臺灣博碩士論文加值系統

壓光機之產品品質優劣取決於壓光滾輪間之變形量控制，在現今工業上，仍以試驗及嘗試錯誤的方法決定各部份尺寸，因此難以得到最佳化之參數值，壓光滾輪是一個分佈參數系統的模型，如我們所知，分佈參數系統中任一點之振動及位移是由無限多個振態疊加而成的，目前工業操作上更改不同條件時，振動的產生及抑制振動是不可避免的，由本研究所設計之控制器可有效地控制滾輪變形達到目標而没有增加滾輪的振動量。本文主要的目的是如何在機台上有效控制壓光滾輪中間之變形量，而不需要停機拆滾輪來研磨滾輪的中凸量，能使產品在生產過程中達到平整的壓光平面，本研究結果希望能設計出精確的控制器及參數，適用於壓光滾輪系統。
首先利用牛頓法推導壓光輪之動態方程式，利用特徵函數求得壓光輪系統之開迴路轉移函數，接著選擇PID類型之控制器，並利用根軌跡法分析其軌跡變化情形，進行時域上之數值模擬分析。最後設計系統第一模態性能，以符合系統主極點之需求，本研究使用工業界慣用的PID控制器，但傳統PID類型控制器雖使系統之響應速度增快，但是卻也造成輸出響應在到達目標值前，會有振盪產生，即在穩定之前會有最大超越量(maximum overshoot)發生。因此利用輸入訊號波形改變方法，以避免最大超越量的發生，除消除振盪的部分，並獲得更快速的穩定時間。振盪消除控制對於系統暫態性能的提昇有很大的助益，並且能降低系統的最大功率與減少因振盪現象而造成機械壽命縮短。
最後，在紡織工業中，具有連續性加工之特性，壓光機基於大量生產的要求，必須先將布匹的末端及下一匹要加工布匹的前端車縫接合，但此縫合之區域為不可加工處，壓光機在加工過程中，自動檢測出縫目時將自動跳開布匹接縫處而連續加工，滾輪系統必須在一個短暫的時間內改變壓力，當布匹接縫處通過時，希望能在最短的時間內達到目標的壓力，本文設計一類神經控制自動縫目檢出裝置，避免對此縫接區域加工，配合所設計之可程式連續加工控制系統，自動進行下一布匹表面加工，保持完整的自動化連續加工性能。實際運用係以光電元件作為自動偵測布匹厚度變化之裝置，並將此厚度變化轉換為電氣訊號送入類神經控制器，進而判斷出接縫處之位置及長度，再由類神經控制器輸出訊號，自動跳開布匹接縫處而連續加工。

Putting deformation amounts of the pressing and pressed rollers under control is crucial to enhance calender’s product quality. Currently in the industry, trial and error is still used to determine each part of dimension of these two kinds of rollers, which makes it difficult to determine the optimal design parametric values. In this paper, the calender pressed roller is modeled as a distributed system. It is well known that, in general, the vibration and displacement at any point of a distributed system can be represented as the superposition of an infinite number of vibration modes. In the industrial operations, the capability of suppressing all the vibration modes during different operating conditions is necessary from the designed controllers that can effectively control the machine to accomplish its task without exciting excessive mechanical vibrations. The major objectives of this paper is to control the calender pressed roller, no need for both of the pressed and pressing rollers, to reach the amount of convexity adjusted online, which in turn leads up to the leveling of the contact surface between the calender pressing roller and the calender pressed roller during processing. In the transient response of the calender roller system, the settling time is expected to be short, reach the non-overshoot controller design, and arrive good tracking property. Fast tracking performance and settling times, and low energy consumption while minimizing vibrations must be provided by a high performance calender roller system. The different kinds of PID control strategy were designed to improve the performance of calender roller system. The traditional PID (proportional pus integral and derivative) controller was designed to expedite the response of the calender roller system without steady-state error. But the overshoot has been existed. The shaping input scheme together with the designed PID controller was employed to regulate the calender roller system input, so as to generate system non-overshoot and shorten the settling time. A regulating reference input technology combined with the aforementioned PID type control scheme is proposed for robustness and dominant mode vibration suppression. This combining scheme possesses the advantages of simplicity and effectives, and because no additional sensor and actuator are required. Finally, In the textile industry, the fabrics must accept various kinds of processing steps in order to meet the mass production requirement. To cope with this demand, the tail end of the fabrics must be sewed together with the front part of the next fabrics to be processed. In this paper, the neural system controlled automatic stitch detection device is successfully designed. This system can avoid operating on the sewed area and work with the programmable continuous processing control, so that the surface of next fabrics can be processed in order to maintain an integrated function of automatic continual processing control. A photoelectric component is used to automatically detect the changes of fabric thickness and transform the thickness variation to electrical signal for sending to the neural controller so as to determine the position and the length of such sewing area. After being fed from the neural controller, the signal skips the sewing area and thus achieving the unceasing processing purpose.

