As part of the activities in a project developed years ago for an Oil and Gas company in Mexico, a Velocity Model was built to perform the time-depth conversion and establish, in the depth domain, the structural-stratigraphic framework of the (carbonate) deposit; the final objective was to identify new prospective zones in the field and support the development plan. It was a requirement of the project that the Petrel Software Platform was used to construct the velocity model.
To create velocity models, the Petrel Software Platform has a standard workflow. However, given the tectonic complexity of the reservoir under consideration (which includes the presence of a complex-shape saline body), it was necessary to design and implement an Out-of-the-Box methodology based on advanced geomodeling techniques that take into account the main structural features and therefore allow carrying out a reliable time-depth conversion, aiming to obtain a reliable structural-stratigraphic framework of the reservoir. Additionally, this methodology can be deployed REGARDLESS of reservoir type, namely, carbonate or clastic. So, this post will mainly emphasize the GEOMODELING techniques deployed to tackle the reservoir tectonic complexities and NOT on the geophysical technicalities. So, this post will be focused on the GEOMODELING techniques deployed to tackle the tectonic complexities and NOT on the geophysical technicalities. These will be discussed in another post.
The starting point is the visualization and validation of all required inputs. Any inconsistency in the inputs was reported and fixed by the geophysicist assigned to the project.
The inputs were the following: (1) boundary polygon of the area of study (approximately 77 km2), (2) basic information and deviations from 4 wells (shown on the base map in the figure above), and their TZ time-depth tables, (3) time interpretation of five horizons (plus one horizon for the datum and another horizon for the base of the model), and (4) surfaces to laterally delimit the interpreted saline body. The faults present in the reservoir were not included since, in the process of populating the model's grid, it will only be required to follow the stratigraphy established by the already mentioned seven horizons.
Regarding the latter, the horizons to be used will only be, following the geophysicist expert criteria, associated (i) with well-picks and (ii) with markers that delimit zones with significant speed changes, the so-called inversion zones. The figure below shows the plot of the interval velocities of the wells as a function of depth Z; In this, the two Mesozoic well-picks Ks and JSK are highlighted (correlated with gamma-ray and resistive logs, shown as an example in the PARETO-1 well to the right of the figure), and the seismic markers V1, V2, and V3 ( in the Tertiary) for three interpreted inversion zones. The stratigraphic column, into which the Velocity Model was vertically subdivided, was defined with the horizons associated with these five markers (plus one for the datum and other horizon for the base).
The next step is the construction of the model's grid; it integrates the results of the geophysical characterization: interpretation of horizons (in time), the well-picks, and associated seismic markers mentioned above. The surfaces with which the interpreted saline body was laterally delimited were not incorporated yet (they'll be used later). The image below schematically illustrates what was described in previous paragraphs; the six resulting velocity regions are shown to the right.
The Interval Velocity Volume is built in two stages, as shown schematically in the figure below (vertical cutaways have been made to visualize volumes interiors):
grid population
modeling of the complex-shape saline body
1. Grid population: typically done on Petrel using the Petrophysical modeling process. First, the Interval velocities in the wells are scaled to the resolution of the grid (Scale up well logs process). Next, a variographic analysis is carried out using the Data analysis process to characterize the spatial continuity of the interval velocity data, and then a Kriging algorithm is applied. Unfortunately, in the present case, and as a consequence of the sparsity in the data available (4 wells with TZ in an area of more than 77 km2), it was not possible to carry out the variographic analysis that drawn reliable results. Therefore, to interpolate the velocities throughout the grid, a Moving average algorithm was applied (following the established stratigraphy). The latest results were reliable enough - from the geophysicist point of view- to be used later in the workflow.
2. Modeling of the complex-shape saline body: using the grid built previously and the Geometrical modeling process in Petrel, a volume was built to which a facies template was associated with the codes 0 “no salt” and 1 “salt”; next, the interpreted salt body is “painted” in this volume, using the Facies modeling process and the surfaces that laterally (and vertically) delimited it. The central image in the figure above illustrates the result.
With these results and using the Properties calculator available in Petrel, the Interval Velocity Volume was generated. A velocity equal to 4500 m/s was assigned to the saline body. The image to the top right in the figure above shows this result.
Finally, the Volume of Mean Velocities Vavg is constructed using the Volume of Interval Velocities obtained so far and the formula (implemented in the Property calculator) shown in the figure below: the j-interval velocities vj in the volume are averaged, weighting them with the corresponding cell heights hj. The result is illustrated in the image to the bottom right in the figure below.
Finally, in Petrel's Make velocity model process (shown in the figure below), the Vave, the geophysical characterization, the well-picks, and the seismic markers are integrated. The process is found within Petrel's Geophysics in the Processes tab. By clicking OK and waiting a few minutes, a Velocity models folder is created within the Models tab. Everything is set to perform the time-depth conversion.
The time-depth conversion of maps, surfaces, fault sticks, grids, etc., can be done for each object or for all of them together. In the first case, as illustrated in the figure below, right-click on the item, and in the drop-down menu, select Domain convert by active velocity model.
Rigorous quality control of the results is imperative. It was carried out by the geophysicist assigned to the project. In the event of any inconsistency in the time-depth conversion, the cause must be detected and fixed. The process was repeated several times until all discrepancies were eliminated.
As a final cross-check, an NW-SW-NE cross-section is depicted in the figure below, showing a satisfactory match between the seismic, the well-interval velocities (PARETO-1) and those derived from the velocity model (shown in transparent colors). Notice the good lateral continuity along the speed regions. It is also anticipated that due to the sharp contrast between the velocities of the salt body and its surroundings, the structures present in time will be substantially modified when performing the time-depth conversion.
Indeed, in the depth map at the JSK level shown in the figure below, possible structural highs around the salt are evident (blank on the map). The figure also includes the limits established with a previous model and those proposed recently based on the new geophysical characterization of the area. Particularly striking is the area just to the right of the salt (indicated by the red arrow). Also of interest is the area to the NE, in the vicinity of the PARETO-1 well.
In this way, what concerns the construction of the Velocity Model is concluded. Write your comments and suggestions below and share. I'll be glad to answer all your questions.
I hope you have enjoyed reading this post as I did writing it. Techniques and methods like some of those presented here could be adapted, as I have already done in previous posts, to be applied in different knowledge domains: MINING INDUSTRY, PUBLIC SECURITY, RETAIL, etc.
Comentários