Reverberation Intensity Modeling Under Ice And Research On Roughness Parameters Inversion

The Arctic Ocean is covered by a large amount of ice.The speed of sound in the North Pole conforms to a positive gradient distribution,so the sound rays will continuously bend upward during the propagation process.The rough ice layer has a strong scattering effect on these sound rays,resulting in strong under-ice reverberation.When active sonar works under polar ice,under-ice reverberation is a typical interference.Strong under-ice reverberation interference seriously inhibits the detection performance of active sonar.Therefore,in order to predict and evaluate active sonar when it works under ice,it is very necessary to establish a suitable under-ice reverberation intensity model to effectively predict the under-ice reverberation intensity.The rough information under the Arctic ice is of great significance for fully understanding the characteristics of polar ice layers and for predicting global climate change from polar ice layers.Sailing and in-situ measurements are usually used,but these two methods have disadvantages.The sailing method is limited by the ship's ability to sail in the ice zone.The in-situ method is time-consuming,labor-intensive,and inefficient,and cannot quickly obtain rough information under polar ice in a large area,which makes it extremely difficult to conduct efficient and accurate scientific investigations in the Arctic.Acoustic method to obtain the rough information under ice has the advantages of simple measurement method,unlimited route,large measurement range,and high efficiency.Therefore,how to obtain the rough information of the ice layer by using appropriate acoustic methods has become an urgent problem to be solved.In view of the above research purposes,this paper analyzes the Arctic subsurface ice roughness spectrum,combines ice layer scattering and subsurface ice propagation models,and establishes the Arctic subsurface ice reverberation intensity model to predict the intensity of the reverberation and analyze its influencing factors.Based on the under-ice reverberation intensity model,combined with the inversion theory and optimization algorithm,three methods of inversion-based roughness spectrum parameters inversion based on the under-ice reverberation intensity are constructed to quickly and effectively obtain Arctic ice physics on a large scale.By analyzing the measured data over the years,it is found that the arctic roughness under the Arctic follows the Von Karman spectrum,so that the acoustic scattering models of rough ice layers at different frequencies and in different years are established.For different frequency bands,combining ray theory and normal wave theory,under-ice reverberation intensity models are established.In order to meet the requirements of fast calculation in engineering,the analytical expression of the reverberation intensity under the horizontal bistatic sonar under ice is derived.For three types of reverberation intensity models under the ice,calculation examples of each model are given,and the effects of ice layer acoustics and physical parameters on the reverberation intensity under the ice are analyzed.The measured reverberation intensity model in this paper is verified by using the measured arctic reverberation data.The results show that the theoretical calculation results agree with the experimental data,and the root mean square error is less than 3d B.Based on the reverberation intensity model under the ice,the neural network,support vector machine,and particle swarm method are used to inversely study the roughness parameters of the ice.The idea of inversion method and the relevant parameter value strategy are given.The inversion performance of roughness parameter inversion method is analyzed.The results show that all three inversion methods can inverse the roughness parameters.When the observation error is small,the accuracy of inversion of the roughness parameters by the neural network and support vector machine method is high.When the observation error is large,the particle swarm method is more accurate and reliable.The inversion method of roughness parameters is verified using the measured arctic reverberation data.The results show that the inversion roughness parameters were very close to the measured values of the Arctic,and the relative errors were within 4% and 8%,respectively.The research results of the under-ice reverberation model in this paper can provide a theoretical basis for the design and performance prediction of active sonars under the ice.The inversion method of under-ice roughness parameters based on this model can provide an effective means for the Arctic scientific expedition to obtain the Arctic ice physical information in a large area.