Estimation De La Porosité Par L’application De Quelques Algorithmes De L’intelligence Artificielle Cas Champ De Hessi Tarafa Sud Algérien

Résumé: The objective of this thesis will focus on the estimation of the porosity which is of great importance in the characterization of the reservoir and the estimation of the reserves in place, using several techniques of artificial intelligence using logs. of wells and direct measurements in the laboratory, case of the compacted reservoir (El Hamra quartzite) of the Hessi Tarfa basin of Oued M'ya in southern Algeria. The main artificial intelligence techniques used are (multiple linear regression, hybrid genetic algorithm with multiple linear regression, artificial neuron network). The artificial neural network technique for the estimation of porosity shows a better correlation R2 = 0.6933, for the prediction of porosity in heterogeneous reservoirs. The RLM multiple linear regression technique for estimating porosity shows a better mean squared error RMSE = 2.09 The results of this brief confirm the effectiveness of artificial intelligence tools for the prediction of reservoir parameters including heterogeneous ones with notable performance which saves time on the one hand and is totally economical on the other hand. , because it minimizes the costs of coring and measurements in the laboratory.

Publié dans la revue: