A. PSA framework of HTRs in USA, Germany and South Africa

In the probabilistic safety analysis report of the modular high temperature gas-cooled reactor (MHTGR), designed by GA, USA, Levels 1 and 2 PSA are integrated [12]. The release frequency and consequences are as the target of event tree analysis. Frequency of the release categories are analyzed quantitatively by the events sequence [12].

Germany proposed the 200 MW pebble bed modular HTR (HTR-MODULE) design, and its probabilistic risk assessment is done. The initiating event, event tree and release categories are the framework of PSA. The release categories frequency is quantified by the events sequence [13].

The pebble bed modular reactor (PBMR) is developed by South African. The PSA of PBMR used the event tree analysis to model the initiating events, plant response, and the reactor building response. And the development of accidents is modeled by PSA. The initiating events are divided into four categories, with 16 groups, and six release categories are defined as the final states [14].

B. PSA framework of HTR-PM in China

The pebble-bed modular high-temperature gas-cooled reactor (HTR-PM) is developed in China. The safety goal is that the cumulative frequency should be less than d-6 reactor^-1year^-1, for all beyond design basis accident sequences which lead to off-site (including the site boundary) individual effective dose beyond 50 mSv [15].

Figure 2 shows the framework of HTR-PM’s PSA analysis [9, 10]. The initiating event, event tree and the release categories are as the framework of HTR-PM’s PSA. It begins with the initiating event analysis, and ends with the release categories analysis. The event tree considers release scenarios, rather than the plant response and operator intervention. Figure 2 shows the framework in the dotted line and the support information outside the dotted line.

Compared with the framework of the pressurized water reactor (PWR, Fig. 1), the boundaries between Levels 1 and 2 PSA is cancelled, without CDF analysis, and the release categories are directly used as the final state of events sequence. This is because of the special characteristics of HTR-PM, which uses spherical fuel element, termed as tristructural-isotropic (TRISO) fuel, and low-pressure containment.

Fig. 1.Full-screen View

PSA framework of light water reactor.

TRISO is UO₂ particles coated with the silicon carbide layer, the buffer layer, and the inner and outer pyrolytic carbon layer. The coated particles ensure that there is a gap between the maximum core temperature and the temperature limit of the outermost layer in a reactor accident. So the core damage frequency is reduced greatly. Even it happens, the damage is inside the coated particles, which are independent of each other, without affecting other fuel particles. So, the radioactive release is limited greatly, hence no need of CDF calculation in PSA [10].

The modular design of HTR-PM is of low power TRISO fuel. In an reactor accident, the decay heat is transferred outside the pressure vessel by radiation and conduction. Even if the coolant is lost, the core damage is of low frequency and the early radiation release is of low probability. The system of negative pressure ventilation and filtration in the containment reduces the radiation emission to the environment. The helium coolant has excellent neutron properties and thermal stability, so the accident phenomena and processes are clearer.

C. PSA framework of sodium cooled fast reactor

From the safety review, the probabilistic safety goals for a fast reactor should be the same as the statement in the HAD 102/17 (The safety evaluation and verification of nuclear power plant). This means the CDF of less than d-5 reactor^-1year^-1, and the frequency of large radioactive release of less than d-6 reactor^-1year^-1 [16]. For the sodium-cooled fast reactor, the PSA method and framework are similar with PWR, but different greatly from HTR-PM.

Based on PSA analysis in China Experimental Fast Reactor (CEFR), framework of the sodium-cooled fast reactor will be discussed. In the CEFR, Level 1 PSA was done, but Levels 2 and 3 PSA are in the stage of planning. The framework in CEFR is almost the same with PWR, with main differences in detailed system analysis [17-19]. As shown in Fig. 3, for the sodium-cooled fast reactor, the PSA framework is divided into three levels, too. The analysis goals of each level PSA are the same as those of PWR, and analyzed the same as those in Table 1 [17, 18]. The core damage modes, which are closely related to designs of the reactor types, safety protection and accident mitigation system, can be divided into not meltdown (OK) and core damage (CD).

Fig. 3.Full-screen View

PSA framework of Sodium-Cooled Fast Reactor.

