Skip to content

Latest commit

 

History

History
30 lines (27 loc) · 2.23 KB

File metadata and controls

30 lines (27 loc) · 2.23 KB

Endorse software

The package provides stochastic simulation tools for the characterization of the safety of an excavation damage zone of a deep geological repository of radioactive waste. The safety is described by the safety indicator – the maximum of a simulated concentration on the boundary of the computational domain. The Darcy flow and transport of the contaminant are calculated by the Flow123d simulator. The simulator uses the discrete fracture-matrix (DFM) approach that combines a network of (random) fractures and a 3D continuum. The safety indicator is considered a random variable dependent on the random fracture network and uncertainties in other parameters of the model. The stochastic properties (mean, variance, …) of the safety indicator are estimated using the Multilevel Monte Carlo (MLMC) method. The whole stochastic calculation is executed through the endorse-mlmc script described below.

The hydraulic conductivity and porosity on the excavation disturbed zone (EDZ) are the key parameters affecting the safety indicator. These parameters are substantially increased in the vicinity of the tunnels of the repository compared to the intact rock, partly due to damage attributed directly to the excavation method and partly due to deformations caused by changes in the stress field. However, the response to the stress changes is not immediate due to the presence of water. The continuous measurement of the pore pressure close to the excavated tunnel is used to determine the parameters of a poroelastic problem describing the relaxation of the EDZ. The Bayes inversion is used to obtain modified fields of hydraulic conductivity and porosity on the EDZ as the set of random samples. These can later be used as the input to the stochastic prediction of the safety indicator. The Bayes inversion is realized through the endorse-bayes script described below.

More details about the context, the used numerical models, and stochastic methods can be found in the methodology for characterization of the safety of EDZ. Currently only in Czech: “Posouzení vlivu zóny ovlivněné ražbou na bezpečnost hlubinného úložiště pomocí výpočetních metod”, link at the main repository page.