where E(x,y) = A(x,y) expj<t>(x,y) represents the aperture illumination law. For the illumination laws which we want to take as models, E(x,y) is not with separable variables and it is not possible to achieve an analytical calculus of the field. The methods of direct computation of the two-dimensional integral (gauss, ludwig, romberg) are all very expensive in computation time, especially for apertures of very large dimensions in relationship with the wavelength. The computation method by discretisation of the aperture is very interesting for it allows computation of the field radiated by plane apertures, whatever may be their shape and illumination law. Moreover, to take into account the effect of localized disturbances, it prevents from having to achieve again the global computation of the field radiated by the aperture, by writing that this field is the sum of: - the field radiated by the nondisturbed aperture. - the field radiated by the disturbed part. This method (Fig. 4) consists in subdividing the integration area A into elementary squares Ay inside which it may be considered that the complex amplitude Etx,^) keeps a constant value. In those conditions, one may write formula [2] by taking this term out of the two-dimensional integral, which presents the interest to make this latter with separable variables and even to allow its analytical calculus. We have indeed, successively:
RkJQdWJsaXNoZXIy MTU5NjU0Mg==