Variations of temperature, salinity and current in the southern tidal passage of the Hangzhou Bay, China

Field surveys covering a spring-neap tidal period were conducted to investigate the characteristics of tidal dynamics within a curved channel in the southern Hangzhou Bay, China. The channel has a maximum depth of more than 100 m with an average tidal range of 2.5 m, serving as the main tidal passage in the southern part of the Hangzhou Bay. Water salinity, temperature and velocity data were collected from the ship-based transects and mooring measurements. During flood tide, the tidal current intrudes into the Hangzhou Bay through the northern side of the channel with a maximum velocity of about 2 m/s, while retreats through the southern side during ebb tide with a maximum velocity of 1.8 m/s. Due to the pressure, density gradients, the Coriolis force and centrifugal effect, a lateral exchange flow is generated as the tidal current relaxes from flood to ebb. Salinity and temperature data show that the water in the channel is weakly stratified during both spring and neap tides in summer time.However, mixing in the middle region will be enhanced by the lateral circulation. Mooring data indicate that the temperature and salinity are varying at a frequency similar to tidal current but higher than sea level oscillation.Our results support the hypothesis that the high frequency salinity and temperature variations could be generated by combination of the tidal current and the lateral exchanging flow.

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.