This research aimed to analyze the effects of ultrasound on the quality characteristics of white wine when processed by two different systems, i.e., ultrasonic bath and ultrasonic probe. In this regard, the multivariate statistical analysis and artificial neural network (ANN) techniques were used. Additionally, the efficiency of high power ultrasound (HPU) combined with sulfite and glutathione (GSH) treatments was explored during 18 months of bottle storage. Regarding ultrasonic bath experiment, the higher bath temperature caused the degradation of volatile compounds, precisely esters and higher alcohols, while the ultrasound effect on phenolic composition was much less pronounced. Interestingly, a combination of larger probe diameter and higher ultrasound amplitude showed a milder effect on phenolic and volatile composition in ultrasonic probe experiment. Both, ultrasonic bath and probe experiments did not cause great changes in the color properties. Moreover, implemented ANN models for flavan-3-ols, higher alcohols and esters resulted in the highest prediction values. HPU processing after 18 months of storage did not affect wine color. However, it modified phenolic and volatile composition, with greater effect in wines with lower concentration of antioxidants. In addition, there was no significant difference in the phenolic and volatile composition among sonicated low-sulfite-GSH wine and the one with standard-sulfite content. Therefore, a combined HPU and low-sulfite-GSH treatment might be a promising method for production of low-sulfite wines.

这项研究旨在分析超声对超声酒和超声探头两种不同系统处理白葡萄酒质量特性的影响。在这方面，使用了多元统计分析和人工神经网络（ANN）技术。此外，在瓶存放的18个月中，探索了高功率超声（HPU）结合亚硫酸盐和谷胱甘肽（GSH）处理的效率。对于超声波浴实验，较高的浴温会引起挥发性化合物，准确地说是酯和高级醇的降解，而超声波对酚类成分的影响要小得多。有趣的是 在超声探头实验中，较大的探头直径和较高的超声振幅相结合，对酚和挥发性成分的影响较小。超声浴和探针实验均未引起颜色性质的巨大变化。此外，针对黄烷-3-醇，高级醇和酯的已实施ANN模型得出最高预测值。储存18个月后的HPU加工不会影响酒的颜色。但是，它可以改变酚类和挥发性成分，在抗氧化剂浓度较低的葡萄酒中效果更好。此外，超声处理的低亚硫酸盐-GSH酒与标准亚硫酸盐含量的酒之间的酚和挥发性成分无显着差异。因此，