SMOGN,MFO,and XGBoost Based Excitation Current Prediction Model for Synchronous Machine

The power factor is the ratio between the active and apparent power,and it is available to determine the operational capability of the intended circuit or the parts.The excitation current of the synchronous motor is an essential parameter required for adjusting the power factor because it determines whether the motor is under the optimal operating status.Although the excitation current should predict with the experimental devices,such a method is unsuitable for online real-time prediction.The artificial intelligence algorithm can compensate for the defect of conventional measurement methods requiring the measuring devices and the model optimization is compared during the research process.In this article,the load current,power factor,and power factor errors available in the existing dataset are used as the input parameters for training the proposed artificial intelligence algorithms to select the optimal algorithm according to the training result,for this algorithm to have higher accuracy.The SMOGN(Synthetic Minority Over-Sampling Technique for Regression with Gaussian Noise)is selected for the research by which the data and the MFO(Moth-flame optimization algorithm)are created for the model to adjust and optimize the parameters automatically.In addition to enhancing the prediction accuracy for the excitation current,the automatic parameter adjusting method also allows the researchers not specializing in the professional algorithm to apply such application method more efficiently.The final result indicated that the prediction accuracy has reached“Mean Absolute Error(MAE)=0.0057,Root Mean Square Error(RMSE)=0.0093 andR2 score=0.9973”.Applying this method to themotor control would be much easier for the power factor adjustment in the future because it allows the motor to operate under the optimal power status to reduce energy consumption while enhancing working efficiency.

Influences of acoustic field parameters on welding arc behavior in ultrasonic-MIG welding

By applying ultrasonic-MIG welding as research object, the behaviors of welding arc were analyzed with varied ultrasonic parameters in welding using arc images recorded by high-speed camera. The influences of the current by exciting ultrasonic and the height and shape of ultrasonic radiator on welding arc were studied. Results showed that when the current was 150 mA, ultrasonic showed most distinct compressive effect on arc. The compressive volumes of arc length at different heights were calculated by adjusting the height of ultrasonic radiator continuously from 10 mm to 35 mm, there were three maximum points. The compressive degrees of them reduced successively. By utilizing different shapes of ultrasonic radiator, it revealed that ultrasonic radiator with spherical crown surface showed better compressive effect in a larger welding standard scope. When radius of radiator increased, axial compressive volume of arc enlarged, while an increasing curvature radius led to mare distinct radial compression of arc.

Parameters analysis and application of the differential excitation detection technology

A differential excitation probe based on eddy current testing technology was designed. Sheet specimens of Q 235 steel with prefabricated micro-cracks of different widths and of aluminum with prefabricated micro-cracks of different depths were detected through the designed detection system. The characteristics of micro-cracks can be clearly showed after signals processing through the short-time Fourier transform（ STFT）. By changing the parameter and its value in detecting process,the factors including the excitation frequency and amplitude,the lift-off effect and the scanning direction were discussed,respectively. The results showed that the differential excitation probe was insensitive to dimension and surface state of the tested specimen,while it had a high degree of recognition for micro-crack detection. Therefore,when the differential excitation detection technology was used for inspecting micro-crack of turbine blade in aero-engine,and smoothed pseudo Wigner-Ville distribution was used for signal processing,micro-cracks of 0. 3 mm depth and 0. 1 mm width could be identified. The experimental results might be useful for further research on engineering test of turbine blades of aero-engine.

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.