WELL CONTROL / INTERVENTION

Well control during well intervention

Part 5 – Planning a pumped fluid hydraulics program for CT well intervention A pumped kill fluid solution should be prepared when surface pressure is possible in CT service, with reassessments performed during the intervention providing enhanced design confidence

Alexander Sas-Jaworsky II, Sas Industries, Inc., Houston

or most well intervention services performed through coiled tubing (CT), fluid pumping operations are an essential feature of the prescribed service. In these types of operations, the fundamental principles of well control must be employed throughout the entire well intervention program to ensure that the prescribed pressure balance is maintained for the desired result. Prudent preparation of any CT-conveyed pumping operation requires that an analysis be performed to predict the behavior of the fluids pumped through the CT and within the wellbore annuli.

Where surface pressure is expected to be present during the CT operation, a pumped kill fluid solution should also be prepared in the event that well control measures, in addition to CT well control stack operations, are required. To properly assess the behavior of fluids pumped through CT, care must be taken to obtain all pertinent information for wellbore construction and service equipment components. Further, assessments performed prior to and during the intervention program enhance design confidence when implementing the prescribed well control measures. The following discussion reviews the hydraulic performance and design issues considered for preparing and conducting CT-conveyed well intervention and well control operations.

Wellbore Conditions and Service Details

The wellbore used in this example was drilled as an "S"-type well and produces to an offshore platform through a single gravel packed completion, Fig. 1. The well is completed with 4-1/2-in. tubing from surface to 10,500 ft MD (9,200 ft TVD) and 3-1/2-in. tubing from 10,500 ft MD to 14,600 ft MD (13,000 ft TVD). The completion tubulars below the packer include a section of 2-7/8-in. tubing from 14,600 ft to 14,800 ft MD (13,000 ft – 13,200 ft TVD), with the 2-3/8-in. base pipe gravel pack completion run from 14,800 ft to 15,000 ft MD (13,200 ft – 13,400 ft TVD). The water depth to mudline is 400 ft and estimated bottomhole temperature and pressure are 250°F and 9,200 psig, respectively, taken at the top of perforations at 14,875 ft MD (13,275 ft TVD).

Fig. 1. The example well produces to an offshore platform through a single gravel packed completion. Completion tubulars below the packer include a section of 2-7/8-in. tubing from 14,600 ft to 14,800 ft MD, with the 2-3/8-in. base pipe gravel pack completion run from 14,800 ft to 15,000 ft MD. Estimated bottomhole temperature and pressure are 250°F and 9,200 psig, respectively.

Prior to shut-in, the well produced about 3,000 bopd, with associated gas and water. The maximum anticipated surface pressure (MASP), based on the hydrostatic pressure exerted by a dry gas column on the formation, was estimated to be 7,000 psig, requiring preparation of PSC-2 CT well control equipment for the given well conditions.

The primary CT intervention program is a sand wash through the completion interval with a subsequent circulation kill to be performed in preparation for workover operations. The initial shut-in surface pressure is 4,800 psig. Wireline was run within the well, finding the fluid level at about 1,800 ft and tagging the top-of-sand at 14,300 ft MD.

The proposed wash fluid is a 3% KCl water, with a blend density of 8.45 ppg at a surface temperature of 70°F. With the wellbore circulated full of the 3% KCl fluid, the hydrostatic pressure exerted by a 13,275-ft column of 8.45-ppg "average" density fluid is calculated to be 5,833 psig. However, when taking into account the density corrections due to changes in temperature from surface to TD, the adjusted hydrostatic pressure is predicted to be 5,775 psig. At a reported bottomhole pressure of 9,200 psig, the static surface annulus pressure with the 3% KCl fluid is expected to be 3,425 psig. The changes in the 3% KCl fluid properties (density and viscosity) are shown in Fig. 2 over the range of temperatures expected within the wellbore.

Fig. 2. How temperatures expected within the wellbore affect density and viscosity of the 3% KCl kill fluid.

For the circulation kill program, the "average" liquid density needed to balance the 9,200-psig formation pressure is calculated to be 13.33 ppg. Since the kill program is intended to be temporary, a CaCl₂ / CaBr₂ brine is selected to minimize completion damage. Taking into account the density corrections due to changes in temperature, the predicted surface blending density of the brine must be increased to 13.5 ppg to achieve the "average" 13.33-ppg density required to hydrostatically balance the well. The changes in density and viscosity for the CaCl₂ / CaBr₂ fluid are shown in Fig. 3 over the range of temperatures expected within the wellbore.

