COMMON APPROACHES TO SIMULATE NATURALLY FRACTURED RESERVOIRS (NFRS) ARE DISCRETE FRACTURE NETWORK (DFN) AND THE DUAL CONTINUUM MODELS. THE FORMER SUFFERS FROM INAPPLICABILITY FOR REAL FIELD BECAUSE OF COMPUTATION COST AND THE LATER HAS CHALLENGE FOR APPROPRIATE DESCRIPTION OF THE MASS TRANSFER BETWEEN MATRIX AND FRACTURE. THE RECOVERY CURVE METHOD IS A NEW APPROACH INTRODUCED TO DEAL WITH THESE CHALLENGES. THIS METHOD DECOUPLES THE LOCAL MATRIX DEPLETION CALCULATION FROM THE REGIONAL FRACTURE FLOW AND CAN BE CONSIDERED AS A COMPROMISE BETWEEN THE TWO APPROACHES. THE MENTIONED METHOD ATTACKS THE CHALLENGE OF PROPER DEPICTION OF MATRIXFRACTURE TRANSFER FUNCTION. THIS PAPER PRESENTS AN APPROACH TO DETERMINE MATRIX-FRACTURE MASS EXCHANGE BY HISTORY MATCHING. A COLUMN MODEL FROM A REAL FRACTURED RESERVOIR IS EXTRACTED AND THROUGH HISTORY MATCHING, RECOVERY CURVES FOR WATER IMBIBITION AND GAS GRAVITY DRAINAGE ARE DETERMINED. BASED ON THE RECOVERY CURVE METHOD, THESE CURVES REPRESENT MORE CREDIBLY THE REALITY OF THE MATRIX-FRACTURE INTERACTION AND CAN BE APPLIED TO DESCRIBE MATRIX-FRACTURE COMMUNICATION TO FORECAST THE PRODUCTION. THE RESULTS SHOW VERY GOOD AGREEMENT WITH THE HISTORY OF THE MODEL. HYBRID GENETIC ALGORITHM IS USED AS THE OPTIMIZATION TOOL AND FOR UNCERTAINTY QUANTIFICATION NEIGHBORHOOD-BAYES ALGORITHM (NAB) IS EMPLOYED. THE BAYESIAN CREDIBLE INTERVAL (P10-P90) OF THE PROPOSED METHOD IS ACCEPTABLE.

HISTORY MATCHING IS THE PROCESS OF UPDATING RESERVOIR MODEL USING PRODUCTION DATA. IT IS NECESSARY BEFORE FORECASTING FUTURE PRODUCTION. THE PROCESS IS NORMALLY CARRIED OUT BY RESERVOIR SIMULATORS, WHICH IS VERY TIME-CONSUMING. AS A RESULT, ONLY A SMALL NUMBER OF SIMULATION RUNS ARE CONDUCTED AND THE HISTORY MATCHING RESULTS ARE NORMALLY UNSATISFACTORY. A GENETIC ALGORITHM IS APPLIED TO THE PROBLEM OF CONDITIONING THE ROCK PROPERTIES OF A RESERVOIR MODEL ON HISTORIC PRODUCTION DATA. THIS IS A DIFFICULT OPTIMIZATION PROBLEM WHERE EACH EVALUATION OF THE OBJECTIVE FUNCTION IMPLIES A FLOW SIMULATION OF THE WHOLE RESERVOIR. DUE TO THE HIGH COMPUTING COST OF THIS FUNCTION, IT IS IMPERATIVE TO MAKE USE OF AN EFFICIENT OPTIMIZATION METHOD TO FIND A NEAR OPTIMAL SOLUTION USING AS FEW ITERATIONS AS POSSIBLE. IN THIS STUDY WE HAVE APPLIED A GENETIC ALGORITHM TO SOLVE THIS PROBLEM. IN ORDER TO RUN RESERVOIR SIMULATIONS COMPOSITIONAL SIMULATOR E300Ò APPLIED TO BE COMPARED WITH OBSERVED DATA. THE MEAN SQUARE ERROR IS USED AS OUR FITNESS FUNCTION THAT SHOULD BE MINIMIZED. USING MATLAB GENETIC ALGORITHM TOOLBOX WITH 3 VARIABLES WILL RESULT IN MATCHING BOTTOM HOLE PRESSURE HISTORY OF WELL NO.2 OF THE RESERVOIR.

HISTORY MATCHING USUALLY IS DONE FOR MODEL VERIFICATION. A GOOD HISTORY MATCHING, HOW EVER DOES NOT GUARANTEE A GOOD MODEL. ALSO THE MODEL IS ONLY A REPRESENTATION OR AN ANALOGY OF THE RESERVOIR, BUT CHOOSING REASONABLE PROJECTION CAN TUNE THE MODEL FOR SPECIAL PURPOSES. UNLIKE CONVENTIONAL RESERVOIRS, DIFFUSION CAN AFFECT SIGNIFICANTLY ON AVERAGE RESERVOIR PRESSURE, FIELD WATER CUT AND FIELD GAS OIL RATIO OF FRACTURED RESERVOIR, BECAUSE MASS TRANSFER AND FLUID FLOW IS OCCURRED BY DIFFUSION BETWEEN MATRIX AND THE FRACTURE. IN BLACK OIL SIMULATION DIFFUSION TAKES PLACE WITH IN BOTH GAS AND OIL PHASES. THE CALCULATION OF THE DIFFUSIVITY IS ANALOGOUS TO THE CALCULATION OF TRANSMISSIBILITY, WITH THE PERMEABILITY REPLACED BY THE POROSITY. IN THE CASE OF FRACTURED RESERVOIRS MATRIX-FRACTURE COUPLING DIFFUSIVITIES ARE PROPORTIONAL TO THE OVERALL CELL VOLUME AND TO THE MATRIX CELL BULK VOLUME. IN THIS STUDY A CONSISTENT BLACK OIL MODEL IS USED TO STUDY THE EFFECT OF DIFFUSION ON MODEL PERFORMANCE. FIVE PHASES ARE PRESENTED IN THE MODEL INCLUDING OIL, WATER, GAS, DISSOLVED GAS AND VAPORIZED OIL. THE DIFFUSION COEFFICIENTS AND AVERAGE PHASE MOLECULAR WEIGHTS FOR EACH PHASE HAVE TO BE SUPPLIED. IN THE CASE OF CROSS PHASE DIFFUSION SEPARATE GAS IN OIL, GAS IN GAS, OIL IN GAS AND OIL IN OIL DIFFUSION COEFFICIENTS ARE CONSIDERED AND EFFECT OF EACH PARAMETERS ARE STUDIED.