|dc.description.abstract||In order to be able to take maximum advantage of the remaining lifetime of the components in the distribution network, utilities need to know the condition of the equipment as accurately as possible. When partial discharges (PD) are measured in medium-voltage cable systems, it is often chosen to disconnect the cable length and use voltage sources with frequencies of 0.1 or 50 Hz. Temperature dependence of the conductivity in the joints' field grading materials, as well as expansion and contraction of materials could potentially cause PD characteristics to be load dependent. This again means the method and time of testing may affect the results and subsequent assessment of the equipment's condition. There are few studies available in the literature where XLPE cable joints from service are characterised by PD during load cycling.
The purpose of this master thesis is to study PD activity during load cycling of field aged cable joints at 50 Hz. PD measurements are also performed at 0.1 and 50 Hz in room temperature to simulate the conditions during off-line testing. The insulation resistance of the test objects will be measured through both tests.
The conditions during on-line and off-line PD measurements have been simulated in the laboratory, where three phases from a field-age joint have been tested. During the "off-line" measurements, the inception voltage was measured with 0.1 and 50 Hz voltages in room temperature. The measurements were performed with and without voltage conditioning (50 Hz, 6 kV for 1 hour) prior to the PD measurements. On-line measurements were simulated by use of load cycling to temperatures in intervals up to 65 $^\circ$C and voltage conditioning. The inception voltage of the joints was measured at stable temperature. PDs were recorded at 6 kV during the cooling of the test objects to investigate whether the PD sources are temperature dependent.
Towards the end of the experimental work, PDs in the joints were localised acoustically. Then, the break-down voltages were measured, before the dissection of the joints ended the experimental work.
The PDs that were measured appear to depend more on voltage frequency than joint temperature. In room temperature, inception voltages of 3-6 kV and 1-1.5 kV were measured at 0.1 Hz with and without conditioning, respectively. When PDIV was measured at 50 Hz in room temperature, the results were between 5.7 and 7.1 kV without and 6.1-7.6 kV with conditioning. No clear trend of inception voltages was observed when the temperature was increased. With the exception of one (very high) PDIV, all inception voltages measured during load cycling are between 6.0 and 7.1 kV, i. e. in the same range as measured at 50 Hz voltage in room temperature. The break-down voltages of the joints were found to be 65, 65 and 50 kV.
The inception voltages measured at 0.1 Hz are all less than or equal to the operating voltage, and highly dependent on how recently voltage has been applied to the joint. Based on this, it might be concluded that the measurement results at 0.1 Hz can be difficult for the utility to interpret, thus increasing the risk of misdiagnosis. On the other hand, the off-line measurement results at 50 Hz provide a realistic impression of the condition of the joints, and are also a good representation of the conditions under operation. This is confirmed by the measured break-down voltages.
Weibull analysis of the PD measured during cooling showed a decreasing number of PD sources in 2 of 3 cases. In the joint where a constant number of PD sources were identified during cooling, approximately the same number of sources was detected acoustically. Dissection of the joints after break-down voltage testing showed that 2 out of 3 failures were thermal and that all punctures occurred at the semiconductor cut-off. Traces of discharge spots and ageing due to water ingress were observed by use of microscope.
These findings indicate that the discharge sources may be temperature dependent, but that more investigations are necessary before any conclusions can be drawn.||