Fig. 3. How temperatures expected within the wellbore affect density and viscosity of the CaCl2 / CaBr2 kill fluid.

The CT workstring is comprised of 18,000 ft of 1.50-in. OD, CT-90 grade tubing, with a tapered wall thickness design. The tapered wall thickness design is shown in Table 1, where the outboard end (whip end) has a 0.125-in. wall and the inboard end (reference end) has a 0.175-in. wall. In addition, the CT bottomhole assembly (BHA) consists of a connector, check valve sub and wash nozzle sub with five, 0.1875-in. ID ports.

Prediction and Validation of Hydraulics Model

The most effective means to evaluate the pre-job hydraulic performance of pumped fluids is through nodal analysis of the fluid circulation system. There are numerous commercial programs available within the industry that provide frictional pressure loss (DP) predictions through CT and within wellbore annuli. The nodal analysis model and related calculation methodology published in previous work by the author^2,3 was used as the means for performing the DP evaluation described in this article.

Pre-deployment preparation. With the basic information provided for the CT string design, CT BHA and 3% KCl fluid properties, a DP model case for the CT string on the reel should be prepared to provide a means for comparison and validation through an onsite pump rate test. Conducting an onsite pump rate test over a range of desired rates provides a means for confirming the fluid displacement measured relative to a specific pump gear and engine rpm. The observed plunger strokes per minute (spm) rate is compared to the theoretical plunger displacement in barrels per stroke (bps), yielding an efficiency value for the onsite pump. The efficiency correction for bps, along with confirmed displacement rates in barrels per minute (bpm) provide greater confidence in validating the CT DP model before commencing operations.

The results of an actual pump rate test conducted at four controlled rates is shown in Table 2, illustrating the type of information typically recorded during these tests. In comparison to the prediction model, close agreement with the observed CT DP was obtained using an averaged surface ID roughness of 0.0008 inch for the given string and fluid parameters.

Deployment for CT well intervention. When deploying or retrieving the CT during well intervention operations, DP should be observed, recorded and periodically compared with the model predictions. As the CT string is deployed, total DP will decrease due to reduced flow resistance through the straightened portion of the tubing, changes in fluid properties at temperature and reduced surface pressure held on the choke resulting from increases in annular hydrostatic pressure within the annuli.

The observed surface pressure prior to CT deployment reflects the amount of hydrostatic pressure acting on the formation pressure. For the surface pressure of 4,800 psig used in this example, an estimated 4,400 psig of hydrostatic pressure is being exerted against the formation pressure, yielding an equivalent resident fluid density of 6.37 ppg.

The initial CT fluid pumping program should commence by displacing wash fluid into the wellbore and attempting to reduce the amount of surface pressure attributed to the gas head. During this process, the CT is slowly deployed into the wellbore while pumping, with adjustments in surface choke pressure decreasing in coordination with the increase in hydrostatic pressure, resulting from the pumped fluid. The volume of fluid needed to establish full liquid returns should be recorded and compared with the volume calculated for replacement of the gas column height above the liquid level in the well.

A pump rate of 0.55 bpm was selected as the initial circulation rate, and with the 4,800-psig surface pressure present, the initial pump pressure is projected to be about 6,100 psig. Using the wireline fluid level measurement, an estimated 23 bbl of pumped fluid should be required to replace the gas head volume in the well. Based on the hydrostatic pressure balance relationship, the replacement of the 1,800-ft gas head volume with 3% KCl brine will reduce the surface wellhead pressure to about 4,300 psig, with a corresponding drop in pump pressure to about 5,600 psig.

Once liquid returns are established, CT deployment should be halted temporarily to conduct the primary circulation analysis, confirm the fluid-in / fluid-out ratio and correlate the observed system DP readings with predicted values. Fig. 4 illustrates the choke and pump pressures expected during deployment at the respective CT depth within the well. Evaluation of the choke and pumping pressure predictions from surface to 14,000 ft finds that the decreases are dominated by the increased hydrostatic pressure differential from the pumped fluid, with minimal decreases attributed to straightening the deployed CT.

