The paper presents numerical simulations related to the problem of how to obtain correct results in transonic wind tunnel during tests at high airfoil angles of attack. At this flow conditions, significant pressure losses appear in the test section, what leads to significant errors in measured data. Regarding the possible ways of tunnel reconstruction, we examined three different possibilities of changing the test section configurations: an increase of the test section height, displacement of the airfoil below the tunnel centreline and, finally, introduction of divergent test section walls. It was shown that neither the use of higher test section, nor the change of the airfoil location, gives any significant improvement in reference to the existing tunnel configuration. Only after divergent test section walls were introduced, the distributions of pressure coefficient became well consistent with their expected values.