2:00 PM - 2:15 PM
[MAG39-02] 129mTe source term prediction at the beginning of Fukushima accident calculated reversely from the soil contamination map and atmospheric dispersion calculation, and recommendations for iodine and Cs source terms
Keywords:Fukushima Daiichi Nuclear Power Plant Accident, 129mTe source terms, MEXT soil contamination map, Weather research and forecasting model
Accurate evaluation of the source terms at the time of Fukushima accident is important for analyses of the accident progression and/or the environmental consequent evaluation. The backward calculation of the source terms using the conventional monitoring post and seawater surface concentration data as well as atmospheric dispersion calculation was a prediction based on the trajectory analysis using the point information of the measurement point. For this reason, it was difficult to predict in the case of situation when the wind blew from land to the sea (land wind), and even the latest evaluation, Terada et al. (2020) could not predict the release from Unit 1 before early morning of March 12.
Meanwhile, in this method, Kawashima et al. (1999), we focused on a wide range of soil deposition distributions, weighted the results of the surface hourly deposition distributions obtained from the mesoscale meteorological model (WRF) calculations assuming the unit release, and added them together. The source term can be evaluated with high accuracy by determining the weight (= release rate) so as to minimize the deviation between the added results described above and the cumulative soil deposition distribution measured after the accident. As a feature, even in the case of land wind, most of the microparticles are transported to the sea side, but some of them return to the land side by the sea wind later, which makes it possible to predict the source term.
In the first presentation, we reversely estimated the release time zone of 129mTe from March 14 to March 16 using the MEXT 129mTe soil contamination map, Saito et al. (2015) and WRF. In addition, Takahashi et al. (2019) predicted its origin from changes in the pressure of the containment vessel, and the presence or absence of water in the leakage path. In the second presentation, Hidaka et al. (2021) took notice that most of the 129mTe released from the fuel during a severe accident is once taken into the inner surface of the unoxidized Zr cladding tube and released as SnTe just before the Zr cladding is completely oxidized during the core refilling, and focused on the release during reinjection from March 12 to March 16. However, since complete oxidation of the Zr cladding also partially occurs at the middle or higher location of the cladding when the cladding temperature rises for the first time. Therefore, in the present analysis, a similar inverse estimation was performed for the time between March 11 and March 15 in order to predict the first 129mTe release from Unit 1.
The present analysis showed that the first release from each Unit occurred as follows: Unit 1 at around 19:00 on March 11, Unit 3 from 4:00 to 6:00 on March 13, and Unit 2 at around 19:00 on March 14, respectively. These correspond to the moments when in Unit 1, immediately after the emergency condenser (IC) valve could not be opened due to the loss of DC power, and in Unit 3, HPCI was manually stopped for fire engine water injection but the fire engine water could not be injected. In Unit 2, the time point corresponds to just after the manually decrease in the pressure and water level in the reactor pressure vessel in order to start the fire engine water injection. In each case, although the predicted points described above are before the leakage from the upper flange of the containment vessel, the release to the environment can be explained by the leakage from core instrumentation or the main steam pipe flange.
As described above, the 129mTe source term obtained in this analysis can be almost explained from the records of events on the reactor side, and the prediction accuracy is considered to be quite high. Considering this result and the higher volatility of iodine and Cs than Te, it is considered that the releases of iodine and Cs increased in the late evening of March 11, and early mornings of March 12 and March 13, which had not been evaluated so far. This method was applicable when the release of 129mTe occurred in the first few days of the accident, and when applied to Cs whose release continued for more than one month, it is necessary to solve various problems such as the convergence of calculations.
Meanwhile, in this method, Kawashima et al. (1999), we focused on a wide range of soil deposition distributions, weighted the results of the surface hourly deposition distributions obtained from the mesoscale meteorological model (WRF) calculations assuming the unit release, and added them together. The source term can be evaluated with high accuracy by determining the weight (= release rate) so as to minimize the deviation between the added results described above and the cumulative soil deposition distribution measured after the accident. As a feature, even in the case of land wind, most of the microparticles are transported to the sea side, but some of them return to the land side by the sea wind later, which makes it possible to predict the source term.
In the first presentation, we reversely estimated the release time zone of 129mTe from March 14 to March 16 using the MEXT 129mTe soil contamination map, Saito et al. (2015) and WRF. In addition, Takahashi et al. (2019) predicted its origin from changes in the pressure of the containment vessel, and the presence or absence of water in the leakage path. In the second presentation, Hidaka et al. (2021) took notice that most of the 129mTe released from the fuel during a severe accident is once taken into the inner surface of the unoxidized Zr cladding tube and released as SnTe just before the Zr cladding is completely oxidized during the core refilling, and focused on the release during reinjection from March 12 to March 16. However, since complete oxidation of the Zr cladding also partially occurs at the middle or higher location of the cladding when the cladding temperature rises for the first time. Therefore, in the present analysis, a similar inverse estimation was performed for the time between March 11 and March 15 in order to predict the first 129mTe release from Unit 1.
The present analysis showed that the first release from each Unit occurred as follows: Unit 1 at around 19:00 on March 11, Unit 3 from 4:00 to 6:00 on March 13, and Unit 2 at around 19:00 on March 14, respectively. These correspond to the moments when in Unit 1, immediately after the emergency condenser (IC) valve could not be opened due to the loss of DC power, and in Unit 3, HPCI was manually stopped for fire engine water injection but the fire engine water could not be injected. In Unit 2, the time point corresponds to just after the manually decrease in the pressure and water level in the reactor pressure vessel in order to start the fire engine water injection. In each case, although the predicted points described above are before the leakage from the upper flange of the containment vessel, the release to the environment can be explained by the leakage from core instrumentation or the main steam pipe flange.
As described above, the 129mTe source term obtained in this analysis can be almost explained from the records of events on the reactor side, and the prediction accuracy is considered to be quite high. Considering this result and the higher volatility of iodine and Cs than Te, it is considered that the releases of iodine and Cs increased in the late evening of March 11, and early mornings of March 12 and March 13, which had not been evaluated so far. This method was applicable when the release of 129mTe occurred in the first few days of the accident, and when applied to Cs whose release continued for more than one month, it is necessary to solve various problems such as the convergence of calculations.