In late 2010, Shell began an Eagle Ford appraisal program in the Piloncillo Ranch lease in south Texas. These wells are 8,500- to 9,500-ft-true-vertical-depth (TVD) horizontals, with an average total depth of 14,500‑ft measured depth (MD). Initially, underbalanced-drilling techniques were used to drill the formation. As more wells were drilled, completion fracturing of offset wells began to cause well-control problems as induced fractures were encountered in horizontal sections during drilling. The authors describe the development of well-control techniques that allowed successful drilling operations.
Well Design
The original pore-pressure concept drove the field to a basic two-string casing design. The wells have 9⅝-in. surface casing and 5½-in. production casing. The surface-casing point was designed to be set between 3,000- and 3,500-ft TVD in the Midway shale below the base of the freshwater aquifer, which varies between 2,000- and 2,800-ft TVD. The fracture strength at this point was found to be 17 lbm/gal equivalent mud weight (EMW), which was considered to be sufficient for kick tolerance through the openhole interval to total depth (TD) in the Eagle Ford. The next hole section is then drilled out of the surface shoe to the kickoff point in the Austin chalk and landed in the upper Eagle Ford, where a 5,000-ft lateral section is drilled. The bottomhole assemblies (BHAs) were optimized until one run was possible from drilling out the surface shoe to TD of the well. The fracture strength of the surface shoe, along with the assumed pore pressure of 12.5 lbm/gal EMW in the Eagle Ford, allowed for a two-string design.
This initial appraisal phase consisted of single-well pads with 5,000-ft laterals along a preferred azimuth. With further field development, the plan was set with multiwell pads in order to minimize the footprint and reduce undeveloped space in the reservoir (Fig. 1 above).
Drilling—Phase 1: Initial Wells
The first wells were drilled with a mud weight (MW) of 11.0 lbm/gal and exhibited few well-control challenges. No flow was observed while making connections or when performing flow checks at TD, and minimal background gas while drilling and at bottoms-up upon reaching TD was common. There was no flare during drilling, but the flare at bottoms-up at TD averaged between 10 and 20 ft. Previous experience in the Eagle Ford and other tight gas plays showed this level of gas to be manageable with a conventional rotating control device and a mud/gas separator. Some losses were encountered during the initial 5½-in.‑production-casing cement jobs with a two-slurry design of top of cement at surface and a 12.7-lbm/gal lead and 16.4‑lbm/gal tail. This was attributed to formation weakness in the Olmos sand, which occurred from 5,300- to 6,700‑ft TVD. Equivalent-circulating-density simulations established this formation as a potential weak point, with an estimated fracture strength between 13- and 14‑lbm/gal EMW across the field.
Drilling—Phase 2: Establishing Boundaries
Wells were initially drilled with an MW of between 11 and 12 lbm/gal. A trial was initiated to determine the lower limit for MW in an effort to increase rate of penetration and reduce drill time. While drilling the Piloncillo A 3H well, the MW was reduced to 9.9 lbm/gal where borehole instability was experienced, resulting in twist-off of the BHA and subsequent plugback and sidetrack operations. The sidetrack portion of the well was drilled to TD with a 10.2-lbm/gal mud, and no issues with borehole instability were noted. This was then established as the lower limit for production-hole MW on single-well pads.
By mid-2011, rigs were drilling on multiwell pads with ongoing spacing trials and growing offset production. During this phase of the program, well-control challenges were faced in the form of increased gas while drilling and flow being observed before pulling out of open hole (POOH). The team believed the well-control challenges were caused by drilling into fractures from offset, completed wells. While drilling the Piloncillo D 6HA well, high connection gas was observed with an MW of 11.4 lbm/gal. The MW was increased to 11.6 lbm/gal at TD, and, upon flow check, the well was observed to be flowing at 2.7 bbl/hr. Further increases in MW were not achievable without inducing losses to the Olmos sand. In response, the team developed a technique to use a dual-density system by pumping a weighted pill that would keep the pressure on the Olmos constant while increasing the bottomhole pressure. After pumping out of the hole to the heel, a 16‑lbm/gal pill was spotted between the bottom of the Olmos sand (the previously established weak zone) and the heel in order to obtain a 12.5-lbm/gal EMW on the Eagle Ford without losing returns uphole. With the assumption that the Eagle Ford had a pore pressure of 12.5‑lbm/gal EMW, the pill was expected to stop the well from flowing. After spotting the pill, a flow check was performed and no flow was observed. During the trip out of the hole, however, the well was observed to be flowing at 1.8 bbl/hr. Upon returning to the kickoff point and washing through the weighted pill, the MW was raised to 12.2 lbm/gal in an effort to reduce the flow rate, which was now 2.4 bbl/hr. This demonstrated that incrementally increasing the MW of the entire system in order to provide sufficient hydrostatic pressure on the Eagle Ford would not be possible without exceeding the fracture pressure on the Olmos. A second heavy pill was spotted between the bottom of the Olmos sand and the heel in order to achieve a 13.5-lbm/gal EMW on the Eagle Ford. A flow check was performed above the pill and at the surface-casing shoe, and the well was not found to be flowing. This indicated that the pore pressure was closer to 13.5-lbm/gal EMW, which is much higher than the original pore-pressure prediction from offset wells. At this time, a diagnostic fracture-injection test was conducted and it was determined that well-control problems were more a function of reservoir connectivity through hydraulic fracture or a reduction in skin damage than a function of pressure.
