Enhanced reproducibility of L-mode plasma discharges via physics-model-based q -profile feedback control in DIII-D

Self Cite

This paper describes a real-time capable algorithm for identifying the safe operating region around a tokamak operating point. The region is defined by a convex set of linear constraints, from which the distance of a point from a disruptive boundary can be calculated. The disruptivity of points is calculated from an empirical machine learning predictor that generates the likelihood of disruption. While the likelihood generated by such empirical models can be compared to a threshold to trigger a disruption mitigation system, the safe operating region calculation enables active optimization of the operating point to maintain a safe margin from disruptive boundaries. The proposed algorithm is tested using a random forest disruption predictor fit on data from DIII-D. The safe operating region identification algorithm is applied to historical data from DIII-D showing the evolution of disruptive boundaries and the potential impact of optimization of the operating point. Real-time relevant execution times are made possible by parallelizing many of the calculation steps and implementing the algorithm on a graphics processing unit. A real-time capable algorithm for optimizing the target operating point within the identified constraints is also proposed and simulated.

Section: Algorithm Implementation On a Gpumentioning

confidence: 99%

Toward active disruption avoidance via real-time estimation of the safe operating region and disruption proximity in tokamaks

Boyer¹,

Rea²,

Clement³

2021

Self Cite

“…the current profile evolution [13,14], and rotation profile evolution [15]. These models have enabled the development of non-linear control schemes [16,17], and model predictive controllers [18,19], which have been demonstrated in simulations and experiments. While reduced physics models have the advantage of being interpretable, it can be challenging to derive reduced models that accurately incorporate all of the important phenomena and interactions, e.g.…”

Section: Model Based Control and High Fidelity Closed Loop Simulationmentioning

confidence: 99%

Model predictive control of KSTAR equilibrium parameters enabled by TRANSP

Boyer¹,

Yuan²,

Ahn³

et al. 2020

Self Cite

Due to the complex behavior of tokamak plasmas and the importance of optimizing performance while avoiding instabilities and machine limits, plasma control algorithms are becoming increasingly dependent on sophisticated model-based control approaches. It is anticipated that the use of integrated modeling codes in the model-based control design process will reduce the amount of experimental time needed to implement new control algorithms by facilitating development of control-oriented models and enabling higher-fidelity closed-loop simulations. In this work, a reduced model is developed from a series of TRANSP simulations and is used to develop a model predictive control (MPC) algorithm for controlling important equilibrium parameters in KSTAR [] discharges. The control algorithm uses the KSTAR neutral beam injection system and the target plasma current and plasma boundary as actuators, and optimizes the plasma stored energy, loop voltage, and internal inductance while avoiding constraints that could lead to disruptions. Higher fidelity testing of the control algorithm is performed using a flexible framework for enabling external processes to actively control plasma parameters in TRANSP simulations. Closed-loop simulations demonstrate the ability of the control algorithm to respond to disturbances in density and confinement, handle actuator failures, and move the discharge to high non-inductive fraction conditions.

“…Integrated modeling suites normally consist of many physics modules, spanning magnetic fusion research areas from equilibrium, stability, heating, and current drive to transport and scrape-off layer physics. In other advanced control approaches for fusion plasmas, such as model predictive control in DIII-D [8,9], TCV [10,11], and Tore Supra [12], control-oriented models obtained from more sophisticated models in physics-oriented software are used for control synthesis. On the other hand, the approach introduced in this paper does not require any simplification of the models because they are simply used to generate NB values at which thermal diffusivity profiles are defined as interpolation sources.…”

Section: Introductionmentioning

confidence: 99%

Ion temperature gradient control using reinforcement learning technique

Wakatsuki¹,

Suzuki²,

Oyama³

et al. 2021

Plasma with an internal transport barrier (ITB) is desirable for a steady-state tokamak reactor because of its high confinement quality and high bootstrap current fraction. However, the local pressure gradient tends to be steep and the plasma often becomes unstable. In this study, an ion temperature gradient control system based on neutral beam injection (NBI) is developed using the reinforcement learning technique. The response characteristics of an ion temperature gradient to NBI are non-linear and sensitive to experimental conditions, which makes it difficult to develop a robust control system. Our control system is trained for plasmas with a wide range of ITB strengths. Using the reinforcement learning technique, the system acquires a robust control feature through several thousand iterations of trial and error in an integrated transport simulation hosted by TOPICS. The control system is composed of neural networks (NNs) whose input variables are the ion temperature gradient, the current NBI power, and the NBI powers for several previous control time steps. The trained system can determine a control output which is suitable for the response characteristics inferred from the input variables. The trained control system is tested in the TOPICS simulation using plasma models based on two experimental plasmas of JT-60U with different ITB strengths. It is shown that the ion temperature gradient can be appropriately controlled for both plasmas, which supports the expectation that this system is applicable to real experiments.