"Very-Low-Frequency" Cable Testing:
Combining Cable Quality Analysis and Selective Breakdown of Water Trees
By Olaf Goll
Very low frequency (VLF) AC (0.1 Hz) voltage testing has been used to test medium-voltage cables for several years. Experience and studies indicate that VLF is suitable for testing both new and service-aged power cables. Unfortunately, this type of testing only accounts for locally restricted faults, such as mechanical damage, cavities and individual advanced water trees to be detected. But an assessment of cable insulation with regard to quality and aging is possible in principle using dielectric methods.
A few years after the first installation of polyethylene (PE) and cross-linked polyethylene (XLPE) cables in the late 1960s it became obvious that these cables could not be tested with conventional test methods. In 1985, in close co-operation with universities and energy distributors, Hagenuk developed the first very-low-frequency 0.1 Hz cosine square wave test set, the W204 VLF system.
The German VDE has approved the technique as a verified test procedure not only for XLPE, but also for PVC and paper cables. Many utilities have replaced their conventional DC cable testing with this technology, which is effective for all types of cables, their termination and joints. The technique works at much lower voltages and produces less stress than the comparable DC testing. A second generation of VLF test sets has now been developed in which state-of-the-art technology, integrated into the control unit, enables a considerable reduction in weight and volume.
VLF 0.1Hz cosine square wave testing causes existing large electrical trees and mechanical defects in the cable to grow and to cause a breakdown during the three-phase one-hour test duration. The technique is very fast, reliable, and affects only the damage that is critical to the further service of the cable. With its patented waveshape and technology, this test method requires very little power due to the recycling of energy stored in the cable capacity.
More than 7000 VLF tests and subsequent observation of the cables for three to four years have yielded statistics which prove that this method is very reliable, and that the tested cable operates problem-free in later years. Statistics also show that the VLF system will reveal all critical singular fault spots in the cable.
Methods frequently used for on-site testing of medium-voltage cable systems are largely those that reveal weak points in the cable insulation by means of electrical breakdown. For oil-paper cables, DC voltage testing has established itself over many years, and provides satisfactory results. But this method is not suitable for PC or XLPE-insulated cables, as massive weak points along the cable cannot be reliably detected even when using very-high-voltage levels.
In recent years several research projects have been carried out with aimed at finding comparative test methods with various alternative voltage types. The results of laboratory investigations and on-site cable tests have shown that typical faults in paper-insulated, oil-impregnated, PE and XLPE cables can be detected with the 0.1 Hz VLF test. The authors found that, owing to the low partial discharge inception voltage and relatively high channel growth speed of electrical trees, 0.1 Hz VLF with cosine square wave voltage is the most advanced on-site test method. The possibility of detecting weak points is therefore relatively high and the risk of further damage very low.
Today, a voltage level of 3 volts and a test duration of 60 minutes are recommended. The results of breakdown testing as already described are limited, as only singular fault points can be detected. It is not possible to obtain information on the degree and nature of aging and continued serviceability of the cable. For this reason, investigations into dielectric diagnostic methods are needed. Measurement of dielectric losses at low frequencies, depolarization currents and return voltage are just a few of these investigative methods.
Studies have shown that by using these measuring methods, it is possible in principle to differentiate between cables damaged with water trees to varying extents. Numerous studies of oil-paper-insulated cables and lead-insulated cables have been carried out since 1960 using the return voltage method. So far, there are few results for PE/XLPE cables. The principle of return voltage measurement and first field experiences of PE/XLPE cable will be described in the second part of this article next month.
The VLF Test Set
The 0.1 Hz cosine square wave test allows an AC voltage to be generated with very little loss. The test system was developed with an emphasis on its ability to test high cable capacitance.
Figure 1 shows the main components of the system, including the DC test set, commutator, HV choke and support capacitor. The commutator changes the polarity of the voltage applied to the tested cable every five seconds. The typical shape of the test voltage is shown in Figure 2.
Principles of Operation
A full wave of this voltage shape can be divided into five sections:
a) The HV test unit is connected directly to the cable via the commutator. In this phase, the cable is charged to the negative nominal value of the voltage in approximately four seconds.
b) Waiting time with deactivated test unit: the nominal value has already been reached.
c) By turning the commutator, the HV choke is connected to the cable via the rectifier. This starts a reversal process with the polarity of the voltage in the cable changing. The oscillation frequency depends on the cable capacitance, but is always in the vicinity of a 50 Hz edge. If the nominal positive voltage is reached, the rectifier interrupts the reversal process.
d) Waiting time in the positive half-wave; the cable is disconnected apart from the voltmeter and support capacitor. A slight decline in the charge voltage at the cable can be observed during this interval.
e) Another change of polarity by turning the commutator as described in c).
In other words, the 0.1 Hz cosine square wave is like a 50 Hz AC test voltage that has been chronologically lengthened by DC periods following each polarity reversal. The energy that the DC test unit has to recharge in phase a) is very small because the substantial electrical energy stored in the cable capacitance is not dissipated during polarity reversal, thus avoiding otherwise high electrical losses. Heat production remains low, and it is therefore possible to utilize the output of normal DC test sets very effectively. The system currently available from Hagenuk is able to test cables of up to the 30 kV series with a capacitance of up to 5µF. Its efficiency makes it possible to integrate this system in cable test vans normally used for cable fault location.
Olaf Goll is head of research and develpment at Hagenuk GmbH. Part II will appear in next month's issue. ET