1
Electromagnetic Pulse

Various natural and artificial processes, such as lightning discharges and nuclear explosions, can produce a strong pulse of broad-band electromagnetic radiation called an electromagnetic pulse (EMP). EMP has been the subject of research since World War II, as Fermi anticipated the electromagnetic effects resulting from a nuclear explosion [1]. The large electric fields in such a pulse can cause damage to electronic and control equipment. The generation of EMP during nuclear tests was first observed in the 1950s, where it sometimes resulted in instrumentation failure [2]. EMP occurring in lightning discharges, and during fast switching of high-voltage circuits, is also known to cause damage to electrical and electronics systems. The experimental and theoretical study of different sources of EMP, and their effects on systems, is an active field of study around the world [2].

1.1 Sources of EMP

There are various natural and artificial sources of EMP. A common natural source is lightning. Artificial sources include high-voltage fast switches, power stations and distribution systems, nuclear explosions, ultra-wideband radar, etc. EMP generated by lightning is called lightning electromagnetic pulse (LEMP), while that due to nuclear explosions is called nuclear electromagnetic pulse (NEMP). More details are available in [3].

Figure 1.1 illustrates the basic mechanism of NEMP generation. A nuclear detonation releases a stream of energetic gamma-ray photons. This primary gamma, , produces Compton electrons following a collision with free electrons available in the atmosphere. The current channel formed by the Compton electrons gives rise to a large d, producing NEMP [1].

Figure 1.2 shows the temporal as well as spectral waveforms of LEMP and NEMP. This figure is adapted from [4]. The electromagnetic fields in a NEMP follow a double-exponential temporal waveform given by [5]:

where , and are constants that govern the amplitude, inverse of rise and fall times, respectively. The rise-time and pulse-width of NEMP are of the order of nanoseconds and microseconds. For LEMP, these parameters are typically microsecond and millisecond, respectively. Both have an ultra-wideband nature.

Figure 1.1 Schematic of basic mechanism for NEMP generation.

Figure 1.2 Temporal and spectral waveform of different kinds of EMP [4].

1.2 EMP Coupling and its Effects

We have seen that EMP from different sources covers a broad range of the electromagnetic spectrum, with frequencies ranging from a few hertz to hundreds of megahertz. This corresponds to a wide range of free-space wavelengths. The longer wavelengths can couple to large objects such as overhead transmission lines, while small wavelengths couple to small objects such as control equipment and semiconductor devices. The coupling mechanism can be divided into two broad types, viz. "front door" and "back door" coupling. Front door coupling refers to energy that enters through the antennas of systems containing a receiver or transmitter. Back door coupling denotes energy that leaks into systems through apertures and seams in their enclosures [6].

The amount of front-door coupling depends upon the design frequency of the antenna and is maximum around its bandwidth. Back-door coupling through apertures and vents is maximum for wavelengths of the order of the aperture size and falls off steadily with increase in the wavelength. Figure 1.3 shows a schematic of front- and back- door coupling of EMP generated following a nuclear detonation, to electronic and electrical equipment [3]. EMP can enter the enclosure through overhead and underground transmission lines, telephone lines, windows, as well as utility ducts.

The high-intensity transient voltages and currents induced in electrical/electronic appliances can cause damage. The damage can be either temporary or permanent, depending upon the intensity of the incident pulse and the hardness of the exposed system [3].

1.3 EMP Simulators

A number of laboratories around the world have developed EMP simulators that can produce pulses of different types, with the objective of testing the susceptibility of systems exposed to EMP [7]. These are generally driven by a high-voltage pulsed-power source, e.g. a Marx capacitor bank. These simulators are used in two ways. The first is for assessing the effects of EMP on systems. The second is for testing the effectiveness of shielding ("hardening") of these systems.

Simulators can be divided into two broad categories: bounded-wave (closed) and radiate-wave (open). In a radiate-wave simulator, an ultra-wideband antenna, e.g. transverse electro-magnetic (TEM) horn, is used to radiate the electromagnetic field. This type of EMP simulator is used when systems to be tested are spread over a wide area [8]. Bounded-wave type simulators, with which this study is concerned, concentrate energy within the workspace of the system itself [9].

Figure 1.3 Schematic showing EMP coupling to electrical and electronics. This schematic is taken from [3].

(Source: Ghose [3]. 1984, Don White Consultants.)

A bounded-wave EMP simulator, in its simplest form, consists of two electrically conducting triangular plates, making up a TEM structure, separated by a parallel plate region [10]. This is illustrated schematically in Figure 1.4.

The front plate, which displays a near-constant impedance over a wide frequency range, plays a significant role in determining the EMP waveform, while the middle and rear plates serve to guide the signal [10]. The object to be tested is mounted in the bounded volume of the parallel-plate region.

There are several variants of the geometry shown in Figure 1.4. Some simulators do not have the rear plate, while others dispense with the parallel-plate section as well. The tapered section could also have some other shape, e.g. conical.

Figure 1.5 shows the setup of a bounded-wave EMP simulator, details of which have been reported in Ref. [9]. Only the tapered section of the simulator is shown here - the test section consists of a parallel-plate section, several meters in length.

Figure 1.4 Schematic of parallel-plate transmission-line type of EMP simulator.

(Source: Adapted from Giri et al. [11].)

Figure 1.5 Experimental setup of a bounded-wave EMP simulator.

(Source: Adapted from Schilling et al. [9].)

1.4 Review of Earlier Work

In this section, we examine earlier work in different areas relevant to modeling of EMP simulators.

We first consider earlier work involving overall analysis of bounded-wave simulator performance. Several time- and frequency-domain models have been reported. However, these analyses are based on several simplifying assumptions. For example, the conducting plates of a simulator have been approximated by wire grids or meshes. The current induced on the wires has been solved in the time-domain using a space-time-domain technique [12]. It has also been solved in the frequency domain using the method-of-moments (MOM) [13].

The transient electromagnetic field distribution inside a simulator has been studied through a space-time-domain technique [12]. The problem was formulated in terms of the radiation of a transient waveform from perfectly conducting wires, which involves the computation of the induced currents on the wires. This has been solved using the space-time-integral equation [14]. King and others [10, 15] have theoretically analyzed the transient behavior of a rhombic EMP simulator. In their simplified model, simulator plates are approximated by a hexagonal wire structure which is located along the edge of the metal plates. This approximation is based upon the fact that the largest current density in the parallel-plate simulator is found along the edge of the metal plates. Klaasen [16] has numerically analyzed the transient behavior of a bounded-wave simulator using a space-time-domain technique. The basic waveform of the current induced in the wire has been taken from [10]. As compared to King and workers [15], that study more accurately models the plates by increasing the number of wires in the calculation.

Hoo [13] has used MOM for the numerical analysis of a transmission line EMP simulator using a known waveform of electric field excitation. The electromagnetic field structure inside the bounded volume of the simulator was calculated by approximating the current waveform through a triangular basis function. As a check on the numerical method, the input impedance of a triangular dipole was calculated, which shows a fairly good match with experiments.

These methods are not suited for detailed analysis of simulators with test objects, for two...