Numerical Estimation of the Electric Field Distributions due to Mobile Radio in an Aircraft

Cabin Based on Large Scale FDTD Analysis

Takashi Hikage, Shinya Hiraiwa, and Toshio Nojima Graduate School of Information Science and Technology,

Hokkaido University Sapporo, Japan

{hikage, hiraiwa, nojima}@wtemc.ist.hokudai.ac.jp

Shunichi Futatsumori, Akiko Kohmura, and Yonemoto

Electronic Navigation Research Institute Tokyo, Japan

{futatsumori, kohmura, yonemoto}@enri.go.jp

Abstract—The aim of this study is to develop an accurate and reliable method that can estimate the electromagnetic field distributions in aircraft while contributing to the field of numerical EMI assessment. This paper describes large-scale numerical simulations of the electromagnetic fields excited by 800 MHz cellular radio inside an aircraft. We use a typical aircraft model and employ the parallel FDTD technique to estimate field distributions throughout the cabin. Computational results of the 1-dimensional electric field distribution inside the cabin agree well with measured values. In order to estimate the effect of absorption by the passengers’ bodies inside the cabin, cumulative probabilities for field distributions are derived.

Keywords- cellular radio; aircraft; parallel FDTD analysis; propagation characteristics; cumulative probability

I. INTRODUCTION The usage of wireless communication devices has extended

to more environments such as buses, trains, and aircraft. Recently, some airlines have begun allowing in-flight voice calls. However, some papers have reported that active radio transmitters can impose electromagnetic interference (EMI) on aircraft systems [1, 2]. The aim of this study is to develop an accurate and reliable method of estimating the EMF distributions in aircraft cabins. Given the rapidly increasing variety of mobile communication devices, comprehensive measurements cost too much, and it is difficult to carry them out precisely. Therefore, we proposed to apply large-scale numerical simulations to examine the EMF created by mobile radios [3]-[6]. The FDTD technique is an efficient way to solve Maxwell’s equations for complex structures [7]-[9]. In addition, a large-scale parallel computing technique based upon several node partitions of a supercomputer is used because of its memory and speed capabilities [10]. It can give us a good perspective within reasonable timeframes. This paper uses the parallel FDTD analysis technique to estimate the propagation characteristics of the cabin of a Boeing 777-200 aircraft model. In order to validate the analysis results, the calculated electric field distribution is compared to actual field measurements in a Boeing 777-200 aircraft. Furthermore, to estimate the effect of the absorption by the passengers’ bodies, we derive the cumulative probabilities of electromagnetic distributions.

II. ESTIMATION METHOD AND AIRCRAFT MODEL FDTD analysis is applied in order to derive spatial EMF

distributions throughout the cabin of an aircraft that contains lossy materials. The problem space is quantized by Yee cells. The cell size must be small enough to obtain accurate computation results, and generally it is less than one-tenth of the minimum wavelength. Given the size of modern aircraft, FDTD has significant computational resource requirements. Therefore, we employ a supercomputer to estimate the EMF in a Boeing 777-200 model. Figs. 1 and 2 show the FDTD model and the aircraft cabin configuration used in this study, respectively. A 800 MHz cellular radio simulator, a vertical polarized half-wavelength dipole antenna located 1.0 m above the floor, is assumed to be placed at the front of the cabin as shown in Fig. 2. The dimensions of the aircraft used in the analysis model are as accurate as possible. Because our research interest of this paper is EMF distribution inside the cabin, the wings are not modeled in the analysis. The dimensions of the aircraft model are: length of 52.3 m, width and height of 6.12 m. Metallic parts of the aircraft model are made of perfect electric conductor (P.E.C.). The window panels are made of 10 mm thick plastic. The cabin is lined with 10 mm thick internal panels. The back and sides of the seats consist of P.E.C. and lossy material. The cabin partitions, lavatories, galleys, and ceiling luggage racks are made of P.E.C. Tables 1 and 2 summarize the FDTD parameters and the parameters of the materials, respectively. The total problem space, including absorbing boundary condition, is 673×673×5298 cells. The memory required to execute the analysis is about 600 GB. Each node of a supercomputer can, when operating individually, be assigned about 100 GB of main memory. Therefore, 6 nodes are used to carry out the EMF analysis of the entire aircraft cabin.

Naruto

Proc. of the 10th Int. Symposium on Electromagnetic Compatibility (EMC Europe 2011), York, UK, September 26-30, 2011

523

(a) Boeing 777-200 model

(b) FDTD Model

Figure 1. FDTD model for Boeing 777-200 Aircraft. (a) shows a typical aircraft model. In this paper, the wings are not modeled. (b) is 2-dimensional

view of the FDTD analytical model.

TABLE I. FDTD PARAMETERS

Figure 2. Cabin configuration.

