系統識別號 U0026-1508201714395200
論文名稱(中文) 智慧型最佳換檔地圖與硬體迴路實證
論文名稱(英文) Intelligent Optimal Shift Map and HIL Experiments
校院名稱 成功大學
系所名稱(中) 機械工程學系
系所名稱(英) Department of Mechanical Engineering
學年度 105
學期 2
出版年 106
研究生(中文) 趙俊傑
研究生(英文) Jun-Jie Zhao
電子信箱 n16044886@mail.ncku.edu.tw
學號 N16044886
學位類別 碩士
語文別 中文
論文頁數 353頁
口試委員 指導教授-蔡南全
中文關鍵字 換檔地圖  動態規劃演算法  支持向量機  能量管理策略  神經網路滑模控制  硬體迴路 
英文關鍵字 Gear Shift Map  Dynamic Programming  Support Vector Machine  Energy Management Strategy  Neural Network Sliding Mode Control  Hardware-In-the-Loop 
中文摘要 對於主要動力源為內燃機引擎(Internal Combustion Engine, ICE)之車輛,在引擎轉速與扭矩的物理限制下,須透過變速箱(Transmission)的轉速/扭矩轉換以達到車輛之實際動力需求。 而現今市面上大多數的自動變速系統皆屬於離散性齒比(Discrete-ratio)的變速系統,故換檔會造成引擎操作點發生大幅度的改變,進一步影響油耗表現及駕駛性能。 因此,該如何決定換檔時機並設計一套換檔策略(Gear Shift Strategy)是一個重要課題,其中又以製作換檔地圖(Gear Shift Map, GSM)為目前各大車廠最常使用的方法。
有鑑於此,本研究針對傳統汽油車(Conventional Pure ICE Vehicle)與配置皮帶式馬達發電機(Belt-driven Starter Generator, BSG)之輕度混合並聯式油電混合動力車(Hybrid Electric Vehicle, HEV)各設計一套換檔地圖,其針對「燃油經濟性(Fuel Economy)」以及「駕駛舒適性(Driving Comfort)」進行最佳化,利用動態規劃演算法(Dynamic Programming, DP)找出最佳的檔位點; 接著使用聚合式階層分群法(Agglomerative Hierarchical Clustering, AHC)處理DP計算獲得的資料點; 最後使用分類演算法(Classification Algorithm)-支持向量機(Support Vector Machine, SVM),找出各檔位之間的最佳換檔超平面(Shift Hyperplane),藉此獲得兩檔位之間其自動換檔時機隨設計參數變化的規則。 另一方面,油電混合車之性能表現不僅受變速箱的檔位變換所影響,亦會與能量管理策略(Energy Management Strategy, EMS)息息相關; 因此,該如何利用馬達與內燃機引擎間的互補特性來改善車輛性能是另一個重要課題。 本研究採用神經網路滑模控制(Neural Network Sliding Mode Control, NNSMC)作為BSG油電車的能量管理策略,作者利用兩組徑向基底神經網路(Radial Basis Function Neural Network, RBFNN),即: RBFNN #1與RBFNN #2,並搭配滑動模式控制(Sliding Mode Control, SMC),構成一線上可實現之即時控制策略(Real-Time Control Strategy)。 首先,將動態規劃(DP)計算所獲得的最佳動力分配比(Power Split Ratio, PSR)當成RBFNN #1的訓練樣本,並藉由此離線(Off-line)訓練完成的神經網路架構,於線上辨識出車輛在特定扭矩需求下所需之動力分配值。 然而,行車型態(Drive Cycle)對於油電車之各項性能影響甚大,故額外加入RBFNN #2作為線上(On-line)之神經網路架構,並根據所遇到的路況來更新參數,以適當調整RBFNN #1辨識得出的動力分配值,使整體控制策略更具強健性,藉此適應現實之各種駕駛狀況並穩定系統之電池電量(State Of Charge, SOC),再搭配本研究設計完成之最佳換檔地圖,進一步改善油耗並提升駕駛舒適性。
關於本研究所設計的“換檔控制策略”與“能量管理控制策略”之初步驗證工作,即利用車輛模擬軟體ADVISOR (ADvanced VehIcle SimulatOR)與MATLAB/Simulink建立的後視模型(Backward-facing Model)與前視模型(Forward-facing Model)進行電腦模擬與分析; 另外,為了評估本研究所提出之控制策略在實務面之有效性,將設計完成的控制策略寫入嵌入式控制器(Embedded Controller)中,並採用目前已被廣泛應用於車輛系統的控制器區域網路(Controller Area Network, CAN or CANbus)作為控制器的溝通橋樑,藉此導入真實世界駕駛至其中以進行硬體迴路(Hardware-In-the-Loop, HIL)實驗。 本論文共選用十種行車型態來驗證研究成果,由電腦模擬結果可得知: (i)於傳統汽油車的部分,燃油經濟性之平均改善率為5.86 %,駕駛舒適性之平均改善率可高達16.18 %。 (ii)在BSG油電車的部分,燃油經濟性之平均改善率可高達20.31 %,駕駛舒適性之平均改善率可達17.18 %。 最後,由硬體迴路實驗得知,實驗結果與電腦模擬結果之改善趨勢及幅度相當一致(兩種驗證方法之誤差值低於3.5 %),也進一步驗證了本研究所提出之“換檔控制策略”與“能量管理控制策略”不管在理論面還是實務面皆能有優越的成效,因此極具潛力將它們應用於實際車輛上。
英文摘要 Nowadays, most of transmissions used in passenger vehicles are of discrete-ratio. Each action of gearshift causes a sudden change by the engine operation points and results in an undesired fuel-deficient or poor-drivability threat to the vehicles. Therefore, an optimal gearshift strategy becomes the core factor of intelligent control on any discrete-ratio transmissions. In order to achieve optimal performance on gearshift and shorten the corresponding calibration cycle, a static shift map has been widely adopted as a