Research On The Methods Of Transient Radiated Noise Analysis

With the persistent development of vibration and noise reduction technology in the field of underwater acoustic, acoustic stealth performance of underwater vehicle has been improved remarkably. When the underwater vehicle is navigating by water, the sudden change of operational configuration induces the underwater vehicle to produce underwater transient radiation noise, such as steering signal, variable speed navigation signal and water entry signal. The transient radiation noise contains abundant identifying signature of underwater target, and the detection of transient radiation noise is the major research area in the field of passive sonar's signal processing. The thesis consists of three parts: mechanism analysis and modeling of underwater transient acoustic signal, measurement method of underwater transient acoustic signal and the detection algorithm for underwater transient signal.To begin with, various types of mathematical model are established based on the basic characteristic analysis of transient radiation noise in time domain and frequency domain. Thus,the simulation of steering signal, variable speed navigation signal, water entry signal and torpedo ejecting signal is closer to the measurement signal in practical.Secondly, the measurement method of underwater transient signal is investigated.According to the measurement of signal duration, maximum peak, average peak and the summation of energy for the signal duration, the characteristic information of transient noise can be extracted more completely. The measured data of steering signal and water entry signal in the experiment are analyzed, and the effectiveness of the proposed measurement method is validated.At the end, owing to the characteristic diversity of underwater transient radiation noise,CFAR Power-Law detector and WT Page Test detector are proposed based on Power-Law detector and Page Test detector. The spectrum signature of the transient radiation noise is detected in time-frequency domain by the combination of time-frequency analysis and line detection principle of radon transform. The effectiveness of the proposed method is confirmed by the research of the simulation signal. The result of experimental data processing proves that the method proposed in this thesis is reliable.