Control House And Relay Design Considerations For EMP .

6m ago
961.89 KB
12 Pages

Control House and Relay DesignConsiderations for EMP ResiliencyRoy Mao1Harsh Vardhan1Aaron Ingham2 Barry Howe2 Sarah Pink2 Curtis Birnbach3Mark Adamiak41-GE Grid Automation 2-Trachte 3-Advanced Fusion Systems 4-Adamiak Consulting LLCKeywords: EMP, Control House Design, EMP Signal Levels, HEMP, NNEMPAbstract: In the evolving world of politics, the threat of an Electro Magnetic Pulse (EMP – nuclear or manmade) and its potential effect on the electric power grid has recently received much attention by governmentofficials as well as the power industry. Specifically, an EMP event is identified as a low probability / highconsequence event. There are several types of EMP events that pose this threat, specifically: NuclearElectroMagnetic Pulse (NEMP), including High-altitude ElectroMagnetic Pulse (HEMP), Non-NuclearEMP (NNEMP), Radio Frequency Weapons (RFW), Lightning, and plasma waves (similar to GeoMagnetic Disturbances - GMD).Of particular concern for electronic equipment is what is defined as an E1 wave which is the leading edgeof the EMP wave. The E1 wave with which the power industry is concerned emits an electric field with afast rise time (about 1 nsec) with a significant radiated field strength – estimated at 50 kV/m. Such a pulsehas been demonstrated to cause catastrophic damage to electrical and electronic equipment by either directradiation or by coupling with conductive materials (e.g. field wiring and transmission lines) and conductingextremely high energy electrical pulses into those equipments. Such conductive materials include metalcontrol houses, equipment enclosures and housings, and power and signal conductors (wires and cables)entering and exiting such equipment.There are a number of mitigation strategies available to the utility industry including hardened control housedesigns with highly-conductive coatings, gasketed entry ways, application of shielded cable, shielded cableentry ways, and EM filtered air intakes. Additionally, application of process bus technologies enables theoptical substation where all entries into the control house are fiber optic. Devices themselves can behardened through the application of fast high-energy absorption devices.EMP OverviewAn Electro Magnetic Pulse (EMP) can best be described as an Electro-Magnetic Tsunami released as aresult of the rapid release of a large amount of energy. The typical association of EMP is with the detonationof a nuclear device – resulting in what is known as a Nuclear EMP or NEMP. A NEMP event results inthree different types of electro-magnetic waves with significantly different time horizons as shown in Figure1 (from the Homeland Security document1 on this topic). The first wave to be emitted is termed the E1wave. This wave has a rise time of less than 1 nsec and an electric field strength of 50,000 V/m which isalso the reference value in the IEC Generic Standard2 on this topic. The follow-up wave, known as the E2wave has the characteristics of lightning which is a wave front 1000 times slower than the E1 wave oraround 1 µsec. The third wave resulting from a NEMP detonation is known as the E3 wave and is an almostDC wave of plasma that couples with power lines and looks much like the Coronal Mass Ejection (CME)1

from the sun that results in almost DC coupling to power lines – similar to Geomagnetically InducedCurrents (GIC).Figure 1EMP Waveforms2The scenario of concern is the detonation of an atomic weapon at a high altitude. At a high altitude, theresulting “pulse” can effect a wide area. Such an event is a very low probability but extremely high impact,In addition to a Nuclear EMP, Man-made EMP exists. With Man-made EMP, the focus is on generationof the E1 wave. In this paper, an E1 wave is generated in order to test the resiliency of power systemequipment to the subject event. It should be noted that the Intelligent Electronic Devices (IEDs) on thepower grid today are designed to withstand a lightning strike so the E2 wave is not a concern. Similar toNEMP, the Man-made (Non-nuclear EMP - NNEMP) starts with a source of stored energy which is thenrapidly discharged through a tuned antenna. Design of the generator must take into account wave guideand antenna effects in order to optimize the E1 wave generation. A picture of an E1 pulse is shown inFigure 2. In order to perform these tests, a large Faraday Cage, (80’L x 40’W x 20’H in size) was used tocontain the E1 pulse. The chamber was instrumented with multiple Tektronix 72504 Digital Oscilloscopes(25GHz instantaneous bandwidth on 4 simultaneous channels each) were used to connect to the D dot andB dot probes (for measurement of e-field and b-field respectively) in the chamber and as located inside thecontrol house.2

