Time Depth Convertion
Time Depth Convertion
Abstract The current paper describes an innovative methodology for time to depth conversion based on the construction of a velocity 3D grid. This method addresses many of the geometric limitations that characterize conventional time to depth conversion methods based on storing of 2D grids what makes it particularly suitable for time to depth conversion of regional as much as field data in complex salt tectonic structural environment. Both the remarkable reduction in work time and increase in accuracy of the velocity model evidenced by the quality control have demonstrated the advantages of the proposed methodology. Conventional methodology for time to depth conversion is based on a 2D grid velocity model that allows to structure in an organized way the individual contribution of the different stratigraphic units to the general velocity model. The logic laying under this methodology is that the contributions of the different stratigraphic units to the general velocity model are contained by the surfaces that limit these units, for this reason it is necessary that each of these surfaces are defined in a continuous way covering all the work area In the case of salt tectonics, the interpretation of the top and base of salt bodies are not continuous over all the work area so the conventional methodology for time to depth conversion can only be applied after these intrusive salt top and base surfaces have been extrapolated to all the area of the velocity model (assigning null thickness to the areas where salt was not present). The alternative solution proposed on this paper is based on the use of a 3D velocity grid, this solution starts with the construction of a simple 3D grid and a surface model excluding the top and base of the salt bodies. Based on this surface model a structural framework is built. This structural framework is populated with a 3D velocity model based on well and seismic velocity information. The fact of having the different velocity data upscaled to a 3D grid facilitates their handling, editing and integration. On the other hand a 3D flag grid is generated, this grid flags the cells contained in the salt bodies and is created based on the Salt top and Salt Base original interpretations or on seismic attributes filtering. Later an algebraic operation allows to assign salt velocity to all the flagged cells while keeping the initially populated velocity to all the other cells (without salt). This final interval velocity model has integrated all the geologic complexity. Finally the interval velocity 3D grid is transformed into an average velocity 3D grid for each cell of the volume. This last interval velocity volume allows to convert from time to depth all the structural framework as much as every individual object (surfaces, polygons, seismic, etc) intersecting this velocity volumes. The proposed workflow was tested on two key projects, The first project is a 30.000 square kilometers regional study involving 140 data wells, 35 regional cross sections and 15 horizons affected by many different salt intrusions at offshore Mexico. Here the salt bodies have been mapped but due to their lack of continuity their integration in the velocity model was a difficult task. The application of this methodology allowed to reduce the time for the depth conversion to a 20% of the originally planned time with a remarkable increase in accuracy. The second project was a Mesozoic mature oil field in Mexicos South Region which shows a highly complex structural setting, salt tectonic has played a fundamental role in deposition and deformation, complex shaped salt bodies are present inserted within the tertiary and Mesozoic column. To avoid shadow effects during the previous depth conversion process the interpreter had to extend the respective top and base of anomalous velocity bodies for the whole area of interest, it implies the building of fake horizons and a very time consuming grid merging process. The oversimplification of not considering those velocity anomalies embedded within the key zones implies the failure in the depth prognosis process. The application of this innovative methodology helped to solve this problem. A 3D velocity grid is modeled considering the main structural and stratigraphic features. Interval velocities from Seismic and from Check Shots are combined and populated. The anomalous velocity bodies are interpreted in the vicinity via manual tracking or seismic amplitude filtering. After being isolated, the bodies are related to different discrete properties as a facie, later the process of appending the connected bodies starts via conditional operations. The result is a new Interval Velocity 3D grid that has considered the complex spatial geometry of high velocity shallow sands, as well tertiary and Mesozoic interspersed salt bodies. This Interval Velocity grid is vertically integrated and mathematically operated to be transformed into the Average Velocity 3D grid we used in the Depth Conversion process. Finally the error in the prognosis process was considerably reduced to a maximum of 100 m for wells with average total depth of 5000m.
CONVENCIO NAL
NO CONVENCIO NAL
TRADITIONAL TIME TO DEPTH CONVERSION, IS BASED ON THE INTEGRATION OF VELOCITIES BETWEEN SURFACES
THE CONVENTIONAL 2D APPROACH REPRESENTS ALL THE VELOCITY VARIATIONS BETWEEN Hz A AND Hz B WITH ONLY ONE VALUE.
