Nonlinear Levitation-Guidance Coupling Force Prediction for HTS Pinning Maglev Under Arbitrary Motion Based on Gated Recurrent Unit

By the merits of self-stability, low energy consumption, and no-mechanical friction, high temperature superconducting (HTS) pinning magnetic levitation (maglev) occupies a key position in high-speed rail transit. As the guarantee to achieve its stable operation, the accurate description of nonlinear levitation-guidance coupling force for HTS pinning maglev is the most crucial. However, existing dynamic studies always adopted hyperelastic model to illustrate coupling forces, simplified the vertical-lateral coupling effect and constrained the moving range of HTS bulk to a regularly hysteresis cyclic motion. This simplification makes the original expression inaccurate under practical operation, namely, the random motion under track irregularity excitation. Hence, based on the factors affecting HTS pinning maglev forces under random motion, 12 basic motions (BMs) are extracted from either complicated arbitrary trajectory. All of collected forces information for 12 BMs are inputted into Gated Recurrent Unit (GRU) as training data. Then, data from two totally different random trajectories produced by random walking model and K-means clustering method are used to test GRU's performance. The relevant results verified the superiority of this GRU-based coupling forces model, with faster learning time 900 s and higher accuracy 0.877 $R<^>{2}$. This article realizes the accurate depict for the coupling-forces under HTS maglev random motion, which is anticipated to provide a reference for the future engineering research.