The project of the Central Solenoid Model Coil (CSMC) was launched in the Institute of Plasma Physics Chinese Academy of Sciences (ASIPP) to develop the manufacturing technology of large-scale superconducting magnet for the China Fusion Engineering Testing Reactor (CFETR). The maximum magnetic field of the CFETR CSMC will be 12 T when the running current is 47.65 kA, and the maximum field change rate will be 1.5 T/s. The CSMC consists of two coaxial Nb3Sn coils in the internal high field and a stack of three NbTi coils in the external low field. All the coils are electrically connected in a single circuit by joints. To verify the electrical performance of the joint, a prototype Nb3Sn-NbTi joint sample has been designed and manufactured in ASIPP. The sample is composed of two different conductor legs, the tested joint of praying hands structure to be placed in the high field region, and two upper terminals to be connected to the transformer of the test facility. The prototype sample has been tested in the SULTAN facility at Villigen (Switzerland). In this paper, the sample design, the techniques used for the manufacturing steps of the different conductor legs, and the joint assembly are described in detail. The test results show dc resistance below 5 nΩ criterion, and ac losses below 30 J/cycle with the amplitude of 0.2 T and frequencies in the range of 0.03-1 Hz.

The Central Solenoid Model Coil (CSMC) project of China Fusion Engineering Test Reactor (CFETR) has begun in 2014 to develop and verify the manufacture technology of large-scale superconducting magnets. CFETR-CSMC coil is mainly composed of inner Nb3Sn windings and outer NbTi windings. There are 12 joints in total including six inter-winding joints and six transition joints. A simple test facility is required for joint experimental evaluation. In this paper, the test facility, sample assembly, installation and calibration of Hall sensor are described in detail. The first test results of NbTi inter-winding joint are also given. The resistance of the measured NbTi inter-winding joint is about 0.55 n $\Omega$ and changes slightly in the current range from 10 kA to 20 kA.

Leveraging discrete sequence prediction for manipulation skill learning has demonstrated immense potential for complex long-horizon tasks. However, guaranteeing precise and robust execution under complex scenarios remains a core challenge for efficient autonomous robotic operation. This demands execution policies capable of generating precise and efficient action sequences based on perceptual feedback. To address the loss of fine-grained action details induced by action discretization and the difficulty in preserving long-horizon temporal dependencies in long-horizon tasks, this work proposed a hierarchical manipulation skill learning method based on residual finite scalar quantization (RFSQ). At the action representation level, a robust RFSQ with an adaptive linear scaling mechanism was introduced. It effectively mitigated residual decay in multi-level action quantization and realized accurate discrete reconstruction of continuous action sequences. At the action sequence generation level, a hybrid hierarchical architecture integrating Mamba2 and Transformer was constructed. Mamba2 efficiently captured temporal dependencies among actions, while Transformer precisely modeled fine-grained residual dependencies across action quantization hierarchies. A multimodal perception fusion module was further incorporated to strengthen instruction-semantic alignment on language-instructed tasks. A series of experiments were conducted on the LIBERO and Meta-World simulation environments. Experimental results demonstrated that the proposed method achieved outstanding performance in multi-task and long-horizon task learning.

Robots are now expected to execute increasingly complex long-horizon tasks in unstructured environments. Despite the strong potential of pretrained Vision-Language Models (VLMs) in task planning, their direct application to robotic manipulation is hindered by logical reasoning deviations and inadequate geometric scene perception. This work proposes an adaptive task planning method based on video priors and dynamic scene graphs (ATP-VPDSG). It leverages the VLM to extract manipulation logic from video demonstrations, thus supplementing manipulation priors. Meanwhile, scene graphs were integrated to convert unstructured environments into structured representations with spatial topological relations, compensating for perceptual deficiencies. A dual-track feedback mechanism based on visual expectations was further incorporated to enable failure diagnosis and adaptive replanning in complex environments. Extensive long-horizon robotic manipulation experiments were conducted on the LIBERO-10 benchmark with Qwen3-VL as the core VLM. Results showed that ATP-VPDSG achieved an average task planning accuracy of 91.2% and a task execution success rate of 74.67%, outperforming the selected task planning baselines. Ablation studies verified that video priors and dynamic scene graphs exerted complementary effects on logical constraints and physical feasibility. Furthermore, a real-robot experiment on an industrial slider–rail assembly task demonstrated successful sim-to-real transfer, achieving an 82.0% success rate without task-specific fine-tuning.

The central solenoid model coil (CSMC) project was launched in the Institute of Plasma Physics, Chinese Academy of Sciences (ASIPP). In order to develop and verify the key techniques and manufacturing process of the full-size central solenoid (CS) coil for the China Fusion Engineering Testing Reactor (CFETR). The CFETR CSMC is designed to be a superconducting magnet of 12 T maximum magnetic field when the running current is 47.65 kA, which consists of two Nb ₃ Sn windings in the internal high field and three Niobium-Titanium (NbTi) windings in the external low field. All the windings of the CSMC are a kind of pancake coil continuously wound one-in-hand by cable-in-conduit conductor (CICC), which are electrically connected in series by ohmic joints. To minimize the heat load of the joints to the cryogenic system, the design resistance of the joint is below 5 nΩ in operating conditions. The design of the joint is based on the concept of lap joint. A NbTi-NbTi full-size joint sample has been manufactured and tested. In this article, the design and manufacturing processes of the joint sample are described in detail, and the test results show dc resistance below 5 nΩ criterion.