TABLE II. PARAMETERS OF CABIN MATERIALS

III. FIELD ESTIMATION RESULTS Fig. 3 shows an example of the 2-dimensional electric field

distributions obtained by the FDTD analyses. A vertically polarized wave at -10 dBm at 810.05 MHz is radiated from a half-wavelength dipole, located 1.0 m above the floor in the front of the cabin. Vertical (Ey) polarized electric field distributions on the horizontal plane at the height of 1.0 m from the cabin floor are shown in the figure. In the cabin, the field distribution is very complicated due to the multi-reflection environment.

In order to validate the numerical analyses, field measurements were conducted in an actual Boeing 777-200 aircraft without any passengers. 1-dimensional propagation characteristics inside the cabin were measured. Fig. 4 shows measurement scene and system configuration. The transmitter system consisted of a biconical antenna, band-pass filter, power amplifier, and signal generator. The receiver system consisted of a biconical antenna, band-pass filter, preamplifier, spectrum analyzer, and a computer. The TX-antenna was located 1.0 m above the floor in the front of the cabin, as per the FDTD analysis. Electric field strength was measured by the RX-antenna at the height of 1.0 m from the floor along the length of the cabin. The total number of measurement points was 20.

6.1

m

zy

x

Back Node #1 #2 #3 #4 #5 #6

Problem space Front

Back

Node #1 #2 #3 #4 #5 #6

Problem space

5298 cells

673 cells

z

x

z

y

Top view

Side view

673 cells

Problem space 673×673×5298 cells( x×y×z )

Cell size (cubic) Δ= 10 mm

Frequency 810.05 MHz

Absorbing boundary condition PML (8 layers)

Iteration number 800

Antenna 1/2 λ dipole antenna

Number of guard cells 22( in all directions )

Antenna (at the height of 1.0 m from the cabin floor)

z

yx

window：plastic

partition, rack,galley and lavatory ：PEC

seat：PEC&pad

aircraftbody：PEC

floorPEC PEC

inside panel wall

Media εr σ [S/m]

Free space 1.0 0

Aircraft body - ∞ (P.E.C.)

Seat(metal & pad)

- ∞ (P.E.C.)

2.0 1×10-3

Cabin Partition - ∞ (P.E.C.)

Lavatory - ∞ (P.E.C.)

Galley - ∞ (P.E.C.)

Ceiling luggage rack - ∞ (P.E.C.)

Cabin floor 3.5 5×10-1

Inside panel wall 3.5 5×10-1

Window 2.25 3×10-4

Proc. of the 10th Int. Symposium on Electromagnetic Compatibility (EMC Europe 2011), York, UK, September 26-30, 2011

524

Fig. 5 shows a comparison of the numerical calculation and measurement results. The upper figure shows a 2-dimensional electric field distribution obtained by calculation on the estimation plane. The height of the estimation plane for this figure is 1.0 m from the cabin floor. Here, the lower figure shows the FDTD and measurement results of the 1-dimensional vertical polarized E-field distribution on the red line of the upper figure. Antenna input powers, FDTD and measurements, were normalized to -10 dBm. The vertical axis denotes the electric field strength, whereas the horizontal axis denotes the position from the source on the z-axis. These results confirm that the FDTD calculation is in good agreement with the measured values.

Next, we discuss the effects of absorption created by the passengers’ bodies in the cabin. In order to examine realistic and complicated situations, homogeneous human models were added to our FDTD analysis. Fig. 6 shows the passenger models and seating arrangement examined in this paper. The human model has dielectric constant (ε r) of 50 and loss tangent (tanδ) of 1. The cockpit had 2 pilots. Executive class held 56 passengers, and the premium economy and economy class had 40 passengers and 149 passengers, respectively. Therefore, 247 humans were present, the full capacity. Fig. 7 shows the 1-dimensional electric field distributions inside the cabin containing 247 passengers in a comparison to the results of Fig.5. The frequency, location and input power of the antenna were the same for both cases. The red (blue) line plots the result without (with) passengers. From the figure, we can confirm that the maximum attenuation of the electric field due the passengers is about 25 dB in Economy class, which has a high passenger density.

In addition, cumulative probabilities for electromagnetic distributions inside the cabin both with and without humans are shown in Fig. 8. These are evaluated from 2-dimensional electric field distributions of the whole observation plane, 1.0 m above the floor. From the figure, when 247 passengers are present in the cabin, the field intensity in the area becomes about 25 dB lower than the case with no passengers at the cumulative distribution of 50 %.

IV. CONCLUSIONS We estimated EMF distributions established in the cabin of

a large commercial aircraft due to a 800 MHz mobile radio. Field distributions measured in an actual aircraft were compared to the values yielded by our large-scale numerical analysis. The simulation results, 1-dimensional electric field distribution inside the cabin, agreed well with the measured values. Furthermore, the energy absorption effects of the passengers’ bodies were discussed. Two models, with 247 passengers and without passengers, were developed and used in the numerical analysis of the cumulative distribution. We found that with 247 passengers in the cabin, the electric field was attenuated by about 25 dB from the no-passenger model. We intend to conduct other estimations that consider different positions, more antenna sources, and different types of transmitting antenna.

