Power supplies
Fault analysis in power grids
Time differences of up to 200 ms between protection devices and control technology complicate fault analysis in the Knapsack Chemical Park's power grid. A pilot project for PTP time synchronization of protection devices is intended to help.
The 180-hectare Knapsack Chemical Park is home to six chemical producers and more than 20 service companies, which employ a total of 2,200 people, 860 of whom work for the site operator InfraServ Knapsack. The chemical park's electricity grid consists of 61 km of cables at the medium-voltage level and 53 km of cables at the low-voltage level.
© Infraserv KnappsackThe power supply at the Knapsack Chemical Park in Hürth near Cologne is ensured via 6 kV and 20 kV medium-voltage grids as well as 400 V and 500 V low-voltage grids, which are operated by Infraserv Knapsack (ISK). "In total, there are around 2,000 electrical measuring points and 500 medium-voltage fields on the 160-hectare site," explains Markus Kanonenberg, who is responsible for energy technology planning at ISK. "In addition, the medium-voltage grid is fully connected to a control system and can be monitored and switched from the central control center."
In operational practice, faults also occur in the chemical park's grid, the causes and effects of which must be analyzed and rectified as quickly and comprehensively as possible. The operator is naturally interested in ensuring that its 24/7 producing customers are supplied as uninterruptedly as possible. If an event occurs, protection devices, network analyzers and the control system evaluate the recorded fault records.
Time synchronization is a must
In order to assign these fault records correctly and also to be able to correlate the correct sequence of events at different points in the power grid, all components must be precisely and consistently time-synchronized. However, there have always been problems with this in the past: "Most of the components for fault analysis are currently synchronized via the control system using the IEC 60870-5-103 or -104 protocol," says Kanonenberg. "The time master receives the time information from a GPS-synchronized clock, which is installed in the central control station, and forwards the signal to the respective RTU units in the switchgear." The existing installation of the switches and the different cable lengths to the systems lead to considerable differences in runtime.

Phoenix Contact and Mauell agree on cooperation
The Velbert-based company Mauell is spinning off its network control technology division into a separate company in which Phoenix Contact is participating. The aim of the cooperation is to develop intelligent solutions for the energy market of the future.
The time synchronization of the protection technology is carried out via the IEC-60870-5-103 protocol and NTP from the central control system.
© Omicron LabThese differences are typical for classic telegram-based time synchronization in accordance with IEC 60870-5-103. With this method, which has also been used by the protection devices in the chemical park to date, the signal delay is not compensated. Further delays are added during the distribution and processing of the time signals, so that in reality differences in the range of 100 ms to 200 ms are ultimately possible.
In addition to the protective relays connected via IEC 60870, network analyzers are synchronized via an Ethernet network using NTP. Although the signal propagation time is partially compensated here, the synchronization accuracy depends largely on the number of switches in the network and the network load. This means that the accuracy that can be achieved with NTP varies from a few milliseconds to 50 ms and more, depending on the network infrastructure. As a result, fault records from devices cannot be superimposed and compared without manual post-processing. An exact analysis of the sequence of events in the event of a fault is only possible if they occur at intervals of several 100 ms.
PTP network with decentralized time master
In order to avoid long time delays in the network in future, ISK decided to set up a PTP network as part of a pilot project. PTP and NTP are protocol-based methods of time synchronization and work according to a similar principle. In contrast to NTP, PTP uses special network components so that an accuracy of better than 1 µs can be achieved.
A PTP network was set up as part of the pilot project. The time synchronization of the protection technology is carried out using PTP components from Omicron Lab.
© Omicron LabThe PTP network set up in the chemical park consists of two special synchronization solutions from Omicron Lab - the PTP time reference 'OTMC 100' and the PTP time converter 'Ticro 100' - as well as the PTP switch 'RSP20' from Hirschmann. The 'OTMC 100' acts as a PTP grandmaster in accordance with IEEE 1588 PTP and also serves as an NTP time server. A special feature of the device is that the GPS antenna, the GPS receiver and the actual time reference are located in a weatherproof housing optimized for pole mounting. All data communication and power supply is provided via an IP67 Ethernet connector on the underside of the antenna housing. The Grandmaster is supplied with power via a PoE injector. The signal propagation time between the 'OTMC 100' and the PTP switch is automatically compensated. Thanks to the GPS connection, the device achieves a time accuracy of better than ±100 ns compared to universal time (UTC).
