An ab initio investigation of the low-lying electronic states of BeH

Potential energy curves （PECs） for the ground state （X2∑＋） and the four excited electronic states （A2∏, B2∏, C2∑＋, 4∏） of a Bell molecule are calculated using the multi-configuration reference single and double excited configuration interaction （MRCI） approach in combination with the aug-cc-pVTZ basis sets. The calculation covers the internuclear distance ranging from 0.07 nm to 0.70 nm, and the equilibrium bond length Re and the vertical excited energy Te are determined directly. It is evident that the X2∑＋, A2∏, B2∏, C2∑＋ states are bound and 4∏ is a repulsive excited state. With the potentials, all of the vibrational levels and inertial rotation constants are predicted when the rotational quantum number J is set to be equal to zero （J = 0） by numerically solving the radial SchrSdinger equation of nuclear motion. Then the spectroscopic data are obtained including the rotation coupling constant w e, the anharmonic constant WeXe, the equilibrium rotation constant Be, and the vibration-rotation coupling constant ae. These values are compared with the theoretical and experimental results currently available, showing that they are in agreement with each other.

Excitation curve calibration for the SSRF magnet system

The needed electrical current for the magnet working under different energy loads can be easily calculated once the right relation between the magnet and the electrical current has been found. Therefore the excitation curve calibration for the magnet system is important to the SSRF. The measuring method on the magnet and the result of the excitation curve calibration are presented. The application of the excitation curve calibration for the bending magnet is given, and it is proved that the COD （Closed Orbit Distortion） and the working point of the storage ring are greatly affected by the current of the bending magnet.

A denoising-classification neural network for power transformer protection

Artificial intelligence(AI)can potentially improve the reliability of transformer protection by fusing multiple features.However,owing to the data scarcity of inrush current and internal fault,the existing methods face the problem of poor generalizability.In this paper,a denoising-classification neural network(DCNN)is proposed,one which inte-grates a convolutional auto-encoder(CAE)and a convolutional neural network(CNN),and is used to develop a reli-able transformer protection scheme by identifying the exciting voltage-differential current curve(VICur).In the DCNN,CAE shares its encoder part with the CNN,where the CNN combines the encoder and a classifier.Based on the inter-action of the CAE reconstruction process and the CNN classification process,the CAE regards the saturated features of the VICur as noise and removes them accurately.Consequently,it guides CNN to focus on the unsaturated features of the VICur.The unsaturated part of the VICur approximates an ellipse,and this significantly differentiates between a healthy and faulty transformer.Therefore,the unsaturated features extracted by the CNN help to decrease the data ergodicity requirement of AI and improve the generalizability.Finally,a CNN which is trained well by the DCNN is used to develop a protection scheme.PSCAD simulations and dynamic model experiments verify its superior performance.