Importance Score: 78 / 100 🔴
Miles Space Unveils Innovative Low-Power, Water-Fueled Electric Thruster
SAN FRANCISCO – Miles Space has presented a groundbreaking water-fueled electric thruster, notable for its exceptionally low power consumption. This novel space propulsion technology could transform mission capabilities, especially for small satellites.
Poseidon M1.5 Thruster Shows Promise in Orbit
“We possess a water-vapor-based plasma thruster that operates at such minimal power levels, it challenges conventional expectations,” stated Miles Space CEO Brad Berkson in an interview with SpaceNews.
During trials conducted in September 2024 on a European satellite, Miles Space’s Poseidon M1.5 thruster, designed for one-unit cubesat integration, validated its effectiveness for applications including deorbiting from low-Earth orbit, a maneuver requiring extended thrust duration, according to an independent engineer who reviewed raw telemetry data.
Key Performance Indicators
The compact, 10-centimeter-cube M1.5 thruster achieved a thrust of 37.5 millinewtons over a five-minute interval, demonstrating a specific impulse of 4,800 seconds while drawing only 1.5 watts of power. This level of efficiency marks a significant advancement in electric propulsion systems.
In comparison, traditional hydrazine propulsion delivers significantly higher thrust, but at a reduced specific impulse. Hall-effect thrusters, while capable of comparable thrust, typically demand substantially greater power resources, potentially hundreds of times more.
“This technology unlocks the potential for high Delta V missions previously unattainable due to limitations in either excessive propellant mass or prohibitive thruster power requirements,” explained the engineer, who requested anonymity due to employer restrictions on discussing the propulsion technology.
AI-Driven Development
Wesley Faler, chief technology officer at Miles Space, based in Florida, spearheaded the development of the M1.5 thruster over several years, leveraging artificial intelligence in the design process. After defining parameters for the reaction chamber, including dimensions, inlet positions, exit voltage, and timing, Faler employed genetic algorithms to simulate numerous configurations.
“The simulations yielded thrust and fuel-efficiency data,” Faler elaborated. “I then evaluated the data points, identifying the most promising results, and the system utilized a neural network to interpret my feedback. Subsequently, this trained neural network autonomously conducted simulations in my absence.”
Faler routinely reviewed the most encouraging simulations, providing iterative feedback every few days to refine the design.
“Genetic algorithms iteratively refined the thruster’s shape,” Faler clarified. “Their performance was assessed by a three-layer neural network that learned from human input. Neither of these individually represents state-of-the-art AI as defined today, but their integration was crucial for rapid development.”