Intelligent Optimization Algorithm For High Temperature Superconducting Levitation And Its Applications

The high-temperature superconducting (HTS) magnetic levitation has the advantages of self-stable, green and energy-saving, and thus has wide applications in magnetic levitation bearing, flywheel energy storage system and rail transportation. The HTS Maglev vehicle system consists of permanent magnetic guideway (PMG) with translational symmetry and HTS bulks of Y-Ba-Cu-O levitated above former. The HTS levitation consituent cost is being one of the main obstacles hindered its engineering application, for the rising price of rare-earth elements contained in permanent magnet (PM). Therefore, design optimization is an indispensable task to promote HTS Maglev application. Generally, present design optimization works can be classified into experiments and numerical computation. As numerical computation can take mass quantitative simulations and analysis, showing superior than quality analysis by experiments. However, most numerical optimization works only coupled with analytical models or did not employ global optimization tools. In this investigation, we first employed a global and intelligent optimization tool with 2D model of HTS Maglev.In Chapter 2, we first introduced a 2D model of HTS Maglev based on magnetic vector potential and deduced the spatial and time discretizations for the control equation of electromagnetic characteristics of HTS bulk. We also presented JFNK method for solving the nonlinear equations briefly. To confirm the validity the above-mentioned 2D model and numerical method, the numerical and experimental results under Maglev system with both a traditional PMG and a Halbach PMG were compared and discussed.In Chapter 3, we introduced and compared several usual intelligent, stochastic and global optimization tools. Then we chose the genetic algorithm to optimize the HTS Maglev system and presented the associated detailed parameters. To reduce the optimization time,we applied OpenMP, a parallel programming model to modify the serial optimization program, using the native parallelism and master-slave parallel model in genetic algorithm.We optimized an HTS Maglev system with translational symmetry with a single YBaCuO bulk to validate the above-mentioned optimization model. Optimization results reveal that both the levitation force and vertical stability are enhanced and the optimization time is reduced to a quarter of the serial computation time by the parallel program.In Chapter 4, we conducted sets of research on optimizations of both HTS Maglev systems with a single HTS bulk and multiple HTS bulks, using the above optimization tool.It is found that the finite-element method parameters, including the discretization density of spatial and time domain and critical current density of HTS bulks influence the final value of optimized levitation force and does not determine the optimized geometry dimensions of two kinds of Halbach PMGs. The working-height (WH) and geometry dimension of HTS bulk influence the optimized PMGs directly. In summary, as the PMG volume increasing, the geometry dimensions of PMs that below the HTS bulks' working domain always keep stable and the heights and widths of the other PMs is enlarging accordingly. To search a best cost-effective HTS Maglev system, the HTS bulk width should be enlarged and its height should be held a best value. The PMG's volume and geometry dimension should be adapted proportionately according to the bulk's volume and geometry. For the case of multiple HTS bulks, the guidance force of HTS Maglev system can be enhanced similarly, even when we take the levitation force maximization as the optimization objective.Multi-seeded YBaCuO bulks with large dimensions are the best choice in HTS Maglev application. However, it is hard to simulate intergrain and intragrain currents existed in a multi-seeded bulk. In this paper, we suppose a multi-seeded bulk to be an assembly of single domains that are merely magnetically coupled to investigate the effect of intergrain current loop on levitation force. Then the validity of this model is confirmed by comparing the numerical and experimental results. For the design for an high-load HTS Maglev system, we proposed that increasing the number of YBaCuO bulks and poles in Halbach PMGs. We also globally optimized various kinds of Halbach PMGs by genetic algorithm and reduced the cost.