The microwave nondestructive testing (MNDT) technique is an attractive candidate for defect detection in the dielectric structures as microwave signals can penetrate inside and interact with the insides of the dielectric material. In this research, a novel MNDT technique is presented for defect detection combining the time domain reflectometry (TDR) technique, the wavelet decomposition analysis (WDA) technique, and an artificial neural network (ANN), which has good generality and can be applied to various dielectric materials with different thicknesses. This method is based on scanning the dielectric material with an open-ended rectangular waveguide (OERW) operating from 18 to 26.5 GHz with 201 frequency points and analyzing its reflections using the proposed algorithm. The method was validated by simulation and experiments successfully. In the simulations, two typical kinds of dielectric materials were used to verify the generality and accuracy of this proposed method. The accuracies of these two dielectric materials were 96% and 90%, respectively. A 2-D image of detection can be constructed by performing repetition of the process with a certain step on the xy plane of a material, which is illustrated by simulations. Meanwhile, the sensitivity analysis was performed by studying the similarity quantified using the Pearson’s coefficient correlation. Eventually, some FR4 samples were machined to demonstrate the feasibility of this method. The acceptable accuracy is 86% of the experimental result.

This paper presents a new electro-thermal coupling simulation method for evaluating the reliability of IGBT modules, which combines a numerical power loss model and a finite element model. To illustrate the method, a specific case of the Infineon FF50R12RT4 module operated with an SPWM signal is considered. Temperature and stress data are obtained and analyzed via electro-thermal simulation, and the service life of modules is calculated accordingly. The influence of ambient temperature and gate signal characteristics on power loss and service life of IGBT modules is discussed, which can be used for the optimization of work points. In this method, the effect of junction temperature on power loss is considered, and a more accurate electro-thermal simulation is achieved. The proposed method can be used to analyze the thermal and mechanical stresses of IGBT modules, predict the failure location, and assess the service life.

The parameter optimization of a multistage synchronous induction coilgun (SICG) is a time-consuming task. Traditional machine learning methods can accelerate the process by building predictive models, but they require separate modeling for an SICG with different stages, which requires numerous datasets and is a cumbersome process. This paper proposes a method for building a predictive model for an SICG with different stages based on a recurrent neural network (RNN). In this method, the feed time of a 2- to 10-stage SICG is selected from the standard orthogonal design table as the training and test datasets, and the current filament method (CFM) is used to calculate the dataset label. The gate recurrent unit (GRU) neural network is used to study the training dataset, and the predictive model has good accuracy with respect to the test dataset, with an average error of 0.022. The predictive model and a particle swarm optimization (PSO) algorithm are applied to optimize the feed time of the SICG with different stages. The results show that the three-stage SICG can achieve a muzzle velocity of 50 m/s for a projectile, while the maximum muzzle velocity of the three-stage SICG in all datasets is 46.87 m/s.