Figure 2E1 Electric Field Wave ShapeRadiated vs. ConductedAlthough the focus of the design criteria in this paper is on radiated EMP, conducted EMP fromtransmission lines is also possible. Measurement devices such as Current Transformers and PotentialTransformers – due to parasitic capacitances – will conduct EMP into the terminals of an IED. Figure 3aand 3b (below) is a circuit model of a Capacitive Coupled Voltage Transformer (CCVT) and its associatedfrequency response. The plot only extends to 10kHz, however it can be seen in figure 3b that highfrequency signal is still passed through to the relay, albeit, with about a 10x attenuation.Figure 3a CCVT Model3

Modeled CCVT Frequency Response Curve111010000Gain e 3b CCVT Frequency ResponseGeneral EMP Mitigation TechniquesOne mitigation strategy for conducted EMP is the use of optical voltage and current sensors. Thesedevices completely decouple the power line from the IEDs. High frequency signals may bemodulated into light signals, however, when converted to an analog signal, the Anti Aliasing filterin the Merging Unit removes any of these frequencies.Mitigation of EMP and other high level electromagnetic interference signals from the power gridis a decidedly non-trivial activity. It requires close attention to detail. The Faraday Cage, acontinuous electrical shield which completely surrounds a device, room, building, etc., and has noapertures other than those absolutely necessary to bring power, signal., control and data into theshielded is most often used to protect high value electric and electronic devices and systems. Itshould be noted that a Faraday cage is only as good as its implementation and as such if not builtcorrectly, or if cable penetrations are not installed meticulously, will probably leak.A Faraday cage can be made from rolled and crimped steel, however, additional attenuation canbe achieved through continuous welding of all seams in the cage. Even holes as small as 1/32 ofan inch must be sealed. Particular care must be taken around doors and ventilation units. Thereis an excellent specification for the construction of honeycomb EMI filters found in MIL-STD188-1254 and MIL-Std-4615.Electrical lines entering a Faraday cage must be grounded and bypassed or otherwise filtered. Thisis a non-negotiable requirement. Where possible and practical, electronic equipment inside aFaraday cage should be battery powered. This is because it is not unusual for interference signalsto be transmitted into a shielded enclosure by the grounding system.4

