Martin Cherrington, Kenneth R. Weeks
Cherrington Corp.
Sacramento
Harold Y. Anderson
H.Y. Anderson Engineers
San Rafael, Calif.
Successful directionally drilled crossings of three major elevated waterways in California shows that long distance, large-diameter crossings are possible even in such environmentally sensitive areas.
The crossings, part of the PGT-PG&E (Pacific Gas Transmission Co./Pacific Gas & Electric Co.) 900 mile, 42 and 36-in. expansion of its natural gas system from Canada to California, set two world records for crossing distance and diameter.
Contractors successfully placed 42in. product pipe under elevated waterways in the Sacramento Delta, 58 miles northeast of San Francisco (Fig. 1), without breaching any levees or causing an underground fracture. Either event would have resulted in a major flood.
The waterways were the Dutch Slough (2,851 ft), the San Joaquin River (3,549.5 ft), and the Sacramento River (3,863 ft).
The most overriding concern for the crossings was to preserve the fragile levee system. Project management, along with general contractor
Bechtel Corp., elected to use horizontal directional drilling (HDD) to avoid the irreparable damage trenching would cause to the levees.
Even with selection of HDD as the preferred method, flooding remained a risk. Special engineering solutions were necessary to protect the levees and reduce the risk of flooding if an underground fracture to the river bottom should occur during drilling.
LEVEES PROTECT FARMLAND
The Sacramento Delta is the confluence of two major rivers and five secondary tributaries which all ultimately empty into San Francisco Bay. These rivers and their associated waterways are unique in that they are elevated as much as 20 ft above the surrounding land by a system of levees built as early as 1862.
This network of meandering rivers has been confined behind levees to form islands, or tracts of dry land, that today support a major agricultural industry and associated econoMY.
Historically, levee construction was accomplished by crude dredging methods with whatever materials were locally available. Levee stability was a matter of trial and error. No geotechnical engineering was used in the design of these structures.
Efforts to construct and maintain artificial levees on each side of these rivers have continued to the present with varying success.
The levees are continually raised as subsidence occurs. Some levees along ship channels, for example, are protected from damaging scour by rock slope protection materials.
A major concern for levee maintenance is the occurrence of "rat holes": conduits from which seepage gradually becomes a major flow of water through the levee. This seepage quickly and sometimes catastrophically erodes out the peat foundation and dumps the entire river upon an adjacent island and its inhabitants.
GEOLOGY
Sherman and Jersey Islands, site of the three crossings, are covered by a deposit of unconsolidated sand, gravel, clay, and peat of alluvial origin. The peat deposits are up to 25 ft thick along the pipeline route.
The peat is a very soft, saturated, black fibrous mixture of organic materials and silt. Typical properties include a dry density of 15 lb/cu ft and in situ shear strengths approaching 0.
It has very little bearing capacity and would not support foot traffic were it not for a silt crust which forms on the ground surface.
Underlying sandy silts and silty sands are typically loose deposits of fine sediments, gradually becoming medium dense with depths of around 50 ft. Below these poorly consolidated layers are strata of dense sands and firm clays,
Borings taken at location sites revealed high water tables; some borings flowed water. All holes were plugged with cement to ensure their integrity.
Permeable layers of sands cross under the rivers and were believed to be able to provide hydraulic conduits between the river bottom and the surrounding lowlands.
These complex formations caused numerous drilling problems.
Although mud densities typically ranged between 9-9.5 lb/gal, variations in lithology from sands to clays required different funnel viscosities to optimize drilling or hole opening.
Project mud engineers ParcChem Inc., Laurel, Miss., determined that sands drilled best with viscosities ranging between 85-150 sec/quart. Dense clays and silts, on the other hand, drilled best with funnel viscosities between 40-55 sec/quart.
These wide variations in viscosities made it extremely difficult to optimize drilling and avoid sticking the drilling assembly when rapid formation changes took place. Mud additives were used to control water loss and increase lubricity.
Drispac, a water-loss reducing agent, was added 2-3 times/day during drilling. Detergent (soap) was added 3% by volume to increase lubricity and prevent clays from sticking to the drill pipe.
Additionally, active aquifers produced positive flow rates of up to 8 gpm to the surface which caused stuck drill pipe or hole openers or both. Drilling assemblies were stuck and required washover to free them an average of three times/crossing.
Five fishing jobs were necessary before the 42-in. product pipe could be pulled into place.
While an average of four core holes with formation samples were taken along the course of each river crossing, geological control still remained a major operational problem. More detailed geological information would have helped planning of more efficient mud programs and increased drilling efficiency.