Drilling—Phase 3: Distinguishing Flow Source
After encountering several similar well-control challenges, it was deemed necessary to consider either adding an intermediate string of 7-in. casing set across the Olmos sand or establishing a fit-for-purpose well-control technique involving the use of weighted pills. It was determined that alternative well plans that included an intermediate casing string would increase well costs by 18% and would prove unfeasible, whereas a fit-for-purpose well-control practice would strike the balance between safety and environmental concerns and economic concerns. Upon review of seven wells with well-control challenges, it was noted that the well flow did not always originate from drilling into stimulated fractured offsets. Increased flow rates lead to increased MWs; however, it was observed that this caused self-induced losses as a result of breathing and ballooning to the Olmos. The breathing problems had been difficult to identify initially. Most of the losses were attributed to flow going over the shakers with the slurry of cuttings being fed through solids-control equipment. Because some volume of drill fluid was temporarily sequestered from the active system and not recorded accurately, it became difficult to distinguish losses in the form of gradual seepage from drilling fluid that was accompanying cuttings over the shakers and subsequently being conditioned and circulated back to the active system. Because of the difficulty in accurately distinguishing breathing fluid from actual reservoir gas flow, the team used well-flow data to create Horner plots that could help distinguish a slowing or stable trend. For details about the Horner plots and their results, please see the complete paper.
Drilling—Phase 4: Advanced Well Control (Tripping With Flow)
Certain wells were found to be impossible to kill because of limitations of pill weights and the height available between the loss zone (the Olmos sand) and the Eagle Ford reservoir zone. Further field development resulted in the need to drill wells on or near existing pads where the likelihood of drilling into a fracture from a shut-in, offset well was high and unpredictable. Once the original wells were hydraulically fractured, the hydrocarbons would preferentially flow toward the path of least resistance. This was first seen on the Piloncillo E 3HA well, where a flow at TD of greater than 10 bbl/hr was recorded along with a high amount of gas at bottoms-up. The team deemed it necessary to spot a weighted pill in accordance with the field’s standard operating procedures. A weighted pill with an EMW of 13.5 lbm/gal at the landing point was spotted. When a flow check was performed, the well was flowing at 3 bbl/hr. After circulating bottoms-up above the pill, another flow check was performed, and the rate appeared to be increasing. This indicated that the source was the reservoir that had been hydraulically fractured and subsequently shut in. The team made the decision to wash through the weighted pill and spot a larger pill of higher density, for an EMW of 14.0 lbm/gal at the landing point. A subsequent flow check indicated that the well was no longer flowing. The Piloncillo E 3HA well is located on the eastern portion of the ranch, where the Olmos sand exhibits a higher fracture strength and the Eagle Ford lies at a deeper TVD. Both of these factors allowed for a larger and heavier weighted pill without inducing losses to the Olmos sand.
The entire portion of nonproductive time (NPT) associated with spotting weighted pills twice and washing through the same amounted to 200 hours and 20% of well cost. This demonstrated that if well-control challenges similar to those the team experienced on the Piloncillo E 3HA well became commonplace, an increase in NPT and in associated cost would need to be absorbed into the field-development plan.
It became apparent that it might be necessary to evaluate the feasibility of tripping for a limited time with the well flowing. In order to determine if the well could be killed with a weighted pill, the team used the steady-state-flow equation to see how flow rate from the reservoir would reduce with increasing mud weight (for a discussion of this equation and its application, please see the complete paper).
In order to ascertain the full effect of POOH with the well flowing, owing to the inability to increase the MWs of the system sufficiently because of the Olmos sand, the team created a model for well flow on the basis of the data gathered on the Piloncillo E 3HA well, resulting in the mitigation of the risk of a well-control event in this scenario to adequate levels.
For a discussion of washing through dual-density systems, please see the complete paper.
Unconventional Reservoirs, Loss Zones, and Implications for Cementing
The two-string casing design that is being executed currently in the Eagle Ford would not be a feasible design if proper cementation of the wells were not possible. After the first few wells exhibited losses to the Olmos sand, the team re-examined the slurry design and developed a new design with different densities and thickening times, development of better static gel strength, and a more robust centralizer-placement program (please see the complete paper for details). In addition, lost-circulation-material treatments were used to strengthen the fracture gradient of the Olmos sand while drilling through the zone.
Field Performance
On the basis of the evolution of well-control practices and other performance-optimization initiatives in the Eagle Ford to date, the team has managed to safely deliver more than 100 wells during a 2-year period. When one considers averages from 2011 and 2012, the team accomplished a reduction in total well time and cost of 30 and 28%, respectively, once it developed fit-for-purpose well control practices such as those discussed here (and in more detail in the complete paper).
This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 163984, “Eagle Ford Shale Well Control: Drilling and Tripping in Unconventional Oil and Gas Plays,” by K. Ridley, SPE, M. Jurgens, SPE, R.J. Billa, SPE, and J.F. Mota, SPE, Shell, prepared for the 2013 SPE Middle East Unconventional Gas Conference and Exhibition, Muscat, Oman, 28–30 January. The paper has not been peer reviewed.