|
SPE 71480
Development of a Foam Monitor for High Pressure Separators
M.K. Poindexter, SPE, and S.C. Marsh, Nalco/Exxon Energy Chemicals, L.P., and G. Fransen, Agar Corporation
Abstract
Many deepwater platforms installed in the Gulf of Mexico produce large amounts of both gas and oil. Due to the limited space aboard these vessels, some separation equipment may become undersized particularly when record setting production rates are attempted. Additionally, there is increased activity in adding new subsea production to platforms that are within specified production limits. When the increased throughput arrives, separation equipment can be overloaded. In other instances, the addition of a new well (or wells) to existing production can greatly change the foaming characteristics of the overall composite. If the change is towards more severe foaming, an immediate problem can arise.
The high pressure separator (HPS) is where gas/oil separation begins. It is imperative that efficient separation occurs in the HPS, otherwise performance in downstream vessels will likewise diminish. With enough foaming, the platform can be forced to take an unwanted, and sometimes unexpected, shutdown.
Foaming can often be viewed as a two-fold problem. While foam (or liquid carry-over) takes place through the overhead outlet of the HPS, there is generally simultaneous gas carryunder through the bottom outlet of the separator. This "double problem" can often be observed by watching process gauges (e.g. pressure, flow rate, and level monitors). While these monitors are often useful for detecting foam, they are located either after the HPS or often do not respond quickly to an impending foam situation. To develop an immediate response to separator foaming, a probe was developed to monitor the conditions directly within the HPS. Lab development and field evaluation of a probe capable of handling the high pressure and flow rates of a HPS will be reviewed. Examples of using the probe to assist in selection and optimization of antifoams will also be presented.
Introduction
The ability to continuously and economically produce petroleum from reservoir to production facility without interruption has been recently termed as flow assurance.1 Deposition of solids, such as waxes, asphaltenes, hydrates, and scale, is cited as a major area for concern. However, even uninterrupted fluid flow can still experience flow assurance problems. Two common examples include unwanted emulsification and foaming.2 Being able to predict, monitor, and prevent such problems is critical to ensure production remains optimized under any process condition. This paper will focus on one particular aspect of flow assurance, namely a new way to detect and respond to foam inside a HPS.
A number of foam/liquid level monitors are reported in the literature, and some have found use in petroleum and petrochemical applications.3 A few of the more common techniques include capacitance, electromagnetic radiation, neutron-backscatter, differential pressure, and sonic echo (e.g. sonar, sonic, and ultrasonic) devices. Since the density, and other chemical properties, of foams are less easy to distinguish than liquids, it is often more facile to detect the liquid level below the foam than the foam/gas interface.3b
For years, energy absorption (i.e. radio frequency (RF)/microwave-based) probes have been used to effectively monitor the water concentration in oil/water emulsions.4 The probes have found application in both oil-continuous and water-continuous systems. By operating in the radio/ microwave region (specifically, up to 2.45 GHz), the probes in effect detect water concentration. The probe has three main components: a transmitter, antenna, and signal conditioner. The output of the device is a 4-20 mA signal that is proportional to the water/hydrocarbon ratio (where higher milliampere readings correspond to higher water concentrations). From field usage, it has been demonstrated that the probe will also effectively monitor foam where the liquids are highly polar (e.g. amine and glycol units).
Recent lab work confirmed that the probe has additional capabilities namely monitoring systems where water or other polar materials are absent.5 By calibrating the 4-20 mA output signal such that the span is between gas (i.e. 4 mA) and crude oil (20 mA), the probe proved capable of monitoring the relative gas/oil ratio with excellent reproducibility in both lab and field evaluations. Furthermore, this concept could be extended beyond crude oil to any non-aqueous, non-polar liquid where foam detection is needed.
The quality of foam is defined by the volume percent gas of a gas/liquid mixture where dry foams have high ratios and wet foams have low ratios.6,7 Having a monitor capable of instantaneously and continuously distinguishing foam quality (i.e. gas/oil ratio) could be useful in systems where foam is either desired or undesired. By determining real-time foam quality, operators could in effect characterize their process and make necessary adjustments to maintain an acceptable gas/oil ratio. Furthermore, careful placement of the probe (or better yet, series of probes) would permit identification of the "foam front" as it progresses or intensifies through a vessel of interest. This mode of monitoring would allow quantification of the foam quality wherever a probe was present.
For a HPS, where gas separation is critical and accessibility is difficult, an operator using a foam probe (or series of foam probes) could establish the status of the contents within the vessel at any given instant. By ascertaining the process conditions within the HPS, more leadtime would be achieved for making process changes versus using more traditional monitors located downstream of the separator.
Laboratory Foam Studies
Lab work with the probe involved placement of a small scale demo version (12.7 mm OD shaft, 510 mm overall length, see Figure 1) vertically in a 500 mL glass sparge cylinder and then foaming a sample of crude oil so that the probe was surrounded by several different gas/oil conditions. Figure 2 illustrates both the antenna and antenna guard. Foam was generated by sparging nitrogen gas through a glass frit of medium porosity (10-20 mm) that was surrounded by a reservoir of crude oil (46 mL). Figure 3 shows the basic setup for these experiments. The guard is not drawn in Figure 3 for simplicity.
