PROCEDURES CONTROL TOTAL MUD LOSSES WHILE DRILLING IN DEEP WATER

John Dewar
Shell Philippines Exploration BV
Manila

Douglas Halkett
Forasol/Foramer
Manila

Hydrocarbon containment procedures and innovative mud loss cures in a deepwater drilling program minimized the problems from severe lost circulation and the possible subsequent formation of hydrates from gas influx.

In the deepwater (8301,000 m) drilling program offshore Philippines, reefal limestones were encountered in which total mud losses could be expected because of the presence of large fractures. The danger was that a sudden drop in hydrostatic head (resulting from the losses) could allow any natural gas to enter the well bore quickly. The gas could then migrate up the well bore and form hydrates in the blowout presenters (BOPs).

Once hydrates form, they are difficult to remove and can make a BOP stack inoperable.

The situation can quickly get out of control, making well killing and relief operations very difficult and possibly leading to a severe blowout.

To combat this potential problem, containment procedures were developed to cope with these fluid losses. The philosophy behind the procedures was to prevent hydrocarbons from entering the well bore and, if they did enter, to ensure that they did not move up the well a bore and into the riser.

Additionally, procedures were developed to allow drilling to continue during the losses and the curing of losses.

CONTAINMENT PROCEDURES

The detailed operational procedures were based on the following hydrocarbon containment procedures:

The first step was to prevent hydrocarbons from entering the well bore.
This step involved the selection of the mud weight. A mud weight with an overbalance of 1,725 kPa (250 psi) at the top reservoir was chosen based on offset well data. If the mud weight were too high, then the losses would be more severe, the reaction time reduced, and the chance of hydrocarbons entering the well bore increased.
However, a reduction in mud weight was a trade off against one of the primary methods to suppress hydrate formation - the use of a saline mud. The mud weight was controlled with NaCl. The greater the concentration of salt (and as a result the mud freight), the greater the hydrate suppression. 1
The second step in preventing hydrocarbons from entering the well bore involved the quick closing in of the well to prevent the fluid level in the riser from falling and the subsequent loss in hydrostatic control. The driller used a simple, graphical method to tell if the mud level dropped below the critical level (that is, the level that would allow hydrocarbons to enter the well bore).
If hydrocarbons did enter the well, then it was necessary to prevent them from reaching the wellhead.
This step was achieved by monitoring the influx and bullheading it back into the formation with viscous mud pills. Bullheading was not anticipated to be a problem because of the high permeability of the fractured thief zone.
If hydrocarbons did reach the wellhead and BOP, it was necessary to prevent the formation of hydrates
Hydrate formation could seriously affect the operation of the BOPS. The best known method to suppress hydrate formation is the use of glycol.
A glycol/mud mixture was kept on standby in the kill line and surface pits. Whenever pumping was stopped into the well annulus (bullheading), a pill of glycol/mud was pumped into the BOP. In addition, the salinity of the mud was kept as high as possible to help prevent hydrate formation.
If hydrates did form, then it was necessary to eliminate them.
The two methods for eliminating hydrates are the use of chemicals (methanol) or the application of heat until the hydrate "melts" back into gas. Methanol was kept on standby to be pumped down the kill line if required. To increase the temperature in the general area of the BOP sea water at 28 C. was ready to be pumped down the booster line and up the riser. This method was not considered optimum because it did not heat up the main section of the BOP.
Once the hydrates melted, then the resulting gas could be bullheaded back into the formation or circulated out conventionally. This procedure would then leave a pocket of gas trapped between the closed rams and the line used to circulate the gas into or out of the well.
At this point, the gas trapped inside the BOPs would need to be removed.
The well-known method of U-tubing the gas out of the well would be used. In other words, the kill and choke lines would be filled with mud lighter than that in the riser, and the rams would be opened.

PREPARATIONS

These procedures allowed the following to be conducted safely:

Drill into the limestone formation and prevent hydrocarbons from entering the well.
Drill ahead with a mud cap under controlled conditions until sufficient hole had been drilled to justify the curing of the losses.
Cure the losses using gunk (a mixture of bentonite, diesel oil, and cement, or BDOC) followed by thixotropic cement plugs. if the BDOC did not work, then silicate-based solutions could be tried.
Drill through the plums and then drill ahead until the process needs to be repeated or until total depth is reached.