Fig. 4. Expected decreases in choke and pump pressures as a function of CT depth in well (0.55 bpm-pump rate).

Throughout deployment, periodic CT system DP checks should be conducted to confirm the accuracy of nodal analysis predictions. These DP checks are best obtained when the CT is stationary to eliminate the affects of pressure and displacement fluid rate changes due to surge and swab. Adjustments may be needed to account for unknowns such as annuli tube eccentricity, influx fluid properties and/or annuli surface roughness.

If an increase in recovered fluid rate is observed during periodic checks, adjustments to increase choke pressure are typically applied to maintain balance and restrict the fluid influx. However, if the amount of observed fluid influx is excessive for the prescribed intervention program, annuli DP correlations can be made to determine if increasing choke pressure, increasing pump rate (dynamic kill principles) or a combination of both actions will correct the underbalanced condition without adversely affecting the prescribed program.

Observed decreases in recovered fluid rate are typically a function of overpressure in the annulus, displacing a fraction of the well fluids into the formation and reducing the annular velocity. These fluid rate decreases may be the result of excessive choke pressure, increased hydrostatic pressure due to entrained solids loading in the wash fluids or excessive annuli DP. The measured returns fluid rate, along with stable choke pressure readings at the specified pump rate, should be constantly recorded and used to determine whether annuli DP parameters need adjustment to minimize injectivity into the completion interval.

Based on wireline measurements, the top of sand is expected at about 14,300 ft. To obtain additional pumping system DP data points prior to initiating the prescribed wash program, the CT deployment should be halted at approximately 14,000 ft to conduct a pre-wash analysis. For this test, the 0.55-bpm pump rate is maintained to establish a baseline pump pressure (estimated to be 4,650 psig) prior to increasing the pump rate to the desired wash rate. The wash rate selected is 0.912 bpm, increasing the predicted system pump pressure to about 6,450 psig. This 1,800-psig increase in system pressure is significantly dominated by the increased DP through the CT, with slight influence from the increase in annuli DP.

For any solids removal wash program, an effort should be made to model the CT wash program to identify the range of acceptable wash penetration rates. The prediction of apparent wash-fluid-density increase as a function of penetration rate is outside the scope of this discussion, but a brief review is warranted to identify its ability to adversely affect the total system DP.

As solids are entrained within the wash fluid, the hydrostatic pressure corresponding to the volume of solids suspended within the annulus increases. This event may be identified as an increase in pump pressure, with a corresponding decrease in choke pressure for a constant fluid-in / fluid-out displacement period. The change in total system pressure results in an increase in applied pressure on the formation, creating an undesirable pressure balance condition. If the completion interval is not in pressure communication with the wellbore, then CT system pressure increases are effectively contained and no adverse consequences are anticipated. Once the slug or slugs of "dirty fluids" are displaced out of the wellbore annulus, system pressure should return to pre-wash conditions.

However, if the wash program exposes a flow path with an excessive wash fluids hydrostatic differential, the imbalance acting on the formation will likely result in fluid losses to the formation. Since production data is available for this well, the modified Darcy steady-state radial flow equation, can estimate loss of fluid to the formation for a given overbalance condition.

For the 75-ft completion interval height, close agreement with the 3,000-bopd production rate at a drawdown pressure of 500 psig was obtained using an average permeability of 145 md and skin factor of 5.0. Assuming the completion skin factor is unchanged, the relationship of fluid loss per psig of applied overbalance pressure is shown in Fig. 5, with the viscosity for the 3% KCl and CaCl₂ / CaBr₂ brines adjusted to reflect the bottomhole temperature of 250°F. Note that for the 3% KCl fluid, the low apparent viscosity at temperature provides little resistance to flow through the completion. Therefore, when washing into the screen assembly, care must be exercised to ensure that the choke pressure held at surface compensates for the increased hydrostatic pressure of the solids entrained within the "dirty fluids" so that an overbalanced condition is not applied to the completion.

Fig. 5. Comparison of fluid losses due to overbalance pressure.

Once the completion interval is cleaned to desired depth, the CT string is typically reciprocated within the completion tubulars while pumped fluids displace the washed solids to surface. At the 0.912-bpm circulation rate, the predicted annuli DP is approximately 80 psig. If the entire completion interval is open to flow, the high fluid loss index makes it very likely that fluid loss will occur, with the loss rate reaching equilibrium based on the pressure balance achieved for the rate of flow and corresponding annuli DP. Therefore, choke pressure and fluid-in / fluid-out rates must be monitored closely and adjusted to ensure that the resulting circulation rate provides the minimum desired annular velocity to complete the solids wash program.

