Consistent gathering and analyzing of the data relevant to waterflood performance is called waterflood surveillance and it has many forms. Simply measuring and recording phase rates and pressures is an initial step. Unfortunately many waterfloods rely on relatively infrequent well tests of individual wells to allocate production that is gathered and measured at any of several locations with higher frequency. Total oil production is usually known with a fair degree of accuracy. In modern oilfilelds total produced gas can be calculated with reasonable accuracy because flaring and venting is now (fortunately) a rare practice and even gas used for fuel is measured. This is not always the case with historical data. Produced and injected water volumes should also be known with accuracy; historical water data may not be so reliable. Common surveillance methods will be discussed in this and in subsequent blog posts. One of the most fundamental methods of surveillance is understanding voidage replacement. The concept is simple. Compare what goes in to what comes out. In this entry I also discuss the ABC plot.
Voidage replacement refers to replacing the volume of oil, gas and water produced from the reservoir by injected fluids. Voidage replacement ratio is the ratio of reservoir barrels of injected fluid to reservoir barrels of produced fluid. Mathematically (for water injection) it can be expressed as
In this equation, Bα, qα and iα are the formation volume factors, production and injection rates for phase α. GOR is the producing gas-oil ratio and Rs is the dissolved GOR. The third term in the denominator accounts for “free gas” produced that is in excess of that gas in the reservoir that is dissolved in the oil.
The voidage replacement ratio can be calculated on an instantaneous basis using injected and produced fluids over any specific time period (typically daily or monthly) or on a cumulative basis by using the cumulative injected and produced fluids. In the case of the cumulative voidage replacement ratio, it is common to start the cumulative production numbers at the commencement of waterflooding. Some authors have used cumulative data starting at fillup (the point at which injected water is estimated to have filled the available gas saturation in the reservoir).
Values of voidage replacement ratio less than unity are remarkably common in mature waterfloods. In such cases it is also typical to observe declining oil production rates, increasing water cuts and increasing GORs along with declining reservoir pressures. Voidage replacement ratios above 1.0 are necessary prior to fillup and most waterfloods target operating at about this level during the economic course of the waterflooding process. Many waterfloods operate well above a voidage replacement ratio of 1.0 (some as high as 2.0 or higher) for years. Invariably this implies out of zone injection and is often associated with highly fractured reservoirs, poor quality cement bonding or other sinks for water injection other than the desired reservoir.
Voidage replacement maintained at 1.0 does not mean that the reservoir is being operated at constant pressure. Pressure is not conserved in the sense that material balances of fluids are conserved. More importantly, a reservoir wide voidage replacement ratio of 1.0 does not mean that the flood is being effectively managed. Voidage replacement ratio should be tracked not only at the field level but by reservoir, by fault block and even by pattern. Streamline models are often particularly useful to supplement voidage replacement tracking.
Terrado et al (M. Terrado) described a method of quick analysis for a large number of wells undergoing waterflooding that they refer to as the After-Before-Compare (ABC) plot. This simple tool only requires knowing the oil and water rates for a certain group of wells at two different dates. This tool can be used for a wide variety of other monitoring applications; the authors suggest that injection wells can be analyzed by plotting the injection rates and pressures at two separate dates. In practice a fixed time period difference is selected and the ratio of the oil rate at time 2 divided by the oil rate at time 1 is plotted on the abscissa while the water rate at time 2 divided by the water rate at time 1 is plotted on the ordinate as shown in the following example figure.
In the following example the values for time 2 are three months after the values for time 1. In a stable flood there should be many data points near (1,1). In this example there are a lot of data points with Most of the data points fall with approximately constant water cuts (as indicated by being near the unit slope line passing through (1,1). Wells in the upper right quadrant show both increasing water and oil production (increasing total fluid) and roughly constant water cuts. This can indicate a positive response to water injection or changing artificial lift conditions. Wells in the lower left quadrant show decreasing total fluid production and nearly constant water cuts. These wells should be examined to determine if the decreased fluids are due to changes in artificial lift, wellbore damage or other reservoir issue. These wells may indicate easy targets to increase production. Wells in the lower right show increasing water production with constant or decreasing oil rates. While all wells will ultimately show this behavior, sudden jumps in water rates could indicate channeling. Wells in the upper left quadrant show increasing oil rates and decreasing water rates. In mature fields there may not be many wells with this behavior; in EOR projects, this may indicate positive flood response.
In this specific example the wells the production engineer wishes to examine in more detail are shaded in and have a slightly larger diameter. More information can be conveyed in ABC plots by adding different color, symbol types, etc. to indicate geographical areas, completion types, gathering systems, etc. The operator in this particular field (actually a specific fault block) also observed a group of five wells that also appeared to have a unit slope passing through approximately (1.7, 1.0). What might cause this?
Figure 1 ABC Plot showing quick method for waterflood surveillance
M. Terrado, S. Y. Suryo Yudono, and Ganesh Thakur, Chevron Energy Technology Company (24-27 September 2006). “Waterflood Surveillance and Monitoring: Putting Principles Into Practice”, SPE 102200-MS. SPE Annual Technical Conference and Exhibition. San Antonio, Texas, USA: SPE.