PLANNING, PREPARATION
Planning and engineering the project took 12 months; completion took approximately 1 year. Many engineering solutions were evaluated before construction could begin.
Profiles of the 42-in. diameter pipelines consisted of 4,000-ft radius curves on each end of the uniformly level crossing some 50-60 ft below the river bottom. Containment structures, where pipelines enter and exit the drilled hole, were positioned no closer than 500 ft from the levees.
Exceptions included the northwest side of the Sacramento River where the natural ground is some 50 ft above the river; thus no structure was necessary.
Pipe stress during installation due to bending, external pressures, radius of curvature, axial tension or compression, and possible pipe ovalling-was examined for all conditions.
The product pipe, API 5L Grade X65 (specified minimum yield strength = 65,000 psi) and having a 42-in. OD with 0.75-in. W.T. (creating a D/t = 56), was furnished by Napa Pipe Corp., Napa, Calif., in double random lengths.
The coating consisted of 11-12 mils of fusion-bonded epoxy (FBE) covered with 60 mils of Protogal (coal-tar urethane). The minimum radius of curvature for temporary installation was established as 1,060 ft.
Stress in the pipe after installation due to such operating conditions as 0 internal pressure, radius of curvature, external pressures, and other conditions were also investigated.
The minimum radius of curvature for permanent installation was established as 4,000 ft.
Major considerations during drilling included collapse of the hole, loss of drilling mud, blowout to the river bottom and accidental creation of a conduit for river water to flow to lowlands.
All of these factors were examined by mud engineers who paid close attention to maintaining proper mud density and other properties.
The submerged weight of the pipeline during pullback was calculated for various mud densities. Although the pipeline weighed 330.5 lb/ft in air, it had a positive buoyancy of 769.6 lb/ft when submerged in 80 lb/cu ft (10.7 lb/gal) drilling mud and cuttings.
To reduce this upward force to manageable limits, a 28-in. high-density polyethylene pipe was placed inside the 42-in. pipeline. This smaller pipeline was filled with water as it was pulled back into the drilled, mud-filled hole with a 4-in. fiberglass fill line.
The entire pipeline was preassembled by H.C. Price, Dallas, on one side of each crossing, hydrotested, then placed on rollers. An articulated pulling head was attached to the front end of each pipeline so that it would follow the drill string along the curved crossing profile (Fig. 2).
Support structures were necessary to develop the required sag and overbend curves at pipe entry points. Observing 1,000-ft radius minimum breakovers, the 42-in. pipeline was raised some 35 ft off the ground before it headed down into the drilled hole. These structures were built with the use of 10-in. pipe piling for foundations.
CONTAINMENT CELLS, WORK PADS
Containment structures, designed by H.Y. Anderson & Associates, San Rafael, Calif., were necessary for drill string entry and exit. These structures had the capability of controlling accidental flow of river water through the drilled hole.
The structures also provided a convenient way to contain drilling mud as it backed up in the drilled hole during pull-back (Fig. 3).
Work pads were used extensively to distribute equipment loads and provide lateral support for filled containment structures on very soft peat. Access to sites for construction was provided for by A.N. Stevens Co., Rio Vista, Calif., using imported materials (sand and gravel) for roadways and pads. Pile driving was provided by Foundation Constructors, Rio Vista.
The concept developed for each river crossing included an entry and an exit site. At the entry site, the drill rig was positioned; at the exit site, the pipeline string was assembled for the pull back into the drilled hole.
The entire working area was provided with a gravel mat consisting of filter cloth, sand, and gravel to depths of 4 ft, in some cases. At each entry site, a piled foundation was provided to support the drill rig on the very soft peat formation.
At each site, except the high-elevation side of the Sacramento River, a sheet-piled containment cell approximately 60 x 330 ft was constructed. The cell was high enough so that, in the event of flooding, it would spill over its top at greater than 8 ft.
The cell was long enough to allow the drilled hole to be installed just under the tips of the sheet piles on the river side. Each cell had to be designed as a rectangular bath tub which tends to burst at corners and bulge out at its sides when filled with water.
No lateral support was available in the peat formation to resist hydrostatic forces until the effect of the gravel pad surrounding the cell was taken into account. Some sheets were driven to depths of 50 ft before encountering any resistance.
At the Sacramento River entry site, Bethlehem PZ35, Grade 50, sheet piles 65 ft long were selected. These were driven and supported by external wales near the ground level, and II/s-in. diameter Dywidag tie-rods were used to provide the necessary lateral restraint.
The Bethlehem ball and socket sheet pile interlocks had adequate strength to resist pulling forces anticipated at the corners.
