Research On Denoising Method Of Microseismic Signal Based On Improved Particle Filter

Due to the increasing demand for oil and gas in recent years,microseismic monitoring technology which can monitor underground oil and gas fractures has also been widely used.In the process of oil and gas resource exploration,the collected microseismic signals are usually accompanied by a large amount of noise,while the effective microseismic signals have the characteristics of short duration and weak energy,so that most of the collected effective signals are submerged in random noise,and the signal-to-noise ratio is very low,which seriously affects the source location and subsequent research.Therefore,we need to separate the noise from the effective signal in the collected microseismic signal,extract the effective signal and improve the signal-to-noise ratio.The main research content of this paper is based on particle filter,analyzes the characteristics of microseismic effective signal and its noise,and deeply carries out the research of microseismic signal denoising.The main research contents include the following three aspects.Firstly,based on the time series model,the characteristics of microseismic signal are studied,and the method of model establishment is selected based on the characteristics of microseismic signal.Firstly,the characteristics of microseismic effective signal and the classification of noise are analyzed,and the simulated microseismic signal is generated by forward modeling according to the amplitude and phase characteristics of microseismic wave,and the generated simulated microseismic signal is used for denoising simulation experiment;Then,according to the analysis of the signal,the moving average time series model is selected to model the microseismic signal,so that the model-based particle filter method can be used for denoising research.Secondly,a micro seismic signal denoising method based on dynamic threshold particle filter is proposed.Aiming at the problem of particle dilution in traditional particle filter,a dynamic threshold particle filter method is proposed by improving the resampling algorithm.The resampling method in traditional particle filter takes random number as weight instead of collecting particles in sections as weight,and introduces dynamic threshold,which changes the particle dilution problem of traditional resampling method and further improves the accuracy of signal estimation by particle filter.The signal-to-noise ratio of the denoised signal is effectively improved,and the complexity of the algorithm is low,which ensures the effectiveness of processing the real microseismic signal.Thirdly,a particle filter micro seismic signal denoising method combined with ensemble empirical mode decomposition is proposed.According to the characteristics of traditional particle filter algorithm for signal processing,the combination of ensemble empirical mode decomposition and particle filter not only eliminates the defect that the ensemble empirical mode decomposition method damages the effective signal,but also avoids the disadvantage of reducing the estimation accuracy of particle filter caused by excessive noise.This method has better retention effect on the effective signal,removes most random noise and makes the signal in-phase axis continuous and smooth.This paper makes a preliminary exploration on the denoising method of microseismic signal,and puts forward two new denoising methods based on particle filter,which overcomes the defect of the original method in damaging the effective signal,improves the signal-to-noise ratio of the signal,enriches the theory for microseismic signal processing,and provides some technical support for improving the development of oil and gas reservoirs.