The preparations were carried out prior to drilling out of the 9-3/8-in. casing shoe above the limestone reservoir. The schematic is shown in Fig. 1.

CALCULATIONS

The following, calculations were required:

Mud weight
The mud weight necessary to enter the top reservoir with a 1,725 kPa (250 psi) overbalance must be calculated using the expected pressure at the top of the reservoir and the estimated depths.
Critical level
The critical level in the riser is defined as the depth below the drill floor at which the mud column balances the reservoir pressure.
Critical back pressure
The critical back pressure is the pressure recorded on the choke line gauge corresponding to the critical level in the riser when the choke line is full of a biodegradable light fluid such as base oil or diesel.

In addition, the expected reaction time of the people and equipment from the time losses are noticed until the annular BOP is closed on the drillstring must be estimated. This estimate is essential because the operator and contractor must know whether the well can be safely closed in prior to the level in the annulus falling below the critical level.

The equations, nomenclature, and an example calculation are shown in the equation box. Fig. 2 is a graph of the data calculated in that example; thus, the annulus level can be quickly observed by the driller from the choke line back pressure without him having to spend time on calculations. This graph allows him to know at a glance if control has been lost and if an influx is likely.

OPERATIONAL PROCEDURES

To aid in the implementation of these procedures, a simple flow diagram was developed for the rig supervisory staff (Fig. 3). This flow diagram was found to be very useful, and the details of each step are given below.

DRILL AHEAD

Drilling ahead approaching the reservoir must be done with all personnel in a state of readiness and the choke line full with a biodegradable light fluid (base oil/diesel). The BOP valves should be open so that the back pressure can be continuously recorded. If losses occur they will be either partial or total.

PARTIAL LOSSES

Partial losses are defined as those mud losses that still give returns over the shale shakers. The partial losses will be detected by the driller on his "flow-show" and by the mud loggers.

If the losses are less than about 3 cu m/hr (20 bbl/hr), then drilling can continue. Losses greater than this level should be cured using the previously prepared lost circulation material (LCM) pill. The quantity and types of products in the LCM pills may need to be altered according to the success or failure of the premixed pill. The squeeze pressure used should not exceed 50% of the leak-off pressure at the last casing shoe.

TOTAL LOSSES

For total losses, there are to be no returns coming back over the shakers. This condition should be detected by the driller, mud loggers, and shale-shaker hand.

In the event of total losses, the primary aim must be to prevent a gas influx from entering the well. The influx should be prevented by the quick reaction of the personnel detecting the losses and the quick reaction of the driller to close in the well without a visual check.

Visual checking could result in an unacceptable delay in the closing in of the well. The well will be closed by the upper annular preventer while pumping continues through the drillstring. Pumping must be stopped following closure. This procedure will limit the level to which the mud in the riser will fall and hence the reduction in hydrostatic pressure acting on the reservoir.

During well closure, the driller must record the lowest pressure he sees on the choke line back pressure gauge. From this pressure, it can be determined whether the mud in the riser has fallen below the critical level and whether hydrocarbons have entered the well.

A second check is the pressure recorded on the choke line back pressure gauge when the well is fully closed in. If the pressure observed is equal to the critical back pressure, then it can be assumed that a gas influx did not occur.

The pressure recorded is the reservoir pressure less the combination of the head of mud in the well up to the BOP and the head of diesel in the choke line. If the pressure recorded is above the critical back pressure, then an influx can be assumed to have entered the well with some of the mud having been replaced by lighter-density hydrocarbons.

If hydrocarbons have entered the well, they must be removed as soon as possible to avoid migration. (Note: Light fluid, such as base oil or diesel, in the choke line to monitor the annulus level can only be used in deepwater wells where the lines are long enough to allow a sufficient back pressure to be recorded.)

BULLHEADING

The easiest method to remove the hydrocarbons from the well is by bullheading. Approximately 30 cu m of viscous mud should be pumped into the well bore via the kill line and displaced into the formation.

The latter part of the displacement must use the glycol/mud mixture so that the kill line is full of this mud at the end of the displacement. The BOP/wellhead should be spotted with this mixture whenever pumping is stopped as a protection against the formation of hydrates. Following the bullheading of the hydrocarbons an attempt must be made to cure the losses.

