Determination of the subcriticality level using the ²⁵²Cf source-detector method

Abstract

Measurement and monitoring of reactivity in a subcritical state, e.g. during the loading of a power reactor, has a clear safety relevance. The methods currently available for the measurement of k_eff in stationary subcritical conditions should be improved as they refer to the critical state. This is also very important in the framework of ADS (accelerator driven systems) where the measurement of a subcritical level without knowledge of the critical state is looked for. An alternative way to achieve this is by mean of the ²⁵²Cf source-detector method. The method makes use of three detectors inserted in the reactor: two “ordinary” neutron detectors and one ²⁵²Cf source-detector which contains a small amount of ²⁵²Cf that introduces neutrons in the system through spontaneous fission. By observing fissions through the detection system and correlating the signals of the three detectors, the reactivity ρ (and hence the multiplication factor k) can be determined.

Before the actual measurements took place, a suitable data acquisition system was realized in order to process the signals and compute the auto and cross power spectral densities. The measurements were then performed in the VENUS reactor, using the ²⁵²Cf source-detector and two BF₃ neutron detectors. The multiplication factor was determined using the Cf source method and compared with measurements using other methods and with computational results (Monte Carlo simulations). The Cf method was benchmarked at a UOX core to other experimental methods that used the critical state as reference and to calculations. Afterwards, the Cf source technique was analyzed in a MOX core to study the possible impact of a significant intrinsic source on the results. This benchmarking gives the possibility to validate the Cf method as a reliable technique for the measurement of subcritical levels in steady state and for cores with an intrinsic source like MOX or burnt fuel cores.

Introduction

Measurement and monitoring of reactivity in a subcritical state, e.g. the loading of a power reactor, has a clear safety relevance. The methods currently available for this purpose are not satisfactory as some international research efforts (Baeten and Ait Abderrahim, 2003) highlight. The question of monitoring reactivity has received increased actuality in connection of a few recent incidents and accidents such as the criticality accident in Tokaimura in 1999 (NRC, 2000), the erroneous loading of the French PWR Dampierre-4 in 2002 (Laurioux and Deschamps, 2002), and lately the cooling accident in a fuel cleaning facility at Paks-2, Hungary (Hungarian Atomic Energy Agency, 2003). Furthermore, the experimental reactor physics community is faced with the challenge to develop measurement techniques for the accurate determination of the subcritical level (Baeten and Ait Abderrahim, 2003). This need is evidently present in the current research for accelerator driven systems (Aït Abderrahim and D’hondt, 2007) where subcriticality has to be guaranteed at all times by an accurate determination of the reactivity level. However, also a strong interest and need is being expressed by the utilities to have a more accurate reactivity determination during loading procedures which is an essential point for criticality safety.

Most of the traditional experimental methods of measuring subcritical levels need the critical state of the reactor as a reference. Sometimes this is difficult or impossible, as in the case of subcritical core driven by accelerator (ADS). Other techniques, like the reactivity meter (Grachev et al., 1986) make use of reactivity changes and cannot be used in stationary subcritical conditions for the estimation of the subcritical level. Hence we need a method for subcritical level measurements in steady state conditions without knowledge about the critical state.

Nowadays, most of the methods to measure reactivity in a stationary subcritical system with a source and without knowledge about the critical state are based on the statistics of the subcritical chains induced by the intrinsic source neutrons. Such so-called noise methods are the variance-to-mean or Feynman-alpha (Pázsit and Pál, 2007), the Rossi-alpha method (Pázsit and PÁL, 2007, Uhrig and Stapleton, 1975, THIE, 1963, Mihalczo et al., 1990), and the auto-power spectral density (APSD) and cross power spectral density (CPSD) (Uhrig and Stapleton, 1975, THIE, 1963, Mihalczo et al., 1990). These methods have been tested and used extensively at research and zero power reactors (Uhrig and Stapleton, 1975, THIE, 1963, Mihalczo et al., 1990, Pyeon et al., 2008). None of them have been applied extensively at power reactors for subcriticality determination. The problem of all these noise methods is the low signal-to-noise ratio due to the poor correlation within the overall neutron noise signal. In power reactors, e.g. during loading operations, one relies on the subcritical approach procedure to estimate the critical condition without having an experimental value for the subcritical level at each state in the procedure.

To obtain possibly a higher signal-to-noise ratio in the neutron noise signal, another method was proposed some time ago, which is based on the so-called ²⁵²Cf method (Mihalczo et al., 1990). The essence is to use a ²⁵²Cf neutron source, which is built together with an ionization chamber (the ²⁵²Cf is put electrochemically on one of the plates of a parallel-plate ionization chamber). Such an arrangement is called a “²⁵²Cf detector”. The detector identifies each spontaneous fission event leading to neutron emission by measuring the ionized fission products, but without absorbing any neutron. In this way, the time origin of the fission is known and correlated neutrons can be easily discriminated and hence an increased signal-to-noise ratio can be obtained.

The ²⁵²Cf source-detector method for subcriticality measurements is a source-driven noise analysis method for the measurement of the reactivity of subcritical reactor systems. In subcritical systems, fission chains are created whenever a neutron is brought into the system, either by an external source or by spontaneous fission in the system itself. Because of the subcritical nature of the system, such a fission chain will eventually die out. The birth and death of a fission chain are stochastic processes and, consequently, the fission chain multiplication process is subjected to fluctuations (“noise”). However, the average behavior of these fission chains is determined by the system’s geometry and composition, and through observation of the fission chains, certain physical properties of the subcritical system (e.g. the system’s multiplication factor) can be deduced.

The ²⁵²Cf source-detector method provides a way to observe these fission chains and derive from it physical properties of the system. This source-detector, which contains a small amount of ²⁵²Cf, is placed inside or in the vicinity of the system, and through spontaneous fission of these ²⁵²Cf-nuclei, neutrons are injected into the system. The time of injection of these source neutrons can be observed through the detection of the fission fragments by the ²⁵²Cf source-detector. In addition, two common neutron detectors are placed into the subcritical system and, by correlating the detection events of these neutron detectors with the source event, dynamic properties of the system can be deduced.

In this respect, the proposed ²⁵²Cf source method offers advantages over other experimental techniques due to its active interrogation character and its technological simplicity. One of the advantages is that the spectral ratio obtained from the noise measurement does not depend on the detection efficiency of the common neutron detectors. Another distinct advantage is that the prompt neutron decay constant can also be obtained from this measurement and, finally, the technique itself has a high sensitivity to small changes in configurations of fissile materials. Hence, the investigation of this technique in this application area could contribute largely to the safety of current and future reactor systems.

The formulas that give the reactivity of the system in function of these correlations will be briefly discussed in Section 2. Section 3 is devoted to the description of the experimental setup – especially the ²⁵²Cf source-detector – as well as the subcritical system on which the reactivity measurements were performed. The data acquisition procedure and the processing of the measured signals, with its specific challenges, will be addressed in Section 4. Section 5 shows the results of the experiment as well as a comparison of the ²⁵²Cf source technique with other different reactivity monitoring methods. Finally in Section 6 the main conclusions of this paper are summarized.