Evaluation of Fracture Parameters for Stochastic Generation of Fracture Networks

Abstract

The subsurface storage of industrial residues in coal mines underneath densely populated urban areas is important for waste disposal in North-Rhine-Westfalia, Germany. The residues from electric energy plants exist as filter ashes and desulphurization residues which are injected behind the long wall faces of coal mines. The use of coal mines as underground repository requires proof that contaminants, like heavy metalls, will not return via slow groundwater flow to the biosphere. The quantitative risk assessment of such underground storage facilities can only be done with the help of numerical models. We consider the Carbonifereous host rock as a fractured aquifer whose hydraulic behaviour is controlled by faults, fractures and fissures. Within the fractured system advection may occur at much higher flow velocities than within the undisturbed porous matrix. The succesfull modelling of the nearfield of the fractured host rock requires a stochastic fractured aquifer model which considers the following processes (Wendland & Schmid, 1995):

• the generated fracture network should reflect natural conditions,

• the fracture network must be generated in a three-dimensional domain considering a random distribution of fractures,

• the diffusion or adsorption of contaminants from the fracture into the porous matrix has to be approximated mathematically,

• the numerical coupling of the fast advectice transport in the fractures and slow diffusive processes in the matrix has to be solved.