Research On Attitude Solving Algorithm Of Towing Cable Based On Convolutional Neural Network Fusion Extended Kalman Filter

The global energy problem is serious,and the scale of oil and gas resources in the ocean is huge.Therefore,the exploration and mining of deep water oil and gas resources in the ocean has become the focus of energy development.Marine resources are an indispensable component of national rights and interests.Neighboring countries covet the oil and gas hydrates in the East and South Seas of China,and have used the technology of developed countries to steal oil in the Nansha Sea.Due to the shortage of internal resources,the plot of external neighbors and the critical situation,China must further promote the research process of marine science and innovate the exploitation scheme of deep-water resources.Deep ocean towing seismic exploration is the most rapidly developing and widely used method of marine exploration,and it is becoming the most effective and high-precision method for detecting submarine hydrate resources.China is in a critical period of technological breakthrough in the exploration and exploitation of submarine hydrate and faces many challenges.Due to the influence of seawater on seismic waves,the detection resolution and penetration depth of conventional seismic exploration system for deep-sea underwater strata will be reduced.In order to reduce the absorption of seismic signals by sea water,the offshore bottom dragging method is used to lower the source and digital cable and keep them within a depth of less than 100 meters from the seafloor,so as to improve the quality of seismic data collected and to meet the requirements of delineating the spatial distribution of submarine hydrate ore bodies.The exploitation of seabed energy requires high-precision imaging of underground oil and gas resources,and the premise of highprecision imaging is to obtain high fidelity seismic wave field information;Vibrator and streamer are important components of underwater towing system.The source transmits seismic waves,and the hydrophone on the streamer receives the seismic signals reflected by the seabed,and then inverts the seismic wave field information after subsequent processing.Many marine environmental factors affect the hydrodynamic characteristics of seawater,which leads to the deviation between the underwater towing system and the expected design.Therefore,real-time recording of the mechanical information of key components such as source and streamer,solving the attitude of streamer and effectively recovering the seismic wave field information have become the key technologies for the research and development of near seabed deep towing seismic exploration equipment.According to the exploration needs,the purpose of this paper is to obtain effective underwater digital receiving cable attitude information.For this research goal,this paper mainly carries out the following four aspects of work:Firstly,the technical route of multi-sensor fusion for posture solution is designed.By collecting data,analyzing the background significance of the analysis,and combining with the actual needs,clarifying the necessity of the study;For the limitation of the marine environment,an integrated MEMS is selected for full-attitude monitoring.This paper summarizes the current status of posture detection at home and abroad,compares the advantages and disadvantages of the methods,and analyses the methods of multi-sensor data information fusion.Secondly,the theoretical research and algorithm design based on Kalman filter and neural network are carried out.Clarify the relevant theoretical support of carrier attitude detection and establish a unified coordinate system.The quaternion method is selected to represent the attitude;After comparing and analyzing the attitude detection models of gyroscope,accelerometer and magnetometer,the model of three sensor data information fusion is adopted;The extended Kalman filtering method and traceless Kalman filtering method are studied and realized theoretically.The solution of the two projects is analyzed.Finally,the extended Kalman filtering method is selected.Then,aiming at the limitations of extended Kalman filtering method,for the sake of seeking optimization scheme.Deeply understand the principle and advantages of neural network,analyze the model architecture and integrate the advantages,a neural network fusion traditional attitude estimation algorithm is proposed;The selected network models are BP and CNN,which are respectively fused with EKF according to their characteristics to optimize the streamer attitude solution algorithm.Then,the streamer attitude calculation software is designed.Combined with the above research contents and program functions,the appropriate design platform is selected to complete the compilation of streamer attitude calculation algorithm;Optimize the interface of attitude monitoring system and highlight key information.If the equipment is too close to the seabed,early warning will be given to ensure the safety of exploration operation.Finally,the performance test of the tow posture solution algorithm is completed.Based on the algorithm verified by simulated data in the laboratory,the system equipment is used to carry out seismic exploration tests in the deep-sea area of the South China Sea.The proposed scheme is used to process the actual data and evaluate the results,which proves the effectiveness of the scheme.