Initial risk assessment of an in-vessel leakage incident at ITER

Initial risk assessment of an in-vessel leakage incident at ITER

Leakage accident inside the ITER ship

In-vessel leakage events are defined in ITER as leakages from a single tube break to multiple breaks occurred in the FW cooling tubes within the VV during plasma burning.

For the worst case of in-vessel cooling tube breakage (multiple FW tube breakage), which is considered highly unlikely, the resulting water ingress into the VV results in significant VV pressure. The VVPSS was started and the overpressure was quickly suppressed by spraying cooling water. When entering the cooling water, part of the tritium, corrosion products and dust will enter the liquid and remain confined in the liquid pool in the drain tank and funnel tank, while the other part will remain in the tokamak building if the negative pressure remains. For smaller, more likely failures of 1 to 10 FW refrigerant tubes, the VV is compressed by water/steam injected through the break similar to the break of multiple FW tubes. In this scenario, most of the aerosol is trapped in the fluid within the funnel system and drain tank or deposited on the VV surfaces. For loss of vacuum events, operation of the funnel tank venting system (ST-VS) provides negative pressure in the VV and port cell and prevents uncontrolled leaks into the gallery. The masses of tritium and dust transferred to the dial cell are very small. Releases are dominated by ST-VS, which is connected to the permanently operated natural erosion system (N-DS), and the N-DS releases of the port cell are very small.

The terminology of mobilized radioactive source in ITER refers primarily to tritium, activated corrosion products (ACPs) present in the water coolant and tritium and activated dust within the plasma. Those radioactive stockpiles are in those four areas: the tokamak building, the tritium building, the hot cell building and the radioactive waste facility. This paper mainly analyzes the potential radiation release during normal operation and maintenance. Therefore, the scope of the main analysis was limited to the first three areas.

The maximum stocks of radioactive materials in each region are shown in Table 1, and the potential path of release of radioactive materials is shown in Figure 2.

Table 1: Source term and environmental versions4.
Figure 2

Possible radioactive release pathways in a ship leakage accident.

Evaluation of the radiation dose of a leakage accident inside a vessel

ITER Incident Analysis Report4 States that the radiation source terms involved are HTO, corrosion products in the coolant, tritium and activated dust within the plasma chamber4. More than 90% of the FW coolant stock inside the vessel (335 tons for the base case and about 450 tons in the case of an additional failure of 10 DV coolant tubes) was released into the vacuum vessel. The maximum tritium concentration for first wall/blanket (FW/BLK) and DV/LIM cooler is 0.005 g T/m3 As HTO. Thus, about 1.3 grams of T were filled with this coolant for the basic case and about 1.9 grams for the case with an additional coolant pipe failure of 10 DV. The corrosion aerosol mass was packaged as a 2 mm aerosol in the VV atmosphere. Inside the plasma chamber, the amount of mobilized tritium was assumed to be 100% of the 120 g T from the cryopumps, and 100% of the 880 g T in the co-deposited layer of the plasma facing components (PFC). The cooling pump tritium and 50% of the co-deposited tritium layer are supposed to be mobilized immediately after the FW tube breaks and the remaining 440 g of co-deposited tritium over the next 6 hours. Total packed tritium stock is 1000 grams. It is assumed that the high radiation fields inside the vacuum vessel and the presence of steam cause rapid oxidation of the tritium such that this tritium is mobilized as HTO. The activated dust inventory was 100% of the 5 kg assumed to have evaporated during the plasma outage (5 kg of 0.1 mm diameter particles), and 100% of the 1000 kg of tokamak dust accumulated over time (1000 kg with a mass average particle diameter of 2.11 mm and a geometric standard deviation 2.0).

Figure 3 shows the masses of radioactive material (in collectors and deposited), the mass of airborne (packed) dust, corrosion products, and tritium in VV, ST, and DT for two cases of in-vessel multiple breakage: involving only 3 FW/BLK cooling loops with additional failure of 10 DV cooling tubes. Table 1 shows the maximum stock of radioactive material in the VV, in the suppression tank (ST) and in the drain tank for two cases as well.

Figure 3
Figure 3

Distribution of dust, ACP, and tritium within VV, ST, and DT in this incident4.

Evaluate the potential risks of a leakage accident inside the ship

The set of in-vessel leakage events is defined as leakages from a single tube break to multiple breaks occurring in the FW cooling tubes within the VV during plasma burning are examined. In this study, the 10 mm cooling door that referred to this incident refers to the water cooling tube inlaid on the back panel of the first wall. The size of the ITER blanket is 1.415m x 1.005m x 0.45m. The cooling loop is arranged as a double U-shaped circuit, so the cooling pipeline length of the single unit is 1.415m x 4 + 1.005m = 6.665m. There are 440 blankets in ITER, so the estimated total length of the FW inner wall cooling pipeline is 6.665 x 440 = 2932.6 m. According to the failure history of the VV intracooler pipeline, the leakage probability in the ITER intercooler pipeline is about 3.5E−5/h. The gap between two plasma pulses is 1800 seconds, the duty cycle requirement is not less than 25%, and the duty cycle of the plasma burning time is calculated to be 0.25 according to the ITER operation plan. Based on a comprehensive evaluation, the frequency of this event is 0.22 per year. The frequency level corresponds to the ITER frequency range. According to the development process of this incident, an event tree model was created as shown in Figure 4. Descriptions of each sequence are given in Table 2.

