Plasma optical emission spectroscopy based on feedforward neural network

Abstract

Optical emission spectroscopy (OES) has been widely applied to plasma etching, material processing, development of plasma equipment and technology, as well as plasma propulsion. The collisional-radiative model used in OES is affected by the deviation of fundamental data such as collision cross sections, thus leading to the error in diagnostic results. In this work, a novel method is developed based on feedforward neural network for OES. By comparing the error characteristics of the new method with those of the traditional least-square diagnostic method, it is found that the neural network diagnosis method can reduce the transmission of basic data deviation to the diagnosis results by identifying the characteristics of the spectral vector. This is confirmed by the experimental results. Finally, the mechanism of the neural network algorithm against fundamental data deviation is analyzed. This method also has a good application prospect in plasma parameter online monitoring, imaging monitoring and mass data processing.

A neural network model relating extraction current characteristics with optical emission spectra for the purpose of a digital twin of miniaturized ion thrusters

“…In practice, the propagation of noises on different lines can enhance the error of the LSQ method. The LSQ method is generally sensitive to noise, as found previously [37], and may not be suitable for the digital twin method.…”

“…The architecture of the neural network (number of hidden layers and number of neurons in each layer) is optimized by a performance test. According to the experience in [37], this work starts testing from a two-hidden-layer network and finds it does not provide satisfactory performance. By adding hidden layers, better performance is obtained with three-hidden-layer and four-hidden-layer networks.…”

“…Recently, neural network and digital twin methods have been widely developed for the diagnostics, modeling, and control of plasma devices [33][34][35][36][37][38][39]. The NNM developed in this work is a component of the ion thruster's digital twin, which describes the thrust and extraction characteristics of the thruster.…”

A minimized electron cyclotron resonance ion thruster for China’s space-borne gravitational wave detection missions

“…At this point, the ECR discharge is self-sustaining stably although it is tuned to the condition of Xe flow rate q = 0.01 sccm and input microwave power P = 0.5 W. In contrast, the emission intensity is noticeably bright for the discharge and plume at 100 µN in figures 4(c) and (d). In addition, the luminescence intensity could be analyzed quantitatively by the optical emission spectroscopy method [51][52][53], which will be reported in the subsequent works. The maximum beam current increases to 2.3 mA at the operating condition of q = 0.06 sccm, P = 1.5 W and U s = 1000 V. In addition, the maximum ionization efficiency exceeds 65% when q = 0.03 sccm.…”