Research On Modeling, Data Processing And Inversion Of The Time Domain Airborne MULTIPULSE System

In general the raw data of the time-domain airborne electromagnetic (AEM) hasthe features of huge amount o and very noisy. Thus, we all kinds of preprocessingmethods to effectively reduce the size of the raw data and improve thesignal-to-noise ratio. This paper has done researches on time-domainelectromagnetic data preprocessing, such as primary and spheric removal, coiloscillation correction, signal stacking, and calculation of B-field and so on and builta data processing platform based on the software development. In this thesis, I willfirst summarize AEM data processing and interpretation, including datapreprocessing and processing, imaging, forward modeling and inversion in bothfrequency-and time-domain, then I assemble these methods into a visual softwareplatform. By using this software I have also worked on filed data processing andinterpretation.The traditional time-domain AEM transmits a single waveform (e.g. a half-sine)to complete the whole survey. However, in this paper I introduced an airborne EMtransmitter with a double waveform (a big half-sine plus a small square wave) at thesame half cycle. In the half cycle,when the big half-sine is almost attenuated, thesystem transmits a small square. The receiver coil receives signal from bothwaveforms. This paper introduces the features of the double waveform from fiveaspects, including the characteristics of the traditional single waveform, theamplitude spectra, calculation of relative anomaly, the inversion of synthetic andsurvey data.In the following, I first introduce the feasibility of double waveform from twoaspects-the characteristics of the traditional waveform’s decay curve and the methods of the data processing. In the amplitude spectra, the double waveformexhibits much higher power than the single half-sine at higher frequencies. Thisindicates that the double waveform has higher near-surface resolution. Then, startingfrom Maxwell’s equations, I derive the forward modeling formula for arbitrarywaveforms. The relative anomaly of different models is calculated to prove that thedouble waveform is more sensitive to near-surface conductors. Finally, I invert boththe synthetic and the survey data and prove that the double waveform can providemore resolution for the near-surface structures, while keeping depth of penetration.