New algorithm efficiently diagnoses shale fracture results from fiber-optic data

Millions of barrels of lipid are produced regular from U.S. shale reservoirs. Yet this quantity is tiny successful examination to the existent magnitude of lipid locked distant successful these subsurface rocks. The lipid and state manufacture has enactment fiber-optic sensing cables downhole to amended recognize hydraulic fracture results and wherefore stimulation and accumulation processes don't escaped the trapped lipid astatine expected rates.

Unfortunately, the streams of information received from these sensors are monolithic and hard to benignant through. A multidisciplinary team, including researchers astatine Texas A&M University and a module subordinate from the Colorado School of Mines, has created an algorithm to cleanable up the subsurface data from fracturing efforts and connection a wide presumption of however and wherever these processes succeeded and failed successful shale reservoir rocks.

"Our quantitative characterization retrieves much accusation astir fracture geometries wrong a reservoir than a elemental qualitative investigation would," said Kan Wu, subordinate prof and Chevron Corporation Faculty Fellow successful the Harold Vance Department of Petroleum Engineering. "We've tested our algorithm and already applied it successful the field."

The results were published Nov. 11 successful the Society of Petroleum Engineers' SPE Production & Operations journal.

Traditional information mentation methods, though adjuvant to engineers, are based strictly connected qualitative accusation oregon probabilities based connected patterns of information. In contrast, the algorithm was developed to stitchery quantitative data that's countable, similar temperature, unit oregon stone deformation changes wrong a reservoir. It recognizes the outcomes that occurred to make the changes and accurately models however acold and accelerated the fractures traveled, what directions they went and however large they became.

Low-frequency distributed acoustic sensing (DAS) information gathering has lone been astir for 5 years, truthful not each accusation received from the wells with fibre optics has been afloat deciphered. Also, each good has its ain scope of characteristics owed to the tremendous variations of subsurface structures. This complexity is wherefore Wu and her colleagues, chap module subordinate George Moridis, prof and Robert L. Whiting Chair, and Ge Jin, adjunct prof of geophysics astatine Mines, needed a sizeable magnitude of clip to meticulously make their algorithm.

First, the researchers tested the algorithm's quality to cleanable the information and construe elemental streams from known fracture processes. That mode they could backtrack oregon reverse the accusation to find the starting constituent of a fracture's growth. As the algorithm was expanded to recognize much analyzable information, they improved its quality to deliberation successful a guardant mode and foretell however caller and analyzable fractures initiate and grow.

Wu is an adept successful stone mechanics, Jin an adept successful geophysics and DAS technology, and Moridis is an adept successful precocious numerical methods and high-performance computing of coupled processes. Because of the multidisciplinary backgrounds of the task team, the algorithm possesses unthinkable flexibility to turn and accommodate to the benignant of information it receives. For instance, Yongzan Liu, the graduate student connected the task for implicit 2 years, is present a postdoc researcher utilizing akin methods and modeling connected fiber-optic information from hydrate-bearing sediments to show earthy state accumulation for the Lawrence Berkeley National Laboratory.

Liu, Wu, Moridis and Jin are the archetypal to make this benignant of algorithm and people results. The extremity of their probe is to yet automate the algorithm truthful that feedback from fracturing events happens successful adjacent existent clip connected a drill site. This way, engineers tin rapidly tailor fracture plan efforts to each well's peculiar composition.

"The manufacture needs this benignant of instrumentality to recognize fracture geometry and to show fracture propagation," said Wu. "The much businesslike it becomes, the amended it volition assistance optimize hydraulic fracture and completion designs and maximize good production."

---

---

More information: Yongzan Liu et al, Quantitative Hydraulic-Fracture-Geometry Characterization with Low-Frequency Distributed-Acoustic-Sensing Strain Data: Fracture-Height Sensitivity and Field Applications, SPE Production & Operations (2021). DOI: 10.2118/204158-PA