scholarly journals A repetitive 800 kA linear transformer drivers stage for Z-pinch driven fusion-fission hybrid reactor

2015 ◽  
Vol 33 (3) ◽  
pp. 535-540 ◽  
Author(s):  
Chuan Liang ◽  
Lin Zhou ◽  
Fengju Sun ◽  
Jiangtao Zeng ◽  
Mingjia Li ◽  
...  

AbstractThis paper presents the design and tests of a repetitive 800 kA fast linear transformer driver (LTD) stage aimed for the Z-pinch driven fusion-fission hybrid reactor (Z-FFR).The LTD stage consists of 34 parallel basic resistor R, inductor L, and capacitor C (RLC) circuits each made up of two 100 kV/40 nF capacitors, a multi-stage gas switch and Metglas magnetic cores. The stage can deliver about 800 kA current pulse with rise time of 100 ns into the matched liquid resistive load at a repetitive frequency 0.1 Hz. A novel method to trigger the stage via a continuous internal trigger bus composed by a single cable has been proposed and demonstrated. The experimental results show that the new trigger method is feasible and reliable. A 140 kV, 25 ns rising time trigger pulse, and a 5.2 kA, 30 μs width pre-magnetization current pulse which can operate at a repetition rate 0.1 Hz were used in this stage to insure the LTD stage generating a 80 kV/800 kA current pulse every 10 s. A multi-stage gas switch that has a lifetime in excess of 10,000 shots and a jitter less than 3 ns one sigma agrees well with the demand of Z-FFR. The electrical behavior of the stage can be predicted from a simple RLC circuit, which can simplify the design of various LTD-based accelerators.

1999 ◽  
Vol 40 (11-12) ◽  
pp. 67-75 ◽  
Author(s):  
Sigrun J. Jahren ◽  
Jukka A. Rintala ◽  
Hallvard Ødegaard

Thermomechanical pulping (TMP) whitewater was treated in thermophilic (55°C) anaerobic laboratory-scale reactors using three different reactor configurations. In all reactors up to 70% COD removals were achieved. The anaerobic hybrid reactor, composed of an upflow anaerobic sludge blanket (UASB) and a filter, gave degradation rates up to 10 kg COD/m3d at loading rates of 15 kg COD/m3d and hydraulic retention time (HRT) of 3.1 hours. The anaerobic multi-stage reactor, consisting of three compartments, each packed with granular sludge and carrier elements, gave degradation rates up to 9 kg COD/m3d at loading rates of 15-16 kg COD/m3d, and HRT down to 2.6 hours. Clogging and short circuiting eventually became a problem in the multi-stage reactor, probably caused by too high packing of the carriers. The anaerobic moving bed biofilm reactor performed similar to the other reactors at loading rates below 1.4 kg COD/m3d, which was the highest loading rate applied. The use of carriers in the anaerobic reactors allowed short HRT with good treatment efficiencies for TMP whitewater.


2021 ◽  
Vol 232 (10) ◽  
Author(s):  
Amanda F. do Amaral ◽  
Alexandre S. A. da Silva ◽  
Rodrigo Coutinho ◽  
Deivisson L. Cunha ◽  
Marcia Marques

2010 ◽  
Vol 22 (5) ◽  
pp. 1163-1166 ◽  
Author(s):  
刘轩东 Liu Xuandong ◽  
孙凤举 Sun Fengju ◽  
姜晓峰 Jiang Xiaofeng ◽  
梁天学 Liang Tianxue ◽  
孙福 Sun Fu ◽  
...  

2012 ◽  
Vol 24 (8) ◽  
pp. 2009-2012 ◽  
Author(s):  
丛培天 Cong Peitian ◽  
孙铁平 Sun Tieping ◽  
邱爱慈 Qiu Aici ◽  
曾正中 Zeng Zhengzhong

2012 ◽  
Vol 60 (3) ◽  
pp. 30801 ◽  
Author(s):  
Xuandong Liu ◽  
Qijun Tang ◽  
Fengju Sun ◽  
Zhuo Yang ◽  
Xiangyu Tan ◽  
...  

1991 ◽  
Vol 70 (2) ◽  
pp. 407-412
Author(s):  
K. ÖZDAŞ ◽  
M. S. KILIÇKAYA
Keyword(s):  

Author(s):  
Fang Dong ◽  
Fanzhang Li

Deep learning has achieved lots of successes in many fields, but when trainable sample are extremely limited, deep learning often under or overfitting to few samples. Meta-learning was proposed to solve difficulties in few-shot learning and fast adaptive areas. Meta-learner learns to remember some common knowledge by training on large scale tasks sampled from a certain data distribution to equip generalization when facing unseen new tasks. Due to the limitation of samples, most approaches only use shallow neural network to avoid overfitting and reduce the difficulty of training process, that causes the waste of many extra information when adapting to unseen tasks. Euclidean space-based gradient descent also make meta-learner's update inaccurate. These issues cause many meta-learning model hard to extract feature from samples and update network parameters. In this paper, we propose a novel method by using multi-stage joint training approach to post the bottleneck during adapting process. To accelerate adapt procedure, we also constraint network to Stiefel manifold, thus meta-learner could perform more stable gradient descent in limited steps. Experiment on mini-ImageNet shows that our method reaches better accuracy under 5-way 1-shot and 5-way 5-shot conditions.


2019 ◽  
Vol 156 ◽  
pp. 371-383
Author(s):  
Komsan Kaewmamuang ◽  
Apirat Siritaratiwat ◽  
Chayada Surawanitkun ◽  
Pirat Khunkitti ◽  
Rongrit Chatthaworn

2008 ◽  
Vol 3 (4) ◽  
pp. 68-73
Author(s):  
Aleksandr V. Akimov ◽  
Petr A. Bak ◽  
Andrey A. Korepanov ◽  
Pavel V. Logachev ◽  
Viktor D. Bochkov ◽  
...  

The circuit of modulator, serving to supply inductive-resistive load in double-pulse mode with currents up to 10 kA and pulse duration of 300 ns, is described. As switching components unheated cathode thyratrons (pseudospark switches) TPI1-10k/50 and TPI5-10k/50 with anode voltage up to 50 kV have been used. The results of tests, confirming possibility of the thyratron reverse dielectric strength recovery within some microseconds after switching of 10 kA forward anode current, are presented.


Sign in / Sign up

Export Citation Format

Share Document