This paper describes thermal modeling and its combination with drilling-fluid analysis to reveal concealed changes in well conditions during various drilling and completion operations. These hidden conditions represent significant changes in the well’s drilling- and completion-fluid temperature, pressure, and density (FTPD) that may help explain wellbore-stability and -integrity issues.
Introduction
In the past, it has been difficult to make a case for temperature modeling. Temperature did not seem to be an issue in drilling, and production engineers could usually assume worst-case scenarios, such as constant bottomhole flowing temperature throughout the production tubing. Deepwater drilling and high-pressure/high-temperature wells have changed that attitude to a certain degree, and the effects of trapped annular pressure have become design issues for well completions. The effect of temperature on cementing has long been recognized, and it is understood that the correct determination of retarder can be critical. Otherwise, temperature modeling during drilling has not seemed to be that important. Typically, intermediate-string cementing is focused on achieving a good cement job and drilling ahead, and not on issues of temperature and pressure. One reason is the extensive use of water-based drilling muds. Water density is not particularly sensitive to pressure and temperature, so surface-measured mud weight usually does not vary much in conventional wells. Oil-based and synthetic-oil-based muds, on the other hand, are much more pressure- and temperature-sensitive. Furthermore, modern deepwater wells are encountering much more extreme temperature and pressure conditions, and maintaining the correct pressure has become more critical because of weak formations. The focus on thermally induced annular pressures has previously been a concern only for casing design, not for wellbore stability or well control. The purpose of this paper is to raise awareness of these issues and their consequences.
Fluid Density
This paper focuses on fluid density because it is generally the principal determinant of annular pressures. Water is the most common base fluid for drilling muds and is probably the most-studied fluid because of its wide use in every aspect of our lives. Water is also difficult to model accurately, but many excellent correlations are available. Water density usually is modeled with hundreds of empirical coefficients.
Closely related to water are brines. Brines are equally difficult to model accurately, and thermodynamics can take one only so far before another list of empirical coefficients is needed to match measured densities. There is a great amount of data on common brines, but there is a need for more data and more correlations for many of the more-exotic brines being considered for use in wells.
The third typical class of drilling muds consists of oil-based and synthetic-oil-based fluids. Diesel oil was commonly used in the past, but environmental considerations have driven the production of a number of synthetic oils. (For a correlation equation that models diesel and synthetic oils effectively, please see the complete paper.)
Most drilling muds contain added solids, typically barite and bentonite, to increase the base-mud density. These solids are suspended by the gel strength of the fluid, but they may settle out over time. This effect is called barite sag. An advantage that brines have over drilling muds is that they are single-phase solutions. For saturated brines, a change in pressure or temperature may cause some of the solute to precipitate, but generally brines do not sag. Oils do not form as strong a gel structure as water-based fluids, so barite sag is more significant in oil-based fluids.
For a discussion of pressures in a sealed annulus, including related equations, please see the complete paper.
Hypothetical Annular-Pressure Situations
Fig. 1 (above) shows the desired pressure distribution in an open annulus as the bit nears the end of the planned depth interval (in this study, we are assuming overbalanced drilling).
The choice of casing-setting depth and mud program is designed to provide the pressure profile shown in Fig. 1. The black line shows the pore pressure, the red line shows the fracture pressure, and the blue line shows the designed annulus pressure, lying between the pore- and fracture-pressure curves over the depth interval. The annulus pressure approaches the facture pressure at the top of the interval and approaches the pore pressure at the bottom of the interval. A specific mud weight is chosen to maximize the depth interval of open hole. (The usual plots are equivalent mud weight vs. depth, but for this study we found that the use of pressure helped clarify some of the examples.)
It is clear that if this carefully selected mud weight varies, there is potential for a lost-circulation problem should the annulus pressure increase, or for a well-control problem should the annulus pressure decrease. In an open hole, the annulus pressures are determined largely by the drilling-mud density. Hotter mud has a lower density, and cooler mud has a higher density.
The higher density of the colder mud results in higher pressures. The annular pressures for the colder mud cross the fracture-pressure curve at approximately 6,000 ft. In this case, there is the potential for a lost-returns problem. The hotter mud has a lower density, resulting in lower pressures. The annular pressure for the hotter mud crosses the pore-pressure curve at approximately 9,000 ft, possibly resulting in a gas kick if there is a producing formation at that depth. How plausible are these results? Because geothermal temperature almost always increases with depth, higher mud temperatures are to be expected. How could the mud be colder? Drilling through a riser in deep water, which is usually colder than surface conditions, could cool mud circulating from the surface. Cold mud would be found in the riser, and in the wellbore for shallow drilling.