Drill rig foundations were designed to be able to resist a pulling force of 300 tons. Again, because of the depth of peat, piles selected for these foundations consisted of 18-26in. diameter steel pipe piles, 8 of which were batter piles.
OPERATIONAL CONSIDERATIONS
NEW EQUIPMENT
Opening a 60-in. diameter hole and installing a 42-in. product pipe necessitated the design and construction of special drilling and pipehandling equipment.
The Dutch Slough crossing was handled by Cherrington's existing rigs which are capable of 300,000 lb of pulling force.
A new rig, however, had to be constructed to handle the anticipated torque (92,500 ft-lb) and pull (up to 750,000 lb) needed to drill and ream under the San Joaquin and Sacramento rivers. Both would be record crossings.
Rig 750 (Fig. 4), as it was called, increased the pulling and torque capacities over conventional rigs by 150% and 130%, respectively. When placed in tandem with a new pipehandling device, called a "pipe thruster," they would provide more than 1.5 million lb of force to move the 42-in. product line the distances required.
The "A" frame-shaped pipe thruster (cover photograph) was strategically located on the opposite side of each crossing to the rig. Secured with 60-ft steel pilings driven 30 ft into competent subsurface formations, it had the capability of supplying 750,000 lb of working force and 900,000 lb of total force to push the pipe.
With the 750 rig and pipe thruster working in tandem, the force necessary to overcome the pipe inertia and friction during installation could be distributed over the entire length of the pipe.
To accommodate the 42-in. product pipe, a 60-in. diameter hole opening device was designed and built (Fig.5). This device has a single row of 2/2-in. wide carbide insert drag-type teeth welded on 4-in. centers along its three arms.
Nozzle numbers 10 to 14 are strategically located adjacent the cutting teeth for maximum lubrication and cuttings removal. Each arm incorporates six nozzles to facilitate cuttings removal.
A 3/16-in. mud screen sieve was placed at the leading and trailing connections of the hole opener in the event mud circulation were required from either side of the crossing.
The long distances and high torque generated by the hole openers required special drill pipe. S135, 21.90 lb/ft, 5/2-in. drill pipe with H90 tool joint connections was used with a tensile strength of 135,000 psi to facilitate the movement of the hole opener as well as pulling the 42-in. product pipe into place on two of the three crossings.
Before the pulling back of the product pipe on the Sacramento crossing, the 5/,-in. drill pipe was replaced with readily available 5-in., S135 19.5 lb/ft drill pipe.
This replacement resulted from the stress fractures found in the 5-1/2.-in. pipe which were associated with the over bending that occurred in a large washout discovered on the exit side of the Sacramento River.
Many downhole problems required moving equipment in and out of the containment cells. Special leak-proof, 10 x 10-ft, doors allowed some small equipment to be easily moved in and out.
Larger pieces of equipment, such as the drilling rig and small cranes for handling drill pipe, had to be lifted with a 4100 Manitowoc crane.
Base rock equipment pads, 60 x 100 x 4 ft, were built in each cell to support the drilling rig and other equipment. Placing -the drilling rig close to the hole entrance ensured better control of drilling assemblies during fishing operations.
HYDRAULICS; CUTTING REMOVAL
Concern for underground fractures during drilling placed severe operating constraints on the job. A critical balance had to be maintained between the mud hydraulic forces exerted during drilling and the need for cuttings removal.
Generating more than 75 psi of open hole hydraulic pressure from the drilling mud risked an underground fracture. Insufficient mud velocity would complicate cuttings removal causing hole problems and sticking of the drilling and hole opening assemblies.
Pilot and intermediate holes were jetted with flow rates of 3-5 bbl/min with mud viscosities of 40-50 sec/qt in clays and 80-150 sec/qt in sands to facilitate cuttings removal.
When the hole was enlarged to 60 in., however, flow rates (up to 20 bbl/min) and mud viscosities (as just cited) only produced annular velocities of 6 fpm, insufficient to move cuttings back to the surface. But higher flow rates would have risked the chance for fracturing into the river bottom.
Air-lift techniques for cuttings removal were employed briefly but with limited success.
Using 9-1/8-in. casing to vacuum the cuttings from the enlarged hole during rearning eliminated the risk of fracturing as a result of hydraulic mud pressure.
While air-lifting could be reasonably accomplished near the entrance to the hole, the technique proved impractical as progress was made to the opposite side.
A decision was made to leave cuttings in the hole, mixed with the drilling mud as a slurry. This, unfortunately, led to pack-offs ahead and behind the hole opener during rearning and caused repeated sticking of the drilling assembly resulting in numerous delays and fishing jobs.
This practice also caused concerns about pulling the product pipe through the hole.
