Optimizing Plasma Characterization and Control Using Dynamic Mode Decomposition: A Data-Driven Approach to Efficient Analysis

The characterization of plasma systems often demands extensive data generation and post-processing, leading to significant computational overhead and time consumption. Traditional approaches, requiring numerous simulations, are particularly tedious when exploring the full parameter space for comprehensive plasma analysis. Pulsed waveforms, such as Gaussian waves, offer notable advantages over continuous waveforms by enabling granular control of the plasma profile within the reactor. However, the sheer volume of data required to study all possible waveform configurations renders exhaustive analysis impractical. Dynamic Mode Decomposition (DMD) offers a robust solution to these challenges. Originally developed for fluid dynamics, DMD is a data-driven algorithm that decomposes complex dynamical systems into spatiotemporal modes, providing deep insights into system behavior. This method has successfully expanded into various fields, including plasma physics. DMD extracts dominant modes that capture the system's essential features by approximating linear dynamics that best fit the observed data. Its data-agnostic nature allows DMD to handle large datasets effectively, uncovering underlying patterns and dynamics without relying on specific physical models. This flexibility is crucial for efficiently exploring various plasma behaviors and operational scenarios. In this study, we leverage DMD to reduce computational overhead in plasma research, specifically targeting the behavior of glow discharge plasmas under different operational conditions. By training the DMD model on a subset of simulation data, we can predict the plasma's response to untested parameter sets, significantly saving resources and accelerating the research process. This approach enables rapid insights into the effects of various parameters, optimizing system performance while maintaining stability. Our findings demonstrate that DMD can predict plasma behavior accurately, making it an invaluable tool for future plasma studies and applications.