In a sealed annulus, the hotter fluid expands, but because it is confined, it must be compressed back to its original volume, causing a pressure increase. This increased pressure will fracture the formation, bringing the pressure down and losing some of the annular fluid to the formation. This is the opposite of the effect seen in the open annulus. However, if this mud cools, there is now potential for reduced pressures, resulting in fluid influx.
The colder fluid contracts, but because the annulus is sealed, the fluid must also expand to fill the annulus, resulting in decreased pressure. The lower pressure will allow fluid influx. If the influx is gas, the bubble will rise to the top of the annulus, increasing the pressure by the U-tube effect. However, this may produce higher pressures, fracturing the formation and ultimately resulting in an underground blowout.
The two sample problems, both based on actual-well data, show that the annular fluid can be either hotter or colder than the undisturbed temperature, depending on the previous drilling history, the well’s static vs. circulating time periods, conditioning before cementing, and the actual cementing operation. Cold fluid results from extensive well circulation with fluids in rig tanks at ambient temperatures, while hot fluid results from restarting circulation in a well that has been shut in too long after operations such as tripping to change bits, logging, running casing, or hole conditioning before cementing, where hot fluid from below is circulated into the annulus.
The effect of barite sag is shown in Fig. 2. Loss of the solid particles results in a lower mud density, so there is less pressure change with depth. However, because the annulus volume is fixed, the average pressure in the mud column must remain the same. The result is a pressure profile that increases at the top of the depth interval and decreases at the bottom. Here, one has the potential for simultaneous fluid influx and formation fracture. If there is a permeable zone at the bottom of the interval, there is the potential for an underground blowout.
Field Test 1
Previously drilled wells showed increased annular pressure after cementing operations. The cause of this pressure buildup was unknown, so there was concern about the well completion, the potential gas influx from the formation, and the cementing operations. Fairly complete data on the next well being drilled in the same location allowed a simulation of the cementing operation with proprietary thermal-simulator and casing-design-analysis programs to model the well and FTPD of the drilling mud, which filled most of the annulus above the top of cement. Unfortunately, previous drilling data were not available, so the initial conditions for the cementing operation were taken as undisturbed geothermal temperatures.
During the cementing operation, the well annulus was cooled below the undisturbed geothermal temperature. During the 12 hours of waiting on cement, the annular temperature gradually increased to a value near that of the undisturbed temperature.
Over the 12-hour period, the annular-fluid volume increased by nearly 14 ft3. In order for this increased volume to be accommodated by the sealed annulus, the fluid had to be compressed, and the casing and formation had to deform in response to this pressure.
Maximum annular pressure was estimated to be approximately 1,700 psi at the time that annular-pressure measurements were taken. According to the operator: “From all indications, the cement job on the (project) went well. After the job, the annulus built up to 1,000 psi in 18 hours. This pressure was bled off to zero, and everything was stable after that with no pressure increase.” The analysis overpredicted the actual measured pressure, but qualitatively matched the field behavior. Reasons for the overprediction include the lack of thermal modeling of the previous drilling process and the possibility of fluid loss caused by formation leakoff. The fact that no gas was found in the annulus and the annular pressure was stable after bleedoff indicates that no other mechanism was producing the annular pressure.
Field Case 2
In Field Test 1, the annular temperature was reduced below the undisturbed temperature, so the heat-up produced increased annular pressure. What would happen if, instead, the annular temperature were above the undisturbed temperature? Because circulation brings up hotter fluid from below, too long a shut-in period before the cementing operation could result in hot fluid in the annulus. The following thermal simulation produced this effect.
Fig. 3 shows the resulting temperatures produced by this operation. The annular temperatures are significantly warmer than the undisturbed geothermal temperature. After completion of the cementing operation, the temperature begins to fall.
As the annular fluid cools, the annular pressures fall. The effective mud gradient also falls. The surface mud weight was 14.3 lbm/gal, but because of thermal expansion of the mud, which results in lower density, the effective mud gradient was less than 14.2 lbm/gal at the time that the annulus was sealed. Further cooling of the mud brought the effective mud gradient to below 13.4 lbm/gal, a loss of nearly 1 lbm/gal. If there were producing zones in this annulus, there would be the possibility of fluid influx because of the lower-than-expected mud gradient. In other words, the modeling reveals hidden FTPD conditions that convert the 14.3-lbm/gal overbalance to a thermally induced underbalance, which could cause a severe, unexpected kick. If not detected in time to shut the blowout preventer, loss of well control (blowout) could result.
This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 163476, “Modeling Reveals Hidden Conditions That Can Impair Wellbore Stability and Integrity,” by Robert F. Mitchell, SPE, Halliburton; Ronald Sweatman,SPE, Consultant; and Gary Young, Encana, prepared for the 2013 SPE/IADC Drilling Conference and Exhibition, Amsterdam, 5–7 March. The paper has not been peer reviewed.