In the design of novel nuclear reactors active systems are replaced by passive ones in order to reduce the risk of failure. For that reason natural circulation is being considered as the primary cooling mechanism in next generation nuclear reactor designs such as the natural circulation boiling water reactor (BWR). In such a reactor, however, the flow is not a controlled parameter but is dependent on the power. As a result, the dynamical behavior significantly differs from that in conventional forced circulation BWRs. For that reason, predicting the stability characteristics of these reactors has to be carefully studied. In this work, a number of open issues are investigated regarding the stability of natural circulation BWRs (e.g. margins to instabilities at rated conditions, interaction between the thermal-hydraulics and the neutronics, and the occurrence of flashing induced instabilities) with a strong emphasis on experimental evidence. The prototypical Economical Simplified BWR (ESBWR) design from the General Electric Company was thereby taken as the reference natural circulation BWR. Two experimental facilities located at the Delft University of Technology were used for that purpose: the GENESIS facility which uses Freon as working fluid and the water-based CIRCUS facility.
First of all, the stability of the ESBWR under nominal conditions was studied. The purpose was to experimentally determine the ESBWR stability characteristics as accurately as possible and to compare the results with numerical results of different origin. In order to study the ESBWR stability under less severe conditions than the nominal ones, a downscaled facility, called GENESIS, was designed and constructed based on a fluid-to-fluid scaling approach. Since the rods in the facility are electrically heated, an artificial void reactivity feedback mechanism was implemented.
Both the thermal-hydraulic and the neutronic-thermal-hydraulic stability performance of the ESBWR for a wide range of conditions were investigated. From the analysis of the results it was found that the GENESIS facility (representing the ESBWR at nominal conditions) is very stable and exhibits a large margin to instability. From the comparison between numerical simulations performed by using the system codes TRACG and ATHLET, a significant discrepancy was observed in the predicted decay ratio at nominal conditions. This finding indicates that limitations still exist in the numerical estimation of the stability performance of nuclear reactors involving complex two-phase flows. For this reason both numerical and experimental tools should be used for such a task.
The GENESIS facility was also used to perform a study in which a number of parameters such as the steam separator friction (the steam separator in the ESBWR is located at the top of the chimney section), the void-reactivity feedback coefficient and the axial position of the feedwater sparger inlet were varied. As a result, it was observed that the characteristic resonance frequency of the thermal-hydraulic mode is found to be much lower (~0.11 Hz) than in forced-circulation BWRs (~1 Hz), indicating a static head dominated phenomenon since it corresponds well with typical frequencies of density wave oscillations traveling through the core-plus-chimney sections. In addition, it was experimentally found that the position of the feedwater sparger inlet influences the stability of the thermal-hydraulic oscillatory mode.
Some experiments and analyses indicated that thermal-hydraulic oscillations may occur under certain low pressure and power conditions during the startup of a natural circulation BWR. For that reason the stability of the ESBWR at start-up conditions was investigated by using both numerical and experimental tools.
In the ESBWR, the chimney is split up into many parallel channels; hence, coupling effects between the channels are of relevance. Two different cases were therefore studied with the help of the CIRCUS facility: the single channel configuration and the two parallel channels configuration for which study CIRCUS was especially modified.
For the single channel configuration, detailed experiments were performed on flashing-induced oscillations. Both the axial temperature and axial void fraction temporal evolution were used to investigate the mechanisms causing the flashing-induced oscillations. A novel way of plotting the axial temperatures in the channel was thereby used. The experimental results were also used to validate a proposed lumped parameter model. Then the code was used to further investigate the influence of the core inlet friction and chimney exit restriction on the stability. It was found that increasing the core inlet restriction stabilizes the system at high subcooling values and destabilizes the system at low subcooling values. In addition, the model predicted a destabilizing effect of the chimney exit restriction coefficient at low subcooling values and no changes in the high subcooling range.
From experiments performed with the two parallel channel configuration, it was found that reverse flow plays an important role in the spatial and temporal evolution of the temperature and vapor production in both parallel channels. Four different behaviors were found, depending on the operational conditions: (i) high subcooling, stable flow, (ii) in-phase oscillations, (iii) a-periodical oscillations which are attributed to multi-fractal deterministic chaos, and (iv) out-of-phase oscillations.
The results from the one and two parallel channel configurations also showed that vapor can be produced while the reactor remains stable. In this way, the reactor can be pressurized without encountering instabilities.
The work presented in this thesis shows that a natural circulation BWR can be safely operated from start-up to nominal conditions. Despite the discrepancies between the experimental and numerical results, it was shown that the ESBWR has large margins to instability at rated conditions. These discrepancies also emphasize the importance of using both (validated) numerical and experimental tools during the design phase of future nuclear reactors. This work indeed shows that the results from these complementary tools facilitate the comprehension of the system dynamics.