Mitigating Transients on PLC-Based Control Systems

The programmable logic controller (PLC) is an important component in most modern automation and process control systems today. It's also susceptible to lightning-induced transients. To verify the extent of this susceptibility, we conducted laboratory tests on five different PLC brands (generically labeled A, B, C, D, and E). We subjected each PLC to transient conditions with and without power conditioning to understand their respective response. We loaded each PLC unit with a common control algorithm, used 120VAC input, output, and power supply modules, and wired each in an identical scheme.

To perform the lightning-induced transient test in a laboratory environment, we used a Keytek 801-S surge generator to create ANSI-standard combination waveforms. The generator delivers the combination wave by applying a 1.2/50µs voltage wave across an open circuit and an 8/20µs current wave into a short circuit. We can determine the exact waveform delivered by the equipment impedance to which the surge is applied.

For these tests, we used IEEE C62.41 Cat. B surge magnitudes because this category pertains to electrical loads connected to bus and feeders in industrial plants. The maximum value of the Cat. B test waveforms for voltage and current are defined at 4kV and 2kA, respectively. These two waveforms have substantial energy-deposition capability and provide representative stresses to the surge protectors and commercial electronics connected to the power system.

We conducted these surge tests in line-to-neutral and line-to-ground configurations. The first series of tests were with power conditioning, as per the test setup shown in Fig. 1. The power conditioning consisted of a 500VA constant-voltage transformer (CVT).

Test Results

A 4.27kV line-to-neutral transient surge had no effect on any of the five tested PLCs with the CVT in place. An example of this surge waveform is shown in Fig. 2, on page 28. Channel 2 is the output of the CVT. Notice that the surge transient (Channel 1) is not reflective in Channel 2. This indicates that the surge transient at this magnitude does not affect the output of the CVT or the PLC.

The line-to-ground surge transient affected the primary side of the CVT. Fig. 3, on page 28, displays a 3kV surge, which caused a “flashover” on the primary side of the CVT (voltage exceeded the scope's measurement scale.) The secondary of the CVT was unaffected by this flashover event.

We performed additional transient surge tests, this time without power conditioning in front of the PLCs. The test setup is shown in Fig. 4.

We subjected PLC Models A and B to lightning strike transients without the benefit of external power conditioning (Fig. 4). Tested in this mode, only internal surge suppressor protection from metal oxide varistors (MOVs) provided mitigation. PLC A was able to survive surges up to 3kV, but the power supply suffered critical damage from a 3.5kV line-to-neutral transient surge (waveform captured in Fig. 5, on page 30).

Unlike PLC A, PLC B was unaffected by transient surges up to 4kV line-to-ground. As shown in Fig. 6, on page 30, the PLC power supply module clamped the 4kV transient to 1.24kV.

We conducted further examination of the MOV surge suppression scheme in PLC A and B's power supplies following the test to determine why PLC A's power supply failed and PLC B's power supply did not. The photos, on page 30, show the MOV placement in the two supplies. In PLC A, the designer chose to use two MOVs wired in a line-to-neutral and neutral-to-ground scheme. Furthermore, the incoming power is routed into the power supply solder trace past several electronic elements before it reaches the surge suppression devices. Inspection of the power supply after the unit failed revealed that the main switching transistor was damaged and a nearby resistor had opened. This probably occurred due to flashover of incoming surge voltage.

In contrast, PLC B's surge suppression design is more thorough. Three MOVs are used in the power supply to suppress line-to-neutral, neutral-to-ground, and line-to-ground surge events. Furthermore, the MOVs are in very close proximity to the incoming AC power source. Therefore, the MOV provided surge mitigation with a lower risk of exposing other power supply components to the transient. The MOVs in this power supply were marked with known designations. Therefore, it was possible to determine that the devices were designed to clamp at around 800V at 2000A.

Conclusion

The lightning-induced transient tests indicate that PLCs are susceptible to these surge events, possibly resulting in damage to the I/O rack power supply. The design of the PLC power supply, along with the surge suppressor and filter protection scheme, can lead to increased survivability for the system. However, if you can mitigate the transient prior to reaching the PLC system, the odds for surviving such events increases dramatically.

A constant voltage transformer acts as an isolation transformer and provides voltage sag, swell, and transient mitigation. Laboratory test demonstrated the ability of an off-the-shelf CVT to suppress both capacitor switching transients as well as lightning induced transients up to 4kV line-to-neutral and 3.5kV line-to-ground. Although the surge event led to a flashover between the line and ground conductors at the 3.5kV level, the output of the CVT remained steady.

Based on the results of this testing, it's apparent that although a PLC with well designed surge mitigation scheme can survive transient events without damage to its power supply, it does not ensure that the other control system components such as sensors, relays, and contactors will not be damaged or lead to a process aberration. For this reason, we recommend a comprehensive strategy to provide mitigation to the PLC power supply and the control power for the system. Using a power conditioner such as a CVT in this manner will improve the ability of the control system to go unscathed when transient events occur.

We also think future testing be performed to determine the effect of ground referencing and unregulated potentials between I/O card wiring and the power supply. Finally, we believe research needs to be done to understand the different mode of transient coupling on data lines to verify the effectiveness of different data line surge protectors for control systems application.

Mark Stephens is an engineering manager at EPRI PEAC Cor. He can be reached at [email protected] or via phone 865-218-8022.

Kermit Phipps is a senior technician at EPRI PEAC Corp. He can be reached at [email protected] or via phone at 865-218-8021.

For more information about EPRI PEAC Corp., visit www.epri-peac.com or www.f47testing.com.

Acknowledgments

This series is based on work conducted by EPRI and funded by the California Energy Commission. The information presented would not have been possible without the diligent efforts of Promad Kulkarni of the commission.