Upper plenum mixing

This work was carried out in the frame of the project "Coolant mixing in pressurized water reactors" (registration number: 1501216) funded from 01.04.1998 to 31.05.2002 by the German Federal Ministry of Economics and Labour.


The reactor core of light water reactors consists of fuel elements with different enrichment and exposure. This circumstance is responsible for differences in the heat release of the single fuel elements. As a consequence, the fuel element outlet temperatures differ, too. Differences of more than 30 K are reached in the stationary state at full power. It is known, that these differences in the outlet temperatures do not fully vanish during the travelling of the coolant through the upper plenum. Thus, a certain temperature profile is still existing over the cross section of the hot leg nozzle.

Methodology of the ROCOM experiments

The core simulator in the ROCOM test facility consists of 193 pipes (one for each fuel assembly) with an inner diameter of 30 mm. The pipes are a direct connection between core inlet and outlet. This construction allows to carry out experiments, where the distribution of the coolant in the upper plenum from one certain fuel element could be measured. An injection lance can be inserted into a certain fuel assembly of the core through a penetration in the head. The lower end of the injection lance with the outlet for the tracer is located about 20 cm above the core support plate. At running pumps, salted tracer solution is injected into the fuel assembly. During the moving-up, the tracer is completely mixed with the surrounding coolant in the pipe. The mixture goes into the upper plenum and leaves the vessel through the outlet nozzles. In all four outlet nozzles, wire-mesh sensors measure the time dependent conductivity of the coolant. Each wire-mesh sensor has a grid of 216 measurement points evenly distributed over the cross section of the nozzle. These measured conductivity values can be transferred into a dimensionless mixing scalar, representing the share of the coolant from the pipe with injection at the given position, relating the local conductivity measured by the sensors to the conductivity at the outlet of the pipe.

The goal of the presented here experiments was the derivation of stationary mixing coefficients, representing the time-averaged mixing scalars on a stationary concentration level at each grid point in the cross section of the outlet nozzles. For that purpose, tracer has been injected into the fuel assembly during quasi-stationary flow conditions nominal state until a quasi-stationary concentration profile has been established.

Partly, the experiments are of generic character, because the plain vessel head of the ROCOM test facility does not correspond to the original and no internals are installed in the upper plenum of the test facility. The construction of the outlet of the fuel assemblies also deviates from the original.


Time-dependent tracer distribution to the four outlet nozzles

Experiments were carried out with tracer injection into each fuel assembly of a quarter of the core. Two of them are presented here as avi-Files. The first one shows the results for tracer injection into fuel assembly A08, the second one into fuel assembly G15, marked by a red square in the corresponding movie. It is clearly to be seen, that the tracer goes preferably to the two nearest outlet nozzles. Further, high-frequent fluctuations of the tracer over the whole cross section of the affected nozzles can be observed. These fluctuations were present in all experiments, independently from the position of the tracer injection. Altough the tracer is fluctuating in the whole cross section of the outlet nozzle, it can be seen from the time-dependent visualisation that there are regions where the tracer is registered more often.

Fig. 1: Injection into fuel assembly A08 (avi-file:1.9Mb)

Fig. 2: Injection into fuel assembly G15 (avi-file:1.6Mb)

Plateau-averaging of the tracer concentration

For the quantification of the experimental results, all measured instantaneous values of the tracer concentration at the quasi-stationary plateau, what establishes shortly after start of the injection are averaged. Fig. 3 shows the results of this averaging for three typical experiments. In the left part, the results for fuel assembly A08, located at the symmetry line between outlet nozzle 1 and 4, are shown. The tracer is distributed nearly symmetrically between these two nozzles. The sensors in the outlet nozzles on the opposite side of the vessel did not register any tracer at all. The maximum of the tracer concentration has been found in the lower part of the outlet nozzle. The results for fuel assembly F11 in the middle of fig. 3, show a complete different behaviour. Nearly the whole tracer passes the measuring plane in outlet nozzle 1, only small amounts are registered in outlet nozzle 2. Contrary to the fuel assembly A08, the maximum tracer concentration is found in the upper part of the outlet nozzle cross section. The experimental results for fuel assembly G15, shown on the right part of the fig., confirm the conclusions for the other two assemblies. The tracer distributes between the two nearest outlet nozzles, only a small part is going to the opposite side of the reactor. The maximum tracer concentration is again found in the lower part of the nozzle cross section.

Fig. 3: Plateau-averaged distribution of the normalized tracer concentration for injection into three different fuel assemblies (marked in the core pattern; the numbers at the colorscale show the minimum and maximum values in the respective experiment in %)



S. Kliem