Electrical Sounding Intelligent Optimization Inversion Theory

Dc resistivity sounding is an applied widely and important technical method ofgeophysical exploration methods. It is playing a more and more important role on theenergy and mineral exploration, also on the hydrology and engineering survey. At the sametime, the people put forward higher request about its exploration and interpretationaccuracy. So inversion as a core technology of the interpretation of electrical sounding datahas become the hot spot to study for the geophysical workers. Along with the developmentof automatization and intelligence of computer technology, the inversion and interpretationof electrical sounding material has entered into a new stage. How to overcome its ambiguityand accurately show its quantitative results by intelligent optimization technology hasbecome the main research direction of geophysical workers. At present, the mostcommonly used inversion methods have damped least-square method, gradient method andvariable metric method, they all belong to linear local optimization method. They not onlyeasily make solution into the local minimum value, and heavily reliance on initial modelselection. They have difficulty achieving satisfactory fitting inversion of local anomaliescaused by thin layer underground. Therefore, developing nonlinear and intelligent inversionmethods which have strong global search capability and are not dependent on initial modelhas the vital significance to the development of electrical sounding technology.Intelligent optimization algorithms belong to nonlinear algorithms which are randomsearch algorithms inspired by the nature biological communities behavior，such as antcolony algorithm，particle swarm algorithm, bacteria foraging algorithm and artificial fishalgorithm, etc. They embody the self-organization and adaptive ability of biological systemin some measure. The artificial colony algorithm (ABC) in this paper is a new heuristicbionic algorithm based on the behavior of colony foraging. It is a new and excellentintelligent optimization algorithm. Because of its fast convergence rate, high qualitynon-inferior solution and strong robustness, the ABC is recognized quickly in theinternational optimization calculation areas and widely be used in the actual problems.This paper combines the artificial colony algorithm with the inversion and explanationof DC resistivity sounding data for the first time. First of all，the basic theories of electrical sounding are analyzed, the modeling theory model and related resistivity modelingformulas of electrical sounding are studied. After that， the principle、 basic structure andcharacteristic of ABC are completely grasped and then the one-dimensional nonlinearoptimization inversion model of electrical sounding curves is established. Through thetesting comparison of theoretical model, a group of reasonable inversion parameters and anoptimal selection strategy is selected. Taking advantage of Matlab platform, the automaticinversion and curves fitting function of resistivity sounding is realized. At last, the programis used in the sounding data inversion of typical hole has been known，and its generalizationability is assessed. Through the comparison between intelligent inversion method andexisting inversion methods，we find that： The artificial colony algorithm has obviousadvantage in improving the precision and speeding the inversion speed. And it occupieslesser memory resource when calculating. The results of this subject will open up a newfield of nonlinear geophysical inversion，and the research results can further be used into2D or3D electrical sounding inversion.