May 16, 2017
in
Ground-fault

Transient EF-Relay - How to test correctly?

W

ELCOME Dear friends of protection and control engineering! In our new three-part series of technical papers, we devote ourselves intensively and extensively to the issue of earth-fault protection. Ground fault protection systems may not be one of the most complicated protective functions, but in practice there are always problems. There are many possible sources of error when it comes to verifying the correct directional decision. This and the two following articles provide simple tips on how to avoid errors when checking the ground fault direction detection.

In our second part, we will show you the most common methods for testing transient earth fault relays.

Part 2: Testing of transient earth fault relays

Ground fault relays detect the direction of an earth fault on the basis of the transient transition process from the fault-free network to the ground-fault-affected network. In this case, they use, in particular, the first deflection of the current and the voltage.

The figure above shows the trend of the current sum of all three phases (red) as well as the course of the star-point displacement voltage (blue). It's quiet easy and clear to see, that the highest current and voltage peaks are created directly at the start of the earth fault. Usually, the direction measurement is startet when the sum current and the displacement voltage exceed a certain threshold value. Then the polarity of the first deflection is compared. If both have different signs, the error is in the forward direction. However, if the sum current and the displacement voltage have the same sign, the error is in the reverse direction.

For the current measurement, transient earth-fault relays are often connected via "Holmgreen-connection" (see part 1 of this series). Since the Holmgreen-connection circuit is realized by the parallel connection of the three phase current transformers, every phase current transformer has to be of the same construction and tolerance-tested. This is not so critical for transient EF relays, because at the moment of earth-fault-start the currents are many times higher than the final continuous earth current. For this reason, transient EF relays can also be used in networks and systems, in which the use of a split core current transformer is very complicated or even not possible at all (for example in 110 kV overhead line networks).

The voltage measurement is usually realized either by measuring the open delta winding or by connecting all three phase-to-earth voltages. During the test, care must be taken to connect the voltage signals as well as the current signals correctly.

Testing with static variables

In order to check the directional decision of transient earth-fault relays, for example, static values can be used. This means that the amount, angle and frequency of the output variables are set manually. At first glance, this procedure appears simple. But the correct definition of these variables is what it makes challenging.

The signal trends at the moment of the earth fault do not have the fundamental frequency of the network (50 Hz) but are composed out of a superposition of high-frequency oscillations. Even if the protection device can be moved with a certain signal to the expected response, this can be quite different for another type of device. This is because the detailed realization of the protection function of existing protective device types differs. Testing with static variables is, in principle, possible, but not always the simplest and most reliable method.

Network Simulation

Alternatively, the test signals can also be generated via a network simulation. On the basis of the simulation, we are able to correctly calculate the transient response. For this purpose, however, the electrical parameters of a network must be known. The network must be at least off:

🌐 An infeed (including the Petersenspule/Coil)

🌐 A line / cable (including phase-earth capacitances)

🌐 An element that summarizes all the remaining capacities of the network

A network model is then created from these elements. Suitable network calculation programs, e.g. Such as PowerFactory, PSS Sincal, ATPDesigner, etc. We are now able to calculate the signals and then export them in Comtrade format. In this case, it is very important to use a sufficiently high sampling rate to correctly map higher-frequency oscillations. For the purpose of checking the earth-fault relay, the last step requires that these exported signals be played by a testing device.

In order to get to the parameters of the network, we can work well with experience in practice. Thus, the parameters of each individual network section need not to be known. An actual disadvantage of this method, however, is the fact that the simulation with network calculation programs is not quite trivial. Furthermore, protection engineers often do not have direct access to these programs, even if they are used in the company.

Therefore, it is a good idea to use tools that are tailored to the protection test, making them easier to use for the test engineer. These tools usually allow the direct output of the signals with a connected test device, without having to complete further intermediate steps. One option here is e.g. the OMICRON transient ground fault test module. The transient response of a ground fault can be reliably simulated using the module. Again, the creation of the network parameters does not appear very easy, but appropriate signals can already be generated via the default settings. Important: This network is limited to a source with a transformer, a faulty line, a fault-free line and the rest network.

For more flexibility, the RelaySimTest test program from OMICRON can also be used. In RelaySim test, any network can be created and calculated. For users who do not have access to the necessary network parameters, executable templates are available here. In addition, this program is capable of simulating transient ground fault scenarios, such as short ground fault wipers, or earth faults, which are self-extinguishing, but always returning, without great effort.

Conclusion

Considering the above-mentioned aspects, it can be summarized as follows: While it is possible to test the directional decision of ground fault relays with static signals, the test is often more reliable by means of transient network simulation. Here, above all, tools are available that are easy to operate and which are, in the best case, integrated directly into the test software. Good templates and useful default values ​​are helpful. This makes it possible to test complex error sequences with little effort in the field and on site.

For the sake of completeness, it should be mentioned that there is also the possibility to reproduce real earth faults. If records of previous earth faults are present, these can also be played back and fed in, using modern test facilities.

Here ends our 2nd part of the test of earth fault relays. Part 3 then proceeds with the "Optimum testing of wattmetric earth-fault relays"