Modeling and optimization of argon-based floating helix electrode cold plasma.

Area of Science:

• Biomedical Engineering
• Plasma Physics
• Machine Learning

Background:

• Cold atmospheric pressure plasma (CAP) offers non-thermal treatment crucial for biomedical applications.
• Existing CAP devices require precise control to maintain the non-thermal regime.

Purpose of the Study:

• To develop a machine learning-based model for process optimization of a novel floating-helix electrode CAP device.
• To ensure the device operates within the safe cold plasma regime for biomedical use.

Main Methods:

• Developed an artificial neural network (ANN) model to correlate process parameters (supply voltage, frequency) with performance metrics (power consumption, jet lengths).
• Employed the composite desirability method for multi-response optimization.
• Utilized logistic regression machine learning models (ANN classifier, K-NN, SVM) to classify discharge types.

Main Results:

• The ANN model demonstrated generality and robustness, validated through experimental data and predictions.
• The machine learning classifiers successfully identified discharge types within the cold plasma operating range.
• The study established a framework for optimizing CAP devices for biomedical applications.

Conclusions:

• The coupled machine learning and statistical approach effectively models and optimizes CAP devices.
• The validated models ensure operation within the non-thermal plasma regime, critical for safety in biomedical applications.
• Further biological testing is required to confirm the potential of this CAP system for biomedical use.