A short-term wind power prediction model based on BiLSTM–CNN–WGAN-GP (LCWGAN-GP) is proposed in this paper, aiming at the problems of instability and low prediction accuracy of short-term wind power prediction. Firstly, the original wind energy data are decomposed into subsequences of natural mode functions with different frequencies by using the variational mode decomposition (VMD) algorithm. The VMD algorithm relies on a decision support system for the decomposition of the data into natural mode functions. Once the decomposition is performed, the nonlinear and dynamic behavior are extracted from each natural mode function. Next, the BiLSTM network is chosen as the generation model of the generative adversarial network (WGAN-GP) to obtain the data distribution characteristics of wind power’s output. The convolutional neural network (CNN) is chosen as the discrimination model, and the semi-supervised regression layer is utilized to design the discrimination model to predict wind power. The minimum–maximum game is formed by the BiLSTM and CNN network models to improve the quality of sample generation and further improve the prediction accuracy. Finally, the actual data of a wind farm in Jiuquan City, Gansu Province, China is taken as an example to prove that the proposed method has superior performance compared with other prediction algorithms.

Sandwiched piezoelectric transducers are widely used, especially in high power applications. For more convenient analysis and design, a PSpice lossy model of sandwiched piezoelectric ultrasonic transducers in longitudinal vibration is proposed by means of the one-dimensional wave and transmission line theories. With the proposed model, the resonance and antiresonance frequencies are obtained, and it is shown that the simulations and measurements have good consistency. For the purpose of further verification the accuracy and application of the PSpice model, a pitch-catch setup and an experimental platform are built. They include two sandwiched piezoelectric ultrasonic transducers and two aluminum cylinders whose lengths are 20 mm and 100 mm respectively. Based on this pitch-catch setup, the impedance and transient analysis are performed. Compared with the measured results, it is shown that the simulated results have good consistency. In addition, the conclusion can be drawn that the optimal excitation frequency for the pitch-catch setup is not necessarily the resonance frequency of ultrasonic transducers, because the resonance frequency is obtained under no load. The proposed PSpice model of the sandwiched piezoelectric transducer is more conveniently applied to combine with other circuits such as driving circuits, filters, amplifiers, and so on.

For achieving the power maximum transmission, the electrical impedance matching (EIM) for piezoelectric ultrasonic transducers is highly required. In this paper, the effect of EIM networks on the electromechanical characteristics of sandwiched piezoelectric ultrasonic transducers is investigated in time and frequency domains, based on the PSpice model of single sandwiched piezoelectric ultrasonic transducer. The above-mentioned EIM networks include, series capacitance and parallel inductance (I type) and series inductance and parallel capacitance (II type). It is shown that when I and II type EIM networks are used, the resonance and anti-resonance frequencies and the received signal tailing are decreased; II type makes the electro-acoustic power ratio and the signal tailing smaller whereas it makes the electro-acoustic gain ratio larger at resonance frequency. In addition, I type makes the effective electromechanical coupling coefficient increase and II type makes it decrease; II type make the power spectral density at resonance frequency more dramatically increased. Specially, the electro-acoustic power ratio has maximum value near anti-resonance frequency, while the electro-acoustic gain ratio has maximum value near resonance frequency. It can be found that the theoretically analyzed results have good consistency with the measured ones.

For a broken rail detection system based on ultrasonic guided waves (UGW), the multimodal and dispersion of UGW degrade signal-to-noise ratio (SNR) and range resolution. To improve SNR of the received signals and range resolution, the pulse compression technique based on 13-bit Barker code is presented in this work. Through a PSpice model of the pitch-catch setup, as well as performing field tests, it is shown that coded UGW signals can efficiently improve SNR by 5 dB and have strong noise immunity. As the detection distance increases, the mainlobe width increases linearly while the sidelobe peak levels remain basically unchanged. In addition, to correctly and quickly identify the corresponding transmissions at the receivers, an adaptive peak detection algorithm is proposed, which is based on a digital bandpass tracking filter, moving averaging filters and Hilbert transform. By using some field tests under different detection distances, it is found that compared to the previous works, the proposed adaptive peak detection algorithm has stronger robustness and better anti-noise performance. In addition, the proposed method is easy to integrate into a real-time detection system by proper software design. INDEX TERMS Peaks detection, barker code, pulse compression, long rail breakages detection, UGW.