Synchronization of non-PTP-capable components
However, most components currently installed in energy technology, such as protection devices or fault recorders, do not support time synchronization using PTP. However, they often support synchronization using time signals such as DCF77 or IRIG B. With the PTP time converter 'Ticro 100', these devices can be integrated into the PTP network. The Ticro 100 is connected to the Grandmaster via PTP and provides the locally required time signals. Depending on how the connected device is synchronized, the time converter provides the required time reference signal at two electrical, two optical and one optocoupler output. Each input can be configured individually so that one time converter can synchronize several protection devices from different manufacturers using different time signals. The time converter itself can be integrated into the PTP network by wire or fiber optic via an LC connection. This allows distances of 100 m (electrical) or up to 2 km (fiber optic) to the next PTP switch to be bridged. The signal propagation times are compensated here using PTP.
This means that a delay difference only occurs over the short distance between the time converter and the protection device. The DCF77 supply line to the protection devices in the Chemical Park Knapsack has a total length of approximately 24 m, so that the distances between Ticro 100 and the protection devices are between 2 m and 16 m. The time offset of 10 ns to 80 ns that occurs over this distance is irrelevant and negligible in relation to the entire system. The third essential component is the PTP switch. It makes the PTP network expandable.
Maximum 1 ms time difference
The 'Ticro 100' time converter enables existing equipment to be connected to the PTP network.
© Omicron LabFor the pilot project, a medium-voltage switchgear with nine protective relays was initially synchronized by a 'Ticro 100' using a DCF77 time signal. With the help of a second time converter, grid analyzers were also synchronized via the IRIG-B time signal. Another advantage of the IRIG-B and DCF77 signal is that it cannot be used for bidirectional communication and only transmits the time information. This means that a time signal can be taken from a secure network into a device without opening the network. As part of the PTP network configuration and commissioning, the settings for the OTMC 100, the PTP switch and the two Ticro 100s were first made via the web interface. The corresponding IP addresses were assigned and the protocols for time synchronization were set up. "To test the accuracy of the system, we simulated simultaneous grid faults on various protection devices and compared the time stamps of the fault messages to determine the time deviation," explains Florian Fink, application engineer for protection test technology at Omicron electronics. During the test, the time synchronization of some protection relays was switched to the DCF77 signal, while other protection devices remained synchronized with the control system. Using the CMC test device and the 'RelaySimTest' software, a fault event was triggered synchronously in three protection relays. The fault messages from the protective relays were then read out and compared. "This showed that synchronization using IEC 60870 works quite well in a single system. Time differences of a maximum of 2 ms were detected," says Fink. "However, a comparison with the absolute time - UTC - showed a difference of up to 17 ms." These differences in the range of one period of the mains voltage can already cause problems when analyzing events.
Another test focused on how good the synchronization via the control system is in two different switchgears. "Testing the central synchronization of the protective relays using IEC 60870 revealed a significant time offset of up to 82 ms between the two switchgear systems," explains Fink. "A comparison with the absolute time showed a difference of up to 172 ms." For verification purposes, tests were also carried out in which the protective relays in both systems were synchronized via PTP using Ticro 100. A temporary PTP network consisting of OTMC 100, RSP20 and Ticro 100' was set up for this purpose. "A barely visible time difference of 1 ms was determined between the two systems. This time difference is due to the time resolution of the protection devices used. This was confirmation for us that the effort to switch to PTP was worthwhile," explains Kanonenberg.
Author:
Wolfgang Schenk works as a sales and application engineer for IEEE 1588 PTP Timing Solutions at Omicron Lab in Klaus, Austria.













