Tham khảo tài liệu 'earth sciences part 7', khoa học tự nhiên, công nghệ sinh học phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | 230 Earth Sciences Nikravesh M. 1996 Zuluaga E. 2000 . The ability of ANN systems is to spontaneously learn from examples reason over inexact and fuzzy data and provide adequate responses to new information not previously seen. It consists of a large number of simple interconnected artificial neurons. As shown in Fig-3 the artificial neuron i takes information from other neurons x1 x2 x3 .xn performs very simple operations on this data and passes results yi on to other artificial neurons. Every artificial neuron meets the following equations n si 2 wjxj ei 2 j i ui g si 3 yi f ui 4 Where equation 2 is the accumulating potential value of the artificial neuron i after synapse Bi is the processing elements threshold and wi is the interlayer connection weights. Equation 4 is the relational expression between input and output values in which ui is the state of the artificial neuron i . Neural networks operate by virtue of many artificial neuron data in this manner. A Fig. 3. Sketch Diagram of Artificial Neuron The specific operation has two steps. One is the training process and another is the testing process. In the process of training the ANN model has to be trained to recognize the relationships between the input and the desired output values by adjusting the connection weights between the different neurons. This process continues until weights converge to the desired error level or the output reaches an acceptable level. In the testing process the developed ANN model is tested with several sets of experimental values which are not used in the training of the model to judge its performance. The developed model can memorize the correct output once input data is given. Compared with other methods the ANN method has many advantages and has been introduced to predict the formation sensitivity. 231 Mechanisms and Effective Prevention of Damage for Formations with Low-Porosity and Low-Permeability Development of the predicting model Determinate the influencing factors
