The analysis of flow through fractured rock is a central problem of "Enhanced
Geothermal Systems", EGS. Most data are derived from field experiments. There are
only a limited number of laboratory experiments available and even less on a
well-observable in-situ scale. The latter one require rather low rate injection rates
(<0.1 l/s, depending on transmissivity) for observing the flow behaviour of a
fractured system whereas high rate injections (>>1 l/s) are used for testing the rock
behaviour and fracture propagation under field conditions. The present study focuses
on an experiment performed by the Swiss National Organisation for Nuclear Waste
Disposal (NAGRA) in the Grimsel Rock Laboratory. At the BK site of this tunneling
system an in-situ salt and heat tracer (SHT) experiment VE520 was performed from
February 22 to March 24, 1993. The primary goal of NAGRA was to extract all possible
information on transport properties by evaluating the significance and
characteristics of transport parameters. The experiment was performed in a steady
state dipole flow field between the two boreholes (BK05 and BK15).
The advantages of the flow experiment in our actual study were: 
*	A good observation of an in-situ experiment was enabled by the design as a
mid-scale experiment (~10 m) at the BK site investigated by a total of 12 boreholes
that could be used for observation.
*	Heat is as a non-reactive tracer which does not disturb the chemical equilibrium of
the aqueous phase in the fracture network
*	The two independent salt and heat datasets were collected by one experiment in the
same spatial domain at very reasonable costs
The finite element code FRACTure was used for the interpretation of combined solute
and heat transport processes. Different model assumptions from single jointed to
multiple heterogeneous conditions were assumed. The results indicate strongly
heterogeneous transport properties with dispersion lengths in the same order of
magnitude like the dipole field. The thermal match the VE520 dataset requires several
partly independent flow paths, each with much larger surface than individual single
fracture models. This provides clues that the heterogeneous flow in a fractured
medium slows down a thermal front. Under realistic field conditions in a fractured
medium a thermal breakthrough can be much slower and its occurrence is less evident
in EGS type reservoirs than idealized model might predict.

14473 Potsdam, Telegrafenberg
Germany