UNCERTAINTY PROPAGATION FROM WELL-LOGS TO PETROPHYSICAL PARAMETERS USING CLUSTERING ALGORITHM AND FUZZY ARITHMETIC - Université Rennes 2 Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

UNCERTAINTY PROPAGATION FROM WELL-LOGS TO PETROPHYSICAL PARAMETERS USING CLUSTERING ALGORITHM AND FUZZY ARITHMETIC

Résumé

Each well-log record is an average over the volume of investigation. Therefore the well-log records have spatial uncertainty. In conventional reservoir characterization, this spatial uncertainty is usually overwhelmed. A hybrid clustering-fuzzy method is here developed to analyse the uncertainty of petrophysical parameters, using well-logs. The proposed method is applied to the carbonate Sarvak Formation in five oil wells in an Iranian anticlinal onshore oil-field. In the first step, the porosity well-logs (neutron porosity, bulk density and sonic transit time) are clustered. The cluster range of neutron porosity is considered as the range of porosity. In each cluster, neutron porosity is calibrated by the core tests. The calibrated clustering-based porosity is at least 33% more accurate than conventional porosity estimations. Analysing the generalization ability of the porosity estimators of the oil wells, a homogeneous porosity zone is determined in the northern part of the oil-field. In the next step, the uncertainty of porosity is converted into the uncertainties of irreducible water saturation, then permeability, using fuzzy arithmetic. Comparing to the crisp irreducible water saturation, the fuzzy number of irreducible water saturation has less overestimation. The fuzzy number of permeability is compatible with core tests (except in well S1). The incompatibility could be related to the tectonic complexity at the location of S1. The validation of fuzzy numbers is done by two criteria: (i) an average of α-cut of core values, which is desired to be higher, (ii) the optimum α-cut which searches the highest α-cut with the least uncertainty interval. For porosity and permeability studies, it is found that the α-cut higher than 0.90 is the most appropriate.
Fichier non déposé

Dates et versions

insu-01713428 , version 1 (20-02-2018)

Identifiants

  • HAL Id : insu-01713428 , version 1

Citer

Masoudi Pedram, Tahar Aifa, Hossein Memarian, Behzad Tokhmechi. UNCERTAINTY PROPAGATION FROM WELL-LOGS TO PETROPHYSICAL PARAMETERS USING CLUSTERING ALGORITHM AND FUZZY ARITHMETIC. 7 éme COLLOQUE MAGHREBIN DE GEOPHYSIQUE APPLIQUEE CMGA 7, Feb 2018, Alger, Algeria. pp.41-42. ⟨insu-01713428⟩
75 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More