commonly employed approach in automotive industries. To further upgrade, an innovative design approach to integrate Dynamic Programming (DP), Agglomerative Hierarchical Clustering (AHC) and Support Vector Machine (SVM) is proposed by this thesis to construct a new Gear Shift Map (GSM) for the discrete-ratio transmissions utilized either in the conventional pure ICE (Internal Combustion Engine) vehicles or mild-parallel-hybrid electric vehicles equipped with Belt-Driven Starter Generators (BSGs).
In addition, the performance of Hybrid Electric Vehicles (HEVs) is affected not only by gearshift strategy but also by Energy Management Strategy (EMS). How to appropriately apply the mutually complementary characteristics of ICE and BSG is an important issue. In this thesis, a novel EMS based on Neural Network Sliding Mode Control (NNSMC) for a BSG Mild HEV is proposed by combining the advantages of DP, Radial Basis Function Neural Network (RBFNN) and Sliding Mode Control (SMC). The optimal datasets of Power Split Ratio (PSR) between ICE and BSG are evaluated by DP algorithm at first, and then an off-line RBFNN is employed to get these datasets trained and recognize the current required torque online in order to determine a suitable value of PSR at once. Most importantly, SMC strategy and another RBFNN, which possesses on-line learning capability, are applied to further fine-tune the value of PSR. As mentioned above, the proposed NNSMC can adaptively learn to satisfy the realistic driving situation such that under any driving pattern, the SOC can be steadily retained within a relatively narrow range and the consumed fuel is minimized simultaneously.
At design stage, the simulation softwares, ADVISOR (ADvanced VehIcle SimulatOR) and MATLAB/Simulink, are employed to verify the proposed control strategies. To be more practical, the proposed control strategies are programmed into the embedded micro-processor to conduct the Hardware-In-the-Loop (HIL) experiments. It is worthwhile to mention that a CAN (Control Area Network) PCI (Peripheral Component Interconnect) card is equipped such that the modules of embedded controllers can exchange data via CAN bus. According to the computer simulation results, (i) the mean of improvement degrees of fuel economy and driving comfort are up to 5.86% and 16.18% respectively for the conventional pure ICE vehicles; (ii) the mean of improvement degrees of fuel economy and driving comfort are up to 20.31% and 17.18% respectively for the BSG Mild HEVs. Finally, the HIL experiments show that the proposed control strategies can perform superiorly and the experimental results are pretty close to the computer simulations undertaken earlier. The discrepancy between theoretical and experimental approaches is below 3.5%. It implies that the control strategies (mainly including GSM and EMS) proposed by this thesis can be potentially applied to the real-world vehicles directly in the future.

Keywords: Gear Shift Map, Dynamic Programming, Support Vector Machine, Energy Management Strategy, Neural Network Sliding Mode Control, Hardware-In-the-Loop.
