Study On Key Techniques For Signal Detection Of Electromagnetic Seismic Vibrator

Vibratory seismic exploration has been applied widely in seismic exploration due toits advantages of controllability, security and absence of environmental pollution.Compared to the hydraulic seismic vibrator that has been commonly used in oilexploration, electromagnetic seismic vibrator that is driven by the principle ofelectromagnetic induction has more broad prospects in high resolution seismicexploration. For independent research and development of Portable High frequencyVibrator System (PHVS), surrounding the seismic signal detection theory, method,numerical simulation and field experiment, this paper studied on key techniques ofseismic signal detection for electromagnetic seismic vibrator.In original vibrator seismic record, the excitation signal is a long time FM signal,reflected waves and other interference signals are mixed together, so the arrival timeof reflected wave could not be distinguished. And as the design for portability, theenergy of excitation signal of electromagnetic seismic vibrator is weak, so Theeffective signal of original record has been submerged in strong noise, thesignal-to-noise ratio (SNR) is very low. Therefore, we must use the appropriateseismic signal detection technique to detect the seismic wavelet similar to pulsesource. So the study for high precision seismic signal detection method of vibroseishas great significance.Now, for the signal detection of electromagnetic seismic vibrator, the conventionalmethod is based on the control-signal or the near base-plate signal as reference signalin correlation detection. Compared to the hydraulic seismic vibrator, in complexgeological condition or the base-plate is poorly coupled with the ground, theconventional method will cause problems such as wavelet arrival time error and falsemultiple waves. This paper analyzes the problems and then proposes a correlationdetection method based on a reconstructed excitation signal (CDMBRES). First, thedirect wave is extracted from the near base-plate signal by a matched filter. Second,the extracted direct wave is used to reconstruct the excitation signal as a referencesignal in correlation detection. Finally, the quality of the seismic record is improvedby spectral whitening. In simulation data, the correlation coefficient between the reconstructed excitation signal and the ideal excitation signal is high linear correlation.The seismic record by CDMBRES is consistent with the model. In the fieldexperiment data, the detection result of PHVS500is consistent with MiniVib T15000(it is100times power to PHVS500) not only regarding wavelet arrival time but alsowaveform characteristics. Moreover, the average SNR of near offset shallow partincreases about1.75dB. The event is more continuous, identifiable horizon is richer,so the resolution is improved obviously. According to the above results, CDMBREScan effectively solve the problems caused by conventional method of electromagneticseismic vibrator in complex geological condition, provides the key technique for highprecision signal detection of portable seismic vibrator.However, with the exploration cost increasing, single source method has been unableto meet the needs of economic development for energy. In order to improve theproduction efficiency of seismic exploration, this paper studied the simultaneoussource sweep method and its seismic signal detection technique. First we research theprinciple of high efficiency acquisition technique as Flip-flop Sweep, Slip-Sweep andDSSS. In the field experiments, compared to single source method, the productivenessis improved and the signal detection result is consistent by high efficiency acquisitiontechnique. Then we research the simultaneous source sweep technique. We designedthree different acquisition method and corresponding blended seismic signal detectionmethod.In the separated detection method based on inversion and space-geometry filter, nseismic vibrators are arranged. The method encodes phase and set random delay foreach sweep signal. The vibrators sweep nearly simultaneously n times and blendedseismic records are acquired. The single seismic record is separated by linearinversion first and detection by space-geometry filter then. But the method mustconsider the vertical stack of each shot points, the number of vibrator could not be toomuch, so the improvement of productiveness is limited.In order to improve the production efficiency max, we research the separateddetection method based on iterative filtering for blended data. The method arrangesany seismic vibrators and set same sweep signal for each one. The vibrators sweepnearly simultaneously with random delay. For the blended seismic data, first weconvert it from common shot gather to other gather, the blended noise is converted from continuous style to random style. Then we use iterative filtering bymultidirectional median value filtering. Finally the filtered result is converted tocommon shot gather and the single seismic record is separated detected. But themethod is complex and if there are many vibrators, the detection effect will be poor.In order to simplify the algorithm, we research the separated detection method basedon random segment sweep signal. The method also could arrange any vibrators, andset segment of frequency and time for each sweep signal based on principle of random.So each vibrator has strong independent signal and sweep simultaneous withoutinterferer from other vibrator. The seismographs are recording continuously to acquireone super blended data. Based on matched filtering technique, we separate and detectthe blended data to get single seismic record.Finally, we apply the three methods to simulation and filed experiment data. They allcould improve the productiveness. Compared to the single source method, theseparated detection result is consistent. The wavelet arrival time is same, waveformcharacteristic is same, the consistent coefficient is high and the SNR decreased a little.These results could demonstrate that the separated detection method the paperdesigned is effective. The method is simple and efficient, but it is strict to the vibrator.The quality control should be done well.In summary, for the electromagnetic seismic vibrator, this paper researches the keydetection technique of seismic signal for single source and simultaneous source. It isof great significance for high resolution portable vibratory seismic exploration.