The Applying of Improved BP Neural Network in the Recognition of Nuclear Fusion's MHD Pattern
暂无分享,去创建一个
Controlled nuclear fusion is an important direction to solve the shortage of energy resources in the future. The commercial nuclear fusion core needs higher temperature, bigger density, stronger constraint efficiency plasma. The plasma’s electric current, pressure distribution and magnetic field etc shift MHD pattern, and make further efforts to cause the split of plasma which will result in disasters. So the recognition of MHD pattern becomes the most significant task. BP neural networks have attracted considerable research on the effect of algorithms and network structures, as well as multiple solutions problem, constriction rate and hide nodes numbers. Experiments shows existing algorithms do not suitable for nuclear fusion MHD pattern detection. This paper builds up an improved BP neural network to recognize the MHD pattern. The experimental evidence strongly suggests this model has obtained a favorable constriction rate and discrimination precision.
[1] Lin Hui. Target Attribute Identification Simulation Based on BP Neural Network Model , 2007 .
[2] F. W. O. Aduol,et al. ROBUST GEODETIC PARAMETER ESTIMATION THROUGH ITERATIVE WEIGHTING , 1994 .
[3] Z. S. Ji,et al. The distributed control and data system in HT-7 tokamak , 2002 .
[4] Robert A. Jacobs,et al. Increased rates of convergence through learning rate adaptation , 1987, Neural Networks.