To prepare for steady-state operation of future fusion reactors (e.g. the International Thermonuclear Experimental Reactor and China Fusion Engineering Test Reactor (CFETR)), experiments on DIII-D have extended the high poloidal beta (βP) scenario to reactor-relevant edge safety factor q95 ∼ 6.0, while maintaining a large-radius internal transport barrier (ITB) using negative magnetic shear. Excellent energy confinement quality (H98y2 > 1.5) is sustained at high normalized beta (βN ∼ 3.5). This high-performance ITB state with Greenwald density fraction near 100% and qmin ≥ 3 is achieved with toroidal plasma rotation Vtor ∼ 0 at ρ ≥ 0.6. This is a key result for reactors expected to have low Vtor. At high βP (≥1.9), large Shafranov shift can stabilize turbulence leading to a high confinement state with a low pedestal and an ITB. At lower βP (<1.9), negative magnetic shear in the plasma core contributes to turbulence suppression and can compensate for reduced Shafranov shift to continue to access a large-radius ITB and excellent confinement with low Vtor, consistent with the results of gyrofluid transport simulations. These high-βP cases are characterized by weak/no Alfvén eigenmodes (a.e.) and classical fast-ion transport. At high density, the fast-ion deceleration time decreases and Δβfast is lower; these reduce a.e. drive. The reverse-shear Alfvén eigenmodes are weaker or stable because the negative magnetic shear region is located at higher radius, away from the peaked fast-ion profile. Resistive wall modes can be a limitation at simultaneous high βN, low internal inductance, and low rotation. Analysis suggests that additional off-axis external current drive could provide a more stable path at reduced q95. Based on a DIII-D high-βP plasma with large-radius ITB, two scenarios are proposed for CFETR Q = 5 steady-state operation with ∼1 GW fusion power: a lower- ( ∼ 0.66) and a higher- ( ∼ 0.75) case. Using a Landau closure model, multiple energetic particle (EP) effects on the a.e. stability are analyzed modifying the growth rate of the a.e.s triggered by the neutral-beam-injection EPs and alpha particles, although the stabilizing/destabilizing effect is weak for the cases analyzed. The stabilizing effects of the combined EP species β, energy, and density profile in CFETR need further investigation.

DIII-D physics research addresses critical challenges for the operation of ITER and the next generation of fusion energy devices. This is done through a focus on innovations to provide solutions for high performance long pulse operation, coupled with fundamental plasma physics understanding and model validation, to drive scenario development by integrating high performance core and boundary plasmas. Substantial increases in off-axis current drive efficiency from an innovative top launch system for EC power, and in pressure broadening for Alfven eigenmode control from a co-/counter-I p steerable off-axis neutral beam, all improve the prospects for optimization of future long pulse/steady state high performance tokamak operation. Fundamental studies into the modes that drive the evolution of the pedestal pressure profile and electron vs ion heat flux validate predictive models of pedestal recovery after ELMs. Understanding the physics mechanisms of ELM control and density pumpout by 3D magnetic perturbation fields leads to confident predictions for ITER and future devices. Validated modeling of high-Z shattered pellet injection for disruption mitigation, runaway electron dissipation, and techniques for disruption prediction and avoidance including machine learning, give confidence in handling disruptivity for future devices. For the non-nuclear phase of ITER, two actuators are identified to lower the L–H threshold power in hydrogen plasmas. With this physics understanding and suite of capabilities, a high poloidal beta optimized-core scenario with an internal transport barrier that projects nearly to Q = 10 in ITER at ∼8 MA was coupled to a detached divertor, and a near super H-mode optimized-pedestal scenario with co-I p beam injection was coupled to a radiative divertor. The hybrid core scenario was achieved directly, without the need for anomalous current diffusion, using off-axis current drive actuators. Also, a controller to assess proximity to stability limits and regulate β N in the ITER baseline scenario, based on plasma response to probing 3D fields, was demonstrated. Finally, innovative tokamak operation using a negative triangularity shape showed many attractive features for future pilot plant operation.

Knowledge distillation is a critical technique to transfer knowledge between models, typically from a large model (the teacher) to a more fine-grained one (the student). The objective function of knowledge distillation is typically the cross-entropy between the teacher and the student's output distributions. However, for structured prediction problems, the output space is exponential in size; therefore, the cross-entropy objective becomes intractable to compute and optimize directly. In this paper, we derive a factorized form of the knowledge distillation objective for structured prediction, which is tractable for many typical choices of the teacher and student models. In particular, we show the tractability and empirical effectiveness of structural knowledge distillation between sequence labeling and dependency parsing models under four different scenarios: 1) the teacher and student share the same factorization form of the output structure scoring function; 2) the student factorization produces more fine-grained substructures than the teacher factorization; 3) the teacher factorization produces more fine-grained substructures than the student factorization; 4) the factorization forms from the teacher and the student are incompatible. 1

Leakage is an important factor affecting the safety of the dam. In the past, manual inspection is a significant way to monitor leakage risk. However, it is time-consuming, inefficient and difficult to quantitative evaluate such as the leakage area. A semantic segmentation method based on the fully convolutional network is proposed to replace the manual inspection for the dam leakage automatic detection. Thirty-eight high-resolution images of dam leakage are collected. FCN-8s and VGG16 backbone are adopted. The results indicated that the FCN-8s achieves the mIoU to 0.59 on the test set, which proves to be an efficient way to detect the dam leakage.

No abstract