Operating fan underwater sound signal feature extraction method based on adaptive stochastic resonance

Abstract

The invention discloses an operating fan underwater sound signal feature extraction method based on adaptive stochastic resonance, and relates to the field of offshore wind plant underwater noise. A self-adaptive stochastic resonance method is adopted, a time-frequency-amplitude comprehensive index established based on support vector regression serves as a fitness function, a balance optimizer serves as an optimization algorithm, system parameters are adjusted in a self-adaptive mode according to the ocean background noise intensity, and self-adaptive extraction of the characteristic frequency of the underwater sound signal in the operation period of the offshore wind turbine is achieved. The time-frequency-amplitude comprehensive index adopted by the method can reflect the information of the signal in the three aspects of time domain, frequency domain and amplitude domain, the problem of extracting the unknown frequency of the underwater sound signal in the operation period of the offshore wind turbine under the changing ocean background noise is solved, and the method has high accuracy. Besides, the balance optimizer adopted by the method can realize synchronous optimization of a plurality of system parameters, so that the operation time is greatly shortened, the efficiency of extracting the line spectrum frequency of the underwater sound signal in the operation period is improved, and the method has the characteristics of simplicity and easiness in implementation.

Description

technical field [0001] The invention relates to the field of underwater noise detection of offshore wind farms, in particular to a method for extracting features of underwater acoustic signals of operating wind turbines based on adaptive stochastic resonance. Background technique [0002] ([1] Bai Xu. Current Situation and Prospects of China's Offshore Wind Power Development [J]. Ship Engineering, 2021, 43(10): 12-15.) While offshore wind power brings us economic benefits, its 20 ~30 years of operation will continue to radiate sound waves into the water. This acoustic signal is typically continuous, low-intensity, and characterized by one or more line-spectral components below 1000 Hz, and may have effects on marine life, especially for acoustic-sensitive Touya fish, such as masking The acoustic communication between marine organisms interferes with foraging, spawning, and triggers physiological responses. In addition, the underwater acoustic signal during the operation pe...

AI Technical Summary

Problems solved by technology

[0005] The traditional stochastic resonance finds the optimal parameters through the grid search method, which is computationally intensive, time-consuming and inefficient

Images