Reservoir characterization

Understanding Well, Reservoir Performance With Modern Analytical Techniques

Analytical tools are useful for reservoir management and can provide simplicity while capturing information derived from events occurring at smaller time scales, which are ordinarily sacrificed in numerical simulations to keep run times reasonable.

jpt-2014-09-165995-hero.jpg

Analytical tools are useful for reservoir management and can provide simplicity while capturing information derived from events occurring at smaller time scales, which are ordinarily sacrificed in numerical simulations to keep run times reasonable. This paper demonstrates the use of a modified Hall analysis (MHA) to evaluate real-time performance of water injectors; reciprocal-productivity index (RPI) to identify time-varying-injection support in a maximum-reservoir-contact (MRC) producer from nearby water injectors; deconvolution and rate-transient analysis (RTA) to determine permeability, skin, and drainage area by use of real-time permanent-downhole-gauge (PDHG) data; and a Y-function based on the Buckley-Leverett (BL) equation and the assumption of a semilog relationship between the oil/water relative permeability ratio and saturation.

Introduction

Reservoir simulation provides a single platform where both static and dynamic data—such as geological, geophysical, and reservoir-fluid and -rock properties—are seamlessly integrated in a numerical model and used for performance prediction. Therefore, numerical simulation is the norm and the preferred tool for reservoir management. Despite many advantages, reservoir simulation normally is not able to account for dynamic changes occurring at smaller scales—for example, changing well conditions such as damage, plugging, and fracturing. In recent years, there has been an increase in the use of analytical tools with simulation models using surveillance data to understand well and reservoir performance.

PDHGs are being installed increasingly across the world, especially in long horizontal and MRC wells for reservoir surveillance and management because of their improved reliability and trouble-free long-term performance. Besides reducing ambiguity and uncertainties in the interpretation, long-term data also provide an insight on how reservoir parameters (e.g., reservoir pressure, effective permeability, and skin) may change as the reservoir is produced. Use of the high-frequency long-term data expands the traditional time scale from snapshot approach to a continuous evaluation without interrupting production. This can lead to real-time production and management decision, resulting in sustained production, recovery optimization, reduced operational expenditure, and maximized life-cycle economics.

Modified Hall Analysis

Plots are generated from modified Hall integral formulations for diagnostics of the matrix injection, plugging, fracturing, and channelized (highly conductive layer) flow. Please see the complete paper for the modified Hall integral formulation. In the plots, no separation of Hall integral and derivatives is indicative of continued matrix injection; upward separation means formation plugging; downward separation of derivative from integral indicates formation fracturing; and parallel separation of Hall integral and derivative indicates the presence of a high-permeability channel.

Performance Monitoring of an MRC Well

A 10,000-ft pilot MRC producer was drilled in an approximately 5- to 7-md area in a large carbonate reservoir. The well is completed with 6⅝-in. perforated liner with openhole swelling packers placed between the perforated liner sections. A PDHG was installed for continuous monitoring of pressure and temperature for well surveillance. The objectives were to monitor well performance continuously by knowing flowing pressure and productivity index; to evaluate pressure support from offset water injectors, including a line-drive 10,000‑ft pilot MRC injector drilled parallel to the MRC producer; to update the well model; and to determine permeability, skin, and drainage area. Understanding the performance of the well and water-injection support is critical from a field-development perspective because it will dictate drilling of additional MRC wells with a line-drive water-injection recovery mechanism in relatively better-permeability areas.

Understanding Voidage Balance and Water-Injection Support. An RPI plot was used to assess voidage balance conditions. See the complete paper for the formulation. In an RPI plot, positive slope indicates weak injection/aquifer support, negative slope means excessive voidage replacement, and a plateau indicates voidage balance.

Deconvolution Analysis. Deconvolution is a technique for converting the pressure and rate data obtained from a well operating under variable-rate conditions (long production-rate-data history with a relatively short-term set of pressure-buildup data) into a much simpler form of constant-rate drawdown-pressure-response function for the entire test duration, and it yields the corresponding pressure derivative normalized to a unit rate. This new, expanded set of pressure/time data can be interpreted, analyzed, and history matched for a well/reservoir system by use of pressure-transient-analysis. The main benefit of deconvolution is a larger area of investigation from the entire test history, and it allows recognition of more-distant reservoir features such as boundaries that may not be visible in individual buildup or flow periods. Deconvolving pressure data over variable time ranges has resulted in similar deconvolved derivatives, indicating no changes occurring in well/reservoir behavior. This type of analysis, therefore, should work as a surveillance alarm suggesting the need for closer examination of PDHG data by time-lapse well-test analysis should differences in the deconvolved derivatives appear. The differences at early time are wellbore-storage-related.