LCM PILL

Once the pressure has stabilized, a very coarse lost circulation material (LCM) pill must be pumped across and above the loss zone. If the losses are because of medium-sized fractures in the limestone, this method may work. If the losses are because of cavernous-type structures, it is unlikely that LCM will cure the problem. Drilling ahead should be attempted with a floating mud cap until it is thought that the loss zone has been drilled.

FLOATING MUD CAP

The practice of drilling a few meters, setting a cement plug, and repeating the process to make progress in the well is not recommended because it is extremely time consuming and can ruin a well for logging and future productivity. This practice is also not safe because of the inherent risks of each additional cement job. Instead, the practice of drilling with a floating mud cap under controlled conditions is recommended.

Drilling ahead should be carried out by opening up the well, monitoring the choke back pressure, and pumping sea water down the drillstring and mud down the annulus. Mud is preferred to sea water in the annulus because the mud helps stabilize the shales above the limestone loss zone.

The annulus pump rate should be controlled such that the annulus level is kept at a safe distance above the critical level. The drillstring pump rate should be large enough to keep the bit cool. Viscous pills should be pumped intermittently into both the drillstring and annulus to aid cuttings removal from around the bottom hole assembly and to hinder any gas migration.

Drilling with the mud cap should be carried out until it is felt necessary to attempt to cure the losses with BDOC and cement. This decision will be based on the following criteria:

The mud use necessary to keep the annulus level above the critical level relative to the mixing and storage capacity of the rig
The reduction in the penetration rate which may indicate that the fractured vuggy formation has been passed and that a more normal reservoir rock is being drilled.

BDOC AND CEMENT PLUGS

Bentonite, diesel oil, and cement (BDOC) is a gunk that gels quick]N- when it comes into contact with either mud or water. A quantity of bentonite and diesel oil should be prepared in advance in a batch tank. The cement will be added to this mixture at the cement unit and pumped into the well through the drillstring.

During the displacement of the BDOC (which should be isolated from the drilling fluid with a preflush and postflush of base oil or diesel) the annulus should be continuously filled with mud. The downward movement of mud in the annulus allows the mud to mix with the BDOC at the bit and form gunk. The pumping rates in the drillstring and annulus must be minimized during the mixing process to ensure good segregation and to prevent any gunk from going into the drillstring annulus.

The BDOC plug must be immediately followed by a cement plug (thixotropic cement, if possible). The purpose of the BDOC is to supply a firm base for the cement, which has a much higher compressive strength and which should therefore minimize the chances of reopening the loss zone. Thixotropic cement has a low viscosity during mixing and displacement, but the viscosity increases rapidly when the slurry becomes static. The resulting effect is that the cement quickly holds its own hydrostatic head to minimize any additional pressure on the weak loss zones.

When the cement is displaced out of the drillstring, mud should be pumped intermittently into the annulus to ensure that the thixotropic cement does not enter the drillstring annulus. At the same time, the contamination of the cement with the mud is reduced.

As a safety precaution, the drillstring should be pulled or stripped back into the shoe (depending on the losses) after the thixotropic plug is set.

Should neither the BDOC nor the cement cure the losses, then silicate-based solutions should be considered. The mixing of certain silicate-based chemicals creates a gunk-type substance very similar to BDOC.

DRILLING THROUGH PLUGS

Following the successful curing of the losses with either of the plugs, the plugs must be drilled out. In many cases during drilling, the losses will recur and additional plugs will be required.

When the losses are finally cured, consideration must be given to setting a conventional cement plug so that the cement around the well bore has an even higher compressive strength. This cement will improve the overall strength of the well bore for continued drilling and also for the setting of a casing string or liner at a later stage.

LOSSES NOT CURED

Certain types of losses will not be cured by these methods. In such cases, it is probable that large subsurface caverns have been encountered, and these caverns cannot be filled with gunk or cement effectively. Should this happen, consideration must be given to plugging back the well and starting a sidetrack. Fortunately, such caverns are unlikely to extend over a large area.

