<< previous page

IRI/TUD Activities

IRI/TUD
Interfaculty Reactor Institute
Delft University of Technology
Mekelweg 15
NL-2629 JB Delft, Netherlands

contact persons: 
Dr. Ir. J.L. Kloosterman
Prof. Dr. Ir. A.H.M. Verkooijen

Tel. +31-15278-1191, Fax  +31-15278-6422
e-mail: j.l.kloosterman@iri.tudelft.nl

Who are we?

The Interfaculty Reactor Institute (IRI) is part of the Delft University of Technology, and is the Dutch national centre for education, training and scientific research on nuclear reactors and applications of ionising radiation. The fundamental research covers the fields of reactor physics, neutron physics, radiochemistry, radiation chemistry, nuclear detection methods, as well as condensed matter physics. The institute's facilities comprise a 2 MW swimming-pool nuclear reactor, a 3 MV pulsed electron accelerator, positron beams, and two-phase boiling loops. The institute employs some 220 persons.

What are we doing regarding accelerator driven systems?
Within the research project on accelerator driven systems (ADS), the emphasis is on the dynamics and thermal-hydraulics of ADS, on the fundamental aspects of sub-critical reactor physics, and on the measurement and calculation of the reactivity and kinetic parameters. The latter research is performed in collaboration with the other partners in the MUSE project.

The thermal-hydraulics phenomena of ADS are studied by the use of Computational Fluid Dynamics (CFD) models. Thermal stresses and their consequences for the integrity of the reactor vessel and its internals are being analyzed under normal operating conditions and under accident conditions. These stresses are expected during transients caused by a loss of power or by uncontrolled changes in the accelerator power. Where possible new improved concepts for the design are generated. Sub-critical reactor physics concentrates on new techniques for the calculation of higher modes in nuclear reactor systems, and the implementation of these techniques in existing and/or newly developed codes. Because each mode has its own dynamic characteristics, knowledge of the higher modes is important to assess the overall dynamics behaviour of ADS. Furthermore, knowledge of the higher modes is important to correctly interpret the experimental results.

Within the framework of the MUSE project, neutron noise experiment are performed at the Masurca Facility in Cadarache (F). The measurements include pulse-counting experiments applied to the reactor in both critical and sub-critical state, continuous current measurements applied to the critical reactor, and pulsed neutron source experiments applied to the sub-critical accelerator-driven configuration.

In the pulse counting experiments, depending on the detector efficiency, we either record the number of neutron detection in fixed time intervals with a width of a few microseconds, or the actual time of neutron detection. Both formats can be used in the Feynman-a, Rossi-a, and auto correlation and cross correlation analyses. The special-purpose data-acquisition software has been written in the LabView G-Language, while the data-analysis software has been coded both in the G-Language and the PV-Wave language.

In the continuous current measurements, the ionisation current from the detectors is converted and after various filtering and amplification steps sampled and stored on disk. Analysis software written in the PV-Wave language is used to calculate the auto power spectral density (APSD) and the cross power spectral density (CPSD).

A special version of MNCP containing digital signal processing routines (MCNP-DSP) is used to simulate and analyze the experiments. Unfortunately, the analogue Monte Carlo method applied by MCNP-DSP is not suitable for large reactor systems and low efficiency detectors, as is the case in MASURCA. Therefore research focuses on improvements. The first one is to unlink the actual Monte Carlo calculation from the digital processing routines. In this way, multiple experimental methods can be simulated from the data collected during only one Monte Carlo transport calculation, or, when the Monte Carlo calculations are made separately for each source, the experimental result for different combinations of sources can be generated. Furthermore, input parameters to the experimental method like sampling frequency and the measuring time, and the input parameters to the source like relative strength and the source pulse frequency can be varied in the post-processing phase instead of being fixed in advance of the Monte-Carlo calculation.

The second improvement is the introduction of non-analogue techniques to increase the efficiency of the Monte Carlo transport calculation. Although it seems trivial to apply standard available variance reduction techniques, it is not if one wants to preserve the higher-order moments needed to simulate the experiments like the variance in Feynman-a measurements. The new methods developed improve the efficiency of the Monte Carlo calculations to a huge extent (typically a factor of 100), which enables us to simulate pulse-counting experiments in large reactor systems.

to top

<< previous page