論文目次 摘要 I
英文摘要 IV
延伸摘要 VII
誌謝 XII
表目錄 XIX
圖目錄 XXII
符號表 XLII
第一章 緒論 1
1.1 前言 1
1.2 介紹 2
1.2.1 手排變速箱 3
1.2.2 自排變速箱 3
1.2.3 自手排變速箱 4
1.2.4 雙離合變速箱 4
1.2.5 無段自動變速箱 5
1.3 文獻回顧 8
1.4 研究動機與目的 17
1.5 論文架構 19
第二章 BSG油電車之建模 21
2.1 模擬方法介紹 23
2.1.1 前視模擬法 23
2.1.1 後視模擬法 24
2.1.3 混合模擬法 24
2.2 模型概述 26
2.3 行車型態 27
2.4 車體動態模型 31
2.5 動力總成模型 32
2.5.1 車輪模型 33
2.5.2 終端傳動模型 35
2.5.3 變速箱模型 36
2.5.4 離合器模型 39
2.5.5 扭矩耦合器模型 41
2.5.6 內燃機引擎模型 43
2.5.7 馬達模型 49
2.5.8 電池模型 53
2.5.9 機械周邊負載與電力周邊負載模型 61
第三章 最佳換檔地圖設計 64
3.1 傳統汽油車之最佳換檔地圖 67
3.1.1 傳統汽油車之最佳檔位點分析 67
3.1.2 傳統汽油車之最佳升檔地圖 72
3.1.3 傳統汽油車之最佳降檔地圖 82
3.1.4 小結 92
3.2 BSG油電車之最佳換檔地圖 92
3.2.1 BSG等效油耗地圖 93 馬達模式 94 發電機模式 97
3.2.2 BSG油電車之最佳檔位點分析 100
3.2.3 BSG油電車之最佳升檔地圖 107
3.2.4 BSG油電車之最佳降檔地圖 117 BSG油電車之強制降檔地圖 117 BSG油電車之一般降檔地圖 127
3.2.5 小結 137
第四章 控制策略設計與模擬驗證 138
4.1 傳統汽油車之控制策略設計與模擬驗證 139
4.1.1 換檔控制邏輯 140
4.1.2 比較基準 142
4.1.3 電腦模擬結果 143 後視模擬法之模擬結果 143 前視模擬法之模擬結果 158
4.1.4 小結 168
4.2 BSG油電車之控制策略設計與模擬驗證 172
4.2.1 換檔控制邏輯 173
4.2.2 針對HEV的能量管理控制策略 175 人工神經網路 176 滑動模式控制 176 神經網路滑模控制 177
4.2.3 比較基準 185 比較基準一(Benchmark #1) 185 比較基準二(Benchmark #2) 188
4.2.4 電腦模擬結果 190 後視模擬法之模擬結果 190 前視模擬法之模擬結果 216
4.2.5 小結 232
第五章 硬體迴路實驗架設與驗證 237
5.1 硬體迴路介紹 238
5.2 實驗設備配置與平台架設 241
5.2.1 硬體迴路平台架設 241
5.2.2 數位訊號處理器 246
5.2.3 控制器區域網路介面卡 249
5.2.4 DSP-CAN通訊協定 252
5.2.5 油門與煞車踏板 254
5.2.6 排檔桿座 255
5.3 實驗結果 258
5.3.1 傳統汽油車實驗結果 260 利用電腦駕駛對於控制策略之驗證結果 261 利用人為駕駛對於控制策略之驗證結果 266
5.3.2 BSG油電車實驗結果 274 利用電腦駕駛對於控制策略之驗證結果 275 利用人為駕駛對於控制策略之驗證結果 282
5.3.3 小結 295
第六章 結論與未來展望 298
6.1 結論 298
6.1.1 最佳換檔地圖設計結論 298
6.1.2 控制策略電腦模擬結論 300
6.1.3 硬體迴路實驗結論 303
6.1.4 總結 304
6.2 未來展望 306
參考文獻 312
附錄A 車體阻力推導 319
附錄B 車輛模擬軟體工具 328
附錄C 換檔地圖設計方法 336
附錄D 類神經網路 347
附錄E 滑動模式控制 350
自述 353
參考文獻 [1] J. MacDonald, "Electric vehicles to be 35% of global new car sales by 2040," https://about.bnef.com/blog/electric-vehicles-to-be-35-of-global-new-car-sales-by-2040/.
[2] H. Naunheimer, B. Bertsche, J. Ryborz, W. Novak, Automotive Transmissions: Springer-Verlag Berlin Heidelberg, 2011.
