Study of TRICO II Reactor Startup and Shutdown Operations Using the OpenMC Calculation Code

The Commissariat Général à l'Energie Atomique (CGEA) in the Democratic Republic of the Congo has a TRIGA Mark II research reactor, which went critical for the first time in March 1972. This 1 MWth reactor, known as TRICO II (TRIGA Mark II in Congo), has advanced research and development in nuclear science, as well as its peaceful applications in industry and agriculture across the country. Designed by the American firm General Atomics for research, training, and radioisotope production, this type of reactor is renowned for its intrinsic properties, which make it easy to control, and offers perfect safety conditions by design, independently of conventional mechanical and electronic devices.

As is the case for all nuclear reactors worldwide, safety is an utmost priority. Indeed, nuclear safety is an essential concept in reactor operation. The parameters required to achieve the reactor safety must be defined, analyzed, and monitored in a controlled manner, in order to minimize the risk of failure or accident during reactor operation [1]. The aim of control system is to manage the reactor operation, in particular through power adjustments (increasing or reducing power), complete shutdown or restart of the reactor. These approaches are different: they no longer involve progressive control to compensate for natural fuel changes, but rather rapid intervention to modify the neutron balance in the core. This is achieved by introducing neutron-absorbing elements such as boron carbide or indium–cadmium alloys [2,3]. These elements are inserted into the core in such a way as to effectively interrupt neutron reactions within the assemblies. These antireactivity devices, which are inserted into or removed from the core, as required, control the reactivity margin needed to operate the reactor at different power levels.

where $ϕ(r,t)$ is the neutron flux at point r and time t (neutrons cm⁻² s⁻¹), $υ$ gives the neutron velocity (cm s⁻¹), $D(r)$ represents diffusion coefficient (cm), $∑a(r)$ stands for macroscopic effective absorption cross section (cm⁻¹), and $S(r,t)$ designates the neutron source term (neutrons cm⁻³ s⁻¹).

with $∑f(r)$ the macroscopic fission cross section (cm⁻¹), and $v$ represents the average number of neutrons produced per fission.

Reactor startup and shutdown are among the main operations affecting nuclear reactor safety. To handle a nuclear reactor correctly, reactor operators must first draw up a checklist of handling operations, then read each of these procedures carefully. Before taking any action, neutron engineers must present the general condition of the reactor core. The information required are as follows [6]: (i) neutron flux behavior during startup, (ii) neutron flux behavior during shutdown, (iii) core coolant cooling efficiency, (iv) void reactivity coefficient behavior, and (v) neutron flux range density behavior in each position inside the core (in-core positions) or outside the core (out-of-core positions, etc.).

Reactor startup and shutdown are basic operations that reactor operators and neutronics engineers have to guarantee at an earlier stage, to ensure smooth plant operation. This can be achieved by trial-and-error methods or by using Monte Carlo-based neutronics calculation codes, for example.

In the present study, using Monte Carlo methods, models are developed to simulate various possible scenarios during startup and shutdown of the TRICO II reactor. These scenarios may include variations in key parameters, potential failures of control systems, or emergency situations. Generally based on random samples, simulations enable the probability of occurrence of different events to be assessed, and the risks associated with reactor startup and shutdown quantified. This enables informed decisions to be taken on the procedures to be followed, risks to be minimized.

Each control rod plays a specific role in the reactor core. They not only start and stop the reactor but also increase or decrease its power and core reactivity [7]. The present focuses on the first two above-mentioned roles of control rods. The aim is to simulate the variation in criticality during startup and shutdown of the CGEA TRICO II reactor, using the OpenMC code. To this end, a number of specific objectives have been set, including: (i) complete 3D modeling of the CGEA TRICO II reactor; (ii) validation of the reactor 3D model; and (iii) simulation of reactivity variation during reactor startup and shutdown.

OpenMC was chosen not only because it is free and open source but also because of its flexibility for complex calculations and simulations. It has been designed using a modern software engineering approach. OpenMC is a simulation code that uses the Monte Carlo method to follow the history of neutrons in a circumscribed environment of solid geometry [8].

The TRICO II reactor has a cylindrical core 44.18 cm in diameter, bounded by two aluminum alloy upper and lower grids 64.8 cm apart and 1.905 cm thick. It is surrounded by a graphite reflector ring (lateral reflector), 30 cm thick and 55.88 cm high. This reflector ring is protected by a 0.64 cm-thick layer of aluminum. The TRICO II core is located at the bottom of a 198 cm-diameter, 7.5 m-high vessel, entirely filled with demineralized light water (H₂O). It comprises of:

• 70 fuel elements, including 66 standard fuel elements, three control rods with fuel follower, and one thermocouple fuel element, all enriched to an average of 20% uranium-235 (U-ZrH_1,6). The active part of the fuel elements measures 38.1 cm in length and 3.33 cm in diameter.

