以粒子尋優最佳化為基礎之磁浮運輸系統即時控制設計

Real-Time Control Design for Maglev Transportation System Via Particle Swarm Optimization

莊坤倫(Kun-Lun Chuang)

指導教授 ： 魏榮宗

元智大學/電機通訊學院/電機工程學系/碩士(2008年)

摘要

一般而言，磁浮運輸系統主要可區分為磁力懸浮與推進裝置兩部分，由於電磁鐵所產生的吸引力具有非線性特性且懸浮平台的姿態控制問題亦有待克服，是故磁力懸浮裝置為目前極為熱門研究主題之一；另一方面，移動平台懸浮時所產生的正向力往往會對推進裝置造成相當程度的干擾，導致於磁浮運輸系統耦合動態呈現高度非線性且時變特性，因此本論文發展以粒子尋優最佳化為基礎之磁浮運輸系統即時控制設計。首先，採用全域滑動模式控制之精神來設計一具有指向性之粒子尋優最佳化控制系統以達成磁浮運輸系統穩定平衡及循軌控制之目的，於此控制架構中，粒子尋優最佳化演算法被用來設計一主控制器，設計過程中不需要特殊限制和詳細系統資訊，而其收斂性可以藉由全域滑動模式控制間接被保證，為了有效的加速粒子尋優最佳化演算法搜尋速度，於是更引入了監督控制機制，以有效加速搜尋到目標。然而此方式仍需要使用部份系統參數和控制轉換，因此本論文更進一步採用步階迴歸控制之設計精神，並結合由里亞普諾穩定理論所設計之適應性法則來組成一步階迴歸粒子尋優最佳化控制系統，其穩定度可以直接由里亞普諾穩定理論證明，且無須系統任何資訊、控制轉換及其他輔助控制器即可達成理想之控制性能。最後本論文利用數值模擬及實作結果來驗證所提出控制系統之有效性與強健性。

關鍵字

磁浮運輸系統 磁力懸浮 推進裝置 全域滑動模式控制 監督控制 步階迴歸控制 里亞普諾穩定理論 粒子尋優最佳化

並列摘要

In general, a maglev transportation system contains two parts including magnetic levitation and propulsion mechanism. The subject of inherently unstable electromagnetic force and nonlinear attitude control of the moving platform in the maglev mechanism is one of interesting research topics at present. On the other hand, the corresponding control performance of the propulsion mechanism is influenced easily by the normal force produced by the maglev mechanism so that the coupled dynamic model of the maglev transportation system is highly nonlinear and time varying. Therefore, this thesis presents the real-time control design for a maglev transportation system via particle swarm optimization (PSO). First, a total sliding-mode-based PSO (TSPSO) control system is investigated based on the direction-based PSO with the spirit of total sliding-mode (TS) control to achieve the stable balancing and tracking control of the maglev transportation system. In this control scheme, a PSO algorithm is developed to be the major controller, and the convergent property can be indirectly ensured by the concept of TS control without strict constraint and detailed system knowledge. In order to further directly stabilize the system states around a predefined bound region and effectively accelerate the searching speed of the PSO algorithm, a supervisory scheme is embedded into the TSPSO control to constitute a supervisory TSPSO (STSPSO) control strategy. Unfortunately, partial system knowledge and control transformation are still required in the design process of the STSPSO control. In addition, a backstepping-based PSO (BSPSO) control system is further constructed to achieve the stable balancing and tracking control of the maglev transportation system and reform the shortcoming of the STSPSO control system. In this control scheme, adaptation laws derived from Lyapunov stability analyses are utilized to adjust appropriate evolutionary steps, and the system stability can be directly guaranteed without the requirement of auxiliary compensated controllers, strict constrains and control transformations. Finally, the effectiveness and robustness of the proposed control strategies are verified by numerical simulations and experimental results under the possible occurrence of uncertainties.

並列關鍵字

Maglev transportation system Magnetic levitation Propulsion mechanism Total sliding-mode control Supervisory control Backstepping control Lyapunov stability theorem Particle swarm optimization

參考文獻

[1] H. W. Lee, K. C. Kim, and J. Lee, “Review of maglev train technologies,” IEEE Trans. Magnetics, vol. 42, no. 7, pp. 1917?{1925, 2006.

[2] I. Boldea and S. A. Nasar, “Linear electric actuators and generators,” IEEE Trans. Energy Conver., vol. 14, no. 3, pp. 712?{717, 1999.

[3] P. Holmer, “Faster than a speeding bullet train,” IEEE Spectrum, vol. 40, no. 8, pp. 30?{34, 2003.

[4] F. J. Lin, R. J. Wai, W. D. Chou, and S. P. Hsu, “Adaptive backstepping control using recurrent neural network for linear induction motor drive,” IEEE Trans. Ind. Electron., vol. 49, no. 1, pp. 134?{146, 2002.

[5] R. J. Wai and W. K. Liu, “Nonlinear control for linear induction motor servo drive,” IEEE Tran. Ind. Electron., vol. 50, no. 5, pp. 920?{935, 2003.