Also, drilling mud was considered a hazardous material and stored in frac tanks on location. Surplus drilling mud could be conveyed to either side of a crossing by a separate, parallel 7-in. diameter hole drilled under each crossing. Mud could be circulated in either direction when needed.
This was not the case on the Sacramento crossing, however. Vacuum trucks were used to haul mud back and forth to each site as it was needed. They were also used throughout the project to haul off used mud and cuttings to disposal sites.
ENVIRONMENTAL MEASURES
Twenty environmental regulatory agencies examined each facet of the operation. These agencies included the California Office of Emergency Services, the National Response Center, the California Public Utilities Commission, the California Department of Fish & Game, four reclamation districts, several irrigation districts. PG&E and Bechtel also examined the operation.
Surface drilling operations were tightly controlled by environmental and regulatory inspectors during all operations. A "no leak" policy was imposed at every worksite.
More than 1 acre of MAXX Select Sorbent, manufactured by National Sorbents Inc., Cincinnati, was employed to prevent hydraulic oil leaks and spills from contaminating the soil and ground water.
This absorbent cloth, made from high absorbency fibers and supplied in 3-ft wide rolls, 0.375-in. thick and 150 ft long, can absorb 25 times its weight. Each work pad was lined with the Mirafi paper and covered with gravel and decomposed granite.
Each piece of stationary hydraulic machinery required a layer of absorbent cloth directly under it as an extra added precaution. Disposal of the used material required placement in lined hazardous waste dumpsters which were removed to acceptable dumps off location.
SACRAMENTO CROSSING PROBLEMS
Although all three crossings had operational problems, the 3,863 ft Sacramento River crossing proving to be the most challenging.
During opening of the hole to 60in., midway between the entrance and exit under the Sacramento River, a sudden reduction in torque from 60,000 ft-lb to 24,500 ft-lb and sudden mud pressure loss (nominal mud pressure = 500-600 psi) suggested that the pipe had separated between the hole opener and the drilling rig on the entry side.
The driller quickly stopped the drilling operation and slowly moved the pipe back until it met resistance. As the pipe was slowly rotated during pumping, the torque momentarily built up to 35,000 ft-lb as the mud pressure increased to 400 psi. This suggested that the separation may have occurred at a connection and that a box had split or cracked (Fig. 6a).
Only pulling the remaining pipe out of the hole, however, could verify that it was a bad connection and not severed pipe. If a connection had separated due to a cracked boy then a special fishing head might be configured to reestablish the integrity of the failed connection in order to continue the hole opening operation.
In preparation for withdrawal of the remaining string of pipe, the entire downhole drilling assembly needed to be pulled back approximately 7 ft toward the exit side. This was necessary so that the drill pipe connected to the rig's draw works on the entry side could be broken with its built-in breakout tools.
Once the rig connection was broken, 9-5/8-in. casing could easily be run over the severed drill pipe so it could be removed and ultimately guided back to the severed connection downhole.
The rigs on the entry and exit side of the crossing worked carefully in unison to move the entire downhole assembly back the necessary 7 ft. As tension was applied to the trailing I string of pipe, it suddenly separated i between the back side of the hole opener and the exit rig (Fig. 6b).
The 60-in. hole opener now lay mid channel with separated drill pipe ahead of it and behind it. With the project close to completion, it now faced two major fishing operations.
Fishing operations were launched on both sides of the crossing to regain control of the downhole assembly. The 7 ft of drill pipe, on the entry side, was cut off, and 9-1/8-in. casing was run over the separated leading string of pipe.
On the exit side, the trailing string was removed. An attempt would be made to fish the severed end behind the hole opener by reentering the 60in. hole with a special fishing tool.
While work was in progress on the entry side to reconnect the suspected cracked boy a special fishing tool was configured to fish the separated trailing string on the exit side. This special fishing tool consisted of 5-1/2in. S135 drill pipe with a 2-1/8-in. OD pipe placed inside it.
Both the 5-1/2 in. and 2/8-in. pipe were bent approximately 5 about 4 ft from their respective ends. These two bent pipes could then be rotated together or independently so that the smaller pipe could be positioned by trial and error for insertion into the severed end of the trailing string.
FISHING TRIP
Attempts to reenter the exit hole with the fishing assembly failed, however (Fig. 7a). Each time the drill string was run in, it stopped 90 ft into the 60-in. hole. It was reasoned that the hole had washed out beyond the 60 in. within the underlying soft peat formation.
Repeated trips in and out of the 60in. hole with the 5-1/2-in. drill pipe had easily eroded the surrounding soft peat formation near the surface. As the washout became larger, the 5-1/2in. drill pipe, not confined to the 60in. diameter hole, apparently exceeded its bending limits which lead to its fatigue and ultimate failure (Fig. 7b).