[3] S. Breukel, "File: Manual transmission clutch.png," https://zh.wikipedia.org/wiki/File:Manual_transmission_clutch.png.
[4] 李承儒, "【特別企劃】寡難敵眾 ─ 雙離合器自手排變速箱與其他變速箱優缺點比較," http://www.carstuff.com.tw/group-test/item/6635-2014-02-22-08-05-48.html.
[5] 蔡高坡, “AMT變速箱同步換檔耐久測試台的研究,” 機械電子工程, 合肥工業大學, 2009.
[6] M. Robinet, Global Light Vehicle Forecast: Readying For The Next Stage, IHS AUTOMOTIVE, London, 2015.
[7] W. C. Shen, H. L. Yu, Y. H. Hu, J. Q. Xi, “Optimization of Shift Schedule for Hybrid Electric Vehicle with Automated Manual Transmission,” Energies, Vol. 9, No. 3, Mar, 2016.
[8] B. Škugor, J. Deur, V. Ivanović, “Dynamic Programming-Based Design of Shift Scheduling Map Taking into Account Clutch Energy Losses During Shift Transients,” 2016.
[9] V. D. Ngo, T. Hofman, M. Steinbuch, A. Serrarens, “Gear shift map design methodology for automotive transmissions,” Proceedings of the Institution of Mechanical Engineers Part D-Journal of Automobile Engineering, Vol. 228, No. 1, pp. 50-72, Jan, 2014.
[10] A. Fofana, O. Haas, V. Ersanilli, K. Burnham, J. Mahtani, C. Woolley, K. Vithanage, “Multi-Objective Genetic Algorithm for an automatic transmission gear shift map,” IFAC-PapersOnLine, Vol. 49, No. 3, pp. 123-128, 2016.
[11] X. Yin, W. Wang, X. Chen, H. Lu, “Multi-Performance Optimization of the Shift Schedule for Stepped Automatic Transmissions,” 2013.
[12] H.-D. Lee, S.-K. Sul, H.-S. Cho, J.-M. Lee, “Advanced gear-shifting and clutching strategy for a parallel-hybrid vehicle,” IEEE Industry Applications Magazine, Vol. 6, No. 6, pp. 26-32, 2000.
[13] W. Wang, Q. Wang, X. Zeng, “Automated Manual Transmission Shift Strategy for Parallel Hybrid Electric Vehicle,” 2009.
[14] D. Le Guen, T. Weck, A. Balihe, B. Verbeke, “Definition of Gearshift Pattern: Innovative Optimization Procedures Using System Simulation,” SAE Int. J. Engines, Vol. 4, No. 1, pp. 412-431, 2011.
[15] T. Zhu, Y. Wu, Z. Wei, R. Zhao, “A new approach of shift schedule optimization for AMT vehicle based on optimal theory,” in Proceedings of the International Conference on Artificial Intelligence and Robotics and the International Conference on Automation, Control and Robotics Engineering, Kitakyushu, Japan, 2016, pp. 1-5.
[16] H. Qin, S. X. Long, K. Yu, “Simulation Research on the Shift Schedule in the Auto with Automated Manual Transmission Based on Cruise Gear Shift Program,” Advanced Materials Research, Vol. 712-715, pp. 2160-2163, 2013.
[17] A. Brooker, K. Haraldsson, T. Hendricks, V. Johnson, K. Kelly, B. Kramer, T. Markel, M. O’Keefe, S. Sprik, K. Wipke, M. Zolot. "ADVISOR Documentation," http://adv-vehicle-sim.sourceforge.net/advisor_doc.html.
[18] T. Markel, A. Brooker, I. Hendricks, V. Johnson, K. Kelly, B. Kramer, M. O'Keefe, S. Sprik, K. Wipke, “ADVISOR: a systems analysis tool for advanced vehicle modeling,” Journal of Power Sources, Vol. 110, No. 2, pp. 255-266, Aug, 2002.
[19] S. Onori, L. Serrao, G. Rizzoni, Hybrid Electric Vehicles: Springer-Verlag London, 2016.
[20] R. N. Jazar, Vehicle Dynamics: Springer-Verlag New York, 2014.
[21] R. Rajamani, Vehicle Dynamics and Control: Springer US, 2012.
[22] A. Lew, H. Mauch, Dynamic Programming: Springer-Verlag Berlin Heidelberg, 2007.
