Wednesday, November 22, 2017

High-Permeability Stimulation

A Classic fracture stimulation creates narrow conduits that reach deep into a formation - typically, about 2.5 millimeters wide and up to 1000 ft long. Since the 1940s, relatively low-permeability formations -less than 20 millidarcies (md) - have been successfully fractured to give worthwhile increases in productivity.

However, as formation permeability increases, creating and propagating fractures become more difficult and economically less necesary. In high-permeability reservoirs, formation damage is usually diagnosed as the major restraint on productivity and matrix acidization treatments are prescribed as the solution.

 But matrix acidization cannot solve every problem. The volume of damaged rock sometimes requires uneconomically large quantities of acid. The damage may be beyond the reach of the matrix treatment. Diverting acid into the right parts of the formation may also be difficult. Additionally, the aqueous treatment fluid or the acid itself may threaten the integrity of the wellbore by dissolving cementing material that holds particle of rock together. 

An alternative strategy for stimulating high-permeability wells has therefore emerged : the creation of fractures that are typically less than 100 ft long up to 1 in. (2.5 centimeters) wide after closure. To appreciate how short, wide fracture  stimulate high-permeability formations, one must examine the factors governing postfracture productivity.

The permeability contrast between the formation and the propped fracture is a key determinant of the optimum fracture length. In low-permeability formations there is a large contrast -and therefore a high relative conductivity - and increased fracture length can yield improved productivity.

In high-permeabiltiy formations, relative conductivity is about two orders of magnitude smaller. Increasing the length of conventional fractures offers only minimal improvement in productivity and cannot be justified economically. However, the productive performance of the fracture is detrmined by the dimensionless fracture conductivity which is directly proportional to the fracture width.  Conductivity can be raised by increasing fracture width. 

 In 1984, in Purdhoe Bay, Alaska, USA, Sohio (now BP Exploration) fractured a well with a permeability of about 60 md. The overriding aim of the exercise was to stimulate the well while avoiding fracturing into the oil/water contact (OWC) about 115 ft below the lowermost perforation.

In a relatively small fracturing treatment, some 15,000 gal (57 m3) of gelled fluid were pumped at 45 bbl/min, placing 5440 of proppant in the fracture. This treatment was calculated to be sufficient to create a fracture with a propped length of 43 ft, which based on the assumption that one foot of lateral extension would result in one foot of downward fracture migration, left the fracture easily short of the OWC. The treatment was a mechanical success and production increased by 133%  -versus a theoritacal maximum of 160%. 

After this job, attention shifted to the North Sea. The Valhal field, offshore Norway, has a soft chalk reservoir. Amoco production Co. found that, although the formation was not highly permeable (about 2 md) , it was very unstable and conventional stimulation was difficult. After acid fracturing, the acid-etched channels quickly collapsed as pore pressure was reduced. And after a convetional propped fracture, the proppant became embedded in the soft rock, destroying fracture conductivity.

Experience around the world has enabled development of a methodology for selecting wells for tip-screenout treatments. There are three classes of candidate:
  • Reservoirs with significant wellbore damage, perhaps caused by formation colapse as the pore pressure reduce during depletion. Past matrix treatments have failed, and short, wide fractures are designed to bypass the damage and connect the undamaged part of the reservoir with the wellbore.
  • Reservoirs with fine migration. A short, wide fracture can alleviate this by reducing pressure losses and velocities in the reservoir sand near the wellbore.
  • Multiple pay zones in laminated sand-shale sequences. The thin sand laminae may not communicate efficiently with the wellbore until a fracture provides a continous connection to the perforation.

 Candidate selection is a multidisciplinary task. Basic openhole logs detect sands and their bounding shales, and indicate their relative permeability and degree of invasion - gaining an insight into the formation's natural permeability before damage, the depth of invasion, the presence of zones thinner than 5 ft (1.5 m) and the formation strength. Specialized techniques like microresistivity logging may then be used to detect thinner layers of interbedded sand-shale laminae. Logs also detect water-bearing zones which must be considered during the design. Pressure transient analysis is used to identify wellbore damage and quantify the production potential of the well. After a candidate well has been identified, the next stage is to design the treatment, a process that relies on knowledge of the rock's mechanical properties and an estimate of the stresses in the reservoir and adjacent rock. 

Mechanical properties can be derived using cores, logs, and direct in-situ measurements. In many cases, however, retrieving good cores  and then accurately testing them in the laboratory are difficult. Log-derived mechanical properties rely on density and sonic measurements. Both compressional and shear sonic measurements work well in consolidated, fast formations. But in soft, slow formations, conventional sonic tools cannot measure shear wave velocity. However, a recently introduced dipole sonic tool can now make these shear wave velocity measurements in any formation.

In practice, there is rarely a comprehensive collection of core and log data with which to build a model predicting fracture shape, used for treatment design. To plug this knowledge gap, data are collected using stress test.

Stress tests consist of pumping a relatively small volume of ungelled fluid without proppant into the formation at sufficient pressure to fracture the well. In normal, low-permeability stress test pumping is then stopped and the pressure can be monitored during flowback. However, in high-permeability formations, the fluid normally leaks off into the formation rather than flowing back. Stress test are repeated several times and the resulting pressure measurements are used to dtermine the minimum in-situ stress, whic equals the closure pressure of the fracture.

Analysis of data from stress test and larger-volume calibration test-which fracture the formation usually using gelled fluid without proppant- enables choice of the most suitable fracture geometry model and confirmation of the fluid leakoff coefficient.  

The models assume that rock is an elastic material , meaning that its deformation is reversible. Dowell Schlumberger is currently examining whether this assumption holds for soft formations, as it is an important factor when looking at the fracture closure and the stress it exerts on the proppant pack. If closure stress is less than anticipated, the proppant pack could become unstable during production - unless the treatment has included a gravel pack. 

Calibration tests also provide a more accurate way of measuring fluid-loss characteristics of the fracturing fluid than can be devised in a laboratory. Fluid loss depends on the viscosity and wall-building capability of the fracturing fluid, the viscosity and compressibility of the reservoir fluid, and the permeability and porosity of the formation. In a formation with high porosity and permeability, fluid loss can be controlled by increasing the viscosity of the fracturing fluid or enhancing the fluid's wall-building capability on the fracture face by the addition of polymers and properly sized fluid-loss control agents.  

Once the choice of fracturing fluid is confirmed, the next step is to design a pumping schedule capable of delivering the necessary high proppant concentrations. The data generated by stres and calibration test are fed into the chosen fracture geometry model, which calculates the volume required to initially propagate the fracture to a predetermined length. 

No comments:

Post a Comment