A new solution was required to get the drill pipe back into the 60-in. hole, past the large washout, to fish the severed end of the trailing string. This solution employed 200 ft of 13-3/8in. casing, weighing 61 lb/ft empty, and having 19 lb positive buoyancy in 11 ppg mud.
With a pull rope attached to a removable plug in the lower end, it was floated past the washout to provide a conduit to the competent 60in. hole for guiding the fishing tool to the severed pipe downhole (Fig. 7c). The 13-3/8-in. casing was not anchored at the surface.
When the plug was pulled from the lower end, it immediately filled with mud and sank out of sight into the large washout. A clam-shell crane was used to fish the 13-3/8-in. casing resting on the bottom of the washout (Fig. 7d).
After it was found and brought to the surface, it was then anchored securely to the exit side so that access to the exit side could be established.
By now the entry side fishing operation was well under way.
On the entry side, the 9-5/8in. casing had been run over the leading string of drill pipe two joints past the separated connection. This allowed the drill pipe to be retrieved so that a pin could be fitted with a fishing tool. The casing also would act as a guide to make it less difficult to re-make the connection (Fig. 8).
The retrieved pipe showed that a box indeed had failed. The ability of the driller to remake the connection with torque and mud pressure after the connection had failed suggested that the connection integrity could be restored.
A box-tap, manufactured by TriState, was first considered but abandoned because of concerns that the box-tap could potentially, deform or crush the cracked box as well as prevent a true axis from being maintained along the drill pipe to continue the operation.
A preferable solution was to reestablish the integrity of the cracked box to reduce the chance of mud leaking and ultimate pipe washout and to keep the axis of the pipe true.
A pin on a joint of 5-1/2-in., S135 drill pipe was fitted with a 10-in. long sleeve welded around the greater OD of the connection. The sleeve's ID (6-1/16 in.) was designed to fit tightly over the cracked box upset OD (6 in.) when screwed into it.
The sleeve's OD was kept to 8-1/2 in. to allow it to pass through the 9-1/8-in. casing. The pin, with its special adapter, was run into the hole through the casing, and the integrity of the failed connection was restored.
The entry-side fishing job now made it possible to continue the rearning operation.
CIRCULATION, INSTALLATION PROBLEMS
Before rearning could continue, however, one more problem needed to be resolved. Because the trailing string had been severed, mud circulated through the drill pipe from the entry side would only exit out the severed end. A seal had to be made so that mud would circulate through the hole opener allowing it to function properly.
A '/it,-in. mud screen previously installed behind the hole opener provided an opportunity to pump two 5in. polyfoam pigs down the pipe to jam up against it and thereby make an ideal seal.
But before this could be accomplished, another screen located in front of the hole opener had to be milled out to clear the path for the pigs. Once this was accomplished, the pigs were pumped into place and hole opening operations success-fully resumed.
On the exit side, 5-in., 19.5 lb/ft drill pipe with a steering tool and jet bit bottom-hole assembly was now run in the hole through the 13-1/8-in. casing to follow close behind the 60-in. hole opener to completion.
The 5-in. S135 pipe would be used to pull the 42-in. product pipe back through the 60-in. diameter hole for installation. A difficult but not impossible situation had been resolved.
The hole had been opened to 60 in. and was now ready for installation of the 42-in. product pipe under the final leg of the project.
The product pipe was welded together and installed in one continuous length.
On rollers, side booms and cranes maneuvered the 42-in. product pipe into the containment cell to the pipe thruster (Fig. 9).
Inside the containment cell, the pipe thruster, securely positioned with 28in. pilings through its legs (cover photograph, provided the force to handle the large pipe and work in unison with the 750 Rig on the opposite side.
A 42-in. bull nose adapter with a swivel connector was welded on the leading end of the product pipe (Fig. 2). This sealed the pipe and provided a means of attaching to it the 5-in. pull-back pipe for pulling it back through the 60-in. hole.
With less than 300 ft to go before the pipe installation would be completed, the pipe-thruster clamp failed to hold the pipe while applying 400,000 lb of force. The clamp had to be removed from the pipe thruster, quickly welded, and re-installed.
The only obstacle to completion was overcoming the inertia and friction of the resting pipe after it had sat for 5 hr during the repair of the clamp. It took 1.1 million lb of combined force from the 750 rig and the pipe thruster to overcome the friction that had developed over the period.
The product pipe was pulled 7 hr later into its final resting place under the Sacramento River (Fig. 10).
Copyright 1993 Oil & Gas Journal. All Rights Reserved.
