BEGIN:VCALENDAR VERSION:2.0 PRODID:-//linuxsoftware.nz//NONSGML Joyous v1.4//EN BEGIN:VEVENT SUMMARY:Machine learning for optimizing plasma resource utilization on Mar s DTSTART:20231129T140000Z DTEND:20231129T160000Z DTSTAMP:20260622T012928Z UID:5c090edb-e6a4-4e01-b924-7627d2eacf74 SEQUENCE:1 CREATED:20231127T152710Z DESCRIPTION: Abstract: On Mars\, the atmosphere plays a pivotal role in th e In-Situ Resource Utilization (ISRU) perspective\, mainly due to the abun dant atmospheric CO2. Plasma technologies\, especially the application of low-temperature plasmas (LTPs)\, have been proposed to decompose the exist ing CO2\, facilitating the recovery of O2 and CO to support life\, fuel\, and agriculture. However\, much remains to be done regarding accurate plas ma modeling\, which relies on the availability of precise reaction rate co efficients. Determining these constants is a non-trivial task\, primarily addressed by scientists with extensive experience in the field.In this the sis\, we investigate the use of machine learning techniques to predict the se parameters. This is done with an automated and systematic approach\, ba sed on the heavy-species densities of the gas'\;s final state. The tra ining data is generated using the Lisbon Kinetics simulation tool (LoKI).B oth Support Vector Regression (SVR) and Artificial Neural Network (ANN) mo dels were trained using an oxygen (O2) plasma kinetic scheme\, encompassin g 11 species and 53 heavy-species reaction processes. When evaluated on a test dataset\, the top performing SVR and ANN models achieved mean relativ e error values of 0.10% and 0.22%\, respectively. Although both models exh ibited significant accuracy\, the SVR model emerged as superior\, offering both simplicity and heightened performance for our specific low-dimension al dataset. These two machine learning approaches serve as a successful pr eliminary step towards more comprehensive models to automate the predictio n of reaction rate coefficients. LAST-MODIFIED:20231127T152710Z LOCATION:Anfiteatro GA2\, Pavilhão Central\, 0\, Alameda URL:http://df.vps.tecnico.ulisboa.pt/pt/eventos/machine-learning-for-optim izing-plasma-resource-utilization-on-mars/ X-ALT-DESC;FMTTYPE=text/html:

Abstract: O n Mars\, the atmosphere plays a pivotal role in the In-Situ Resourc e Utilization (ISRU) perspective\, mainly due to the abundant atmospheric CO2. Plasma technologies\, especially the application of low-temperature p lasmas (LTPs)\, have been proposed to decompose the existing CO2\, facilit ating the recovery of O2 and CO to support life\, fuel\, and agriculture. However\, much remains to be done regarding accurate plasma modeling\, whi ch relies on the availability of precise reaction rate coefficients. Deter mining these constants is a non-trivial task\, primarily addressed by scie ntists with extensive experience in the field.

In this thesis\, we investigate the use of machine learning te chniques to predict these parameters. This is done with an automated and s ystematic approach\, based on the heavy-species densities of the gas'\ ;s final state. The training data is generated using the Lisbon Kinetics s imulation tool (LoKI).

Both Support Vector Regression (SVR) and Artificial Neural Network (ANN) models were t rained using an oxygen (O2) plasma kinetic scheme\, encompassing 11 specie s and 53 heavy-species reaction processes. When evaluated on a test datase t\, the top performing SVR and ANN models achieved mean relative error val ues of 0.10% and 0.22%\, respectively. Although both models exhibited sign ificant accuracy\, the SVR model emerged as superior\, offering both simpl icity and heightened performance for our specific low-dimensional dataset. These two machine learning approaches serve as a successful preliminary s tep towards more comprehensive models to automate the prediction of reacti on rate coefficients.

END:VEVENT END:VCALENDAR