Could 4D Unlock Your Field’s Value?

4D interpretation is now regularly applied to many assets by major oil companies. Its wider application in the industry is justified using modest assumptions about the reservoir management benefits of the new information derived. The extension of simulation to include the matching of time lapse seismic data can identify flow barriers and the reserves that are compromised by them. The additional data reduces the uncertainty in predicted production further than is possible by matching well performance data alone. Maximum value is obtained from TLS by tailoring an integrated 4D interpretation process for easy repetition as information is derived from the new data. Here Neil Dunlop (Neil.Dunlop@enscitech.com), Director of Engineering at Energy Scitech, Richard Margesson (Richard.Margesson@hydrosearch.co.uk) a Geophysicist with Hydrosearch and Chris Dawson, a Principal Petroleum Engineer with Energy Scitech look at the benefits that time lapse seismic data can bring to reservoir management.

Introduction

Time lapse seismic (TLS) is often misunderstood as being 4D. It is better to characterise 4D as the process that combines all historical data and new observations together to understand the behaviour of the reservoir under uncertainty. The objective is to build a calibrated tool with which to predict reservoir behaviour. TLS is a key additional source of historical data with a unique volumetric distribution that is typically available at only one time. The other observations used in reservoir simulation history matching remain relevant to creating a coherent predictive model.

TLS alone is not a technique for directly providing an estimate of saturation change. A 4D interpretation is required. It goes beyond seismic and is the integration of many disciplines: geology, geophysics, petrophysics and reservoir engineering. It is a key tool for uncertainty quantification and reduction. There is a strong focus on understanding fluid displacements to support decisions impacting inter-well flow behaviour and new well locations and interventions. Some of the most important outcomes sought from any 4D study are the identification of flow barriers and paths and the location of bypassed hydrocarbons. It is this information that can make 4D a cost effective contribution to reservoir management by helping to more precisely locate new wells, side tracks or re-drill locations. Rapid derivation of information from newly acquired TLS data is often critical as the value of most interventions is reduced by delay. This encourages careful preparation and an iterative interpretation procedure that provides more fully resolved results as it proceeds.

A rational decision to perform TLS data acquisition and 4D analysis requires an understanding of the potential benefit and of the technical feasibility in each case. The economic case can be simply illustrated by considering the expected monetary value, EMV, of 4D in a field where a new well is planned. As Figure 1 shows, using the modest assumptions explored later in this article, 4D substantially reduces the risks of the well, reducing the target payback oil required for a proposed well from 600 to 275 thousand barrels.

TLS data acquisition and seismic interpretation

Time lapse seismic accompanied by 4D interpretation is now an established technique for reservoir monitoring. It involves analysis of multiple reflection seismic surveys combined with reservoir description and reservoir simulation to track and predict the movement of fluid saturation fronts in a reservoir. This provides key information for optimum reservoir management.

The use of TLS is based on the principles that the response of the earth to seismic waves is a combination of the response of the solid rock and the fluids within them. After hydrocarbons have been produced from a reservoir, the fluids will have changed, but the rock is assumed to remain constant (which may not be entirely the case). These fluid changes may take a number of forms. There may be movement of gas, oil or water giving rise to changes to vertical contacts (GOC, OWC, GWC), or horizontal fluid fronts. Gas may come out of solution to form secondary gas caps. There may be changes in saturation in any layer due to pressure changes, as well as removal of hydrocarbons. Pressure changes may be up or down depending on whether production is by natural depletion or by injection. Changes in fluid saturation, pressure and/or temperature change the properties of the rock system and its seismic response.

If a baseline 3D seismic survey is carried out before production followed by one or more monitor surveys, separated by sufficient time for fluid movement due to production to have taken place, differences in the seismic response due to fluid changes may be identified. An example of the seismic differences in cross section is shown in Figure 2. The magnitude of the fluid effect on the seismic response can be small compared to other effects but may be measurable as a difference if all other causes of variation are minimised. It is generally assumed that the rock geological properties remain the same. In order to reduce other causes of variation, it is desirable that the acquisition and processing of the two surveys should be as similar as possible. This can sometimes result in a conflict as seismic technology inevitably will have advanced between the surveys and the optimum 3D monitor survey may be non-optimum for comparison with the original baseline survey.

Rock physics describes the response of the rock-fluid system to seismic waves. The seismic differences may be observed in a number of ways, among which the most important and detectable are amplitude and time changes. There will be a change in acoustic impedance as both compressional velocity and, sometimes to a lesser extent, density change with fluid. This impedance change will affect the amplitude of the reflection events in and around the reservoir. Change in velocity will also affect the travel time of events below the reservoir. This time shift may also be seen as a change in wavelet shape of the base reservoir and subsequent events. The AVO response may be changed by production, particularly on the far offsets, which are sensitive to fluid changes.

4D based Reservoir Performance Prediction

Effective reservoir management is improved by deriving a reservoir model that identifies the range of reservoir parameters that can match production history and using them to predict future reservoir performance under uncertainty. The 4D inversion workflow (Figure 3) is an extension of the conventional history matching workflow routinely employed in reservoir simulation studies that match well behaviour. Reservoir parameters are updated iteratively until the flow model results conform to all the observations including seismic. It is important to use a 3D geological model built from well and seismic information, both to resolve the seismic observations for matching, and to validate proposed matches from the flow modelling in the seismic domain. Once validated, the range of reservoir models that are consistent with the data allows calibrated predictions of future fluid movement under uncertainty.

The broader arrows in this workflow diagram highlight the inner interpretation loop that is iterated to derive the final reservoir model used in predictions. The close integration between reservoir engineering and geophysics is emphasised by the colour code.