Figure 4
Figure 4

Event tree of in-vessel leak events.

Table 2: Event tree sequence of in-vessel leakage events.

According to the title event conditions, success criteria and task time in the event tree analysis, the reliability of the VVPSS was analyzed using the fault tree analysis method according to the existing design scheme. The failure rate of other systems involved in the event tree analysis was cross-referenced to system failure data with similar functions and operating environments.

VVPSS is designed to reduce VV internal pressure. In cases of coolant loss from components inside the vessel, loss or other vacuum accidents, the VV internal pressure is limited to an absolute 0.15 MPa by opening the rupture discs to allow steam or non-condensable gas from the VV to flow into the VVPSS-ST.

The VVPSS consists of a large linear tank approximately 46 m long, with a circular cross section of 6 m diameter. The cylindrical jacket wall is generally 30 mm thick, and contains enough room temperature water (about 675 tons at below 300°C) to condense the vapor created by the more harmful coolant leaking into the vessel. The tank is connected to the discharge vessel through two H&CD neutral beam channels provided with the VVPSS boxes. From these locations, one main vent pipe is routed to the VVPSS tank, and the vent pipe has dual rupture disc assemblies that form the vacuum boundary between the vacuum vessel and the room temperature funnel tank during normal operation. Numerical studies predict that a total of at least 1.0 m of vent piping is required2, to limit the VV pressure below 0.15 MPa during a Class IV coolant leak. The VVPSS vent line includes two sets of rupture discs connected in series and these rupture discs open during a Class IV refrigerant leak. The relief line also includes a rupture disc bypass system, consisting of side pipes containing isolation valves designed to open when the discharge vessel pressure (about 0.94 bar absolute) is less than the rupture disc opening pressure.

As shown in Figure 5, during coolant leakage into the vessel, the VVPSS system acts in concert with the VV drainage system: the preceding discharges develop vapor into the funnel tank where it is condensed; While the latter facilitates timely drainage of water from the VV to limit the amount of steam that the funnel tank has to condense. The VV drain system is operated automatically by opening the rupture discs in the VV drain lines for a large coolant leak, and by opening the drain valves for a small one. In addition, the VVPSS is connected to the depletion system (DS), liquid and gas distribution system and vacuum monitoring system. The VVPSS system has a provision to deal with the gaseous exhaust that can arise during a coolant leak in the VV, by extracting this gaseous exhaust from the VVPSS tank body and transferring it to the Vent Disposal System (VDS).

Figure 5
Figure 5

PFD for vacuum pressure suppression system.

The fault tree model is shown in Figs. 6, 7, 8 and 9.

Figure 6
Figure 6

Error tree-1 in VVPSS.

Figure 7
Figure 7

Error tree-2 in VVPSS.

Figure 8
Figure 8

Error tree 3 in VVPSS.

Figure 9
Figure 9

Error Tree-4 in VVPSS.

The fault tree model generated in this study was analyzed by RiskA, a large-scale integrated probabilistic safety assessment program independently developed by the Institute of Nuclear Safety Technology of the Chinese Academy of Sciences. Now, RiskA3.0 has developed to a relatively mature stage, and has been successfully applied in many engineering software, such as TQRM, ITER-TBM, EAST, etc.10. Reliability data used in the analysis are presented in Table 3.

Table 3: Input data for fault tree analysis for VVPSS.

The analysis determined that there were 44 lower-order cut-sets that could lead to the upstream event, including 27 first-order cut-sets and 17 second-order cut-sets. The top 10 combinations that contributed most to the occurrence of the main event are shown in Table 4.

Table 4: Cutoff set for VVPSS.

Table 4 shows that the top 10 parts groups that have the greatest impact on the overhead event are all first-class parts groups, among which the top three groups are connected safety valves for overpressure protection tanks and other systems. In the subsequent process, more attention should be paid to connecting safety valves. Pipeline leakage and burst disk are a low-probability event, but in the long run, their cumulative failure cannot be ignored; The fault of the sensor inside the overpressure protection tank is also the focus of attention in the subsequent process.

According to the results of quantitative analysis, the probability of the highest event occurring is 5.06E−5.

For other systems involved in event tree analysis, the input conditions for fault tree analysis are not available at present, and the probability of system failure consistent with its function is taken as the basis for judgment. Table 5 shows the probability of failure of the relevant system.

Table 5: Partial reliability data used in this study.

Radiological risk analysis of a leakage accident inside a ship

In order to quantitatively assess the radiation risks of a leakage accident inside a ship. It is necessary to calculate the harmful dose caused by different working conditions. The maximum personal effective dose of 1 gram of tritium and 1 gram of dust at an altitude of 800 meters outside the plant boundary was noted as “fusion safety issues and their impact on design and R&D needs.”14. The dose conversion factor is shown in Table 6. The radiation dose per unit mass of ACP is based on the maximum radiation dose of 3.1E−3 mSv/g given in the ITER accident analysis report for the most severe meteorological environment. Table 7 shows the radiological consequences and release frequencies of an intra-ship leakage event under average meteorological conditions.

Table 6: Dose conversion factor indicates in this study.
Table 7: Frequency and design of each incident sequence for an in-vessel leakage incident.

You may also like...

Leave a Reply