But there are some differences between the sodium-cooled fast reactor and PWR, such as the neutron energy and coolant, good thermal capacity of sodium and active chemical properties, etc. These make the details of PSA in sodium-cooled fast reactor different from PWR. For example, the initiating events, success criteria, event tree analysis, and fault tree analysis are quite different. Also, the sodium-cooled fast reactor is designed as non-pressure, with low pressure in its primary and secondary loops. Therefore, the pressure-bearing containment is replaced by the airtightness confinement. This is the difference in concrete analysis of PSA [17, 18].

A. Characteristics of FHR

The general characteristics of FHR are as follows. TRISO fuel pellets are used as fuel and the coolant is mixture of molten fluorine salts. It uses passive cooling safety systems are designed and supercritical water circulation system. Its maximum fuel temperature is about 1250 ℃ under normal operation conditions. In the primary loop, the coolant is 2LiF-BeF₂ molten salt mixture; while in the secondary loop, the coolant is FLiNaK molten salt mixture. The maximum pressure in the reactor vessel is less than 0.5 MPa. The TRISO in the core is randomly arranged and the average temperature of core is around 600 ℃. Other systems include the protective gas system, molten salt cleanup system and monitor system, etc. [23-27].

FHR has many outstanding features. Its particular coolant and fuel will have an impact on its PSA framework. In FHR the operating temperature of molten salt is far lower than the boiling point of 1430 ℃ for 2LiF-BeF2, which means a large gap as safety margin. In normal pressure, the fluoride salt is molten in high temperature, and the environment does not need to maintain a high pressure state. The TRISO fuel particles fail at about 1600 ℃, which is much higher than the boiling point of the coolant temperature and average temperature of core. So the probability of core meltdown due to a temperature rise is greatly reduced [25].

B. PSA framework of FHR

In PWR and sodium-cooled fast reactor, the fuel structural is fuel pellets and cladding composition. The damaged fuel cladding will result in radioactive leaks and fissile material. This will cause core damage. Therefore, Level 1 PSA of the two reactor types is to quantify the core damage and obtain the core damage frequency.

In FHR, the CDF caused by temperature rising up is reduced. The TRISO failure temperature of 1600 ℃ is much higher than the peak coolant temperature of 700 ℃, and higher than the boiling point of coolant (∼ 1400 ℃). So there is a large heat margins to extend the artificial non-intervention time and extend the accident processing time. From the physics, the TRISO damage is of such a lower frequency that it can be hardly happen.

The TRISO fuel particles are independent of each other. So damage of one TRISO fuel particle does not affect other fuel particles. Because of its limited volume, the release of radioactive fission products is limited. Therefore, similar to the HTR-PM, the large core damage and large radiation release of FHR is of lower frequency.

As an experimental reactor, the FHR’s power density is lower than a power plant reactor. The containment in a power plant reactor can be replaced by the confinement. The PSA analysis method will be revised with this.

Because of the low CDF, the PDS analysis in Level 2 PSA of PWR can be combined into the event tree quantification step in PSA of FHR, and Levels 1 and 2 PSA be combined in FHR for the source term analysis.

The PSA framework of FHR is similar with HTR. The boundary of Levels 1 and 2 PSA is not clearly in FHR PSA analysis. They can be combined. The beginning is the initiating events analysis and grouping, followed by the event tree analysis and quantification. The release category and source term are the end of this part PSA analysis. Also the results of source term analysis are input of Level 3 PSA. The framework and detail process is shown in Fig. 4. The analysis target and beginning of different level PSAs in FHR are shown in Table 2.

Fig. 4.Full-screen View

PSA framework of FHR.

The analysis and grouping of initiating events must follow the general standard, and this must be evaluated according to the reactor type and design. In the FHR’s PSA analysis, the event tree will not only consider a range of usually level 1 PSA, but also should expand to the serious accident and relief. The performance and response of the confinement should be also considered. The corresponding fault tree will also expand its model range which the event tree covered. Because there is similar physical state of coolant between the FHR and sodium cooled fast reactor, the initiating events and fault tree development could be referenced.