper, and the preliminary validation of the revised model, show that the derived correction terms can influence both the design field and the life line estimation for cables having different insulation thickness and lengths. Eventually, it is highlighted how much and in which way the dimensional effect issues can impact on the results of prequalification and type tests made according to the present IEC and CIGRE specifications.

This work is performed to assess the role played by some quantities affecting space charge measurements with the Pulse Electro-Acoustic (PEA) technique in HVDC power cables. In particular, the possible disturbances to the PEA signal arising when measuring space charges with a commercial PEA cell not optimized for the cable system to be measured, which may affect consequently the estimation of trapped charge and the field profile within the insulation, are analyzed. The results show that in particular, cable length, cell arrangement and exposed semicon length play a role and may affect the PEA signal significantly. Thus they have to be optimized for the cable to be measured.

Recent experiments on the use of the PEA method for testing dielectric materials in specimens and comparison with a detailed model provide an insight of the phenomenon and suggest the need of adopting similar models also for cables. What is said is even more important considering the possible future adoption of the PEA methodology to test DC cables for Pre-Qualification and Type Tests. The use of an accurate model of the PEA cell used for testing specimens and related experiments prove that the thickness of the different parts composing the PEA setup is a basic element for providing accurate charge reading and interpretation of the phenomenon. Both simulation and experimental results, carried out for flat specimens, show the importance of the PEA cell components and sample features in the output charge profile. In this work, we demonstrate that these considerations for flat samples are also valid for cable specimens. This aspect, and the other critical issues encountered in the experimental test on a Full-Size cable, have been reported in this paper, as well as some possible solutions.

This paper describes some critical issues arising during the partial discharge (PD) testing of a 500m delivery length of an XLPE-insulated MV cable. At first, conventional PD measurements were performed on the whole length by using a standard acquisition system; PDs activity was detected, but the system allowed the operator to localize the PDs only on a length of 70 m. A visual inspection of this section highlighted the presence of an excoriation in the outer sheet. Aiming at a more accurate pin-pointing of the defect, more measurements were carried out in the proximity of the outer defect using a portable PD detection system. These measurements showed PDs activity, whose source was located and identified as consisting in internal defects on the inner semicon in proximity of the external excoriation. This proves how PD detection portable instruments are able, in cases like this, to locate the PD source and provide indications for its correct classification.

A novel Partial Discharge detection system is here described. The system employs a special wideband antenna as a sensor, allowing to sense either Partial Discharges pulses either the AC supply voltage at a distance from the component under test, so avoiding the need of direct connection or component shutdown. Thanks to its wireless nature and small physical dimensions, the system enables to perform accurate on-line Partial Discharge diagnosis even in circumstances where conventional instrumentation would be impractical to use, so greatly extending the applicability of this diagnostic method.