ADVERTISEMENT

Water-Injection Techniques Increase Recovery in Conventional Gas Reservoirs

You have access to this full article to experience the outstanding content available to SPE members and JPT subscribers.

To ensure continued access to JPT's content, please Sign In, JOIN SPE, or Subscribe to JPT

Primary gas recovery for a volumetric reservoir ends when the reservoir pressure declines below the value required to flow gas to the surface at the sales-line pressure. Secondary gas-recovery techniques can then be used to increase the recovery. The most common of these techniques is gas compression, but another feasible technique that is rarely explored is water injection. This paper evaluates the incremental benefit of water injection in a conventional gas reservoir when compared with gas compression.

Introduction

An operator hopes to increase the recovery in Reservoir TM-20 in the Sorrel Field offshore Trinidad and Tobago by use of water injection once the technique is deemed more economical than gas compression. The wells in the reservoir are drilled from the Tamarind Platform, which also acts as a hub for the separation of fluids from fields in the immediate vicinity. Gases and liquids are piped through separate lines, with the liquids heading to the onshore liquid-processing facility. However, the company can modify the platform separators to operate in three phases and use the produced water for injection. Company records show that the average water-production rate from the platform is approximately 5,000 BWPD. If the shut-in wells were to be opened, roughly 10,000 BWPD of produced water and 15 MMscf/D of gas would be added for each well until liquid loading, thus increasing gas output significantly.

Field History

The Sorrel Field is a retrograde gas-­condensate field located in the Columbus Basin, which has been producing from the TM-20 reservoir since 1998. The reservoir is a faulted sandstone formation of good quality that is divided into two hydrocarbon-bearing segments and underlain by an active aquifer. Segments 1 and 2 have been producing from Wells X and Y and Wells A and B, respectively, as shown in Fig. 1.

Fig. 1—Reservoir TM-20.

 

Unlike Segment 1, primary gas recovery in Segment 2 ended because of a decline in the reservoir pressure, thus making it a suitable choice for water injection. Segment 2 is a wedge-shaped structure bounded by two faults, with both faults being interpreted as sealing. Production in Segment 2 began in April 1999 with a reservoir pressure of 4,732 psig. It ended in January 2014 because of an inability to flow at the sales-line pressure of 950 psig.

Wells A and B are both slightly deviated, passing through the entire formation, and are on opposite ends of the segment. The completions type in both is cased-hole gravelpack with wire-wrapped screens. Log data for Well A showed the sand in that region was clean and blocky, while data for Well B showed the sand was interbedded with shale. This implied that there was degradation in sand quality from the northwest to southeast and toward the aquifer sands. The average porosity within the segment was 25%, with permeabilities ranging between 200 and 450 md. The water saturation averaged 10% within the segment and showed an increase with depth and proximity to the aquifer. Core samples showed the formation to be highly compressible.

Methodologies

Material-Balance Method. This is the most common means of determining drive mechanisms. The main drive mechanisms for most gas reservoirs are volumetric drive and water influx from an underlying aquifer. Once there is enough production and pressure history available, a mole balance of the materials entering and exiting the system can be used to generate graphs that provide an indication of the classification and strength of the drive mechanisms within that system.

One such graph shows the ratio of reservoir pressure (P) to gas-deviation factor (Z). This P/Z plot is derived by replacing the moles in the material balance with their gas-law equivalent. Extrapolation of the plot to atmospheric pressure provides an estimation for the original gas in place. When the plot deviates from the typical linear relationship, it indicates that water influx is taking place. If the deviation is clear, it is possible to align the plot with the early-time data, when the effects of encroachment are minimal, to obtain a rough approximation of the original gas in place.

The Cole plot is another such graph, derived by expressing the material balance in terms of the volume changes of material in the system. It is used primarily as a diagnostic tool for distinguishing between volumetric and water drive. Extrapolation of the plot for a water drive to the vertical axis also yields an approximation for the original gas in place, because effects of water influx have not begun to take effect.

Building a Reservoir Model. The reservoir model for Segment 2 was built using a commercial software program. After the production history was entered, the P/Z and Cole plots were generated. In the P/Z plot, a gentle protuberance at the center of the graph was observed, which gave some indication of compaction drive. In the Cole plot, the graph appeared to have the signature of a moderate water drive that subsided with time. This seemed unlikely, however, and it was inferred that the changing formation compressibility with reservoir pressure, as seen with the core samples, had an influence on production in this segment.  

