RESEARCH OF DENOISING OIL-DRILLING TESTING DATA BASED ON WAVELET

To provide accurate real-time information for engineer,a new method based on wavelet transform is proposed to eliminate the noise from signals which is sampled by the test device in oil drilling.First,a denoising flow based on wavelet is established according to the model of the oil-drilling testing data which contained much noise.Second,on the basis of analyzing the transforming characteristics of noisy signal based wavelet,a reasonable level of wavelet decomposition is calculated by using self-correlation function to estimate the white noise verification in wavelet scale spaces.Third,compared with different methods in wavelet threshold selection,an optimal denoising threshold is calculated using generalized cross validation(GCV) to hold the important features of signal and remove the noise from the signal.At last,the paper provides detailed steps of denoising the oil-drilling testing data based on the wavelet transform,which has obtained favorable results.The performance of pressure data and torque data confirms the usefulness of this approach,which provides a reliable data support to engineers for further analysis in drilling engineering.