Study On Inversion And Interpretation Of The Transient Electromagnetic Method Based On Artificial Neural Network And Its Applications

Being a method of time-domain electrical magnetic surveying, the transient electromagnetic method (TEM) that is used to observe and investigate the secondary field (pure anomaly) on condition that the primary field does not exist can greatly simplify investigation of anomaly of buried geologic bodies. It is so wide to be used with powerful resolution capacity and high working efficiency in the field of resource prospecting and engineering exploration that the method has been become conspicuous one of geophysical prospecting, especially in coal mining areas with hydrogeological investigations into the most effective means of detection. However, as the survey precision required improvement in the transient electromagnetic methods presented new research on technology issues.Firstly, this dissertation discuss the basic theoretical fundament of Transient electromagnetic methods and the basic characteristics of TEM fields. Secondly, it analyzes the present situation of the inversion and interpretation of the Transient electromagnetic methods, and has made the detailed introductions on several most commonly used methods of inversion and interpretation. There are some defects of nowadays traditional methods of electromagnetic information inversion in TEM, in the first place, it is difficult to master the theory of electromagnetic information inversion;secondly, it is difficult to analyze data and model in the data processing, in the third place, only expert of electromagnetism and geology can do it.Artificial neural network is particularly effective in the information processing which is uncertain and non-structure. A great deal of information in geophysical exploration and exploitation is like this. So, it does well in researching the electromagnetic information inversion. In the same, it avoids detailed and knotted electromagnetic computation. Once after selecting appropriate swatches and trainingor learning, it can solve complicated practical problem effectively. Furthermore, it has excellent capacity to study and memory. Through the systemic training and working, it can enhance the continuity and inheritance in the field of TEM inversion and is easy to be generalized.In this dissertation an artificial neural network expert system is designed which uses the arithmetic of the Back-Propagation (BP) and has made the improvement to it, and the practical inversion procedure being established. Finally some tentative research in actual TEM curve inversion with this artificial neural network(ANN) expert system has been done which with the union of known drill hole and the electricity log information, and has made the inversion results and the traditional inversion methods results the comparison and has carried on the preliminary error analysis and the appraisal. The results shows that the method applying ANN to TEM curve inversion is feasible in the practice.