Application of Artificial Neural Networks, (ANN( inProviding of self organization feature Map?? (SOFM) for Exploration of mines Stages
The pore throat size distribution plays an important role in formation evaluations and applies in many domains such as calculation of porosity, permeability, seals evaluating.
For prediction of this parameter, the ability of Artificial Neural Networks was used. Well logs and core analysis results are available from seven wells in six fields of Zagros basin, south Iran. The well logs, core porosity and permeability along with measured pore throat size were utilized to train the networks pore throat size prediction. The 25 percent of data were used for validation and 25 percent, to verify the developed networks predictions. The results indicated that the new method could be a useful method for predicting the pore throat size without using the mercury injection data.
Application of Artificial Neural Networks, (ANN) has gained considerable favor in 1980. These applications include mining, geosciences, natural resources and related sciences. An application of their usage is subject of remote sensing. Previously interpretations of remotely sensed data usually was done using manual methods which was a time taking and tedious work accompanied by serious errors and inaccuracies. For this reason many workers in the field put forward much effort to obtain the needed data for geologic and geographic- natural resources maps by usage of the signals received directly from satellites. For the first time, in the second Conference on Artificial Neural Network convinced in 1992 in Finland, the subject of ?self organization feature Map? (SOFM) was subject of debate and presentations. The remarkable results of this application were also presented for the two case studies carried out in southern France. From then on this subject picked up a good momentum in developed countries which omployed the technique for mapping purposes in mining exploration. In present work by usage and aid of (ANN) satellites? signals with noise filtration and deletion were converted into surface image maps of the earth. These maps reveal all kinds of natural and artificial features of the earth?s surface and are usable as exploration tools in reconnaissance and prospecting stages. In this paper first a summary of A.N.N.s structures and architecture is presented, and then the algorithms used in design of such networks are discussed Finally, the evolution of the Neural Networks in preparation of (SOFM) and the details of their interpretation is investigated and presented in this work.