Latest posts by Fiona Richardson (see all)
- FEKO Website Content Migration to Altair Locations - March 25, 2018
- Designing EMC Compliant Products for a Connected World - February 26, 2018
- Why Join the Electromagnetic Support Forums? - September 20, 2017
An analysis of electromagnetic pulse radiation in indoor environments.
Behaviour and propagation of electromagnetic fields in indoor environments are relevant to a range of technologies including through-wall radar devices, shielding of critical equipment in hospitals and the analysis of lightning strikes or electromagnetic pulses (EMP) incident on buildings.
In this white paper we will investigate the behaviour a high power EM source typical of an EMP incident on a two storey building. Various simulation approaches are discussed such that the building structure can be analysed accurately and efficiently. Results are presented in frequency and time domain to give insight into how the pulse propagates through the structure and identifies areas in the building which may be susceptible to higher fields.
The geometry of the building is shown below. Three different materials are used for the simulation, namely concrete, wall board and metallic reinforcement bars (rebars).
The size of the building is 25 x 18 x 10.5 meters. A plane wave excitation is used as the EMP source to estimate the EMP behaviour of the structure.
In order to reconstruct the time domain pulse accurately, a frequency spectrum up to 200 MHz is chosen (< -35 dB) with 21 discrete frequency points. The wavelength at 200 MHz is 1.5 m, implying that at the highest frequency the building is tens of wavelengths large. Initial simulations are performed using Method of Moments, applying the thin dielectric approximation to the walls and floors of the building. However, simulation with the finite difference time domain (FDTD) method leads to a much more efficient approach:
- a single FDTD simulation is sufficient for the time domain pulse analysis
- the voxel mesh is more suitable to capture the geometric details of the building especially when including the rebars
- applying GPU acceleration enables the structure to be simulated in several minutes
While FDTD remains the most efficient approach here, applying different methods for cross validation purposes builds confidence in such a numerical analysis.
An average voxel mesh of 10 cm is used, resulting in a total grid size 26 million FDTD cells, which runs in 4 min 26 sec on an NVIDIA Tesla K20c.
The E- and H-field are calculated at different planes within the building (a spatial resolution of 0.5 m is used for the near-field requests).
The tendency is that at lower frequencies the fields are reflected, but as the frequency increases there are more creeping and propagating waves around and through the building. Above 100 MHz, higher fields are seen inside the building.
After the simulations are run, the time analysis is performed in POSTFEKO using the defined pulse. The advantage of this approach is that it is possible to change the pulse waveform definition without needing to rerun the simulation. The following two movies show the animations of the pulsed E- and H-fields respectively, illustrating how the field propagates through the building.
This pulse exhibits a waveform that is typical for EMPs . The time domain waveforms of the pulse are shown below.
pulsed E-field animation for EMP analysis
pulsed H-field animation for EMP analysis
For a more quantitative comparison, the pulsed field can be probed at specific locations in and around the building and plotted as a function of time. The influence of scattering and superposition is more prominent at some probe locations than others. Constructive and destructive interference occurs.
An efficient approach was illustrated how a building tens of wavelengths in size can be simulated, capturing frequency and time domain aspects of how a typical EMP signal will propagate through the structure.
1. M. D’Amore et al., Theoretical and Experimental Characterization of the EMP Interaction with Composite-Metallic Enclosures, IEEE Trans. EMC, vol. 42, no. 1, February 2000.
Visit the website for more information