Passive long-range sonars on board submarines are often limited in performance due to low frequency selfnoise, which originates from the machinery of the own ship. Obviously, this self-noise has the same characteristics as the noise of possible submarine targets and manifests itself as loud “directional” noise on the flank array of the submarine. It is known from many studies that such noise can be reduced by adaptive beamforming techniques as: MVDR beamforming, Fourier integral method or inverse beamforming. In this study several algorithms for adaptive beamforming have been implemented and tuned for optimum performance. These algorithms are applied to flank array data. The beamforming performance of the different algorithms is compared. Although it is hard to find objective criteria for the performance of totally different beamformers, it is found that in many environments adaptive beamformers outperform conventional beamformers.