Information derived by 4D is often substantial and unexpected but its value usually decays at its greatest rate immediately after the TLS is shot. Interventions like adjustments to offtakes from the existing well pattern, new wells or re-drilling tend to reduce in value and increase in uncertainty as the TLS data becomes older. Shorter projected field life also shortens the time available for decision making. Energy Scitech, with Hydrosearch, has focused on reducing the time required for the history matching loop as it is time consuming using the widely available current technologies. The EnABLE™ uncertainty estimation and history matching system substantially reduces the time to accomplish history matches, often by a factor of four or more. By using this system with existing reservoir simulation packages inside the larger iterations the overall time required to create an effective predictive model is reduced. Because the workflow requires the repetition of broadly similar history matching exercises, EnABLE™’s ability to identify a range of history matches is particularly compelling in a 4D interpretation.

Reservoir modelling incorporates direct measurements from wells but most information between the wells is inferred. TLS data provides an independent, indirect measure of fluid movements between wells highlighting areas which may not be swept optimally, or are bypassed. Areas of gas or water coning round wells may be identified. Better understanding of the reservoir behaviour then allows optimisation of the way in which each well is produced and targeting of in-fill wells.

4D can be used to monitor the movement of fluid fronts, identify potential barriers to flow, whether these are due to faulting, stratigraphic changes, or other phenomena that are unobvious before flow occurs. It may be possible to help understand and clarify the fluid movement in areas where it is difficult to obtain a good history match from the current reservoir model and so more closely predict behaviour. EnABLE™ identifies multiple alternative matched models that provide the basis for estimation of prediction uncertainty (Figure 4).

Is 4D worthwhile?

Decision tree analysis indicates the possible economic benefits of using 4D. The illustrative scenario is one in which 4D is a candidate for reducing the risk associated with drilling a well to access by-passed oil. The cost of one additional seismic survey and the incremental costs of 4D interpretation over conventional reservoir management techniques is assumed to be $2 million. The incremental reserves target illustrated is 1 million barrels of oil. Oil price is taken to be $25/bbl and operator after tax netback as 50%. The cost to drill the well is assumed to be $5mm. The total spend could equally be for the cost of a side-track, a well intervention, workover or re-completion.

Two cases, one using 4D and the other using conventional tools, illustrate how the decision to drill the well or not drill a well is dependant upon the following key factors:

• The data available and the quality of that data
• The interpretation of that data
• The risk of a failed well

In the 4D case it is probable that data quality and the enhanced interpretation will help to reduce the risk associated with drilling the well. Hence a decision to drill the well is more likely and has been weighted more positively in the 4D case at 60:40, against 50:50 in the no 4D case.

4D interpretation can actually identify, validate and visualise the areas where there is by-passed oil. The risk associated with drilling the well is therefore greatly reduced as it is likely to be placed accurately. A split of 80:20 has been assumed, while in the no 4D case it is not unreasonable to assume that even after the decision has been made to drill a well, there is still a high risk of failure on that well. A split of 40:60 has been assumed. The cost of failure in the 4D case includes the cost of the survey and its interpretation, which results in a total cost of $7mm, as against $5mm for the no 4D case. These alternatives are illustrated in the following two decision trees (Figure 5):

The resulting EMV for the 4D case is $4.4mm versus an EMV of $1.0mm for the no 4D case.

Clearly many assumptions in this simplistic analysis can be altered, one of them being reserves size. To gain some insight into the minimum economic reserves (reserves that generate an EMV of $0mm, with all other assumptions remaining the same) a plot was created of reserves against resulting EMV. It can be seen that with the 4D case the minimum economic reserves are of the order of 275,000 bbls and for the non 4D case 600,000 bbls.

The effect of different volumes of target oil expected (or identified by 4D) are illustrated in Figure 6 both for the two cases discussed above and a matching pair of cases whose total cost of a new well and incremental 4D interpretation costs is doubled to $14 million. Both 4D cases show positive EMVs for a modest target oil volume. Without 4D the target oil must be much greater.

Clearly, where the risks of drilling an unsuccessful new well can be substantially reduced through the use of 4D interpretation technology, then the resulting benefits are significant.

References/further reading on reservoir monitoring

1. Petroleum Geoscience Vol. 9, No 1, Feb 2003 is a special issue on 4D seismic technology.
2. Archer, S.H., King, G.A., Seymour, R.H., and Uden R.C., (1993). Seismic Reservoir Monitoring - The Potential. First Break, September 1993.
3. Dunlop, K.N.B. and Uden, R.C. (1989) Improved Reservoir Monitoring by Incorporating Seismic Data. Presented to the Nigerian Association of Petroleum Explorationists, Lagos, Nigeria, September, 1989.
4. Johnstad, S.E., Uden, R.C. and Dunlop, K.N.B., (1992) 3D Seismic Reservoir Monitoring. Presented to the SEG Annual Convention, New Orleans, La, October 1992.
5. Johnstad, S.E., Uden, R.C., and Dunlop, K.N.B. (1993) Seismic Reservoir Monitoring over the Oseberg Field. First Break, May 1993.
6. Lumley, D.E. (2001). Time-lapse seismic reservoir monitoring. Geophysics, 66 50-53.
7. Uden, R.C. and Little A.J.H. (1996): How to make a success of 4D seismic. IBC 5th annual conference on advances in reservoir technology 1996, March 18 1996.
   

Disclaimer: The material available on this website is designed to provide general information only. Whilst every effort has been made to ensure that the information provided is accurate, it does not constitute legal or other professional advice.
Please note: The Department of Trade and Industry cannot be held responsible for the contents of any pages referenced by an external link.