Prototype of an early warning system based on deep learning for the CSNS accelerator
-
摘要: 为了能在中国散裂中子源(CSNS)加速器的部分故障发生前发出预警信息,利用深度学习建立了基于CSNS加速器真空度和漂移管直线加速器(DTL)温度的特征模型,开发了一套CSNS加速器预警系统样机。该样机基于实验物理及工业控制系统(EPICS)架构搭建,主要由训练、识别和信息发布3部分组成,采用Python进行程序设计开发,实现了训练样本获取、深度学习网络设计和训练、在线识别和信息发布等功能。测试结果表明,该样机对基于CSNS加速器真空度和DTL温度历史数据生成的测试集的准确率达98.4%,且能根据实时数据识别出CSNS加速器真空度和DTL温度的异常,并能发出预警信息,证明了其可行性和有效性。
-
关键词:
- 中国散裂中子源 /
- 加速器 /
- 预警系统 /
- 深度学习 /
- 实验物理及工业控制系统
Abstract: To send out early warnings before some failures of the China Spallation Neutron Source (CSNS) accelerator, the feature models of the CSNS accelerator vacuums and drift tube linac (DTL) temperatures have been established based on deep learning, and a prototype of an early warning system has been developed. This prototype of an early warning system was built based on the experimental physics and industrial control system (EPICS) architecture, and it is mainly composed of three parts: training, recognition and information release. Python was adopted for program design and development, and functions such as training samples acquisition, deep learning networks design and training, online recognition and information release have been realized. The test results show that the accuracy of this prototype can reach 98.4% for the test set generated based on the historical data of the CSNS accelerator vacuums and DTL temperatures, and the anomalies of the CSNS accelerator vacuums and DTL temperatures can be recognized based on the real-time data, and the early warnings can be sent out, which proves its feasibility and effectiveness. -
表 1 图形工作站的主要信息
Table 1. The main information of the graphics workstation
CPU Memory GPU Operating system Python TensorFlow Intel Xeon E5-2678v3×2 128 GB RTX 2080Ti×2 Ubuntu 18.04.1 Anaconda 2019.07 1.14.0 -
[1] Wei Jie, Chen Hesheng, Chen Yanwei, et al. China Spallation Neutron Source: design, R& D, and outlook[J]. Nuclear Instruments & Methods in Physics Research Section A—Accelerators Spectrometers Detectors and Associated Equipment, 2009, 600(1): 10-13. [2] Wei Jie, Fu Shinian, Tang Jingyu, et al. China Spallation Neutron Source—An overview of application prospects[J]. Chinese Physics C, 2009, 33(11): 1033-1042. doi: 10.1088/1674-1137/33/11/021 [3] Wang Sheng, Fang Shouxian, Fu Shinian, et al. Introduction to the overall physics design of CSNS accelerators[J]. Chinese Physics C, 2009, 33(S2): 1-3. doi: 10.1088/1674-1137/33/S2/001 [4] Fu S N, Chen H S, Chen Y W, et al. Status of CSNS project[C]//Proceedings of IPAC. 2013: 3995–3999. [5] Chen H, Wang X L. China's first pulsed neutron source[J]. Nature Materials, 2016, 15(7): 689-691. doi: 10.1038/nmat4655 [6] Liu Huachang, Peng Jun, Gong Keyun, et al. The design and construction of CSNS drift tube linac[J]. Nuclear Inst and Methods in Physics Research A, 2018, 911: 131-137. doi: 10.1016/j.nima.2018.10.034 [7] LeCun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 2015, 521(7553): 436-444. doi: 10.1038/nature14539 [8] Lecun Y, Kavukcuoglu K, Farabet C. Convolutional networks and applications in vision[C]//International Symposium on Circuits and Systems Nano-Bio Circuit Fabrics and Systems (ISCAS 2010). 2010. [9] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84-90. doi: 10.1145/3065386 [10] Lecun Y, Bottou L, Bengio Y. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324. doi: 10.1109/5.726791 [11] Zhang Yuliang, Jin Dapeng, Zhu Peng, et al. The accelerator control system of CSNS[J]. Radiation Detection Technology and Methods, 2020, 4: 1-14. doi: 10.1007/s41605-019-0147-6 [12] Zhang Yuliang, Kang Mingtao, Jin Dapeng, et al. The run management system for CSNS[J]. Radiation Detection Technology and Methods, 2019, 3(3): 35-37. doi: 10.1007/s41605-019-0116-0 [13] 康明涛, 黄涛, 张玉亮, 等. CSNS加速器真空控制系统的设计与实现[J]. 强激光与粒子束, 2020, 32:084001. (Kang Mingtao, Huang Tao, Zhang Yuliang, et al. Design and implementation of vacuum control system of China Spallation Neutron Source[J]. High Power Laser and Particle Beams, 2020, 32: 084001 [14] 何泳成, 李刚, 金大鹏, 等. CSNS漂移管直线加速器水冷联锁系统设计[J]. 核电子学与探测技术, 2017, 37(6):585-589. (He Yongcheng, Li Gang, Jin Dapeng, et al. The design of csns drift tube linac water cooling interlock system[J]. Nuclear Electronics & Detection Technology, 2017, 37(6): 585-589 doi: 10.3969/j.issn.0258-0934.2017.06.006 [15] 赵籍九, 尹兆升. 粒子加速器技术[M]. 北京: 高等教育出版社, 2006.Zhao Jijiu, Yin Zhaosheng. Particle accelerator technology[M]. Beijing: Higher Education Press, 2006 [16] Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2015: 1-9. [17] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016: 770-778. [18] Goodfellow I J, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets[C]//28th Conference on Neural Information Processing Systems (NIPS). 2014. [19] Sainath T N, Kingsbury B, Saon G, et al. Deep convolutional neural networks for large-scale speech tasks[J]. Neural Networks, 2015, SI(64): 39-48. [20] Sanner M F. Python: A programming language for software integration and development[J]. Journal of Molecular Graphics & Modelling, 1999, 17(1): 57-61.