Two Applications of Machine Learning: Plasma Dosimetry and Optimizing Plasma-Assisted Nitrogen Fixation

报告题目：Two Applications of Machine Learning: Plasma Dosimetry and Optimizing Plasma-Assisted Nitrogen Fixation

报告人：卢新培教授

报告时间：2024年11月28日上午8:00

报告地点：物理科技楼409

报告摘要：In recent years, the application of machine learning in low-temperature plasma research has garnered significant attention. By leveraging limited experimental data, machine learning enables us to uncover complex relationships between multidimensional adjustable parameters and experimental outcomes. This talk presents two case studies: plasma dosimetry and optimizing energy efficiency in plasma-assisted nitrogen fixation (NF). In the first part, we propose a machine learning-based method for modeling Equivalent Total Oxidation Potential (ETOP). Using Laser-Induced Fluorescence (LIF), we collected reactive oxygen and nitrogen species (RONS) density data under various conditions. An Artificial Neural Network (ANN) was then trained to predict ETOP values based on input parameters such as voltage, gas flow rate, oxygen concentration, and humidity. This method facilitates efficient ETOP prediction across diverse conditions, paving the way for standardization and clinical applications in plasma medicine. In the second part, we trained a Multi-Layer Perceptron (MLP) model using experimental data to predict NF energy efficiency and NOx concentrations. The MLP model's hyperparameters were fine-tuned using the Tree-structured Parzen Estimator (TPE) algorithm to ensure high predictive accuracy. Additionally, gradient analysis was employed to evaluate the influence of feature variables on the predictions. The results demonstrate that this approach effectively addresses the challenge of predicting NF performance under multiparameter conditions, offering a novel pathway for optimizing NF processes. This work highlights the transformative potential of machine learning in advancing plasma research and applications.

报告人简介：卢新培，华中科技大学教授，长江学者特聘教授，国家杰出青年基金获得者。以第一作者和通讯作者发表SCI论文180余篇，包括Physics Reports 2篇（影响因子22.9），17000余次引用，H因子65。7篇ESI高被引/热点论文。入选爱思唯尔(Elsevier)2014至今中国高被引学者榜单。由CRC出版社出版编著和专著各一本。应IEEE Trans. On Plasma Sci.邀请发起和主持了第一、二、三期《等离子体射流及其应用专刊》。研制出的国际上首个干电池驱动空气等离子体手电、国际上唯一能在牙齿根管内产生等离子体的装置等成果多次分别被 Nature、Science、美国物理协会等作为研究亮点报道。在牙齿根管内产生等离子体的射流装置被写进等离子体医学的专著《Plasma Medicine》。应威斯康辛大学麦迪逊分校 J. Shohet 教授邀请撰写了《等离子体技术百科全书》中“等离子体用于根管治疗”一章。担任IEEE Tran. Radiation Plasma Medical Sci.，Plasma Processes and Polymers, Plasma Research Express, Applied Sciences, Fundamental Research, Plasma Science & Technology, High Voltage编委，IEEE Tran. On Plasma Sci.客座编辑，Europe Physics Letter客座编辑。