Periodic Signal Recognition Technology Based on Framing Window Adaptive Scaling Algorithm and Trajectory Tracking Algorithm: A Case Study of Aerospace Loose Particle Detection Signal

In this article, a novel algorithm combining adaptive scaling of framing windows and pulse trajectory tracking is proposed for the detection of internal loose particles in aerospace sealed electronic components. The proposed algorithm can be used to identify whether the detection signal has general or local periodicity and to distinguish particle signals from component signals. The algorithm uses adaptive scaling of the framing windows length, which can effectively reduce the influence of the signal periodic instability caused by the change of the signal frequency. In order to evaluate the performance of the algorithm, 600 sets of data were collected on the particle impact noise detection (PIND) platform. The single-component signal, loose particle signal, multicomponent signal, and mixed signal were verified, respectively, with the accurate rate close to 95%, and the recognition effect was great. In addition, compared with results using Fourier transform, the identification results of signal type using the proposed algorithm are more intuitive.