Method Of Harmonic Noise Elimination Based On Its Estimation

In the process of vibroseis acquisition, due to the nonlinear effects of vibrator's mechanical system and hydraulic servo system and the coupling with the earth, the harmonics are unavoidable. Particularly when high efficiency techniques are used to speed up the vibroseis acquisition, such as the slip sweep technique and V1 technique and so on, one vibrator group starts sweeping without waiting for the previous group's sweep to terminate, results in two or even several records from different shots are overlapped together. Through the traditional correlation process, the harmonics of the current shot will contaminate several previous shots, and can cause a remarkable decrease of the signal-to-noise ratio.Therefore, eliminating or suppressing the harmonics through some certain method will have great significance for the following data processing. This paper firstly gives d detailed analysis of the causes and the features of the harmonics. Research shows that for the slip sweep data the harmonics can be divided into two parts: the harmonics containing in the current shot itself and the harmonics interference from adjacent shots, and the energy of the latter is much larger that of the former, so in the processing we mainly suppress the harmonics of the latter. Based on the previous research, we study several methods of harmonics elimination, which can be divided into three categories: methods of time-frequency analysis, methods of subtraction on the base of estimation and methods of deconvolution. Assuming the medium stratum is linear and elastic, and the plate is stiff which means the bending force can be ignored, this paper proposed the method of suppression harmonics based on estimation. This method uses the harmonics prediction operator (HPO) from the ground force to estimate the harmonic components, and then subtract the harmonics from the record. This method can deal with the uncorrelation data and the correlation data, also this method will not damage the effective signals when eliminating the harmonics. The tests of synthetic data and the real data show that the method is stable and efficient, and can meet the demands of the current field production.