[23] O. Sundstrom, L. Guzzella, "A generic dynamic programming Matlab function." pp. 1625-1630.
[24] O. Sundstrom, D. Ambuhl, L. Guzzella, “On Implementation of Dynamic Programming for Optimal Control Problems with Final State Constraints,” Oil & Gas Science and Technology-Revue D Ifp Energies Nouvelles, Vol. 65, No. 1, pp. 91-102, Jan-Feb, 2010.
[25] A. Fernandez, S. Gomez, “Solving non-uniqueness in agglomerative hierarchical clustering using multidendrograms,” Journal of Classification, Vol. 25, No. 1, pp. 43-65, Jun, 2008.
[26] C. Fernandez-Lozano, C. Canto, M. Gestal, J. M. Andrade-Garda, J. R. Rabunal, J. Dorado, A. Pazos, “Hybrid Model Based on Genetic Algorithms and SVM Applied to Variable Selection within Fruit Juice Classification,” Scientific World Journal, 2013.
[27] V. Vapnik, The Nature of Statistical Learning Theory: Springer-Verlag New York, 2000.
[28] T. Guan, C. W. Frey, "Using ensemble of decision trees with SVM nodes to learn the behaviour of a transmission control software." pp. 1323-1328.
[29] R. Wang, S. M. Lukic, "Dynamic programming technique in hybrid electric vehicle optimization." pp. 1-8.
[30] C. Kirtane, S. Ghodke, D. S. Kurode, D. P. A.K., Dr.D.N.Malkhede, “Gear Shift Schedule Optimization and Drive Line Modeling for Automatic Transmission,” in 1st International and 16th National Conference on Machines and Mechanisms (iNaCoMM2013), India, 2013.
[31] D. Kriesel, "A Brief Introduction to Neural Networks," U. o. Bonn, ed., 2005.
[32] J. Moreno, M. E. Ortuzar, J. W. Dixon, “Energy-management system for a hybrid electric vehicle, using ultracapacitors and neural networks,” IEEE Transactions on Industrial Electronics, Vol. 53, No. 2, pp. 614-623, 2006.
[33] Y. Shtessel, C. Edwards, L. Fridman, A. Levant, Sliding Mode Control and Observation, New York: Birkhäuser Basel, 2014.
[34] J. J. Niu, Y. L. Fu, X. Y. Qi, “Design and Application of Discrete Sliding Mode Control with RBF Network-based Switching Law,” Chinese Journal of Aeronautics, Vol. 22, No. 3, pp. 279-284, Jun, 2009.
[35] F. J. Lin, S. Y. Lee, P. H. Chou, “Intelligent Integral Backstepping Sliding-mode Control Using Recurrent Neural Network For Piezo-flexural Nanopositioning Stage,” Asian Journal of Control, Vol. 18, No. 2, pp. 456-472, Mar, 2016.
[36] Y. Tao, J. Q. Zheng, Y. C. Lin, “A Sliding Mode Control-based on a RBF Neural Network for Deburring Industry Robotic Systems,” International Journal of Advanced Robotic Systems, Vol. 13, Jan, 2016.
[37] B. Adel1, Z. Youtong, S. shuai, “Parallel HEV Hybrid Controller Modeling for Power Management,” World Electric Vehicle Journal, Vol. 4, No. 1, pp. 190-196, 2010.
[40] T. H. Christoph Filser, Rolf Jung, “V-model development of safety application,” CAN Newsletter, 2013.
[41] National Instruments. "硬體迴路 (HIL) 測試系統架構," http://www.ni.com/white-paper/10343/zht/.
[42] Softing AG, "CAN-AC1-PCI CAN-AC2-PCI Hardware Notes," SOFTING AG, 2006.
[43] 郝瑞祥, DSP原理與應用, 北京交通大學 電氣工程學院, 北京, 2010.
[44] 陳明昌, “FlexRay/CAN 車載網路之閘道器封包驗證系統,” 電資學院電資碩士專班, 國立台北科技大學, 台灣, 2015.
[45] “Recommended Practice for Measuring the Exhaust Emissions and Fuel Economy of Hybrid-Electric Vehicles, Including Plug-in Hybrid Vehicles,” SAE Standard J1711, 2010.

  • 同意授權校內瀏覽/列印電子全文服務,於2020-09-01起公開。
  • 同意授權校外瀏覽/列印電子全文服務,於2021-09-01起公開。

  • 如您有疑問,請聯絡圖書館