Print this page
Click to view article in PDF format.
Development of a Methodology for the Location of New Producer Wells from the Computational Modeling based on Streamline Simulation*
Olga Patricia Rueda1, Miguel Jose Bernal1, and Elkin Rodolfo Santafe1
Search and Discovery Article #40697 (2011)
Posted February 21, 2011
*Adapted from oral presentation at AAPG International Conference and Exhibition, Calgary, Alberta, Canada, September 12-15, 2010
1Industrial University of Santander (UIS), Alternative Technologies for Hydrocarbon Investigation Group (GITAH), Bucaramanga, Colombia ([email protected])
This work focuses on a specific application of estimation or allocation of potential productive zones which leads to the establishment of infill wells, and comparing the advantages of streamline simulation over conventional finite differences simulation. The streamline simulation has had great reception in the last fifteen years, being accepted as a complementary technology to conventional numerical simulation techniques which are only based on finite differences.
The implementation of a methodology for the allocation of new producer wells (where a history match is embedded) will be shown, in which weighted property maps are compared with the model streamlines, realizing that it is possible to see through a property called “time of flight” in which zones the injection front is found, and what zones are part of the drainage volume at the end of the production history. This leads to the evaluation of new producer well inclusion influence.
Copyright � AAPG. Serial rights given by author. For all other rights contact author directly.
|
Streamline and Time of Flight Definition
The utility and uniqueness of streamline simulation is situated under the context of the development of the main reservoir simulation tasks as the upscaling, the displacement efficiency quantification, the computational speed, the history matching and the field optimization. The streamlines are defined as instantaneous curves in the space along each of its points that are tangent to the local velocity vector. The time of flight is the time which an infinitesimal particle spends to reach a distance, s, along each streamline. Time of flight concept in flux modeling is used in oil reservoirs and mathematically is defined as:
Streamline Simulation Algorithm
A simulator based on streamlines has an IMPES solving procedure (implicit pressure, explicit saturation) and it undergoes the following sequential steps: (I) Initial conditions reading; (II) Pressure field computation from rock properties, fluid properties and other available data; (III) Tracing of streamlines along of the tridimensional domain; (IV) Definition of saturations along each streamline; (V) One dimensional material balance equations resolution along streamlines, then the updating of streamlines through time; (VI) The saturations that were calculated along the streamlines are re-allocated again to the tridimensional system; (VII) The pressure field is recalculated, then the above steps are repeated until reaching simulation time.
Streamline Simulation Advantages
The streamlines simulation models have significant advantages over the conventional simulations, mainly due to its computational speed, the conceptualization and visualization of flow between injectors and producers, the identification of drainage areas, the facility to classify complex geological and geostatistical models, and the obtaining of more accurate solutions. Besides the latter advantages the streamlines offer the alternative to undergo an automatic and/or assisted history match. Based on streamlines formulation a minimization of an objective function and sensibility coefficients (multipliers) computation can be reached. The sensibility coefficients relate production response changes with reservoir properties changes. Once the sensibilities are calculated they are mapped in the model and the integration begins with the support of an iterative algorithm that minimizes the difference between the observed and the new simulated response.
The methodology proposed for the location of wells in potentially productive zones was built from ideas extracted of applications proposed in the literature, proper experiences and suggestions of specialist in this area. The main components of methodology are the development of a history match, the generation of a unified weighted property map, the analysis of well production performance and the streamline tracing to visualize drainage volumes and invasion fronts. Undergoing the latter components, it is possible to locate prospective zones.
Application Case: Brugge Field
Brugge Field is located in the North Sea and it is basically an elongated and sectioned dome in east to west direction with a fault in the north limit (Figure 1) that is stratigraphically divided into four formations (Broom, Rannoch, Etive, Ness, and Tarbet) with fluvial, shoreface and sandy shelf depositional environments. From the original model we generated a model of 10 injector wells and 10 producer wells upscaled to 139x48x9 gridcells.
For the history match the observed water cut data corresponds to the data obtained from the original model (refining model) of the ten year production history. The region penetrated by the streamlines was selected at different times of flight and the multipliers (sensibility coefficients) were chosen to increase or decrease the permeability in each reservoir region. In zones where the simulated water cut response was higher compared with the observed water cut, multipliers that made the permeability lower were applied. Conversely, in zones where the simulated water cut response was lower compared with the observed water cut response, multipliers that made the permeability be higher were applied.
