Progress of the development of an all-electric control system of a rocket engine

Chopinet, Jean-Noel; Lassoudiere, Francois; Roz, G.; Faye, O.; Gonidec, Serge Le; Alliot, Patrick; Guedron, Sylvain

doi:10.2514/6.2012-3873

Cited by 3 publications

(5 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This failure mode can be averted by smoothing out the Q-function over similar actions. For this, one computes the action that is used to form the Q-learning target in the following way: u (x ) = clip(π(x ; w − ) + clip( , −c, c), x Low , x High ), (9) where ∼ N (0, σ) is noise sampled from a Gaussian process if it is time to update then 10:…”

Section: Reinforcement Learningmentioning

confidence: 99%

“…Although the importance of closed-loop control has been evident for many years, the majority of rocket engines still employ valves which are operated with pneumatic actuators, too inefficient for a sophisticated closed-loop control system. The development of an all-electric control system started in the late 90s in Europe [9]. The future European Prometheus engine will have such a system [10].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Reinforcement Learning Approach for Transient Control of Liquid Rocket Engines

Waxenegger-Wilfing¹,

Dresia²,

Deeken³

et al. 2021

IEEE Trans. Aerosp. Electron. Syst.

View full text Add to dashboard Cite

Nowadays, liquid rocket engines use closed-loop control at most near steady operating conditions. The control of the transient phases is traditionally performed in open-loop due to highly nonlinear system dynamics. This situation is unsatisfactory, in particular for reusable engines. The open-loop control system cannot provide optimal engine performance due to external disturbances or the degeneration of engine components over time. In this paper, we study a deep reinforcement learning approach for optimal control of a generic gas-generator engine's continuous start-up phase. It is shown that the learned policy can reach different steady-state operating points and convincingly adapt to changing system parameters. Compared to carefully tuned open-loop sequences and PID controllers, the deep reinforcement learning controller achieves the highest performance. In addition, it requires only minimal computational effort to calculate the control action, which is a big advantage over approaches that require online optimization, such as model predictive control.

show abstract

Section: Reinforcement Learningmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A Reinforcement Learning Approach for Transient Control of Liquid Rocket Engines

Waxenegger-Wilfing¹,

Dresia²,

Deeken³

et al. 2021

IEEE Trans. Aerosp. Electron. Syst.

View full text Add to dashboard Cite

show abstract

“…In Q-learning, one starts from an arbitrary initial Q-function Q 0 and updates it using observed state transitions and rewards. The update rule is of the following form: (8) where α k ∈ (0, 1] is the learning rate. The term inside the square bracket is nothing else than the difference between the updated estimate of the optimal Q-value of (x k , u k ) and the current estimate Q k (x k , u k ).…”

Section: Reinforcement Learningmentioning

confidence: 99%

“…Although the importance of closed-loop control has been evident for many years, the majority of rocket engines still employ valves which are operated with pneumatic actuators, too inefficient for a sophisticated closed-loop control system. The development of an all-electric control system started in the late 90s in Europe [8]. The future European Prometheus engine will have such a system [9].…”

Section: Introductionmentioning

confidence: 99%

A Reinforcement Learning Approach for Transient Control of Liquid Rocket Engines

Waxenegger-Wilfing,

Dresia,

Deeken

et al. 2020

Preprint

View full text Add to dashboard Cite

Nowadays, liquid rocket engines use closedloop control at most near steady operating conditions. The control of the transient phases is traditionally performed in open-loop due to highly nonlinear system dynamics. This situation is unsatisfactory, in particular for reusable engines. The open-loop control system cannot provide optimal engine performance due to external disturbances or the degeneration of engine components over time. In this paper, we study a deep reinforcement learning approach for optimal control of a generic gas-generator engine's continuous start-up phase. It is shown that the learned policy can reach different steady-state operating points and convincingly adapt to changing system parameters. A quantitative comparison with carefully tuned open-loop sequences and PID controllers is included. The deep reinforcement learning controller achieves the highest performance and requires only minimal computational effort to calculate the control action, which is a big advantage over approaches that require online optimization, such as model predictive control.

show abstract

“…However, electrical actuators are gaining importance in valve-position control with respect to traditional pneumatic ones. The main reasons are that auxiliary helium-gas consumption and costs can be reduced and throttling efficiency is improved [59]. That is why the new hardware-in-the-loop (HIL) simulation platform by CNES and ArianeGroup [60] includes real valve-internal electric actuators.…”

Section: Sensors and Actuators Considerationsmentioning

confidence: 99%

A survey of automatic control methods for liquid-propellant rocket engines

Pérez-Roca

Marzat

Piet-Lahanier

et al. 2019

Progress in Aerospace Sciences

View full text Add to dashboard Cite

The main purpose of this survey paper is to review the field of convergence between the liquid-propellant rocketpropulsion and automatic-control disciplines. A comprehensive collection of academic works and some industrial developments are summarised and discussed, making the link to the current context of launcher reusability. The main control problem in liquid-propellant rocket engines (LPRE), the target of this survey, generally consists in tracking set-points in combustion-chamber pressure and mixture ratio, their main operating quantities. This goal is attained via the adjustment of flow-control valves while complying with operating constraints. Each aspect of control systems, ranging from modelling to the actual control techniques, is reviewed and subsequently related to the rest of aspects. The comparison of the different approaches points to the general use of linearised models about concrete operating points for deriving steady-state controllers, which in most cases are of PID type. Other more sophisticated approaches present in the literature, incorporating some nonlinear, hybrid or robust techniques, improve certain aspects of performance and robustness. Nevertheless, no fully nonlinear or hybrid frameworks, which may allow the control of a wider throttling domain, have been published. In addition, control during the thermally and structurally demanding transient phases is usually in open loop, and hence with low correction margins. This sort of control enhancements, probably together with more sophisticated monitoring architectures, would serve to extend the life of LPRE, a significant factor in their reusability. Therefore, this survey intends to draw a path for the future control design trends which will certainly be more suitable for reusable LPRE.

show abstract

Progress of the development of an all-electric control system of a rocket engine

Cited by 3 publications

References 0 publications

A Reinforcement Learning Approach for Transient Control of Liquid Rocket Engines

A Reinforcement Learning Approach for Transient Control of Liquid Rocket Engines

A Reinforcement Learning Approach for Transient Control of Liquid Rocket Engines

A survey of automatic control methods for liquid-propellant rocket engines

Contact Info

Product

Resources

About