Contaminated environment from textile industries have attracted global concern owing to the traits of organic pollutions with high stability against light and chemicals attacks. How to improve the preparation process of photocatalysts and reduce the cost is a considerable requirement for the wide application of photocatalytic wastewater treatment. The Ni-doped BiVO₄ (NBVO) process was improved to a facile and reliable hydrothermal method. Ni with a theoretical mass of 0.75% was added to BiVO₄ (BVO), which displayed outstanding photocatalytic activity and stability. Under visible light irradiation, the decolorization rate of 0.75% NBVO to Rhodamine B (RhB) can reach 96% within 75 min, and the removal efficiency can still achieve 86% after four successive cycles. The active ingredient ∙O₂ ⁻ confirmed from capture experiments played an indispensable role and was hired to explain the possible photocatalytic mechanism. In addition, the enhanced photocatalytic activity of 0.75% NBVO should be ascribed to the increase in specific surface area (beneficial for adsorption) and the decline in impedance (favorable for photocarrier migration). 0.75% NBVO as a highly efficient visible-light-driven photocatalyst has a brighter prospect for wastewater treatment in the years ahead.

β-Bi₂O₃/BiVO₄/Mn_xZn_1−xFe₂O₄ (BV/MZF) composite magnetic photocatalyst was first synthesized using the hydrothermal and calcination method. BV/MZF was a mesoporous material with most probable pore size and specific surface area of 18 nm and 17.84 m²/g, respectively. Due to its high saturation magnetization (2.67 emu/g), the BV/MZF composite can be easily separated and recovered from solution under an external magnetic field. The results of photo-decomposition experiments show that the decomposition rate of Rhodamine B (RhB) by BV/MZF can reach 92.6% in 3 h under visible light. After three cycles, BV/MZF can still maintain structural stability and excellent pollutant degradation effect. In addition, analysis of the photocatalytic mechanism of BV/MZF for RhB shows that the p-n heterojunction formed in BV/MZF plays a vital role in its photocatalytic performance. This work has potential application in the future for solving environmental pollution.

The superconducting current limiter (SFCL) is considered as one of the most promising short-circuit current limit device. However, fragile structure of superconductor and refrigeration system fault are leading causes of performance degradation in SFCL. The application conditions of SFCL are harsh and there has not any methods of condition monitoring been reported. In this paper, the design consideration of monitoring system is analyzed in detail. Combined with the running environment of the SFCL, the overall scheme design of the system is carried out. Furthermore, this paper proposes an equivalent heat source to simplify the process of thermal analysis in a large space scale. The design parameters and a simulation model with COMSOL and FLUENT for transient analysis are presented. The temperature change trends of SFCL under different fault conditions are obtained. Fault types can be judged which provides data support for fault diagnosis.

We present Mu$^{2}$SLAM, a multilingual sequence-to-sequence model pre-trained jointly on unlabeled speech, unlabeled text and supervised data spanning Automatic Speech Recognition (ASR), Automatic Speech Translation (AST) and Machine Translation (MT), in over 100 languages. By leveraging a quantized representation of speech as a target, Mu$^{2}$SLAM trains the speech-text models with a sequence-to-sequence masked denoising objective similar to T5 on the decoder and a masked language modeling (MLM) objective on the encoder, for both unlabeled speech and text, while utilizing the supervised tasks to improve cross-lingual and cross-modal representation alignment within the model. On CoVoST AST, Mu$^{2}$SLAM establishes a new state-of-the-art for models trained on public datasets, improving on xx-en translation over the previous best by 1.9 BLEU points and on en-xx translation by 1.1 BLEU points. On Voxpopuli ASR, our model matches the performance of an mSLAM model fine-tuned with an RNN-T decoder, despite using a relatively weaker sequence-to-sequence architecture. On text understanding tasks, our model improves by more than 6\% over mSLAM on XNLI, getting closer to the performance of mT5 models of comparable capacity on XNLI and TydiQA, paving the way towards a single model for all speech and text understanding tasks.