The use of interwell tracer tests is becoming increasingly important to the petroleum industry. In addition, the interwell tracer test has proven to be an efficient tool to investigate reservoir flow performance and reservoir properties that control gas and water displacement processes. Tracer data has been used to reduce uncertainty attributed to well-to-well communications, vertical and horizontal flow, and residual oil saturation. Unfortunately, at present, analysis of the tracer response is still largely qualitative in nature, and most information provided through the monitoring of the tracers is not quantitative. Among the numerous papers found on tracer technology, only a few include history matching of water-tracer test results and only one paper includes history matching of gas tracer test results. This paper describes the development of interwell tracer analysis methods in the petroleum industry, from the first qualitative tracer study in the 1950s to the latest quantitative tracer study in 2000. The results of our study indicate that only a small number of interwell tracer studies employed advanced numerical modelling methods. In addition, tracer analysis methods in the petroleum industry are not well studied. However, they are far better studied in the hydrology industry. Tracer modelling methods deserve to be given more attention, so that petroleum engineers can take better advantage of results from costly interwell tracer tests. Introduction Although tracer tests(1) were developed for tracking the movement of groundwater in the early 1900s, they were neglected by the petroleum industry until the mid-1950s. At this time, petroleum engineers(2, 3) started to conduct tracer tests for determination of fluid flow in waterflooded reservoirs. In the petroleum industry, solvent is sometimes injected into oil or gas bearing formations for the purposes of producing more hydrocarbons. Tracers can be added to the injected solvent to determine where the injected solvents go. The subsurface flow in the reservoir is anisotropic, and the reservoirs are usually layered with significant heterogeneity. As a result, solvent movement in the reservoir is difficult to predict, especially in reservoirs containing multiple injectors and producers. However, the flow paths can be identified by tagging solvents at each injection well with a different tracer and monitoring the tracers that appear at each producing well. Therefore, multiple tracers are often used for interwell tracer tests in the petroleum industry. Interwell tracers can provide information on flood patterns within the reservoir. This information is reliable, definitive, and unambiguous. Thus, it helps reduce uncertainties about flow paths, reservoir continuity, and directional features in the reservoir. Therefore, petroleum engineers can obtain information on reservoir continuity from the amount of each tracer produced from each well. Reservoir barriers can be identified by non-recovery or delayed recovery of specific tracers. At the same time, tracer test data can help determine residual oil saturation. Tracer test results also provide information on fracture characteristics in a naturally fractured reservoir. Interwell tracer tests have been applied in many petroleum producing fields across the world. The majority of these fields are located in North America and Europe.

No abstract

Tahiti prospect in deepwater Gulf of Mexico is a three-way anticlinal structure trapped against salt, with primary pay sands ranging from 24,000 to 27,000 ft TVD. The field contains several hydrocarbon-bearing turbidite sands. The discovery well was drilled in 2002, and two appraisal wells were drilled soon afterwards. Due to significant uncertainties remaining after appraisal, probabilistic methods were used to assess development plan alternatives and reserves. The primary purpose of this paper is to show the integrated earth modeling and reservoir simulation workflow, which incorporates uncertainty analysis and experimental design. The first step of the reservoir simulation workflow was the identification of the uncontrollable parameters that might influence performance predictions. These uncontrollable parameters included the earth model, water-oil contact location, faulting and compartmentalization, reservoir anisotropy, aquifer support, relative permeability curves, pore volume compressibility, rock compaction and dilation, fluid characterization, skin factor and well pressure drawdown. The analysis of available information from Tahiti field and analog fields with simple statistical techniques allowed the unbiased estimation of low, medium and high values of each uncontrollable parameter. The second step was the application of design of experiments to evaluate the impact of each uncertainty on reservoir performance. Monte Carlo simulation was used to estimate P10, P50 and P90 oil recovery and discounted oil recovery for each reservoir. Finally, we developed reasonable P10, P50 and P90 reservoir simulation models that incorporate the full range of each significant reservoir uncertainty. Probabilistic production forecasts for primary depletion and waterflood strategies supported project decisions like number of pre-drills, total well count, bottom-hole locations, waterflood strategies and facility capacity. Reservoir simulation results also supported proved reserves estimation and economic evaluation of the project. Introduction Tahiti prospect is located in Green Canyon blocks 596, 597, 640 and 641, about 190 miles southwest of New Orleans in deepwater Gulf of Mexico as shown in Fig. 1. The field was discovered in April 20021 with the drilling of well GC 640 #1 in slightly more than 4,000 ft of water. Chevron is the operator and holds a 58% working interest. Statoil holds a 25% interest, and Shell holds a 17% interest (Total E&P USA, Inc. purchased Shell's working interest effective January 20062). The field is expected to come on-line in mid-2008. Results from the exploratory well indicated the presence of high quality reservoir sand with total net pay of over 400 ft essentially distributed in three main Miocene turbidite sheet sands at depths ranging from 24,000 to 27,000 ft TVD. A downdip sidetrack and an updip sidetrack encountered net pays of 75 ft and 400 ft, respectively. In 2003 two appraisal wells were drilled simultaneously using two rigs. Each rig drilled a vertical well with an updip sidetrack. One of the appraisals encountered more than 1,000 ft of net pay in high-quality sandstones, confirming one of the most significant net pay accumulations in the history of the deepwater Gulf of Mexico.3–5 Consequently, the field was delineated from three surface locations and seven reservoir penetrations. The main sands were identified as M-21A and M-21B. For simplicity, they are referred as M-21. The drilling of two appraisal wells simultaneously is unusual, particularly in deepwater, but the benefit is that key reservoir information is obtained in half the time, allowing Chevron to begin development planning for the field sooner and ultimately shortening the time to first production.

This paper is the first window model study in the northern area of a large carbonate reservoir in Saudi Arabia. It describes window reservoir simulation with geostatistics to model uneven water encroachment in the southwest producing area of the northern portion of the reservoir. In addition, this paper describes performance predictions that investigate the sweep efficiency of the current peripheral waterflood. A 50 × 50 × 549 (240 m. × 260 m. × 0.15 m. average grid block size) geological model was constructed with geostatistics software. Conditional simulation was used to obtain spatial distributions of porosity and volume of dolomite. Core data transforms were used to obtain horizontal and vertical permeability distributions. Simple averaging techniques were used to convert the 549-layer geological model to a 50 × 50 × 10 (240 m. × 260 m. × 8 m. average grid block size) window reservoir simulation model. Flux injectors and flux producers were assigned to the outermost grid blocks. Historical boundary flux rates were obtained from a coarsely-gridded full-field model. Pressure distribution, water cuts, GORs, and recent flowmeter data were history matched. Permeability correction factors and numerous parameter adjustments were required to obtain the final history match. The permeability correction factors were based on pressure transient permeability-thickness analyses. The prediction phase of the study evaluated the effects of infill drilling, the use of artificial lifts, workovers, horizontal wells, producing rate constraints, and tight zone development to formulate depletion strategies for the development of this area. The window model will also be used to investigate day-to-day reservoir management problems in this area.