The grounding system itself requires substantial attention. There are two schools of thought here. Onestates that the entire enclosure should have a single very low impedance ground (typically a wide copperstrap of six inches or more in width) going to a welded or soldered connection to an underground groundinggrid. Typical electrical practice uses a heavy wire loop buried four or five feet underground that surroundsthe perimeter of the area to be protected. While this does provide signal attenuation to a certain extent, itis a “relatively” high impedance design and does not dissipate fast transients well. A better approach is touse a grid of two-inch-wide copper conductors on two foot centers, with each intersection Heliarc orcadwelded (a thermite based field welding technique). This design is common in data center construction.The Unit under Test was grounded at two corners.Instrumentation is another area where close attention must be paid. It is generally agreed that the use of Ddot probes are best for electric field measurements and B dot probes are best for magnetic fieldmeasurements. These probe designs are well described in the literature, but careful attention needs to bepaid to the data reduction. A common mistake is to take the readings directly from the probes. In bothcases, these probes are differential sensors and their outputs must be carefully integrated to produce anaccurate answer.It is frequently difficult to eliminate unwanted noise from fast transient measurements of high level signals,even when fiber optics are used. If noise is a problem, then the use of a noise channel which is a completesignal channel that has a termination instead of a probe on it, is used and the resulting signal is subtractedfrom the probe signal to produce a clean signal for display or further processing (such as integration).Coaxial cable should be used with extreme caution as most coax leaks signals through its shields. The onlyform of coax that is largely immune to this is coax with a solid continuous outer shield, generally referredto as Heliax. While substantially more expensive and difficult to terminate, Heliax provides excellent noisefree results. The use of double solid shields can lead to noise attenuation in excess of 120 dB, while a singlesolid shield will produce at least an 80dB reduction in noise.IED Design RecommendationsBeyond normal electronic design for EMI protection, there are a several design practice e can behighlighted for the mitigation of EMP effects on IEDs.High Power/Energy Transient Voltage Suppression (TVS) Diodes can be placed across vulnerableentry/exit points in an IED. A TVS diode features a very fast response and ultra-low clampingcharacteristics over traditional Metal Oxide Varistor (MOV) solutions. TVS diodes were designed toprotect against voltage transient events, like lighting surge, switching transient, and Electrostatic Dischargeand are somewhat useful in the protection of EMP events.For an EMP event, because of the extreme induced peak voltages and currents, standard power TVS diodes,which only clamps a few hundred voltage, are not suitable to absorb the power/energy from an EMPevent. There are, however, higher power/energy TVS diodes which are capable of clamping larger peakvoltages which can be used on the AC current and voltage inputs and power supply inputs to preventdamage. Because of high reliability requirements for IEDs, it is recommended that high reliability TVSdiodes be applied in this application.Shielding and Grounding A metal enclosure of the IED components is recommended to providedshielding from radiated EMP. In addition to metal shielding, a adequate grounding and ground wiring withminimum ground impedance is proposed to reduce the common mode impedance. It is to be noted that aflat braid ground wire has lower inductance than a solid copper wire and its use is a best practice in rackground wiring.5

Error-correcting Code (ECC) memory. In modern high-speed processor and memory electronic designs,a processor may run as fast as a few GHz frequency. The memory access speed can be as high as hundredsMHz frequency. A strong electric field from an EMP event may cause a bit error in such high-speedprocessing. Error-correcting Code memory technology can correct any single bit errors thereby providingimproved immunity to the large electric fields emitted by an EMP event and subsequent processor lock-upand required re-boot.Hardware vs. software watchdog. In the case where an EMP event does result in the latching of variouslogics in a processor and where no hardware damage occurs, a re-start of the processor may be able to clearthe event and bring the IED back on line. The function to detect a latch-up is typically know as a Watchdog.If the Watchdog is not reset in a user-defined interval, it is designed to force a re-start of the system in anattempt to recover. It is strongly recommended that this function reside in Hardware so that a transientevent cannot prevent a system re-start. The re-start should include power reset of the various modules ofthe IED.IED Performance CriteriaWhen tested, an IED’s performance can be classified into 3 categories:Category 1: There were NO IED component failures and the IED continued to operate normallyCategory 2: There were NO IED component failures, however, the IED locked-up and had to be resetCategory 3: There were component failures in the IED and normal operation was no longer possible.Control House Design Criteria for IED Resiliency to EMPFor this test, two Control Houses were provided. One of these houses was a shielded design (See Figure4); the other was identical except it was unshielded, providing a control sample. The shielded control houseprovided a six-sided, electrically-contiguous Faraday cage constructed of solid, highly conductive metals.Corner joints at floor, walls and ceiling were electrically bonded by continuous, mechanically-fastened,conductive brackets. Such corner joints can also be augmented with conductive paste, conductive gasketsor continuous welds, depending on the level of attenuation required. The resulting Faraday cage isgenerically referred to as the shield or shield layer. The typical control house (including this test model)will have multiple conductors entering and exiting, this necessitates that the shield be connected at one ormore points to a low impedance earth ground. All conductors entering the test model were shielded cableswhose shields were bonded to the house ground at the point of entry.6

Figure 4 - Shielded Control House3The shielded control house was equipped with an RF personnel/equipment access d