A Study on FCM Inversion Algorithm of 2D Magnetotelluric

Magnetotelluric sounding method is an exploration method based on the difference of the rock's resistivity. In this paper, we develop a fuzzy c-means clustering (FCM) algorithm to invert the 2D magnetotelluric data. First, we add the prior information and the membership to the inversion algorithm and build the objective function which include the central points v and the membership u. Second, by updating the central points v and the membership, the misfit value of the objective function is decreasing and when the misfit value is less than the target misfit value, we get the inversion result (resistivity value).Third, we design some synthetic models and add 5% error to the forward data which regarded as the observed data and is used for the FCM inversion. When I get the FCM algorithm results, we compare the results of the classic Occam algorithm with our FCM algorithm. The inversion result shows that the FCM algorithm is superior to the classic Occam inversion. The resistivity value of the FCM inversion result is close to the true resistivity value which proves the FCM algorithm is effective.

Keywords：magnetotelluric;2D inversion; Occam inversionalgorithm; FCM inversion algorithm