Modeling spatial fracture intensity as a control on flow in fractured rock

Abstract

Spatial fracture intensity (P 32 , fracture area by volume) is an important characteristic of a jointed rock mass. Although it can hardly ever be measured, P 32 can be modeled based on available geological information such as spatial data of the fracture network. Flow in a mass composed of low-permeability hard rock is controlled by joints and fractures. In this article, models were developed from a geological data set of fractured andesite in LanYu Island (Taiwan) where a site is investigated for possible disposal of low-level and intermediate-level radionuclide waste. Three different types of conceptual models of spatial fracture intensity distribution were generated, an Enhanced Baecher's model (EBM), a Levy-Lee Fractal model (LLFM) and a Nearest Neighborhood model (NNM). Modeling was conducted on a 10 9 10 9 10 m synthetic fractured block. Simulated flow was forced by a 1% hydraulic gradient between two vertical x-z faces of the cube (from North to South) with other boundaries set to noflow conditions. Resulting flow vectors are very sensitive to spatial fracture intensity (P 32 ). Flow velocity increases with higher fracture intensity (P 32 ). R-squared values of regression analysis for the variables velocity (V/V max ) and fracture intensity (P 32 ) are 0.293, 0.353, and 0.408 in linear fit and 0.028, 0.08, and 0.084 in power fit. Higher R 2 values are positively linked with structural features but the relation between velocity and fracture intensity is non-linear. Possible flow channels are identified by stream-traces in the Levy-LeeFractal model.

Heterogeneity of structure and stress in the Rotokawa Geothermal Field, New Zealand

“…Wide aperture fractures and high fracture densities commonly correlate to permeable zones in geothermal wells [ Barton et al ., ; Sheridan and Hickman , ; McLean and McNamara , ; Wallis et al ., ; Co , ]. Fracture length and connectivity, their interrelationship, and the relationship between these factors and fracture aperture and density are also important when considering structural permeability in geothermal reservoirs [ Long and Witherspoon , ; Lee et al ., ; Barton et al ., ]. Increased connectivity and density of a fracture network in a reservoir will often allow for higher permeability associated with the presence of longer fractures (in turn associated with wider apertures) [ Long and Witherspoon , ; Gudmundsson et al ., ; Philipp et al ., ].…”

Naturally fractured hydrocarbon reservoir simulation by elastic fracture modeling

“…Lapponi et al (2011) used outcrop data to construct a 3D model in a dolomitized carbonate reservoir rock from the Zagros Mountains, southwest Iran. Lee et al (2011) studied the spatial fracture intensity effect on hydraulic flow in fractured rock. They used outcrop for simulation of spatial fracture intensity distribution.…”

An Assessment on Permeability and Grout Take of Limestone: A Case Study at Mut Dam, Karaman, Turkey

“…This formation is impermeable and about 1 This unit outcrops at the reservoir area and downstream side of the Mut dam site. boreholes showed that this formation has a thickness of approximately 40-50 m alo Regression analyses as statistical methods are widely used in geotechnical and rock engineering practice (e.g., Hoek and Brown [27]; Apte et al [28]; Akca [29]; Uddameri [30]; Sivrikaya [31]; Zorlu et al [32]; Yagiz et al [33]; Yagiz and Gokceoglu [34]; Chen-Chang et al [35]; Valia and Arpa, [36]; Azimian and Ajalloeian R [10]; Farid and Rizwan [15]; Öge and Çırak [37]; Rahimi et al [38]). Average values of rock mass properties of karstic limestone block (Q value, RQD, GSI, JSP, Ap, Lu, SPI, k) and average Gt were calculated (Table 9).…”

“…Function modellings were performed according to equations derived with a simple regression equation and the results of the multiple nonlinear regression analyses are given in Table 12. Considering the VAF, RMSE, R 2 , and MAPE values, Equations (17), (20), (23), (27), (28)-(30), (34), and (35) show the better prediction. Remaining empirical equations also exhibit significant performances.…”

“…A 1:1 line was drawn in order to assess better the placement of the equations. Equations (23), (27), (30), (31), (34), and (35) show better performance according to the accumulations of data along the 1:1 line. On the other hand, all equations also show better performance.…”