2- LIMITATION DESCRIBING DISCONTINUOUS AND VELOCITY ANOMALOUS BODIES (SALT, SHALE, LIMESTONE)
2550m/s
t0
2350m/s
t0
V1 V1
t0 t1 tcs12 tbs12 t2
V1
t1 t2
t1 t2
V2
V2
2130m/s
V2
V3
t3
V3
V3
t3
t3
THIS VELOCITIES ARE USED TO CREATE DATA POINTS THAT ARE USED TO CREATE A 2D VELOCITY GRID
WHEN THIS BODIES ARE PRESENT THEIR BASE AND TOP HAVE TO BE EXTENDED (WITH NULL THICKNESS OUTSIDE THE BODY) TO ALL THE WORK AREA TO PREVENT THE SHADOW EFFECT OVER THE DEEPER HORIZONS. THIS USUALLY MEANS COMPLICATE SURFACE MERGING OPERATIONS
2550m/s
2350m/s
t0
V1
2130m/s
t0
V1
t1 t2
t1 t2 tcs12 tbs12
V2
V2
V3
t3
V3
tcs23 tbs23 t3
THIS SIMPLE MODEL CANNOT HONOR PROPERLY THE VELOCITY VARIATIONS EXISTING IN THE RESERVOIRS. EVENTUALLY AN INTERVAL VELOCITY OR AN APPROXIMATED VELOCITY FUNCTION IN THE FORM V=V0+KZ CAN BE FOUND BUT THIS NOT SOLVE ALL THE METHOD LIMITATIONS
Z3 = 1/2 [(v1(t1-t0)) + (v2(tcs12-t1)) + (vs (tbs12-tcs12) + (v2(t2-tbs) + (v3 (t3-t2))] Z3 = 1/2 [(v1(t1-t0)) + (v2(tcs12-t1)) + (vs (tbs12-tcs12) + (v2(t2-tbs) + (v3 (tcs23-t2)) + (vs (tbs23-tcs23) + (v3 (t3-tbs23)) ]
1- Build 3D grid
DURING THE TUNING OF THIS WORKFLOW WE HAVE WORKED WITH FIELD DATA AND GRIDS IN HUNDRED METERS AND WITH REGIONAL DATASETS WHERE 1500 METER SIDE GRIDS WHERE USED. THE GRID CAN BE A CLASSICAL PETREL GRID OR A SIMPLE VERTICAL GRID. APPROPRIATE LAYERING OF THIS GRID, FOLLOWING THE BASIN FILL STYLE WILL BE IMPORTANT AT THE VELOCITY POPULATION PROCESS.
3- IN THE TWO FIGURES ABOVE YOU CAN SEE THE WELL SONIC CALIBRATED VELOCITY AND THE SEISMIC PROCESS INTERVAL VELOCITY UPSCALED TO THE 3D FRAMEWORK, STATISTICS AND HISTOGRAMS ALLOW TO QC THIS PROCESS.
4- SINCE ALL THE DATA ARE WORKED AS GRID VALUES REPRESENTING A LIMITED VOLUME OF ROCK AND SINCE INTERVAL VELOCITY IS A MORE CLEARLY PERCEIVED VARIABLE IS SIMPLE AND FAST TO DO THE QUALITY CONTROL OF THE INPUT DATA AND TO ANALYZE THE RELATIONSHIP BETWEEN WELL AND SEISMIC DATA SEE AT THE FIGURES ABOVE, HOW THE OUTLIER SAMPLES ARE IDENTIFIED AT THE TIME VS VELOCITY CROSSPLOT AND THEN VISUALIZED IN THE 3D VIEW TO ANALYZE THEIR 3D SPATIAL DISTRIBUTION.
4000 4000
4800 4800
5600 5600
6400 6400
7200 7200
8000 8000
8800 8800
9600 9600
10400 10400
11200 11200
12000 12000
6- THE FIGURE ON THE RIGHT SHOWS THE FINAL VELOCITY GRID THAT HONORS THE WELL HARD DATA AS MUCH AS THE LATERAL VARIATION INTRODUCED BY THE SEISMIC DATA. THIS VELOCITY GRID IS NOT YET CONSIDERING THE ANOMALOUS VELOCITY BODIES.
