EMD Algorithm Research On Marine Electromagnetic Signal De-Noising

Today， Rich in offshore oil and gas resources in the worldand how theexploration of marine resources has become an inevitable strategic choice of acountry’s oil reserves. Marine controlled source electromagnetic detection (CSEM) isan effective means of pre-drill reservoir evaluation, not only supplement the marineseismic exploration, improve the success rate of exploration drilling, it is also animportant means of oil/water storage characteristics of the seabed structureidentification. Empirical Mode Decomposition (EMD) is a nonlinear, non-stationarysignal frequency analysis method.This article focuses on the basic theory of EMD, the basic concepts, includingthe basic principles of analytical signal and the instantaneous frequency of theintrinsic mode function, and EMD algorithm, namely the presence of EMD algorithmadvantages and disadvantages for different EMD problems and gives thecorresponding solutions.Secondly, a brief introduction to several traditional filtering methods, pointingout deficiencies. Since EMD algorithm has high frequency resolution, strongadaptive, highly unique nature of the time-frequency aggregation, put forwardseveral EMD based filtering, such as filtering method based on EMD, EMD-basedde-noising method and scale filter EMD threshold based filtering noise, and thenpropose an improved EMD filtering method, analyzes its advantages.Finally, the previously mentioned EMD based filtering method is applied to thesimulation CSEM signal de-noising using MATLAB, perform simulation and analysisof the differences and advantages of each method. EMD results show that there areother methods can not reach advantage, is currently found in the most suitable methodfor analyzing nonlinear and non-stationary signal analysis, making analysis andprocessing of marine CSEM signal a new breakthrough. The main defect is that thepresence of EMD algorithm mode mixing. The mode mixing of marine controlledsource electromagnetic signals cause low-frequency mixing of the signalcharacteristics of the problem can not tell, causing signal analysis difficult, for this phenomenon, this paper proposes an improved EMD decomposition method-EEMD.EEMD as improved EMD algorithm to eliminate this phenomenon modal aliasingeffect when the spectrum to solve the problem of low-frequency mixing time, tobetter distinguish the frequency components. By contrast, after EMD denoisingsuppress most of the noise, achieve better results, but due to modal aliasinglow-frequency mixed, but after EEMD noising, not only to remove most of thenoise, but also removes the low frequency mixed part-time, and a clear signal featureextraction.In summary, this paper studies the EMD algorithm and simulation EMDalgorithm in CSEM signal de-noising, comparative advantages and disadvantages offive methods, and through simulation experiments, a good simulation confirmed theEMD in CSEM signal de-noising effectiveness.