Due to the number of joints in land-based HVDC cables, which can easily reach over 1000 in a single system, failure rate of joints plays a decisive role in overall system performance. Unfortunately, routine tests in the case of HVDC cable joints may not be sufficient to guarantee high reliability during operation. This is because electric field during operation in some areas inside an HVDC joint can be much higher than seen during routine testing due to the fact that field distribution in such a joint is highly sensitive to temperature. Therefore, to properly cover field distribution in the joint at higher conductor temperatures, additional electrical testing may be needed.
In contrast to AC cable systems, HVDC cable systems allow electrical energy to be transmitted over distances in the range of several hundred or even over 1000 km. Notable examples include HVDC cable systems being planned in Germany that will run about 580 km (i.e. the South-East Link) and about 700 km (i.e. the South Link) respectively. However, due to weight limits during transport, the maximum length of a single cable is usually limited up to only about a kilometre. That means that a joint is needed about each km in a land cable system and reliability of each joint is therefore essential. Experience in HV cable systems for AC has shown that failure of accessories plays a major role in unplanned shutdowns and one of the main reasons for such failures is errors during installation. By contrast, failures caused by impurities or air inclusions inside accessory insulation or production mistakes play a relatively minor role since these are detected during routine tests required for each main joint and termination insulation (e.g. according to IEC 60840 and IEC 62067). The question is whether these kinds of failures in HVDC cable joints would also be detected during routine tests according to IEC 62895.
According to IEC 60840 and IEC 62067, each cable joint must undergo a routine test under AC voltage of 2.5 x U0 for 30 minutes and 2.0 x U0 for one hour. In addition, each joint has to pass a partial discharge test at 1.5 x U0. Similar routine tests for AC joints are required in the IEEE 404 standard. These tests are intended to guarantee that failures in the main joint insulation due to impurities, air bubbles or errors in manufacturing are detected in the factory. However, even with these routine tests, there are still failures in the field that lead to unplanned cable system shut down. For example, a CIGRE investigation on AC cable joints between 220 kV and 500 kV found an annual failure rate of 0.048/100 joints, i.e. roughly one in every 2000 installed joints fails each year. This same investigation determined that the typical time to repair the cable system after such failures is 25 days for this voltage range. The question is whether expected failure rate for HVDC cable joints is similar to experience with AC where most failures are caused by installation errors. But this is far from certain since there are key differences in required routine tests for DC joints that might in fact lead to higher failure rates than typical for AC joints.
IEC 62895 for HVDC cable joints requires a DC voltage test at ambient temperature with a voltage level of 1.85 x U0 for one hour. An additional partial discharge test is optional and test parameters here are agreed to between customer and manufacturer. However, in contrast to AC voltages, electric field distribution in HVDC cable joints is highly sensitive to temperature and can increase locally due to several factors if temperature of the cable conductor increases. An example of such a change in field is shown in Fig. 1 that shows simulated field distributions in joints at different conductor temperatures. These finite-element simulations were performed with a finite element tool that considers change in resistivity with temperature and electric field.
Fig. 1 shows that electric field below the deflector at a conductor temperature of 70°C is much higher than at conductor temperature of 20°C.
This increase is also shown in Fig. 2, where total electric field is calculated along the interface between joint insulation and cable insulation. Below the deflector, electric field at conductor temperature of 70°C is about five times higher than at a conductor temperature of 20°C. Voltages for this calculation were the same as in Fig. 1 and correspond to test voltages for a 320 kV HVDC joint during routine tests according to IEC 62895.
At the upcoming 2019 INMR WORLD CONGRESS, expert Dr. Hansjoerg Gramespacher will present a detailed review of this increasingly important topic.
REGISTER NOW at: www.inmrworldcongress.com