Once the history match was finished, the property weighted maps were built. This was done for permeability, porosity, pressure and oil saturation at the end of the production history (Figure 2). Once the weighted property maps were compared, the unified weighted map was built (Figure 3) whose result came from the sum of all the property weighted maps from each layer and the selection of those regions where the oil saturation was relevant.
Once the unified weighted map was generated, the next step was to undergo the field production distribution analysis. During this analysis it was seen that wells 12 and 15 were shut-in, and well 9 had high and low oscillations due to operational problems but the causes were not registered. The wells which gave the major part of the production were the wells that were located at the top of the structure (east to west direction) which had not been flooded by the injection front at the current time. From this analysis, this zone was considered a target zone for the location of new wells. It also was concluded that locating lower in the dome was not recommendable due to the flooding as shown in wells 12, 15, 18 and 20. Some of them were shut-in as a consequence of high water cut.
After having built the unified weighted map and the analysis of wells production history, the beginning of streamlines took place and so far the analysis of flux dynamics. The first filter was made with the streamline tracing until the final production time (Figure 4). From the gridcells penetrated by the streamlines the zones that were flooded were discarded after ten years of production (Figure 5). The second filter was made for high times of flight, in other words in zones where the sweep efficiency was poor and the injection front did not have the expected effect. Figure 6 shows the zone which had high times of flight (between final simulation time and 30 years). The streamlines traced at high times of flight were compared with the unified weighted map by layer with the objective of choosing the zones and specifically the layers where the new producer wells were perforated.
Figure 7 shows the comparison made for the first six layers. The coordinates of the four wells that were introduced are showed in Table 1. The wells introduced were controlled by minimum production pressure and by a maximum water cut registered in Brugge Field literature.
In Table 2 (where also normalized rates were reported), Figure 8 and Figure 9 it was observed that the recovery factor by well was minimum compared with the model without infill wells. However, that small difference in produced barrels represented approximately 17,600,000 barrels for the infill well #4. Wells which represented the higher recovery factor and the best location (top of the dome) whose near well conditions were convenient (BR-P-1, BR-P-3, BR-P-4, BR-P-7, BR-P-8), due to not flooding, and the accumulated production was the highest.
1. Due to streamlines formulation nature it is possible to employ a dynamic data inversion which allows integration of reservoir properties with production response.
2. The time of flight as the main parameter of streamlines formulation is characterized by having reservoir heterogeneity since it is a function of flux interstitial velocity and at the same time is a function of reservoir properties, pressure gradients, and gravitational effects.
3. For the methodology application it was necessary to work over a real model which had a defined geological structure, and an established flux dynamics.
4. It was determined that the best zone corresponded to the zone where the infill well #4 was located, as it was the well that showed the highest recovery factor increment and the highest instantaneous production peak with an increment of the accumulated production of 17,600,000 barrels.
Datta-Gupta, A,. and M.J. King, 1995, A semianalytic approach to tracer flow modeling in heterogeneous permeable media: Advances in Water Resources, v. 18/1, p. 9-24.
Oyerinde, A.S., 2008, in A.S. Oyerinde, (ed.) Streamline based three phase history matching: Texas A&M University Dissertation, 122 p. Web accessed 7 February 2011 http://repository.tamu.edu/handle/1969.1/85951
Peters, L., R.J. Arts, G.K. Brouwer, C.R. Geel, S. Cullick, R.J. Lorentzen, et al., 2010, Results of the Brugge Benchmark Study for Flooding Optimization and History Matching, SPE 119094: SPE Reservoir Evaluation & Engineering, v. 13/3, p. 391-405. http://dx.doi.org/10.2118/119094-PA
Pollock, D.W., 1988, Semianalytical Computation of Path Lines for Finite-Difference Models: Ground Water, v. 26/6, p. 743-750.
Thiele, M., and R.P. Batycky, 2003, Water injection optimization using a streamline based workflow, SPE 84080.
Rodriguez, R.D., and J.F. Bernal, 2009, Methodology for the Drilling of Infill Wells in a Mature Field with Fluvial Depositional Environment. Colorado Field Application, Industrial University of Santander, Bucaramanga.
|
Abstract 