Implementing CT Circulation Kill Program

The preferred position for the BHA wash nozzle sub during the initial fluid circulation is directly above the top of perforations at 14,875 ft MD (13,275 ft TVD). This position provides a convenient means for obtaining steady-state pressure readings for the observed fluid-in / fluid-out rates. During the post-wash displacement program, the circulation of clean 3% KCl brine also serves as the first "drillers" circulation in a standard CT-conveyed kill program. Once the prescribed volume of KCl brine is displaced, the CT string and wellbore annuli will contain a uniform density fluid, providing a means for confirming kill-weight fluid density and pre-kill DP calculations.

The circulation pressure should be recorded and compared with the DP prediction for the 3% KCl fluid. At this time, adjustments should be made to validate the prediction model and confirm close agreement with observed readings. Accuracy in the DP predictions is a critical component in minimizing the unknowns when the single-salt KCl fluid is displaced by a dual-salt fluid system with significantly different density and viscosity. If CT geometry and relative position within the wellbore is held constant, the density and viscosity values (appropriately adjusted for temperature changes) should be the only variables in the prediction model. Upon completion of the first "drillers" circulation program, the pumps may be shut down to obtain the static surface pressure reading and confirm the desired kill fluid density.

As mentioned earlier, the kill fluid selected is a CaCl₂ / CaBr₂ brine with a specified surface density of 13.50 ppg. Due to the incompatibility of the brine systems, separate tanks must be used to blend the CaCl₂ / CaBr₂ brine. During the circulation kill program, the captured KCl fluid and reserve suction volume (approximately 250 bbl total) must be isolated from the kill fluid blending pits.

Referring to Fig. 5, the CaCl₂ / CaBr₂ brine has a reported apparent viscosity about eight times greater than the KCl fluid at 250°F, substantially reducing the rate of flow through the completion interval for the same amount of overbalance pressure. Therefore, there may be relatively high losses of KCl fluid observed during the initial circulation program, with relatively little loss of the CaCl₂ / CaBr₂ brine during the "kill" circulation.

When planning the circulation kill program, wellbore volumetrics must be calculated to ensure that sufficient volume is prepared for the complete displacement of all CT and wellbore tubulars. For this example, the minimum volume of kill-weight fluid needed to fill the exposed tubulars is calculated to be 186 bbl, with 27 bbl needed to fill the CT string and surface piping. Allowing for fluid losses to the formation and suction tank reserve, the minimum recommended kill-weight fluid volume to be prepared for this case is 300 bbl.

Breakdowns of the CT string and wellbore tubular displacement stages are also prepared to ensure that the DP changes in the circulation kill program are properly anticipated. Figs. 6 and 7 represent predicted DP changes in the prescribed circulation kill program at a kill rate of 0.260 bpm and 0.550 bpm, respectively. A brief explanation of each of the circulation kill stages is offered below. Note that the illustration of the stages in Figs. 6 and 7 are for presentation purposes and are not intended to reflect relative volume increments of pumped fluid needed to complete the prescribed kill program.

Fig. 6. DP prediction for kill stages at a 0.26-bpm rate.

Fig. 7. DP prediction for kill stages at a 0.55-bpm rate.

Stage 1 is the pre-kill circulation pressure (PKCP). This stage relates to the pumping pressure needed to initiate displacement of the kill-weight fluid through the surface pump piping and CT string remaining on the service reel. The desired kill fluid pump rate is initiated and held constant throughout the entire circulation program. Note that as kill-weight fluid is pumped through the CT string, DP will increase. Choke pressure should be held constant between Stage 1 and Stage 2.

Stage 2 is the initial circulating pressure (ICP). The value observed at Stage 2 represents the maximum required pump pressure and identifies the point at which kill-weight fluid displaced through the CT string makes the turn over the tubing guide arch. As kill-weight fluid is pumped through the deployed CT string to the BHA, DP decreases significantly as a function of the difference in fluid densities (kill-weight versus resident). The pressure balance is maintained by holding choke pressure constant between Stage 2 and Stage 3.