• 15 graphite reflector elements, 55.652 cm long and 3.556 cm in diameter,

• one pulse bar,

• one neutron source (Ra–Be), and

• two fixed in-core irradiation positions, with aluminum cladding.

The three fueled follower control rods are 114.3 cm long. The active part is made of 20% enriched uranium-235, and the absorber part is made of boron carbide (B₄C). The pulsation rod has the same B₄C composition as the other three control rods [9].

The complete model was visualized using the OpenMC plotter, as shown in Figs. 1 and 2.

The effective neutron multiplication factor (k_eff) was computed by running a total of 500 cycles with 20,000 neutrons per input, using the endf/b-viii.0 library.

To simulate reactor startup and shutdown operations, the position of the control rods (transient rod, regulating rod, shim rod, and safety rod) was varied relative to the positive z-axis for startup and negative z-axis for shutdown, with the origin of the axes being the plane separating the boron carbide (B₄C) and fuel (U-ZrH_1.6) parts of a control rod, as shown in Fig. 3. For all scenarios [10], a variation in steps of 4 cm over a maximum length of 38 cm, corresponding to the length of the active part of each fueled follower control rod, has been taken into account.

For the startup case, all transient rod positions were assumed to be on the outside (at the upper end of the stroke) to facilitate flexible startup, while establishing a standard startup operation that will allow comparison with other operations. The following startup scenarios were considered:

• Standard (normal startup operation): In this startup operation, the “shim rod” was varied while keeping the regulating rod and safety rod fixed with a portion of B₄C inside the reactor core. This was the method used for startup when the TRICO II reactor was in service.

• RO1 (startup operation 1): Here, the regulating and shim rods vary with part of B₄C in the direction of the positive z-axis, but the safety rod remains fixed, i.e., its part B₄C is in the reactor core.

• RO2 (startup operation 2): The safety rod varies with part B₄C in the direction of the positive z-axis, but the shim rod and regulating rod remain fixed with part B₄C in the reactor core.

• RO3 (startup operation 3): All control rods (safety rod, regulating rod, and shim rod) vary simultaneously with part B₄C toward positive z.

• RO4 (startup operation 4): The safety rod and shim rod vary with part B₄C toward the positive z-axis, but the regulating rod remains fixed with part B₄C in the core.

• RO5 (startup operation 5): The regulating rod varies with part B₄C toward the positive z-axis, but the shim rod and safety rod remain fixed with part B₄C in the reactor core.

For reactor shutdown operations, with the exception of the standard case, all other cases were simulated with the transient rod fully inserted in the reactor core. The following scenarios were considered:

• Standard (normal shutdown operation): In this shutdown operation, all control rods (transient rod, regulating rod, shim rod, and safety rod) were assumed to be inserted simultaneously into the core, i.e., with the absorbing part of B₄C descending toward the negative z-axis. This is the standard method used for reactor shutdown operations, which involves inserting the transient rod into the reactor core. It is also used for SCRAM emergency shutdowns.

• SO1 (stopping operation 1): In this stop operation, it is assumed that the regulating rod and the safety rod vary with part B₄C toward negative z, but the shim rod remains fixed with part B₄C in the core.

• SO2 (stop operation 2): Here, the regulating rod and shim rod vary with part B₄C toward the negative z-axis, but the safety rod remains fixed with part B₄C in the core.

• SO3 (stop operation 3): The regulating rod varies with part B₄C toward the negative z-axis, but the shim rod and safety rod remain fixed with part B₄C in the core.

• SO4 (stop operation 4): The shim rod varies with part B₄C toward the negative z-axis, but the regulating rod and safety rod remain fixed with part B₄C in the core.

• SO5 (stop operation 5): The safety rod and shim rod vary with part B₄C toward negative z-axis, but the regulating rod remains fixed with part B₄C in the core.

• SO6 (stop operation 6): The safety rod varies with part B₄C toward negative z, but the shim rod and regulating rod remain fixed with part B₄C in the core.

This methodology is summarized in the flowchart in Fig. 4.

It is less than, equal to, or greater than 1, respectively, when the system is subcritical, critical, or supercritical [11].

where V represents the volume of reactor core [3].

Validation of any theoretical code requires a comparative analysis test with the actual model, in order to confer credibility on the results produced by the theoretical model. Table 1 and Fig. 5 highlight the comparison of the control bar reactivity results produced by the OpenMC code with those obtained experimentally.

