BP神经网络电子束扫描均匀度校正系统

 电子束的扫描均匀度是工业辐照电子直线加速器的重要技术指标，该校正系统利用人工神经网络中的误差反向传播法可以对扫描电流波形进行校正，以克服扫描磁场引起的电子束形状变化、位置改变等造成的电子束分布的不均匀性，从而使扫描曲线的归一化均方偏差达到1.8%。

Uniformity calibration system of electron beam scan with BP ANN

Scan uniformity of electron beam is an important technological index of industrial irradiation LINAC, especially to semi-conductor device, which has pretty high requirements for scan uniformity. Standard triangle wave current generator cannot get good scan uniformity, because the beam shape at the ends of scan region has significant difference compared to that in the middle. Thus, in order to obtain ideal scan uniformity, corrected scan current wave shape has to be adopted. Error back propagation method in artificial neural network is used in the calibration system in this paper. It conquers the change of electron beam shape caused by scan magnetic field and the change of electron beam position which result in the non-uniformity of electron beam distribution, and the normalized RMS of scan curves can be as small as 1.8%.