RCS Measurement and Simulation of Generic Simple Shapes
Introduction
Woo et. al. published measured and simulated results for a range of simple RCS benchmark targets in 1993, [1]:
 The NASA almond
 Simple ogive
 Double ogive
 Conesphere
 Conesphere with a gap between cone and sphere.
These targets are defined mathematically so it is easy to ensure that the correct shape is simulated during RCS benchmarking. These targets were simulated in CADFEKO as to achieve two goals:
 To show that CADFEKO has the means to easily generate simple and complex mathematical shapes (such as the NASA almond) .
 To benchmark FEKO's RCS abilities against accurately measured data.
All 5 targets that were presented in [1] are simulated and the FEKO results are compared with the measured results. In all cases the FEKO result closely matches the measured result.
A note on model construction
For all of the models that will be described, the following steps are taken in addition to the geometry construction:
 All internal faces of the shape is deleted
 All face normals were set to point outward
 All faces were set to be solved using CFIE (Combined Field Integral Equation)
 All models were solved using MLFMM
The CFIE setting helps the MLFMM to obtain convergence more quickly, which results in a reduction in runtime.
NASA Almond
Constructing the model
The NASA almond is not a simple bodyofrevolution (BOR) shape as it is essentially flattened in the zaxis. The mathematical definition of the curves were used to define curves analytically in CADFEKO that represent the curves of the model. The shape is formed by lofting two adjacent lines to form surfaces; this is repeated for all of the defined curves until a quadrant of the almond has been formed. The faces are stitched together to account for any mathematical inaccuracies and to ensure a closed surface. The resulting shape is then mirrored to form a halfalmond and the surfaces stitched together. The halfalmond is then mirrored again to form the complete model. In this way a precisely defined shape can be created in CADFEKO to almost arbitrary levels of accuracy.
Results
Figure 1: NASA Almond RCS Comparisons  

(a) Geometry  (b) 1.19 GHz (both polarisations) 
(c) 7 GHz (HHpolarisation)  (d) 7 GHz (VVpolarisation) 
(d) 9.92 GHz (both polarisations)  

Simple Ogive
Constructing the model
The simple ogive was constructed using a single analytic curve in CADFEKO that was rotated to form the surface. This can be done since the ogive is a BOR shape.
Results
Figure 2: Metallic Ogive RCS Comparisons  

(a) Geometry  (b) 1.18 GHz (both polarisations) 
(c) 9 GHz (both polarisations)  
DoubleOgive
Constructing the model
The double ogive was constructed in the same way as the simple ogive. The only exception is that two analytic curves were required to describe the front and rear points of the shape.
Results
Figure 3: Metallic Double Ogive RCS Comparisons  

(a) Geometry  (b) 1.57 GHz (both polarisations) 
(c) 9 GHz (both polarisations)  
ConeSphere
Constructing the model
The conesphere was created by using the sphere and cone primitives in CADFEKO.
Results
Figure 4: ConeSphere RCS Comparisons  

(a) Geometry  (b) 869 MHz (both polarisations) 
(c) 9 GHz (HHpolarisation)  (d) 9 GHz (VVpolarisation) 
ConeSphere with Gap
Constructing the model
The gap that is inserted between the conical and spherical components of the shape is represented by a cylinder.
Results
Figure 5: Metallic ConeSphere with Gap RCS Comparisons  

(a) Geometry  (b) 869 MHz (both polarisations) 
(c) 9 GHz (both polarisations)  
References
[1] 
A. C. Woo, H. T. G. Wang, and M. J. Schuh, "Benchmark Radar Targets for the Validation of Computational Electromagnetics Programs," IEEE Antennas and Propagation Magazine, vol. 35, no. 1, February 1993, pp. 84  89. 