As shown in Table 1, the theoretical model shows acceptable deviations from the real model. These deviations can be attributed to approximations made to the geometry of the model studied, to the average consideration of core temperature (considered on average at 298 K) or to other parameters. The difference between the results produced by OpenMC and those of the experiment is small, of the order of 0.53. This difference is acceptable. One can therefore say that the model is satisfactory to further calculations.

Figure 5 shows that at the regulating rod level, the reactivity of OpenMC is almost equal to that of the experiment. High reactivity is observed at the safety rod level, which would be related to its location in the core, where it is surrounded by fuel rods, creating a high neutron flux there. But at the other control rod locations, reactivity is slightly low, as the graphite reflector rod belt reflects neutrons in the opposite direction of the configuration. As shown in Fig. 1, OpenMC results are also close to the experimental results for the control rods of the TRIGA II reactor in Bangladesh [12].

In the TRICO II research reactor, the three control rods modify the K_eff and, consequently, the system's reactivity. Figures 6 and 7 show the direct influence that upward and downward movements of the four control rods can have on the reactivity of the reactor core.

Figure 6 shows all possible startup operations RO1, RO2, RO3, RO4, and RO5. It can be seen that operations RO1, RO3, and RO4 are highly reactive, probably because the quantity of neutrons consumed is small compared with those produced by fission or neutron emission decay. This is to be expected, given the position of the control rods in the reactor core.

This observation can also be explained by the fact that the shim rod varies during the three startup operations (RO1, RO3, and RO4) in the TRICO II core of the model developed. The shim rod does not sufficiently absorb neutrons produced in the core, and its position in the core is eccentric to the left of center (Fig. 1), surrounded by highly enriched fuel elements, which also means it is in a region of high neutron flux. On the other hand, the startup operation of the RO5 reactor shows low reactivity. This is normal, as reactivity increases progressively as the three control rods are lifted. But the RO5 is not as good as the standard either, as the reactor cannot be started by lifting the regulating rod alone, which moderates core reactivity.

To guarantee safe reactor startup, the boron carbide (B₄C) part of the shim rod must be in the core, but the regulating and safety rods remain fixed with the boron carbide part in the core, and the transient rod must be kept out of the core; this is the standard case of startup operation, offering greater safety during startup operation. However, RO1, RO3, and RO4 startup operations can also be used and are just as reliable.

Figure 7 presents SO1 to SO6 shutdown operations. It can be seen that the SO2 and SO5 shutdown operations give acceptable results compared with standard operation, meaning that the SO2 and SO5 give better reactivity prediction and a stable core, although reactivity remains high compared with the standard operation.

The results on the reactor shutdown operation are almost identical, which could be due to the fact that the poison part of all control rods is composed of B₄C, which has a large effective cross section with thermal neutrons. This means that by inserting all the entire poison (B₄C) part of the control rods, one should observe a decrease in thermal neutrons in the core, and therefore a decrease in the ²³⁵U fission chain reaction. Similarly, to ensure a good method of reactor shutdown, it is preferable to use the standard operation, since when all three control rods are fully inserted, reactivity is much lower and it does not take much longer to effectively extinguish core activity and absorb the maximum number of thermal neutrons. This is why standard operation is used as an emergency shutdown method, known as SCRAM.

The safety simulation of the startup and shutdown processes of the TRICO II research reactor was studied using the OpenMC calculation code. The aim was to propose startup and shutdown operations of the TRICO II reactor. After computer simulation of various startup and shutdown operations on the computed model of the TRICO II reactor, it was found that startup operations RO1, RO3, and RO4 gave good results. Among various reactor shutdown operations, the standard operation was found to give an appreciable and better result. This is an effective and safe approach to planning a better reactor startup and shutdown operation. These results may vary from one nuclear reactor to another, given that every nuclear reactor in the world has its own specificity in terms of design or geometric characteristics. This approach will also avoid wasting fuel, electricity, and financial resources on trial-and-error testing. For the future, it would be desirable to determine the ratio of fissile material (U-ZrH_1.6) and moderator (B₄C) in the current reactor core configuration, in order to obtain its startup and shutdown behaviors, and then repeat it on various possible critical configurations.

The authors express their thanks to the Commissariat Général à l'Energie Atomique/Centre Régional de l'Etude Nucléaire de Kinshasa (CGEA/CREN-K) for its support.

• Commissariat Général à l'Energie Atomique/Centre Régional de l'Etude Nucléaire de Kinshasa (CGEA/CREN-K).

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.