Prediction of Some Parameters of Gas-Condensate Slits by Means of Neural Networks

About the author:

Abstract:

Under present-day conditions the accurate measurements of the yield on gas condensate wells are difficult and, as a consequence, prediction is difficult too. Typically, the parameters are measured in the collection areas, but you cannot get a clear picture for each specific well. Nevertheless, based on the history of changes in some parameters of the well, its output parameters can be predicted. In this case, it is necessary to take into account the mutual influence of these parameters and the history of their change.

The study described in this paper is aimed to obtain a mechanism for predicting the flow of a gas-condensate mixture and a gas-condensate ratio with a forecast error of no more than 5%. The practical application of the generalized regression neural network to the problem of prediction of some well parameters is presented. The rationale for using this topology is given. The conducted experimental check showed the acceptable quality of the proposed algorithm. The dependence of the prediction error on the number of learning points is given. Based on the experimental data, the application of GRNN network for forecasting the gas production rate and gas-condensate ratio is considered promising.

Nevertheless, the application of the neural network does not take into account the phenomenon physics, which also negatively affects the accuracy of the selected parameters prediction.

References:

1. Arthur D., Vassilvitskii S. 2007. “k-means++: the advantages of careful seeding”. Proceedings of the eighteenth annual ACM-SIAM symposium on Discrete algorithms. Society for Industrial and Applied Mathematics Philadelphia. PA, USA, pp. 1027-1035.
2. Elman J. L. 1990. “Finding Structure in Time”. Cognitive Science, pp. 179-211. DOI: 10.1016/0364-0213(90)90002-E
3. Fahlman S. E., Lebiere C. 1989. “The Cascade-Correlation Learning Architecture”. Advances in Neural Information Processing Systems 2 (NIPS 1989), pp. 524-532.
4. Rummelhart D. E., Hinton G. E., Williams R. J. 1986. “Learning Representations by Backpropagation Errors”. Nature 323, pp. 533-536. DOI: 10.1038/323533a0
5. Specht D. F. 1991. “A General Regression Neural Network”. IEEE Transactions of Neural Networks, vol. 2, no 6, pp. 568-576. DOI: 10.1109/72.97934