High Fidelity Multi-Objective Design Optimization of a Downscaled Cusped Field Thruster

Abstract

The Cusped Field Thruster (CFT) concept has demonstrated significantly improved performance over the Hall Effect Thruster and the Gridded Ion Thruster; however, little is understood about the complexities of the interactions and interdependencies of the geometrical, magnetic and ion beam properties of the thruster. This study applies an advanced design methodology combining a modified power distribution calculation and evolutionary algorithms assisted by surrogate modeling to a multi-objective design optimization for the performance optimization and characterization of the CFT. Optimization is performed for maximization of performance defined by five design parameters (i.e., anode voltage, anode current, mass flow rate, and magnet radii), simultaneously aiming to maximize three objectives; that is, thrust, efficiency and specific impulse. Statistical methods based on global sensitivity analysis are employed to assess the optimization results in conjunction with surrogate models to identify key design factors with respect to the three design objectives and additional performance measures. The research indicates that the anode current and the Outer Magnet Radius have the greatest effect on the performance parameters. An optimal value for the anode current is determined, and a trend towards maximizing anode potential and mass flow rate is observed.

Particle‐in‐cell simulation of an optimized high‐efficiency multistage plasma thruster

“…The specific impulse with the mass flow of the neutral source and g = 9.81 m/s 2 , can be calculated directly from the obtained thrust. The efficiency η t in the MDO is the anode efficiency with the anode voltage U a and the anode current I a .…”

“…In the MDO the mean ion emission angle, energy transfer coefficients and cusp arrival probabilities of the electrons, were treated as constant input parameters obtained from experimental measurements of a different HEMP-T model. [3,10] The different thruster geometry and operating conditions makes the validity of this approach uncertain. This is confirmed by the differences between the predicted performance parameters from the MDO and PIC, seen in Table 2.…”

“…Each set of these basic parameters defines a different thruster design with a different performance. An MDO study was successfully conducted by coupling evolutionary algorithms with performance analysis incorporating magnetic field simulation, achieving robust global optimization to identify key design parameters and mechanism for HEMP‐T scaling . The performance was estimated by using a zero‐dimensional (0D) power balance equation system for particles, energy, and momentum, but some parameters inevitably had to rely on empirical assumptions, rendering it difficult to ensure self‐consistency in the system.…”

“…In addition other important effects are neglected in the 0D power balance model, especially the influence of neutral gas distribution, ionization efficiency, doubly charged ions, and the details of the ion beam structure. This does not lead to self‐consistent solutions of the MDO, but with the application of empiric correction terms, which take into account the neglected effects, the MDO derived several thruster designs, among which a most promising design, called the S 1 thruster, is considered in the present study.…”

“…An MDO study was successfully conducted by coupling evolutionary algorithms with performance analysis incorporating magnetic field simulation, achieving robust global optimization to identify key design parameters and mechanism for HEMP-T scaling. [10] The performance was estimated by using a zero-dimensional (0D) power balance equation system for particles, energy, and momentum, [3] but some parameters inevitably had to rely on empirical assumptions, rendering it difficult to ensure self-consistency in the system. Empirical assumptions in this case include constant power transfer coefficients, which were originally assumed for a different HEMP-T geometry.…”