It is not the incident plane wave that you can see, it is the shadow of the structure. The presence of the structure creates a shadow behind it. This requires a field of opposite sign to the incident field, and with the strength required to reduce the incident field to the level in the shadow. This field is radiated forward and creates some of the scattering pattern you can see. As it travels forward it spreads out and the shadow becomes dappled and less intense. Eventually it turns into a lobed radiation pattern that you see. There is no way of stopping this. You may be able to calculate what it is or work out how to simulate it in CST without the feature you are looking for. Then you can subtract it out of your results. The pattern is the Fourier transform of the field that must be added to the incident field to produce the field directly behind the object.

It is the cause of forward scatter Radar Cross Section.

The polar diagram of the incident plane wave is a delta function in the forward direction, not a polar diagram with lobes.

It may be there is a reflection from the front face or back faces of your structure, or loss through it. These will generate a reduction in field strength and cause forward scatter.

Thank you for the answer,

But I don't understand if there is a way to remove it by doing some operations on CST. For example getting directly the shadow contribute and then subtracting it from the farfield.

I don't think CST has this as a built in feature, or even as a post-processing macro. I expect you will have to try to do it yourself. The first step I would take is to look at the peak forward scattering to get the gain, I would then simulate the forward scattering from a PEC plate the same shape as your lens, and reduce it to the same forward peak level by division, maybe complex division (if the phase isn't right!), then subtract it from your results. This will give a first try that might be good enough. I think there are lots of reasons why this won't be exact, but it might be a good first try.

Thank you again! I will try it.

Do you think that is a good idea remove it by permorming the following steps?

1- Simulate the full structure and take the Bistatic RCS in Linear values (m2).

2- Simulate just a plate of almost the same width of the structure. (In this way maybe just the contribute of the shadow is obtained)

3- Subtract the 2 from the 1.

I've already performed these steps and I will share my results:

I think that is sensible.

You have managed to cancel the lobe by about 15 dB, which is fair, but not spectacular cancellation. 20 dB is goodish, and 40 dB would be very difficult. I'm not sure how I would go about trying to improve it, though.

The question is "Is it good enough for you?" Being able to explain it and being reasonably sure that it doesn't affect the lobe you want much is also important. It looks as if your correction hardly affected the size of the diffracted lobe at all, and any further corrections should have even less effect on it.

I'm slightly surprised that subtracting the forward scatter hasn't affected the diffracted lobe more. The forward scatter looks about 20 dB down on the diffracted lobe, so has the potential to give about 2 dB ripple.

It does look like you have done your subtraction using the complex data, as you should have done, but it would be worth checking again that you have done it correctly.

I think you should use the same width of structure, in the direction of the lobe, because that will decide the period of the sidelobe structure you need to subtract out. If it is a rectangular structure you can use the other dimension to vary the magnitude to see if you can improve the cancellation a bit.