5. Project‐specific considerations regarding the choice of transmission line
6.4 Overvoltage following line de‐energization
6.4 Overvoltage following line de‐energization
When a compensated transmission circuit is disconnected from the power system, the disconnected circuit will resonate at its natural frequency as energy exchange between capacitive and inductive elements take place. Furthermore, capacitive coupling between OHL conductors and inductive coupling between the UGC and OHL conductors and shunt reactor windings will give rise to slow modulated overvoltage [20] [25] [26].
The magnitude, frequency and duration of overvoltage will depend on circuit parameters which in turn depend on circuit's physical construction, the ratio between UGCs and OHLs, and the degree of reactive compensation.
Another point of interest is overfluxing in shunt reactors connected directly to a transmission line. Flux in the reactor core is proportional to voltage and inversely proportional to frequency. The natural frequency of the de‐energised, compensated line is inversely proportional to the square root of the degree of compensation with natural frequencies of 50 Hz at 100 % compensation, 36 Hz at 50 % compensation and 15 Hz at 10 % compensation. Depending on the voltage/frequency ratio and overfluxing duration, a shunt reactor can be overheated and, in worst cases, damaged. Nonetheless, there is currently only limited published literature that deals with assessment of overfluxing in magnetic components. In this context, CIGRE Electra No. 179 [24] provides TOV withstand envelopes as overfluxing is the main constraint. However, these envelopes are only applicable under power‐frequency overvoltages and therefore not useful for other frequencies.
As the four alternatives for the west coast project all apply different UGC/OHL ratios, it is relevant to determine if any of them give rise to unacceptable overvoltage and in return, reactor flux levels during de‐
energization. This is done in the following sections.
6.4.1 Slow, modulated overvoltages following line de‐energization
As an example, phase‐to‐ground voltages following de‐energization of the 50 km transmission line between Idomlund and Stovstrup constructed as alternatives A, B, C or D are shown in Figure 45.
Figure 45 Phase‐to‐ground voltages after de‐energization of transmission line between Idomlund and
Stovstrup constructed as alternatives A, B, C or D.
The figure shows that the highest voltage occurs in Alternative A, followed by B and C, however, none of the overvoltages being critical. Overvoltage peaks in Alternative A because this has the highest share of OHLs and resulting increased capacitive coupling. Mutual capacitance between phases in a UGC system is negligible, because a metallic sheath is used and the ground acts as an equipotential surface. Therefore, no modulated overvoltages are seen in the Alternative D voltage profile, and voltage is decaying slowly at its natural frequency. In any case, the time constant of decay is dependent on losses in line components. For all four alternatives, overvoltages are generally not critical with the highest phase‐to‐ground voltage for Alternative A at 1.2 p.u. peak in phase B.
The respective flux in the shunt reactor is shown in Figure 46.
Figure 46 Relative flux linkage after de‐energization of the transmission line between Idomlund and Stovtrup
constructed as alternatives A, B, C or D.
Flux in a shunt reactor increases after de‐energization by the grid frequency to natural frequency ratio.
Furthermore, due to modulated voltage in the hybrid line‐based alternatives (A, B and C), flux is further increased when modulated voltage reaches its maximum. Especially in Alternative A, flux is limited as a shunt reactor is saturated.
6.4.2 De‐energization with variable shunt reactors
From an operational perspective, equipping the line with variable shunt compensation instead of fixed compensation, as discussed in Section 6.2, is preferable. This is especially true for hybrid‐based alternatives A, B and C. However, as shown in the previous section these are also the alternatives most prone to shunt reactor overfluxing due to the increased share of long OHL segments.
Decreasing the compensation degree decreases natural line frequency and leads to elevated flux levels resulting in faster heating of the reactor. An example is depicted in Figure 47 where line‐side phase‐to‐
ground voltage, reactor flux and reactor currents are shown following de‐energization of a 25 % compensated line.
Figure 47 Line voltages, reactor current, and relative flux linkage following de‐energization of transmission
line between Idomlund and Stovtrup constructed as Alternative A with 25 % compensation.
Reactor currents become heavily saturated and exceed 1 p.u. for 6‐7 seconds after disconnection, and flux takes an equal amount of time to return to rated levels. This can be critical for a reactor and that risk should be mitigated. One mitigation option for is to add a circuit breaker in line with the shunt reactor so that it can be isolated following de‐energization of the line. However, this introduces other complexities compared to a fixed connected shunt reactor. Disconnection of a shunt reactor following line de‐energization hinders the operation of delayed auto‐reclose (DAR). DAR is an important operational facility that brings a line back following spurious trip of the line. Another option is to increase the nominal flux levels in the reactor core by overdimensioning. This will, however, drive up shunt reactor costs and should be avoided, if possible.
6.4.3 Discussion and conclusion
This section describes how de‐energization of hybrid lines with long OHL sections can result in low frequency modulated overvoltages due to the OHL's mutual capacitance. Shunt reactor flux levels will increase during de‐energization as flux is proportional to voltage and inversely proportional to frequency. Flux levels are strongly dependent on the level of compensation with levels increasing as the compensation degree decreases. Especially for alternatives A and B with low compensation levels, the shunt reactor will be driven heavily into saturation. However, with limited duration and no recognized international standard governing this matter, related criticality is difficult to determine. It is sufficient to say that there is a high risk of overfluxing being an issue and detailed discussion with shunt reactor manufacturers must take place before any categoric conclusions can be made. For UGC‐based lines, de‐energization issues are of less critical importance due to low flux levels and short overflux duration.
In conclusion, introducing variable shunt reactors in the hybrid line design requires special attention for alternatives A, B and possibly C. No such issue exists for Alternative D.