Superconducting Levitation Temperature Detection via Deep Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting temperature rise in superconducting levitation devices are invasive, damaging the superconductor and prone to measurement errors due to magnetic field sensitivity.
Innovation Solution
A deep learning-based method that extracts vibration acceleration features, calculates wavelet band energy, and trains a detection model to predict internal temperature rise without the need for contact sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If temperature sensors are deployed inside the superconductor to directly measure temperature rise, then measurement accuracy is improved, but the structure of the superconductor is damaged, impacting levitation performance
Solution Approach 1:
The patent uses vibration acceleration signals as an intermediary to indirectly infer temperature rise inside the superconductor. Instead of directly measuring temperature with sensors inside the superconductor, the system measures vibration acceleration on the outer surface and uses deep learning to predict internal temperature, thus avoiding structural damage while achieving accurate temperature monitoring
Solution Approach 2:
The patent replaces the mechanical/physical temperature sensing system with a computational intelligence system. Instead of using physical temperature sensors that require contact with the superconductor, the system uses deep learning algorithms to process vibration acceleration signals and predict temperature rise, substituting direct physical measurement with indirect computational inference
2Difficulty of detecting and measuring
If platinum resistance sensors are used for temperature measurement, then temperature detection capability is improved, but measurement errors increase due to sensitivity to magnetic field fluctuation
Solution Approach 1:
The patent introduces vibration acceleration signals as an intermediary measurement parameter that is not sensitive to magnetic field fluctuations. By measuring vibration characteristics on the outer surface and using deep learning to infer internal temperature, the system avoids the magnetic field sensitivity problem of platinum resistance sensors while maintaining temperature detection capability
Solution Approach 2:
The patent changes the measurement parameter from direct temperature measurement (using magnetic-field-sensitive platinum resistance sensors) to vibration acceleration measurement (which is not affected by magnetic fields). This parameter transformation allows temperature inference without the drawbacks of magnetic field sensitivity
3Ease of operation
If the superconductor is drilled to deploy temperature sensors, then temperature measurement access is improved, but structural integrity deteriorates
Solution Approach 1:
The patent uses outer surface vibration acceleration measurements as an intermediary to infer internal temperature conditions. This approach provides temperature measurement access without requiring physical penetration or drilling of the superconductor structure, thereby maintaining structural integrity while achieving the measurement goal
Solution Approach 2:
The patent creates a virtual model of internal temperature distribution by copying and processing external vibration acceleration signals through deep learning algorithms. This virtual copying allows temperature information to be obtained without physical intrusion into the superconductor structure
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides non-contact, high-accuracy real-time temperature rise detection, preserving the levitation performance and eliminating the need for temperature sensors and specific installation space.
Implementation Method 1
performing wavelet band energy calculation for the high-frequency feature parameter set and the low-frequency feature parameter set to obtain a wavelet band energy information
Data Source
AI summary
A method for detecting temperature rise inside a superconducting levitation device based on deep learning is provided. An initial vibration acceleration information, an initial temperature rise information, and a vibration acceleration detection information are obtained. Feature extraction is performed on the initial vibration acceleration information to obtain a high-frequency feature parameter set and a low-frequency feature parameter set. Wavelet band energy calculation is performed for the high-frequency feature parameter set and the low-frequency feature parameter set to obtain a wavelet band energy information. The wavelet band energy information and the initial temperature rise information are input into a preset deep learning network for training to obtain an internal temperature rise detection model of the superconducting levitation device. The vibration acceleration detection information is input into the internal temperature rise detection model to obtain an internal temperature rise prediction information to reflect real-time temperature rise of the superconductor.