The presence of an aquifer within the segment meant that the effects of water drive needed to be included in the reservoir model. The aquifer appeared to act radially from segment geometry and corresponded to the Hurst-Van Everdingen model. By aligning to early- and late-time dates in the P/Z and Cole plots, the gas in place was found to be between 345 and 385 Bcf. Regression was then performed on the aquifer properties. History matching showed that the effect of the water drive contributed little to production. Compaction drive was found to be the main contributor of energy in the system.

The formation compressibility at the initial reservoir pressure, once averaged, was increased nonuniformly for declining reservoir pressures, similar to the trend observed with the core samples, until the history match was obtained. The P/Z plot (overpressured) and the Cole plot (modified), both of which removed the effects of compaction and water drive, showed linear trends. This indicated that the regressions performed and values entered quantified the drive mechanisms in the segment accurately, producing a suitable reservoir model with a gas in place of 365 Bcf. The fact that the aquifer presence had little effect on production compounded the importance of understanding drive mechanisms in a reservoir.

Building Well Models. The well models for Wells A and B were built with commercial software. The inflow-performance relationship for both wells was calculated while Gray’s correlation was used to simulate the vertical-lift performances. Well-test data were used to determine the skin in each well, with Wells A and B having values of 90.02 and 8.49, respectively. Such a high skin in Well A was believed to be the result of possible plugged perforations. Sensitivity analysis showed that the minimum reservoir pressure needed for flow at the sales-line pressure of 950 psig was 1,350 psig for Well A and 1,320 psig for Well B, with production rates of 10 and 8 MMscf/D, respectively.

Production Forecasting. The production forecasting for gas compression and water injection was performed with commercial software. For gas compression, both wells were used for production, with the Darcy flow coefficients for Wells A and B calculated to be 49.42 and 6.46 psi2/Mscf/D, respectively. For water injection, Well B was used for injection because it was closer to the aquifer edge, with the injection index calculated to be 58.52 BWPD/psi. Various production scenarios were simulated for each technique to deduce the optimum solution. For each gas-compression scenario, the maximum production rate was varied to test different compressor sizes, with the minimum tubinghead pressure being 500 psig, common with most projects of this type. For each water-injection scenario, the maximum injection rate was influenced by the hub water-production rates with the shut-in wells open, with 10,000 BWPD deemed the most feasible. If seawater was included, however, the possible injection rate would be much higher.  

Results and Conclusions

Tables 2 through 5 of the complete paper provide production and financial summaries for gas-compression and water-injection scenarios. On the basis of simple economic evaluations and financial estimates derived from projects offshore Trinidad, the following conclusions were made regarding water injection when compared with gas compression. 

  • Water injection yielded similar net cash flows to gas compression over a longer time period, which essentially translates to lower net present values and internal rates of return.
  • Treating seawater to increase the injection rate had a negative effect on economic performance, while using a small compressor over a longer period of time had a positive one.
  • If the gas production revenue from each of the shut-in wells was factored in for water injection, approximately $4.55 million would be added per well if a lifetime of 1 year was assumed.
  • If the savings from avoiding commercial water disposal was factored in for water injection, approximately $90.50 million can be considered as revenue for each water-injection scenario.
  • If a company needs to meet contract demands quickly, then gas compression using a small compressor is the ideal option. If the company is already meeting its contract demands, however, then water injection is a viable option.

For a limited time, the complete paper SPE 191206 is free to SPE members.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 191206, “Exploiting Water-Injection Techniques for Increasing Gas Recovery in Conventional Gas Reservoirs,” by Keval Darvesh Rambaran, SPE, Sarah Tammie Chin Chee Fat, SPE, and Lugard Evans Layne, SPE, University of Trinidad and Tobago, prepared for the 2018 SPE Trinidad and Tobago Section Energy Resources Conference, Port of Spain, Trinidad and Tobago, 25–26 June The paper has not been peer reviewed.

Water-Injection Techniques Increase Recovery in Conventional Gas Reservoirs

01 November 2018

Volume: 70 | Issue: 11

STAY CONNECTED

Don't miss out on the latest technology delivered to your email weekly.  Sign up for the JPT newsletter.  If you are not logged in, you will receive a confirmation email that you will need to click on to confirm you want to receive the newsletter.

 

ADVERTISEMENT

ADVERTISEMENT

ADVERTISEMENT

ADVERTISEMENT

ADVERTISEMENT