To understand how the probe would behave if placed in the upper half (gas phase) or lower half (liquid phase) of a HPS, the probe was placed at three heights with respect to the crude oil reservoir. In the first experiment (see the left portion of Figure 4), the end of the probe was positioned just above the surface of the crude oil when no gas was flowing. Introduction of nitrogen gas through the frit generated foam. Both foam volume (see the left ordinate of Figure 4) and the output of the probe in milliamperes DC, which was acquired using a multimeter (see the right ordinate), were recorded versus time.
The signal resided at 3.60 mA before the start of any foaming, and returned to this value once the foam had collapsed even though there was residual oil wetting the probe shaft, antenna, and antenna guard. While foam was rising up the probe to a maximum value of 155 mL (see blue dashed line), the probe signal changed to 3.63 mA (see red solid line). While the change in signal output was quite small, it was realized that the 4-20 mA signal could be recalibrated to any set of conditions (e.g. air versus oil, as opposed to air versus water).8
The next experiment involved adding more crude oil to the sparge cylinder such that the gas/liquid interface was at the 90 mL mark (see the middle portion of Figure 4). This setup was to represent a condition when the probe was residing at the interface. Sparging was again performed such that the foam height rose to 465 mL. During the course of this run, the probe signal dropped from an initial reading of 3.72 mA to 3.64 mA, and then returned to 3.72 mA after the foam had collapsed. Once again, the small droplets of crude oil remaining on the shaft after foaming did not interfere with reproducibility.
The third experiment was a variation of what was just described, namely, that more crude oil was placed in the sparge cylinder such that the antenna was totally covered with crude (see the right portion of Figure 4). This additional crude oil changed the signal so that the start point was now 3.77 mA. Sparging with nitrogen resulted in a maximum foam volume of 485 mL, at which point the probe signal dropped to 3.66 mA. Cutting off the gas flow and allowing the foam to collapse resulted in the signal returning the original value of 3.77 mA. This third experiment represents probe behavior for gas carry-under conditions.
As seen in Figure 4, the rise and fall of the probe signal is proportional to the foam height in the first study, but inversely proportional to the foam height in the second two studies. The reason becomes clear when one considers that the probe reads a low value for air and a higher value for oil. Foam values would therefore always be somewhere between the air and oil values. In the first study, the probe starts in air and then detects foam. The foam it detects has signal values lower than those from the other two experiments as this foam is less dense (or has a higher foam quality value). The reverse occurs when the probe is initially submerged in oil and then foam is created.
Another purpose of starting the experiments at different crude oil heights was to demonstrate that the probe would be capable of detecting foam formation under any scenario. For example, if the probe was inside a HPS and the liquid level was to change for whatever reason, the probe should still be able to see changes in the oil/gas composition.
Ancillary Laboratory Studies
Antenna. The above studies used an antenna guard with an internal diameter (ID) of 16 mm. A second, larger guard (ID 28.5 mm) was also evaluated. Due to the larger diameter guard, it was necessary to use a different sparge vessel (spec. a 1 L graduated cylinder). In each of the three crude oil depths discussed in Figure 4, the starting signal with the larger antenna guard never wavered from 2.08 mA. Thus, the distance the signal travels is critical to accurately detect foam and the quality of foam.
Mist. Misting is believed to almost always be present in the separator due to the extremely high gas flow rates (often >200 MMSCFD). If the probe were placed in the gas phase and misting occurred to varying degrees, then this phenomenon might interfere with accurate foam detection by signaling a false positive. To determine whether the probe could identify mist, two spray bottles were filled with water (note: the probe is much more sensitive to water than oil), held at a close range, and used to liberally soak the probe. A negligible signal change was detected (+0.02 mA). This result suggested that the probe would not be sensitive to mist composed of hydrocarbon and thus, the presence of mist would not mask the formation of foam.
Temperature. Temperature effects on the probe signal were considered since different wells often possess different temperatures, and temperature fluctuations, though generally minor within the separator, could be envisioned. Additionally, it was deemed advantageous to know if a probe used at one location would behave similarly if placed in a second location. Six different Gulf of Mexico crudes, at 225 mL, were individually added to a 500 mL graduated cylinder. With the probe immersed, readings were recorded when the contents were at three temperatures (21, 42, and 54 °C). As shown in Figure 5, the probe signal was not altered during this temperature increase for all six crudes. The same held true for non-polar solvent toluene; however, for a polar solvent, like methanol, a clear dependence on temperature was observed with the signal being inversely proportional to temperature. This inverse relationship of signal output to temperature mimics how the dielectric constant of polar or polarizable molecules (like methanol, ethanol, or acetone) changes with respect to temperature, whereas non-polar molecules do not show such a strong dependence.9
Field Results
The field probes are a large-scale version of the unit shown in Figure 1. Field units have 31.75 mm OD shafts with lengths that depend on the dimensions of the HPS. Since the probe is an intrusive device and will need to continuously withstand both high pressure and flow conditions within the separator, the probe's structural integrity needed to be determined before installation. Past experience with the probes in similar high pressure environments indicated that probe integrity would not be compromised. However, a tensile pull test was performed with bored through Swagelok stainless steel tube fittings (i.e. those planned for use in the HPS). Separation did not occur until a peak load of 28,718 pounds. Using the cross sectional area of the fitting, the equivalent internal pressure was calculated to be 23,405 psig, a value well in excess of that experienced in field systems.10 Additional compression force and shear stress calculations were performed for several HPS process scenarios.11 The results predicted the probe structure would be well within safety limits. To date, several probes have been in service for more than two years. After repeated field inspections, no structural problems have been noted with any probe.