DEVELOPING PROCEDURES

These procedures were developed over a period of three wells drilled by Shell Philippines Exploration By and several wells drilled by other operators in the same area (OGJ, May 10, pp. 4849). The implementation of these procedures in the field was believed to be critical to the success of the wells. As such, it was important that the people using the procedures were also "asset owners" of them; in other words, the best way to design the procedures was to seek the help of the offshore personnel.

Numerous meetings were held offshore with the following key personnel: company drilling superintendent, drilling contractor operations manager, company drilling supervisors, drilling contractor site manager, drilling contractor drilling engineer, drilling contractor tool-pushers, mud engineer, cementing engineer, and mud loggers. Through these meetings everyone had the opportunity to contribute to the procedures. This policy was very effective.

While drilling the most recent well, the procedures and the offshore personnel's knowledge of the procedures were audited onboard the drilling vessel. The audit results indicated that the personnel had a high level of knowledge. If the procedures had been written by only the base personnel and sent to the rr for implementation, it is predicted that the results would not have been as favorable.

FIELD RESULTS

To date, three deepwater wells have been drilled in which losses were encountered. A summary of the effectiveness of the procedures follows:

WELL NO. 1

In the first well, the procedures were not fully developed. However, severe total losses were encountered causing the annulus level to fall approximately 500 m. The well was closed in, and the annulus level was monitored using the kill line.

At this time, the level was monitored by allowing it to settle in the kill line, closing the BOP kill line valves, and then filling up the kill line. The volume of mud necessary to fill up the line to surface was used to calculate the natural le%-el of the well. The BOP kill line valves were then opened and the process repeated.

The level change would indicate one of the following conditions:

Stationary - The well is in equilibrium with no losses and no influx of hydrocarbon gas.
Increasing - Hydrocarbon gas has entered the well and is migrating to surface.
Decreasing - The well is continuing to reach an equilibrium point.

The level in the annulus/kill line was found to be static. At this point the string was found stuck.

Subsequent investigations proved that the string had not been stuck by the formation, but rather by the BOP rams which had closed on the drillstring when the massive losses were encountered.

An analysis of the ram design and control configuration proved that this was a possibility if the rams were in the reactivated "block" position, as was the case.

The losses were cured on the fourth attempt by the use of a silicate-based solution and thixotropic cement. Drilling eventually continued until sufficient hole was made to allow logging. The well was found to be water bearing. Had the well been gas bearing, then the severity of the losses encountered would have required the full implementation of the procedures developed.

WELL NO. 2

Total losses were encountered as soon as the reservoir was penetrated. Because this .vas another exploration well, the mud weight in use had an overbalance of 4,140 kPa (600 psi) compared to the top reservoir pressure.

This overbalance increased the severity of the losses encountered.

Drilling ahead was conducted successfully for 50 in using a floating mud cap. BDOC and thixotropic cement plugs were effectively used to cure the losses. Drilling then continued through the formation with various degrees of losses, all of which were compared with contentional LCM pill '

This well contained both gas and oil. Because of the use of the containment procedures, those hydrocarbons that entered the well were safety and effectively, handled.

WELL NO. 3

Only partial losses were encountered in this well, partly because the top reservoir pressure was known and the mud weight overbalance was less than that in Well No. 1. Coarse and medium LCM were effectively used to cure the losses.

RESULTS

The danger from severe total mud losses in a deepwater well are much greater than that in a conventional well because of the possibility of hydrate formation from migrating gas.

As such, the hydrocarbon containment procedures must be well understood and strictly followed to prevent the BOP stack becoming inoperable.

The main advantage of the deep water in such a situation, however, is that drilling can continue with a floating mud cap in a controlled environment.

The hydrocarbon containment procedures have been used to various degrees on the three wells drilled to date by Shell in the Philippines, and the procedures have proved beneficial. The use of procedures developed by the personnel who would use them in the field was a key element in the successful application.

ACKNOWLEDGMENT

The authors would like to thank Shell Philippines Exploration By, Occidental Philippines Inc., and Forasol/Foramer for permission to publish this article.

REFERENCE

Lai, D.T., ,and Dzialowski, A.K.,"Investigation of Natural Gas hydrates in Various Drilling Fluids," SPE/IADC paper 18637 presented at the SPE IADC Annual Drilling Conference. New Orleans, Feb. 28-Mar. 3, 1989.