RTA. RTA is carried out by matching overall history of pressure and rate data with a horizontal-well model. A reasonably good history match of pressure and rate data provides confidence in the derived parameters. The RTA-model results corroborate those derived from deconvolution analysis.

Waterflood Analysis With Y-Function

On the basis of the BL theory and the assumption of a semilog relationship between the oil/water relative permeability ratio and water saturation, a consistent analytical solution is derived. See the complete paper for the equation to find the Y-function, a function of fractional flow.

Diagnostic Plots. Normal Displacement. There are two diagnostic plots that can be used for waterflood analysis. One is a log-log plot of Y vs. displacement time. Decline of Y with a slope of –1 is an indication of normal displacement efficiency. Change of volumetric sweep caused by operational changes will change the intercept, and the slope will remain at –1. Y decreasing with a slope less than –1 is an indication of low displacement efficiency caused by formation or operation.

Volumetric Sweep. The second diagnostic plot is a reciprocal time plot of Y vs. the inverse of displacement time. This plot is used to estimate the volumetric sweep. The relative permeability ratio parameter is estimated from simulation data or field performance. The slope of the curve of the Y-function plot vs. the inverse of displacement time after horizontal is used to estimate the volumetric sweep.

Waterflood Channeling (Fracture or High-Permeability Layer) and Premature Breakthrough. When premature water breakthrough happens, injected water channels to the producers without effectively displacing the oil. The new diagnostic method is able to quantitatively differentiate premature water breakthrough from normal water breakthrough in a waterflood. The authors compared the performance of two high-water-cut wells in two reservoirs. The Y-function for the normal-water-breakthrough well has a slope of –1 in the log-log plot. An irregular curve shape with oil-cut function Y at significantly less than 0.25, or any significant deviation from the slope of –1, generally indicates abnormal water production because of water channeling. This behavior is accompanied by large liquid rate. This qualitative assessment has proved useful for surveillance of waterflood production performance.

The Y-function for the premature-water-breakthrough well is lower than that for the normal-water-breakthrough well and has no consistent feature. The premature water breakthrough does not follow the BL frontal-displacement model. Therefore, the Y-function for premature breakthrough does not have any more application than to demonstrate the existence of the premature-water-breakthrough problem. This diagnosis can be the basis for making any appropriate operational change. Remedial action can be planned for premature water breakthrough, either on the offending injector or on the producer, once further analysis and surveillance data recognize the interaction between injector and producer.

Conclusions

The following conclusions were reached from this study:

  • In tandem with the Hall integral, derivative curves (obtained either numerically or analytically) provide unambiguous diagnosis of an injector’s performance—matrix injection, fracturing, wellbore plugging, or interception of a high-permeability channel.
  • PDHG data in the MRC producer revealed voidage balance and good pressure support from offset injectors. There is no permeability impairment in the MRC producer. Good interwell connectivity between MRC producer and injectors suggests that water injection with a line-drive pattern is a feasible development option in an approximately 5- to 7-md area of the field. RTA-model results corroborate those derived from deconvolution.
  • The log-log and reciprocal-of-time diagnostic plots are appropriate indicators of waterflood maturity and displacement efficiency and can be applied to assess volumetric sweep, forecast recovery in a mature waterflood, and distinguish normal displacement from premature water breakthrough caused by channeling.
  • The combined use of a suite of analytical tools can provide consistency and quality assurance in understanding well and reservoir behavior.

This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE 165995, “Understanding Well and Reservoir Performance With Modern Analytical Techniques,” by Lakshi Konwar, SPE, Ahmed Mohsin Al Hendi, SPE, and Syed Tariq, SPE, ZADCO, prepared for the 2013 SPE Reservoir Characterization and Simulation Conference and Exhibition, Abu Dhabi, 16–18 September. The paper has not been peer reviewed.