等离子体高级氧化技术在杀菌灭毒和有机污染物处理等方面具有重要的应用，提高羟基自由基的产率是高级氧化技术的核心。本文利用文丘里管中水力空化作用产生气-液两相体，同时注入氧气进行纳秒脉冲放电，高效生成·OH。通过实验和数值模拟，本文开展了以下三方面的研究。首先，本文研究了文丘里管中两相体的流体特性，发现气液两相体区域随着水流量的增大而伸长，只有当它跨接放电间隙时，才能产生稳定可靠的击穿。而氧气流量的增大将改变两相体的流形和压力，使之从“泡状流”过渡到“环状流”。在上述研究基础上，本文研究了流体特性和电源参数对放电特性的影响，发现并解释了以下实验现象。该两相体击穿电压随着氧气流量的增大呈“V”形变化曲线，并且击穿电压存在很强的极性效应，即正极性击穿电压远低于负极性击穿电压。击穿时约化场强随氧气流量显著变化，约化场强最高可以达到2000 Td。当脉冲电压幅值为20 kV时，随着氧气流量的增大，该两相体可以呈现辉光放电、火花放电和电晕放电三种不同放电模式。即使在1 kHz的高重复频率下,该两相体纳秒脉冲击穿电压也仅略微下降。根据上述研究结果，本文预估了高效产生·OH的最佳实验条件（9.5 L/min水流量、3 L/min氧气流量、45 kV和1 kHz的负极性纳秒脉冲电压）。其次，本文研究了实验条件对·OH产率的影响。相较于无氧气注入，3 L/min氧气注入使·OH产率与能效提升约6倍。其原因是：除了氧活性粒子直接同水分子反应产生·OH之外，注氧放电后产生了高浓度过氧化氢与臭氧，引发过臭氧化反应（Peroxone）而产生大量的·OH。实验还验证了高重复频率（1 kHz）放电不会降低生成·OH的能效。确定了高效产生·OH的最佳实验条件，它们和通过放电特性预估的结果完全一致。最后，本文开展了两相体纳秒脉冲放电处理有机染料与微生物污染的应用研究，该技术降解有机污染物属于典型的一级反应，具有很高的能量效率。在最佳工况下，实现了大肠杆菌污染水体的一次通过灭菌，·OH是大肠杆菌速效杀灭的关键因子，其杀灭动力学满足Double-Weibull模型。

Plasma-based advanced oxidation processes (AOPs) play a significant role in sterilization and wastewater treatment, with the enhancement of hydroxyl radical (·OH) production being central to their efficacy. In this study, we employed a Venturi tube to generate gas-liquid two-phase flow through hydrodynamic cavitation, simultaneously injecting oxygen for nanosecond pulsed discharge, resulting in efficient ·OH production. This research consists of three major components: experimental investigation, numerical simulation, and application study.Firstly, we examined the fluid characteristics of the two-phase flow in the Venturi tube, finding that the gas-liquid interface extends with increasing water flow. Stable and reliable breakdown only occurs when the interface spans the discharge gap. An increase in oxygen flow leads to a change in flow pattern and pressure, transitioning from a bubbly flow to a slug flow. Building on these findings, we explored the influence of fluid characteristics and power supply parameters on discharge properties. We observed and explained the following phenomena: the breakdown voltage of the two-phase flow exhibits a "V"-shaped curve with increasing oxygen flow, and a strong polarity effect is present, with the positive breakdown voltage significantly lower than the negative breakdown voltage. The reduced field strength at breakdown varies with oxygen flow, reaching a maximum of 2000 Td. At a pulse voltage amplitude of 20 kV, the two-phase flow demonstrates three distinct discharge modes—glow discharge, spark discharge, and corona discharge—as oxygen flow increases. Even at a high repetition frequency of 1 kHz, the breakdown voltage of the two-phase flow experiences only a slight decrease. Based on these findings, we estimated the optimal experimental conditions for efficient ·OH production (9.5 L/min water flow, 3 L/min oxygen flow, 45 kV, and 1 kHz negative polarity nanosecond pulsed voltage).Secondly, we investigated the impact of experimental conditions on ?OH production efficiency. The introduction of a 3 L/min oxygen flow led to a sixfold increase in ·OH yield and energy efficiency compared to a condition without oxygen injection. This is due to the direct reaction between oxygen active particles and water molecules to produce ·OH, as well as the formation of high-concentration hydrogen peroxide and ozone, which triggers the Peroxone reaction, generating abundant ·OH. It was also confirmed that a high repetition frequency (1 kHz) discharge does not decrease the energy efficiency of ·OH generation. The optimal experimental conditions for efficient ?OH production were determined, which are in complete agreement with the results predicted from the discharge characteristics analysis.Lastly, we conducted application studies on the treatment of organic dye and microbial contamination using gas-liquid two-phase nanosecond pulsed discharge. This technique exhibits high energy efficiency in degrading organic pollutants, following a characteristic first-order reaction. Under optimal conditions, a single pass-through effectively sterilizes water contaminated with Escherichia coli, with ·OH serving as the key factor for rapid inactivation. The inactivation kinetics adhere to the Double-Weibull model.