Paper Information

Title: 

APPLICATION OF SUPERVISED AND UNSUPERVISED NEURAL NETWORKS IN DECREASE HETEROGENEITY IN CARBONATE RESERVOIR

Type: PAPER
Author(s): ASSADI ALI,KADKHODAIE ALI,SFIDARI EBRAHIM*
 
 *DEPARTMENT OF GEOLOGY, UNIVERSITY COLLEGE OF SCIENCES, UNIVERSITY OF TEHRAN, TEHRAN, IRAN
 
Name of Seminar: INTERNATIONAL CONFERENCE OIL, GAS, PETROCHEMICAL AND POWER PLANT
Type of Seminar:  CONFERENCE
Sponsor:  SHERKATE TAVONI AMOZESHI PAZHOHESHI VESTA
Date:  2012Volume 1
 
 
Abstract: 

ONE OF THE KEY ELEMENTS IN RESERVOIR CHARACTERIZATION IS TO DETERMINE OR SEMI-QUANTITATIVELY MAP THE VERTICAL AND HORIZONTAL DISTRIBUTION OF ITS HETEROGENEITY AND RELATED PROPERTIES. IN GENERAL, THE CARBONATE RESERVOIRS MORE COMPLICATE THAN SILICICLASTIC RESERVOIR AND NEED TO USE NEW METHOD IN THIS RESERVOIR IS ESSENTIALLY. THAT MAKE THEIR LATERAL AND VERTICAL PREDICTION FROM PETRO PHYSICAL LOGS COMPLICATED IN CASE OF AVAILABILITY OF CONVENTIONAL WELL LOG DATA, ARTIFICIAL NEURAL NETWORKS ARE CAPABLE OF RECOGNITION OF NONLINEAR RELATIONSHIP BETWEEN WELL LOG DATA AND RESERVOIR PARAMETERS. FOR SUCH PURPOSED, A TWO STEP APPROACH IS ADAPTED TO CHARACTERIZE AND DECREASE AMBIGUITY IN CARBONATE RESERVOIR STUDIES. IN THE FIRST STEP AN UNSUPERVISED NEURAL NETWORK IS USED BASED ON THE SELF-ORGANIZING MAP TO IDENTIFY AND EXTRACT CONSTITUENT LOGFACIES. THIS CLASSIFICATION DOES NOT REQUIRE ANY FURTHER SUBDIVISION OF THE DATASET BUT FOLLOWS NATURALLY BASED ON THE UNIQUE CHARACTERISTICS OF WELL LOG MEASUREMENTS THAT REFLECT MINERAL AND LITHOLOGICAL RESPONSES OF THE LOGGED INTERVALS. IN THE SECOND STEP, A SUPERVISED NEURAL NETWORK IS APPLIED BASED THE BACK PROPAGATION ALGORITHM TO QUANTITATIVELY PREDICT THE POROSITY AND PERMEABILITY WITHIN EACH PARTICULAR LOGFACIES, DETERMINED IN THE FIRST STEP. BASED ON THE POROSITY AND PERMEABILITY MAP AT DIFFERENT DEPTH LEVELS, THE TARGET RESERVOIR IS RANKED AND INTERPRETED IN TO SIX LOGFACIES (LOGFACIES 5 WITH BEST RESERVOIR QUALITY, LOGFACIES 4 WITH GOOD RESERVOIR QUALITY, LOGFACIES 6 WITH MODERATE RESERVOIR QUALITY AND LOGFACIES 1, 2 AND 3 WITH NON TO BAD RESERVOIR QUALITY).

 
Keyword(s): 
 
Yearly Visit 17   tarjomyar
 
Latest on Blog
Enter SID Blog