Figure 3. Example of FDTD analysis result of 2-dimensional electric field distribution.

Figure 4. Measurement scene and system configuration in the cabin of the actual aircraft.

Figure 5. Comparison of the calculation and measurement results for 1-dimensional electric field distributions.

Figure 6. Passenger model and seat arrangement in the cabin.

HighLow

|E|

Antenna

zx

TX antenna RX antenna

Transmitter System Receiver System

HPF

Pre-Amplifier

Spectrum AnalyzerADVANTEST U3751

PC

Biconical Antenna(RX-antenna)

BiconicalAntenna(TX-antenna)

Power AmplifierTHAMWAY T152-206DA

Standard Signal Generator

Agilent N5181A

BPF

0 10 20 30 4060

80

100

120

140

160

Position (m)

E-fe

ild st

reng

th (d

BμV

/m)

FDTDmeasurement

z

x

Antenna HighLow

Input power 0.1 W (20 dBm)

Homogeneous human models

CockpitExecutive class

Premium economy

Economy class

2 pilots 56 passengers40 passengers

149 passengers

seat arrangement in the cabin(247 humans are present in all)

Proc. of the 10th Int. Symposium on Electromagnetic Compatibility (EMC Europe 2011), York, UK, September 26-30, 2011

525

Figure 7. 1-dimensional electric field distributions both with and without passengers.

Figure 8. Cumulative probability of Electric field distributions for both with and without passengers in the cabin.

ACKNOWLEDGMENT Computations and visualizations were performed using the

computer resources provided by the high performance computing system at the Information Initiative Center, Hokkaido University.

REFERENCES [1] W. Kreitmair-Steck and W. Tauber, “Aircraft hazards by using portable

electronic devices (PED),” in Proc. 2002 Int. Wroclaw Symp. Exhib. Electromagnetic Compatibility, pp. 136–41.

[2] E. Ross, “Personal electronic devices and their interference with aircraft systems,” Langley Research Center, Hampton, VA, NASA/CR-2001-210 866, 2001.

[3] T. Hikage, T. Nojima, S. Watanabe and T. Shinozuka, "Electric-Field Distribution Estimation in a Train Carriage due to Cellular Radios in order to Assess the Implantable Cardiac Pacemaker EMI in Semi-Echoic Environments", IEICE TRANS. COMMUN., Vol.E88-B, No.8, pp.3281-3286, Aug. 2005.

[4] Y.Kawahara, T.Kono, Y.Sumi, T.Hikage, T.Nojima, M.Omiya, S.Watanabe and T.Shinozuka, "EMF Excitation Dependency on the Boundary Condition due to Mobile Radios - Parallel FDTD Analysis on the Radio Environment in a Train Carriage -", The 2005 International Symposium on Antennas and Propagation (ISAP2005), FE2-4, pp.1241-1244, Korea, Aug. 2005.

[5] T.Hikage, T. Nojima, M. Omiya, K. Yamamoto,"Numerical Analysis of Electromagnetic Field Distributions in a Typical Aircraft", Proceedings of the 8th Int. Symposium on Electromagnetic Compatibility (EMC Europe 2008), pp. 303-306, Sep. 2008.

[6] S. Hiraiwa, T. Hikage, T. Nojima, S. Futatsumori, A. Kohmura, N. Yonemoto, "Estimation of the Electromagnetic Fields Distribution due to Mobile Radio in a Typical Aircraft Cabin Using Large Scale FDTD Analysis,"Proc. International Symposium on Antennas and Propagation in 2010, pp.547-550, Nov. 2010.

[7] K. S. Yee, “Numerical Solution of Initial Boundary Value of Problems Involving Maxwell’s Equations in Isotropic Media,” IEEE Trans. Antennas Propagation, Vol. 14, No. 5, pp. 302-307, May 1966.

[8] Taflove, Computational Electromagnetics, Artech House, Boston, 1995. [9] D. M. Sullivan, “A Simplified PML for Use with the FDTD Method,”

IEEE Microwave and Guided Wave Lett., vol. 6, no. 2, pp. 97-99, Feb. 1996.

[10] C. Guiffaut and K.Mahdjoubi, “A Parallel FDTD Algorithm Using the MPI Library,” IEEE Antennas and Propagation Mag., vol. 43, no. 2, pp. 94-103, Apr. 2001.

0 10 20 30 4060

80

100

120

140

160

Position (m)

E-fe

ild st

reng

th (d

BμV

/m)

no passenger247 passengers

Input power 0.1 W (20 dBm)

60 70 80 90 100 110 120 130 140 150

0.0001

99.999999.999

99.9999.9

99

90

70

0.0010.01

0.1

1

10

3050

no passengers247 passengers

Cum

ulat

ive

prob

abili

ty (%

)

Electric field intensity (dB)

Proc. of the 10th Int. Symposium on Electromagnetic Compatibility (EMC Europe 2011), York, UK, September 26-30, 2011

526