Journal of Radiation Research and Radiation Processing, Volume. 41, Issue 4, 040601(2023)
Consequence prediction in nuclear transport explosion accident using long short-term memory network
Figures & Tables(15)
Fig. 2. LSTM unit ("⊙" denotes the hadamard product, "+" denotes the matrix addition)
Fig. 5. Horizontal distribution of Pu-239 concentration at different time (in the XY plane where Z=200 m) (color online)
Fig. 6. Horizontal distribution of Pu-239 concentration at 10 min (in the XY plane where Z =200 m)
Fig. 10. Comparison of test set prediction results in area A and area B
Fig. 11. Comparison of loss convergence results in area A and area B
Table 1. Preliminary data from 780 s to 815 s in area A
View tableView in ArticleTable 1. Preliminary data from 780 s to 815 s in area A
时间 / s 浓度 / (Bq∙m-3) 相对压强 / Pa X轴风速 / (m∙s-1) Y轴风速 / (m∙s-1) Z轴风速 / (m∙s-1) Time Concentration Relative pressure Wind speed-X Wind speed-Y Wind speed-Z 780 19.187 11 0.3369 06 0.300 333 1.239 889 0.040 355 785 20.816 61 0.3372 67 0.300 599 1.239 967 0.040 489 790 22.555 58 0.3376 52 0.300 872 1.240 033 0.040 625 795 24.408 33 0.3379 43 0.301 149 1.240 111 0.040 765 800 26.379 01 0.3381 82 0.301 429 1.240 167 0.040 907 805 28.471 60 0.3384 76 0.301 714 1.240 178 0.041 052 810 30.689 79 0.3386 63 0.301 996 1.240 189 0.041 199 815 33.036 92 0.3387 99 0.302 280 1.240 167 0.041 349
Table 2. Model performance comparison based on grid search
View tableView in ArticleTable 2. Model performance comparison based on grid search
全连接层数
Dense layers
平均绝对误差 / (Bq∙m-3)
MAE
平均绝对百分比误差 / (Bq∙m-3)
MAPE
LSTM层数
LSTM layers
2 4 6 8 2 4 6 8 单元数=4
Units = 4
1 20.348 7 31.069 5 36.874 2 38.786 3 16.607 5 26.246 5 29.982 8 31.631 2 23.620 7 19.986 6 29.247 40.358 3 19.562 9 16.609 6 25.029 4 33.310 1 3 12.362 4.045 3 4.029 15.346 5 10.433 4 3.514 1 3.283 4 12.756 7 单元数=8
Units = 8
1 10.047 4 29.505 7.370 4 12.775 6 8.583 1 24.929 6 6.314 8 10.769 6 2 13.893 8 6.885 7 8.254 3 18.402 9 11.856 8 5.804 6 7.172 1 15.239 7 3 17.251 5 18.299 11.066 9 27.215 3 15.028 9 15.568 3 9.002 23.059 8 单元数=16
Units = 16
1 5.618 7 8.727 5 4.187 9 9.550 3 4.755 7 7.639 3 3.576 9 9.994 5 2 3.481 4 5.687 8 5.962 6 1.812 9 2.949 7 4.923 5 3.998 8 1.561 6 3 4.682 3 9.550 5 6.833 6 4.498 1 3.921 5 8.271 7 5.713 7 5.238 3 单元数=32
Units = 32
1 10.564 5 5.629 7 9.15 7.836 8 9.292 4.905 7 8.067 8 6.929 4 2 7.204 4 3.873 1 9.700 5 6.573 2 5.929 3.357 6 8.520 6 5.676 6 3 8.945 11.334 5 3.335 7 10.249 5 7.743 9 9.892 5 2.789 8 8.959 6 单元数=64
Units = 16
1 4.863 7 8.404 5 7.964 3 5.635 5 4.080 8 7.281 2 6.771 9 4.728 2 6.128 7.228 8 7.526 5.033 1 5.131 6 6.311 5 6.345 9 4.352 6 3 6.610 4 10.830 3 9.958 9 3.846 1 5.643 3 9.118 4 8.681 9 2.523 4
Table 3. Model running time comparison based on grid search
View tableView in ArticleTable 3. Model running time comparison based on grid search
全连接层数Dense layers 时间 / ms Time LSTM层数
LSTM layers
2 4 6 8 单元数=4
Units = 4
1 2 886 4 645 6 284 8 990 2 3 049 4 697 6 216 8 952 3 2 971 4 663 6 326 8 779 单元数=8
Units = 8
1 3 005 4 813 6 411 9 167 2 3 070 4 812 6 543 9 231 3 3 05 4 4 871 6 617 9 277 单元数=16
Units = 16
1 3 464 5 673 7 718 16 380 2 3 396 5 655 7 781 18 768 3 3 349 5 685 9 670 17 667 单元数=32
Units = 32
1 162 171 225 290 247 738 436 999 2 184 227 206 015 254 425 438 195 3 197 060 210 831 263 081 425 058 单元数=64
Units = 64
1 208 019 214 643 420 245 427 670 2 204 623 215 532 421 523 431 274 3 210 504 215 880 419 631 433 215
Table 4. Specific values of test set prediction results and true values in area A and area B
View tableView in ArticleTable 4. Specific values of test set prediction results and true values in area A and area B
区域A Area A 区域B Area B 时间 / s
Time
预测值 / (Bq∙m-3)
Prediction
CFD模拟值 / (Bq∙m-3)
CFD simulation
相对误差 / %
Relative error
时间 / s
Time
预测值 / (Bq∙m-3)
Prediction
CFD模拟值 / (Bq∙m-3)
CFD simulation
相对误差 / %
Relative error
1 620 142.495 140.667 1.30 380 49.901 9 50.252 4 0.70 1 625 138.726 136.552 1.59 385 48.715 49.076 1 0.74 1 630 134.597 132.323 1.72 390 47.487 3 47.875 6 0.81 1 635 130.188 127.99 1.72 395 46.211 5 46.648 9 0.94 1 640 125.486 123.565 1.55 400 44.899 5 45.401 2 1.11 1 645 120.343 119.061 1.08 405 43.535 3 44.139 5 1.37 1 650 114.843 114.492 0.31 410 42.123 8 42.870 4 1.74 1 655 109.237 109.875 0.58 415 40.691 3 41.598 7 2.18 1 660 103.74 105.223 1.41 420 39.255 2 40.327 2 2.66 1 665 98.089 100.557 2.45 425 37.858 8 39.060 2 3.08 1 670 92.5631 95.8882 3.47 430 36.672 5 37.801 1 2.99
Tools
Get Citation
Copy Citation Text
Lingpan RUAN, Chunhua CHEN, Liwei CHEN, Fang RUAN, Xiajuan LI, Jianye WANG. Consequence prediction in nuclear transport explosion accident using long short-term memory network[J]. Journal of Radiation Research and Radiation Processing, 2023, 41(4): 040601
Paper Information
Category: Research Articles
Received: Mar. 1, 2023
Accepted: Mar. 21, 2023
Published Online: Sep. 21, 2023
The Author Email:
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20