Aiming at the problem of multi-target oscillating signal detection in the range of 2.5 Hz～300 Hz in wideband measurement for novel power system

the design method of detection algorithm with high reliability (extremely low false detection rate) as the core performance is studied. Adopting the windowed discrete-time Fourier transform (DTFT) detection algorithm

the optimization design of window function is studied deeply. Firstly

through theoretical analysis

the constraint conditions of high reliability detection on the spectral characteristics of window function are proposed

and the technical index of window function is defined

and the optimal selection criteria of window function is established. Secondly

based on the signal environment and performance requirements of broadband measurement signal detection

specific examples of window function optimization are given for sub-synchronous oscillation band (2.5 Hz～100 Hz) and high-frequency oscillation band (100 Hz～300 Hz) respectively. The simulation results show that the DTFT detection algorithm based on preference window is close to the expected design index

with high reliability and adaptability to complex signal environment

and has important engineering practical value in broadband measurement.