The dominating cabin noise in propeller aircraft consists essentially of strong tonal components at harmonics of the Blade Passage Frequency (BPF) of the propellers. In order to efficiently reduce such low frequency periodic noise, it is advisable to employ an Active Noise Control (ANC) system based on a feedforward controller. This paper presents a set of normalized complex Filtered-X Least-Mean-Square (FX LMS) algorithms. By using different variants of normalization factors the convergence rate, the tracking performance and the steady-state noise attenuation can be improved. The algorithms presented are based either on a single normalization factor for the whole control system (global normalized FX LMS algorithm), or several individual normalization factors (reference-individual FX LMS algorithm) or the novel actuator-individual FX LMS algorithm. The evaluation is performed on noise recorded during flight in the cabin of a Dornier 328, a twin-engine propeller aircraft.