Numerous tests have been performed in various locations in the Middle East and across North America involving XLPE-insulated cables from 66 through 380 kV and with circuit lengths ranging from 3 to 30 km. Typically, those circuits rated at 220 kV and higher have required use of two resonant test systems operating in parallel.
The majority of circuits subjected to such overpotential testing successfully withstood application of the prescribed voltage for one hour. However, in a small number of cases, there were dielectric breakdowns and these failures occurred almost exclusively either in or close to a joint or termination. Typically, defects in the accessories were attributed to problems with installation procedures rather than to deficiencies in design or materials. For example, there was one instance of cable failure where mechanical damage during transport or storage was suspected as root cause.
It may seem attractive to perform PD measurements at the terminals of the cable circuit rather than at individual joints. However, it is well known that problems due to signal attenuation and dispersion limit the lengths of cable where this method could be successfully applied. Detection of PD-related phenomena occurring several kilometres from the detection point requires that the measurement bandwidth be relatively low. Such techniques therefore suffer from problems due to relatively high background electrical interference associated with field measurements. Moreover, although various steps are taken to mitigate the effect of switching noise from the high voltage power supply, these signals will also be present and add to the difficulty of separating PD signals from background noise.
A critical question when making PD measurements relates to establishing pass/fail criteria for their magnitudes. The simplest and most conservative answer is that any detectable level of PD is too high but many practical issues need to be addressed. Among these are the costs and delays needed to replace any accessory or section of cable. One practical example of attempting to define PD-based acceptance criteria for commissioning testing is illustrated in Figure 3.
The left side of Fig. 3 displays data acquired during the off-line test whereas the right side shows data from the on-line test – both conducted at a voltage of U0. PD activity measured during the off-line test has a frequency content ranging from 7 MHz to 8 MHz whereas minimum, mean and maximum PD magnitudes for the off-line data are 5, 7 and 24 mV respectively. In addition, the phase resolved PD plot for the off-line data shows clusters of negative and positive polarity pulses centered near 45° and 225° phase angle with reference to the phase-to-ground voltage. These are classic locations for phase-to-ground dependent partial discharge data.
However, during on-line testing, measured signal activity is quite different, i.e. frequency content of the signals ranges from circa 800 kHz to circa 2 MHz. In addition, the phase resolved PD plot shows the pulses measured to be located throughout the AC cycle. Consequently, the data acquired during the on-line test relates to electrical noise. This figure demonstrates that no PD signals were present during on-line tests since, if present, these would have been detected simultaneously with the noise signals and would have been identifiable in the frequency domain plot of Figure 3. While there are a number of factors that may account for the differences observed above, a decision was made to put the cable in-service in spite of detectable PD during the off-line test. This circuit is currently operating without incident.
One key point emerging from the testing performed to date is that all of the circuits that successfully withstood the high voltage tests and on which no PD was detected have operated without incident since commissioning. Circuits which have only been hi-pot tested without the inclusion of PD testing have experienced some failures. Consequently, the combination of overvoltage and PD testing appears to provide the highest level of confidence in the reliability of the cable, accessories and installation method.
Coordinated development of existing standards for high voltage cable systems allows their effective type test qualification to both IEC and ICEA simultaneously. At the same time, innovations in techniques used to perform standard type testing of HV cables can improve accuracy and efficiency, while also lowering costs.
Numerous solid dielectric transmission cable circuits have been subjected to high voltage and PD testing using this methodology as part of the commissioning process following installation. The majority of cable systems tested successfully met the test criteria. However, a small number suffered dielectric breakdown during the overvoltage test with failures generally located in or in close proximity to accessories. Breakdown of the cable itself is extremely rare.
Where post-failure analysis has been performed, the results are consistent with installation issues or damage to the cable during transport or storage. None of the breakdowns have been attributed to design, materials or processing issues. While there is still ongoing debate regarding the technical and cost issues associated with optimizing PD test procedures for on-site transmission cable testing, experience shows that the combination of high voltage and PD testing of transmission cables is necessary to satisfy industry demands for high reliability.