Near-infrared spectroscopy (NIRS) combined with pattern recognition technique has become an important type of non-destructive discriminant method. This review first introduces the basic structure of the qualitative analysis process based on near-infrared spectroscopy. Then, the main pretreatment methods of NIRS data processing are investigated. Principles and recent developments of traditional pattern recognition methods based on NIRS are introduced, including some shallow learning machines and clustering analysis methods. Moreover, the newly developed deep learning methods and their applications of food quality analysis are surveyed, including convolutional neural network (CNN), one-dimensional CNN, and two-dimensional CNN. Finally, several applications of these pattern recognition techniques based on NIRS are compared. The deficiencies of the existing pattern recognition methods and future research directions are also reviewed.

In this paper, the fault diagnosis problem for power converters is considered. Given that the existing fault diagnosis models rarely address the problems of the data noise and the new faults that are never emerged in the database, thus, an optimized fault diagnosis model for power converters based on temporal convolutional network (TCN) is proposed. Our contributions include: 1) unknown faults can be efficiently distinguished with an optimized classifier; 2) the proposed model has good robustness and reliability under noisy environment without any subsidiary pre-denoising algorithm; 3) it can realize adaptive feature extraction, and the parameters are small. Experimental results on a three-phase voltage inverter platform demonstrate that the proposed approach is efficient and can be adaptively applied to various real applications.

An end-to-end 1D-CNN based on the NIRS technique is proposed to non-destructively and effectively discriminate aristolochic acids and their analogues.

Dual-active-bridge (DAB) converter can connect several ac distribution networks with different voltage levels by a common dc bus to form ac/dc hybrid distribution networks. Its reliability issue has become a hot research topic in academia and industry. In this article, the fault operation mode of DAB is analyzed in detail, and a fault diagnosis strategy based on adaptive threshold denoising for residual networks is proposed on this basis. The main module of the diagnosis algorithm consists of an improved channel attention module and threshold algorithm. The improved channel attention module improves the network’s ability to learn features by using two ways to extract different features of the fault signal separately. The threshold algorithm is embedded into the deep network as a nonlinear layer. The thresholds can be learned adaptively in the established cells, which plays a good adaptive denoising capability. Finally, by building a hardware-in-the-loop system, we compare the signal denoising ability of various methods at different signal-to-noise ratios. The experimental results demonstrate that the proposed method can accurately locate the open-circuit fault and achieve 99.82% accuracy under the original data. It is also more accurate than other methods under different signal-to-noise ratios, which verifies the correctness and effectiveness of the method.