第1章 緒論 1
1.1. 前言 1
1.2. 研究動機與目的 3
1.3. 文獻回顧 4
1.4. 研究流程 8
1.5. 論文架構 9
第2章 滾輪系統動態模式及轉移函數 10
2.1. 系統模式推導 12
2.2. 滾輪系統模型建立 15
2.3. 滾輪系統動態方程式之推導 17
2.4. 特徵值與特徵函數 20
2.5. 滾輪系統垂直振幅轉移函數 23
第3章 滾輪系統控制器設計 28
3.1. PID控制器 31
3.2. 振盪消除理論 38
第4章 壓光機系統之自動縫目檢出器設計 43
4.1. 布厚檢出機構設計 44
4.2. 類神經網路控制器設計 48
第5章 結果與討論 53
5.1. PID控制器模擬 54
5.2. 振盪消除模擬 65
5.3. 自動縫目檢出器實驗結果 76
第6章 結論 84

圖目錄
圖 1 1 研究步驟流程圖 8
圖 2 1 壓光機整體圖 10
圖 2 2 壓光機實體圖 11
圖 2 2 滾輪受力變形示意圖 12
圖 2 3 目前使用之滾輪系統 14
圖 2 4 壓光輪的中凸量控制系統示意圖 14
圖 2 5 滾輪控制系統示意圖 15
圖 2 6 滾輪外力平衡及慣性力圖 17
圖 2 7 模態函數示意圖(n=1~4) 26
圖 2 8 模態函數示意圖(n=5~8) 27
圖 3 1 系統開迴路之極點與零點分佈圖 29
圖 3 2 閉迴路控制系統方塊圖 31
圖 3 3 系統加入比例微分積分控制器之方塊圖 34
圖 3 4 閉迴路極點配置圖 37
圖 3 5 標準二階控制系統 38
圖 3 6 振盪輸出相消圖 39
圖 4 1 布厚檢出立體外觀 44
圖 4 2 非接觸式厚度感測器 44
圖 4 3 布厚檢出器實體 47
圖 5 1 第一模態( =0.6,ωn =20rad/s)之閉迴路系統根軌跡圖 55
圖 5 2 滾輪中點位移響應圖( =0.6, =20rad/s) 56
圖 5 3 控制力輸出響應圖( =0.6, =20rad/s) 56
圖 5 4 PID控制後各節點響應比較圖( =0.6, =20rad/s) 58
圖 5 5 PID控制各節點響應立體圖( =0.6, =20rad/s) 58
圖 5 6 第一模態( =0.4, =20rad/s)之閉迴路系統根軌跡圖 59
圖 5 7 滾輪中點位移響應圖( =0.4,ωn=20rad/s) 60
圖 5 8 控制力輸出響應圖( =0.4,ωn=20rad/s) 61
圖 5 9 PID控制後各節點響應比較圖( = 0.4, =20rad/s). 62
圖 5 10 PID控制各節點響應立體圖( = 0.4, =20rad/s). 62
圖 5 11 比較不同 之滾輪中點位移響應圖 63
圖 5 12 比較不同 之控制力輸出響應圖 64
圖 5 13 比較不同 控制後各節點響應圖 64
圖 5 14 改變輸入訊號之變位響應圖 66
圖 5 15 消除振盪前後之滾輪系統變位響應圖( = 0.6, ωn=20rad/s) 66
圖 5 16 消除振盪前後之滾輪系統控制力比較圖( = 0.6, ωn=20rad/s) 67
圖 5 17 滾輪各點變形位移響應圖( = 0.6, ωn=20rad/s) 67
圖 5 18 消除振盪前後之滾輪系統變位響應圖( = 0.4, ωn=20rad/s) 68
圖 5 19 消除振盪前後之滾輪系統控制力比較圖( = 0.4, ωn=20rad/s) 69
圖 5 20 滾輪各點變形位移響應圖( = 0.4, ωn=20rad/s) 69
圖 5 21 消除振盪後之滾輪系統變位響應比較圖(ωn=20rad/s) 70
圖 5 22 目標改變消除振盪前後之滾輪系統響應比較( = 0.6, ωn=20rad/s) 71
圖 5 23 目標改變消除振盪前及後之滾輪系統控制力比較( = 0.6, ωn=20rad/s) 72
圖 5 24 最終目標減少為一半變形量滾輪系統之響應比較( = 0.6, ωn=20rad/s) 72
圖 5 25 最終目標減少為一半變形量滾輪系統之控制力比較( = 0.6, ωn=20rad/s) 73
圖 5 26 目標改變滾輪中間各點變形位移響應圖( = 0.6, ωn=20rad/s) 73
圖 5 27 最終目標減少為一半變形量滾輪中間各點變形位移響應圖( = 0.6, ωn=20rad/s) 74
圖 5 28 消除振盪後滾輪整體變形立體圖( = 0.6, ωn=20rad/s) 75
圖 5 29 目標改變消除振盪後滾輪整體變形立體圖( = 0.6, ωn=20rad/s) 75
圖 5 30 最終目標減少為一半變形量消除振盪後滾輪整體變形立體圖( = 0.6, ωn=20rad/s) 76
圖 5 31 布匹厚度變化 77
圖 5 32 厚度輸入之正規化值 78
圖 5 33 誤差均方根(MSE)收斂圖 79
圖 5 34 倒傳遞網路連結加權值和閥值 79
圖 5 35 驗證資料之厚度輸入值 81
圖 5 36 倒傳遞神經網路判斷輸出值 81
圖 5 37 厚度輸入、判斷輸出值關係圖 82
圖 5 38 倒傳遞神經網路預測散佈圖 83