7- FLAGGING OF THE ANOMALOUS BODIES IS A KEY COMPONENT OF THIS WORKFLOW, DURING THE TUNING OF THIS WORKFLOW WE HAVE DONE THIS IN TWO WAYS:
A- GENERATING A DIFFERENTIATING FACIES BETWEEN THE TOP AND THE BASE OF THE ANOMALOUS BODIES USING PETREL GEOMETRICAL MODELING TOOL ASSIGN FACIES BETWEEN SURFACES (SEE FIGURE ON THE RIGHT) B- THE SECOND METHODOLOGY THAT WE HAVE APPLIED WAS TO CALIBRATE THE RELATIONSHIP BETWEEN VELOCITY AND SOME SEISMIC ATTRIBUTE (USUALLY ENVELOPE) AND ASSIGN THE HIGH VELOCITY CELLS WITH THROUGH A CALCULATOR OPERATION (SEE FIGURES BELOW).
-4000 -4000
5- ON THE FIGURE AT THE LEFT WE SEE A CROSSPLOT TIME VS INTERVAL VELOCITY FOR THE SEISMIC DATA (GREEN) AND THE WELL DATA (RED) THIS SHOWS A GOOD COINCIDENCE BETWEEN BOTH VELOCITIES AND FACILITATES THEIR INTEGRATION THROUGH A GUIDED ALGORITHM (COKRIGING, ETC). IN OTHER CASES DURING THE CALIBRATION OF THIS WORKFLOW WAS NEEDED TO APPLY CORRECTION COEFFICIENTS TO THE SEISMIC DATA WHEN ENTERING INTO MESOZOIC RESERVOIRS.
-3600 -3600
-3200 -3200
-2800 -2800
-2400 -2400
-2000 -2000
-1600 -1600
-1200 -1200
-800 -800
-400 -400
-1600 -1600 -2000 -2000 -2400 -2400 -2800 -2800 -3200 -3200 -3600 -3600 -4000 -4000
7 (CONT) THE RESULT OF THE FLAGGING PROCESS IS THE CREATION OF A FACIES GRID WHERE THE ANOMALOUS BODIES WERE ASSIGNED A FLAG VALUE TO DIFFERENTIATE THEM FROM THE REST OF THE FRAMEWORK. BASED ON THIS FLAG VALUE CALCULATOR OPERATIONS ALLOW TO ASSIGN SPECIFIC VELOCITY VALUES TO THOSE CELLS ON THE FIGURES IN THE RIGHT CAN BE SEEN THE FLAG CELLS OF A SALINE DOME CREATED TO ILLUSTRATE THIS PRESENTATION 8 ONCE THE ANOMALOUS BODIES FLAG GRID WAS CRATED A SINGLE CALCULATOR OPERATION ALLOWS TO ASSIGN A VELOCITY (IN THIS CASE SALT VELOCITY) TO THE FLAGGED CELLS. 9- WHEN ANOMALOUS VELOCITY HAS BEEN ASSIGNED TO FLAGGED CELLS A CALCULATOR OPERATION ALLOWS TO MERGE BOTH VELOCITY GRIDS. IN THE FIGURE AT THE LEFT CAN BE SEEN THE MERGE BETWEEN THE ORIGINAL INTERVAL VELOCITY GRIDS AND THE SALT VELOCITY GRID ASSIGNED TO THE DOME CELLS.
Vavge = Z/T=vi*ti/ti
10 THE CONVERSION OF THE INTERVAL VELOCITY GRID TO THE AVERAGE VELOCITY GRID IS BASED ON THE EQUATION Vavrg = Z/T = Sum(vi*ti) / Sum (ti) THE ti GRID IS GENERATED AS THE CELL HEIGHT GRID IN GEOMETRICAL MODELING.
THE Sum OPERATION IS THE sum vertical OPERATION, ALLOWED UNDER GRID OPERATIONS.
11 THE TIME TO DEPTH CONVERSION IS A CONVENTIONAL PROCESS IN PETREL. THE DIFFERENCE IN THIS CASE IS THAT INSTEAD OF USING SURFACE LIMITED VELOCITIES THIS MODEL IS CONVERTED BASED ON A VELOCITY VOLUME THAT ALLOWS TO HONOR IN A BETTER WAY THE VELOCITY HETEROGENEITY AND THAT ALLOWS TO MODEL FAST AND IN AN ACCURATE WAY COMPLEX SITUATIONS AS FOR EXAMPLE THIS SALT DOME.
Analysis