When using the Transformer model for wind power prediction, the presence of noise in wind power data and the model’s final layer relying solely on a simple linear output reduces the model’s ability to capture nonlinear relationships, leading to a decrease in prediction accuracy. To address these issues, this paper proposes an ultrashort-term wind power prediction model based on exponential weighted moving average (EWMA) data processing and Kolmogorov–Arnold Network (KAN)-Transformer. First, multiple variable features are smoothed using EWMA, which suppresses noise while preserving the original data trends. Then, the EWMA-processed data is input into the Encoder and Decoder modules of the Transformer model to extract features. The output from the Decoder layer is then passed through the KAN layer, built using a cubic B-spline function, to enhance the model’s ability to capture nonlinear relationships, thereby improving the prediction accuracy of the Transformer model for wind power. Finally, experimental analysis is conducted, and it shows that the proposed model achieves the highest prediction accuracy, with a mean absolute error of 4.38 MW, a root mean squared error of 7.37 MW, and a coefficient of determination of 98.73%.

The traditional model of the electromagnetic force and torque does not take the coil corners into account, which is the major cause for the motor fluctuation. To reduce the fluctuation, a more accurate real-time computation model, which considers the influence of the coil corners, is proposed in this paper. Three coordinate systems respectively for the stator, the mover, and the corner are established. The first harmonic of the magnetic flux density distribution of a Halbach magnet array is taken into account in this model. The coil is divided into the straight coil segment and the corner coil segment based on its structure. For the straight coil segment, the traditional Lorenz force method can be used to compute its electromagnetic force and torque, which is a function of the mover position. For the corner coil segment, however, the numerical calculation method can be used to get its respective electromagnetic force and torque. Based on the above separate analysis, an electromagnetic model can be derived, which is suitable for practical application. Compared with the well-known harmonic model, the proposed real-time computation model is found to have less model inaccuracy. Additionally, the real-time ability of the maglev planar motor model and the decoupling computation is validated by NI PXI platform (Austin, TX, USA).

When using point-by-point data input with former series models for wind power prediction, the prediction accuracy decreases due to data distribution shifts and the inability to extract local information. To address these issues, this paper proposes an ultra-short-term wind power prediction model based on the Z-score (ZS), Dish-TS (DT), and Patch time series Transformer (PatchTST). Firstly, to reduce the impact of data distribution shift on prediction accuracy, ZS standardization is applied to both training and testing datasets. Additionally, the DT algorithm, which can self-learn the mean and variance, is introduced for window data standardization. Secondly, the PatchTST model is employed to convert point input data into local-level input data. Feature extraction is then performed using the multi-head attention mechanism in the Encoder layer and a feed-forward network composed of one-dimensional convolution to obtain the prediction results. These results are subsequently de-standardized using DT and ZS to restore the original data amplitude. Finally, experimental analysis is conducted, comparing the proposed ZS-DT-PatchTST model with various prediction models. The proposed model achieves the highest prediction accuracy, with a mean absolute error of 5.95 MW, a mean squared error of 10.89 MW, and a coefficient of determination of 97.38%.

Wind power is an essential component of renewable energy. It enables the conservation of conventional energy sources such as coal and oil while reducing greenhouse gas emissions. To address the stochastic and intermittent nature of ultra-short-term wind power, a combined prediction model based on variational mode decomposition (VMD) and gradient boosting regression tree (GBRT) is proposed. Firstly, VMD is utilized to decompose the original wind power signal into three meaningful components: the long-term component, the short-term component, and the randomness component. Secondly, based on the characteristics of these three components, a support vector machine (SVM) is selected to predict the long-term and short-term components, while gated recurrent unit-long short-term memory (GRU-LSTM) is employed to predict the randomness component. Particle swarm optimization (PSO) is utilized to optimize the structural parameters of the SVM and GRU-LSTM combination for enhanced prediction accuracy. Additionally, a GBRT model is employed to predict the residuals. Finally, the rolling predicted values of the three components and residuals are aggregated. A deep learning framework using TensorFlow 2.0 has been built on the Python platform, and a dataset measured from a wind farm has been utilized for learning and prediction. The comparative analysis reveals that the proposed model exhibits superior short-term wind power prediction performance, with a mean squared error, mean absolute error, and coefficient of determination of 0.0244, 0.1185, and 0.9821, respectively.

An external rotor transverse flux motor using a combined-type stator core is proposed for a direct drive application in this paper. The stator core is combined by two kinds of components that can both be manufactured conveniently by generic laminated silicon steel used in traditional motors. The motor benefits from the predominance of low manufacturing cost and low iron loss by using a silicon-steel sheet. Firstly, the basic structure and operation principles of the proposed motor are introduced. Secondly, the expressions of the electromagnetic torque and the cogging torque are deduced by theoretical analysis. Thirdly, the basic characteristics such as permanent magnet flux linkage, no-load back electromotive force, cogging torque and electromagnetic torque are analyzed by a three-dimensional finite element method (3D FEM). Then, the influence of structure parameters on the torque density is investigated, which provides a useful foundation for optimum design of the novel motor. Finally, the torque density of the proposed motor is calculated and discussed, and the result shows that the proposed motor in this paper can provide considerable torque density by using few permanent magnets.

In the semiconductor industry, positioning accuracy and acceleration are critical parameters. To improve the acceleration speed of a motor, this paper proposes the moving-coil maglev planar motor with a concentric winding structure. The coordinate system has been built for the multiple degrees of freedom movement system. The Lorenz force method has been applied to solve its electromagnetic model. The real-time solving of the generalized inverse matrix of factors can realize the decoupling of the winding current. When the maglev height changes, the electromagnetic force and torque decreases exponentially with the increase of the air gap. To decrease the influence on control system performance by the internal model change and the external disturbance, this paper proposes an improved active disturbance rejection control (ADRC) to design the controller. This new controller overcomes the jitter phenomenon due to the turning point for the traditional ADRC, thus it is more suitable for the maglev control system. The comparison between ADRC and the improved ADRC has been conducted, the result of which shows the improved ADRC has greater robustness.