1. Engineering feasibility of the FEBEX concept

• The design, fabrication, handling and installation of the various components of the in situ test accomplished a major part of the initial objective of the experiment.
  
• The disposal drift diameter of 2.40 m is quite restrictive. A drift of this diameter is considered to be viable, but will require a very refined design for the handling and transportation equipment.
  
• The influence of relative humidity on the mechanical integrity of the bentonite blocks is important. Adequate control of this humidity is necessary at the storage sites and during the handling and installation of the blocks.
  
• The presence of a film of water on the walls of the drift does not, however, produce any problems from the point of view of installation, as long as this progresses at an adequate rate.
  
• A critical factor during insertion of the canisters is the alignment of the steel liner with respect to the axis of the drift. The steel liner must be aligned in relation to the rails and the insertion equipment.
  
• The design and realization of the concrete plugs require special attention of the form of over-excavation for the key and the concreting procedure.

2. Performance of instrumentation and technical equipment

• The performance of the heating system has so far been excellent, including the remote control system.
  
• The clay barrier and rock instrumentation systems have also shown a high degree of reliability, (higher in fact than was originally expected).
  
• The number of sensor failures after three years of operation has been in the order of 10 %.
  
• However, this good performance does not allow their final duration to be predicted. It was thus decided to extend the experiment duration.

3. Quality assurance

The QA program applied to all research procedures. However, due to the special requirement of flexibility, the quality assurance program must include mechanisms that allow any discoveries and innovations that occur during the procedures to be accommodated.

The numerical model used is CODE-BRIGHT and it was confirmed that the model can reproduce, with reasonable accuracy, the results of the measurements performed during the large-scale tests. Although complete validation is not possible in practice, this check of the model increases the degree of confidence in its capacity to predict the near-field THM evolution of a repository.

Two THG codes, CORE-LE and FADES-CORE-LE, have been developed and verified.

From the application of the codes it can be concluded that they reproduce fairly well the observed patterns of geochemical behavior of a large number of laboratory tests, thus generating confidence in their predictive capability.