Stage 3 is the final circulating pressure (FCP). This value represents the point at which kill-weight fluid has completely displaced the resident fluid within the deployed CT string and has begun to enter the wellbore annulus.

Displacement conditions shown in Stages 4 – 8 represent the fill-up of five distinct segments of the wellbore annuli identified as having differences in borehole geometry, specific thermal gradients, notable deviation or a combination of the above. After kill weight fluid reaches the CT nozzle, the wellbore annulus is displaced in Stages 4 – 8. The CT string pump pressure is held constant at FCP, with the choke pressure adjusted to compensate for increased hydrostatic pressure resulting from kill-weight fluid entering the wellbore. Therefore, to maintain a constant bottomhole pressure kill program, close coordination between the surface pump and choke station is critical.

When selecting the most desirable kill rate, the final annuli pressure losses for the proposed kill rates should be compared with the Darcy steady-state radial flow predictions for the CaCl₂ / CaBr₂ brine. The final annuli pressure represents the minimum overpressure expected upon completion of the circulation kill program.

Referencing Figs. 6 and 7, the final predicted annuli pressure loss for the 0.266-bpm pump rate is 45 psig, as compared with 91 psig for the 0.55-bpm pump rate. In review of Fig. 5, the predicted fluid loss for the CaCl₂ / CaBr₂ brine at an overbalance pressure of 41 psig is approximately 0.07 bpm, with 0.15 bpm for an overbalance pressure of 91 psig. Since it is likely that choke pressure will fluctuate the desired setting during the kill program, fluid losses should be expected. As such, an accurate returns volume record must be kept during Stages 4 – 8 to identify and compensate for deficient kill-weight fluid displacement volumes within the wellbore. In review of the pumping time, the circulation kill program for the 0.55-bpm circulation rate is projected to be 332 min (5.53 hr) as compared to 11.58 hr for the 0.26-bpm rate.

Once the desired circulation kill rate is selected, additional attention must be paid to ensure effective implementation of the kill program. The kill fluid displacement schedule should be prepared on a per-pump-stroke basis to track the predicted DP reduction from ICP to FCP and provide a reliable means for tracking kill fluid placement within the wellbore. As discussed previously, the constant bottomhole pressure kill procedure requires that choke adjustments are made to maintain a constant CT pump pressure once kill weight fluid has reached the nozzle. To aid in diagnostics of model predictions, a choke pressure schedule should also be prepared to allow for comparative analysis of the predicted choke adjustment steps to observed annulus pressure with respect to the volume of kill fluid circulated within the wellbore.

The example above illustrates the complexity of hydraulic and hydrostatic pressure balances expected during pumped fluid well intervention or well control programs. The use of nodal analyses to plan well intervention and kill programs aids in identifying the areas of concern regarding pressure balance and significantly enhances the ability to select the most desirable solution option prior to implementing the prescribed service.

Literature Cited

¹ World Oil’s Coiled Tubing Handbook, 3rd Edition, Gulf Publishing Company, Houston, TX, 1998.

² Sas-Jaworsky, A., and T.D. Reed, "Predicting friction pressure losses in coiled tubing operations," World Oil, September, 1997.

³ Sas-Jaworsky, A., and T.D. Reed, "Predicting frictional pressure losses in CT annuli: An improved method," World Oil, April, 1998.

⁴ McCain, Jr., William D., The Properties of Petroleum Fluids, The Petroleum Publishing Co., Tulsa, OK, 1973.

⁵ Well Control School, Guide to Blowout Prevention, WCS First Edition, New Orleans, LA, 2000.

The author
	Alexander Sas-Jaworsky II is founder and principal engineer of SAS Industries, Inc., specializing in mechanical testing, training and service consultation for all aspects of applied coiled tubing technology. He began his career with Conoco after receiving a BS in petroleum engineering at the University of Southwestern Louisiana in 1982, and he worked for various coiled tubing companies before and while attending college. He worked in several Conoco divisions as a production engineer before transferring to the Conoco Houston Production Technology group in December 1990 as worldwide concentric workover consultant for coiled tubing and snubbing. He is a registered professional engineer in Louisiana and Texas, SPE member, and serves on API committee 3/subcommittee16 as chairman of the well intervention well control task group.

Part 1

Part 2

Part 3

Part 4