2 results about "Superconducting magnet" patented technology

This invention discloses a low-temperature high thermal conductivity polymer composite material and its preparation method, belonging to the field of magnetic confinement nuclear fusion technology. This invention uses a specific process to composite a low-temperature high thermal conductivity filler with a polymer matrix. Compared to the microstructure of conventional polymer thermally conductive composite materials, this material has a quasi-continuous thermally conductive network formed by the low-temperature high thermal conductivity filler surrounding the polymer. At low temperatures (e.g., 4K), when the volume percentage of the thermally conductive filler is 50%, it can reach 100 W·m³. ‑1 ·K ‑1 The low-temperature high thermal conductivity polymer composite material of this invention exhibits significantly higher thermal conductivity than other polymer composite materials under the same thermally conductive filler volume percentage. It can be applied in low-temperature heat and mass transfer fields such as superconducting magnets, and possesses advantages such as lightweight, low-temperature resistance, and ease of processing. Furthermore, the preparation method of the low-temperature high thermal conductivity polymer composite material of this invention is simple and reliable, offering significant advantages in mass production.

The application relates to the technical field of long-time sequence prediction, and discloses a superconducting magnet multi-step temperature prediction method based on an improved whale optimization algorithm, which comprises the following steps: pre-processing original data sets related to the temperature of a superconducting magnet; constructing a whale optimization algorithm according to the logic of three kinds of hunting behaviors of a whale group hunting, and initializing a whale population; improving the whale optimization algorithm by combining Cauchy mutation and double adaptive weights; inputting a training data set into a long-short term sequence model, combining the improved whale optimization algorithm to optimize and adjust the long-short term sequence model super parameters, and updating the best long-short term sequence model super parameters; using the updated best super parameters to directly-recursively perform multi-step temperature prediction on a verification set, comparing the temperature prediction result with actual temperature data to check the prediction accuracy; and using the updated best super parameters to directly-recursively perform multi-step temperature prediction on a test set, comparing the temperature prediction result with actual training data to evaluate the prediction accuracy.