Probes have been installed such that the antenna resided in either the gas or liquid phase. Depending on available entry points, both horizontal and vertical installations have been made. Lab work indicated that the probe would detect changes in either phase.
In two instances, a probe was placed in the gas phase. In both cases, the probe never detected process changes (e.g. well switching or flow rate changes) or antifoam dosage changes. However, probes located in the bottom portion of six different separators have consistently detected process changes often with extreme sensitivity. It is uncertain whether the insertions into the gas phase were at places where foaming was not present or whether foaming in the separator is radically different from that depicted by the static, vertically oriented laboratory setup shown in Figure 3. The high gas flow rates experienced in a HPS would likely take a drier foam (if it ever exists) directly into a mist, while wetter, more dense foams might reside in the lower portion of the HPS. As gas/oil separation progressively diminishes, a wet foam front could be envisioned traveling into the lower portion of the separator until gas carry-under begins to occur into the adjoining downstream intermediate or low pressure separator. Field results indicate that this is what the liquid phase probes are detecting.
Figures 6-9 illustrate instances where the probe detected various process and antifoam changes. In Figure 6, a high gas producing well was removed from the HPS which was producing several commingled wells. The change was readily noted in both the HPS gas rate meter and the liquid phase probe (denoted as B). An ordinate scale is not included as five different parameters are represented in the figure. Four of the five parameters have extremely different scale ranges (only the two probes have the same scale, 0-100% output which corresponds to the 4-20 mA signal output). Meters for both the downstream intermediate pressure separator (IPS) and combined gas rate, as well as the gas phase probe (denoted as probe A), did not detect this process change.
Antifoam optimization is depicted in Figure 7. The liquid phase probe readily tracked three separate antifoam dosage reductions while none of the gas meters were able to detect the chemical reductions. In Figure 8, an antifoam known to reduce gas carry-under was introduced to the HPS. Upon introduction, the liquid phase probe was once again the only parameter able to detect the change. The gas phase probe would periodically rise or fall, as shown in Figure 8; however, these changes never corresponded to any known process or antifoam dosage change. As a final example, Figure 9 represents a production rate increase which lasted for about 80 minutes. The change was readily noted in the combined gas rate increase and slightly noted in the HPS gas rate. The liquid phase probe was also able to follow the increased production, though the signal change was relatively minor. During this process change, no foaming was detected aboard the platform. Thus, there are instances when noticeable production rate increases do not necessarily lead to a critical condition such as foaming.
Water Production. The examples described above involved little to no water in the production fluids. However, breakthrough of water recently occurred at one installation. While the amount of water is still relatively low (up to 6%), the presence of water was noted in a higher value probe reading when there was no foaming. In this instance, the probe was initially calibrated with oil only at 18 mA (not 20 mA as previously described). This procedure was done for several sites to help determine when water breakthrough was occurring. The probe still functions well with regard to detecting foam as the signal drops noticeably under such circumstances; however, the higher value probe readings vary slightly as the water content varies. Future work will focus on determining how much water is needed before accurate foam detection is compromised.
Summary and Conclusions
Locating RF/microwave-based probes directly into the liquid phase of high pressure gas/oil separators permits instantaneous detection of subtle wet foam changes as they occur. While traditional monitors (like pressure, flow rate, and level gauges) are able to follow process conditions to some degree, none are located directly inside the separator. Thus, there is a time delay in what they detect and what is occurring in the vessel. Having immediate access to separator process conditions allows for quicker response to unwanted foaming.
The probe is able to detect process changes such as well switching, when the production ratio of several wells is varied, or when throughput changes are made. Certain well mixtures can be more difficult to process and might need different chemical treatments to maintain desired production rates. Optimizing antifoam addition has also been demonstrated. Before gas carry-under (and corresponding liquid carry-over) becomes a problem, the foam probe is able to see impending foam before the situation worsens.
Future applications could focus on placing a gas phase probe further back in the separator near the gas overhead exit. Due to the lack of entry points, there are often few choices for inserting probes into existing separators. One way to avert a limited number of choices is to design separators with probe designated entry points.
Abbreviations
|
Printable View
