DECOVALEX-2019 Task C. GREET intermediate report

An experiment known as GREET (Groundwater REcovery Experiment in Tunnel) is being conducted at the Mizunami Underground Research Laboratory of the Japan Atomic Energy Agency (JAEA) to evaluate the environmental recovery process around underground galleries in fractured crystalline rock. The experiment has been planned to observe any environmental changes following water-filling in Closure Test Drift (CTD). The baseline hydro – mechanical – chemical (H-M-C) condition was identified prior to excavation of CTD. Then excavation of CTD and isolation by the water-tight plug, and subsequent flooding with groundwater were conducted. Environmental disturbance and recovery were observed for more than 3 years. DECOVALEX-2019 Task C aims to develop modelling and prediction methods using numerical simulation based on the water-filling experiment to examine the post drift-closure environment recovery processes. The Task consists of following three Steps; Step1: Modelling and prediction of environmental disturbance by CTD excavation Step2: Modelling and prediction of environmental recovery by CTD isolation Step3: Modelling and prediction of long-term environmental condition after CTD isolation In this intermediate report, the results of Step 1 are summarized from each of the research teams (JAEA, SNL, TUL). Groundwater inflow rates to the tunnel during the excavation, hydraulic drawdown, and variation of chlorine concentration at monitoring boreholes in the vicinity of the tunnel were chosen as comparison metrics for Step1. JAEA team constructed a reference model to check the viability of the original simulation code "COUPLYS", and then examined the sensitivity of the various parameters to the simulation results. The target parameters were estimated by using a DFN model and an ECPM model. SNL team developed a DFN model based on the fracture data (size distributions, orientation, volumetric intensity, the relationship between the aperture and permeability and the radius). Simulations of homogeneous and fractured system models were conducted using the coupled "DAKOTA-PFLOTRAN" codes. TUL team applied a multidimensional concept (discrete fractures + equivalent continuum) using the code "Flow123d" for modelling. Consequently, prior to tunnel excavation in fractured granite, it is likely to be possible to foresee the scales of inflow rate and hydraulic drawdown by current simulation techniques. On the other hand, the predictions of precise drawdown location and chemical variation have large uncertainties using the current models and associated process understanding. (author)