Study On The Suppressing Noise Of Correlation Identify Method Based On M-Sequence Applications In Controlled Source Electromagnetic

In practice,the strong disturbance had critical whittle the effectiveness of CSAMT method and bring down data quality, classic de-noising method unable to overcome this dilemma. At the same time, correlation identification method based on pseudorandom signal develop well in other application for the purpose of identify weak signal and suppress disturbance. This thesis comes up in such a background and study at the sound effects of suppress strong noise using pseudorandom sequence as transmit signal in CSAMT explore method. At first test how to select a series of suitable parameter though a known signal system, the result show that the period-length of m sequence should be more 5 times than the system response duration time; clock rate is rely on object frequency range, while expect for high frequency electromagnet spectrum,we should choice large clock rate, otherwise,choice small one; the repeated cycle does not matter on the result. In the third chapter, consider one dimension model forward and the frequency-time domain transform algorithm based on DCT transform in order to simulate time series really record, and then, identify the impulse response under kinds of condition including period, random impulse, discharging disturbance, show that correlative identify method get a well result. Chapter four carry out the study at three dimension modeling of CSAMT method with electric dipole instead by a finite volume region distributed with current and the current density is according to pseudo detal function, this treatment can omit the time spend on calculate background field. Last chapter simulate the response of three model and identify the electromagnetic spectrum in the condition with real noise get form MT station, invers the identify result with WingLink software,the inversion section well consistent to the model, indicate that correlate identify done well in complicated situation.