表目錄
表 2 1 系統物理參數值 16
表 2 2 系統特徵值與自然頻率 25
表 3 1 系統開迴路之極點與零點值 29
表 5 1 各模態系統參數表( =0.6, =20rad/s) 55
表 5 2 各模態系統參數表( =0.4, =20rad/s) 60

1.Usoro, P. B., Mahil, S. S., and Nadir, R., A Finite Element/Lagrange Approach to Modeling Lightweight Flexible Manipulators, ASME Journal of Dynamic Systems, Measurement, and Control, 108(3), 198-205, (1986).
2.Skelton, R. E., Hughes, P., and Hablani, H., Order Reduction for Models of Space Structures Using Model Cost Analysis, Journal of Guidance and Control, 5(4), 351-357 (1982).
3.Hickin, J., and Sinha, N. K., Model Reduction for Linear Multivariable Systems, IEEE Transactions on, Automatic Control, 25(6), 1121-1133, December (1980).
4.Chang, R. Y., and Wang, M. L., Model Reduction and Control System Design by Shifted Legendre Polynomial Functions, Transactions of the ASME, Journal of Dynamic Systems, Measurement, and Control, 105(1), 52-55 (1983).
5.Crawley, E. F., Use of Piezo-Ceramics as Distributed Actuators in Large Space Structures, Structures, Strctural Dynamics and Matherials Conference, Orlando, FL, AIAA Paper No. 85-0626, 126-133 (1986).
6.Miller, D.W., Crawley, E. F., and Ward, B. A., Inertial Actuator Design for Maximum Passive and Active Energy Dissipation in Flexible Space Structures, Structures, Strctural Dynamics and Matherials Conference, Orlando, FL, AIAA Paper No. 85-0777, 536-544 (1986).
7.Kuo, C. Y., and Huang, C. C., Active Control of Mechanical Vibrations in a Circular Disk, ASME Journal of Dynamic Systems, Measurement, and Control, 114, 104-112 (1992).
8.Rao, S. S., “Mechanical Vibrations” , Addison-Wesley Publishing Company, New York, (1992).
9.Kuo, C. F. J., and Lin, S. C., Discretization and Computer Simulation of a Rotating Euler-Bernoulli Beam, Mathematics and Computers in Simulation, 52(2), 121-135 (2000).
10.Kuo, C. F. J., and Lin, S. C., Modal Analysis and Control of a Rotating Euler-Bernoulli Beam, Part I: Control System Analysis and Controller Design, An International Journal of Mathematical and Computer Modeling, 27(5), 75-92 (1998).
11.Kuo, C. F. J., and Lin, S. C., Modal Analysis and Control of a Rotating Euler-Bernoulli Beam, Part II: Residual Vibration Control, An International Journal of Mathematical and Computer Modeling, 27(5), 93-97 (1998).
12.Meirovitch, L., “Dynamics and Control of Structures” , John Wiley & Sons, New York, (1990).
13.Gevarter, W. B., Basic Relations for control of Flexible Vehicles, AIAA Journal, 8(4), 666-672 (1970).
14.Franklin, G., Powell, J., and Emani-Naeini, A., “Feedback Control of Dynamic System”, Addison-Wesley Publishing Company, (1991).
15.B. C. Kuo, “Automatic Control systems”, Prentice-Hall, (1997).
16.Cetlinkunt, S. and Yu, W. L., Closed-Loop Behavior of a Feedback-Controlled Flexible Arm: A Comparative Study, Int. Journal Robotics Research, 10(3), 263-275 (1991).
17.Ogata, K., “Modern Control Engineering”, Second Edition, Prentice-Hall, NJ, (1990).
18.Kuo, C-F. J., Wang, C. C., and Hsieh C. T., Theoretical Control and Experimental Verification of Carded Web Density, Part Ⅰ: Dynamic System Analysis and Controller Design. Textile Research Journal, 68(12) 873-880 (1998).
19.Kuo, C-F. J., and Hsieh C. T., An Entire Strategy for Precise System Tracking Control of a Revolving Thin Flexural Link, Part Ⅱ：System Output Modifiction, Mathematical and Computer Modelling, 29 115-121 (1999).
20.Signer, N. C., and Seering, W. P., Preshaping Command Input to Reduce System Vibration, AMSE Journal of Dynamic System, Measurement and Control, 112 76-82 March (1990).
21.Singhose, W., Crain, E., and Seering, W., Convolved and Simultaneous Two-Mode Input Shapers. IEE Proceedings Control Theory Applications, 144(6) 515-520 November (1997).
22.Craig, F. C., and Lucy, Y. P., Control Using Equal Length Shaped Commands to Reduce Vibration. IEEE Transactions on Control Systems Technology, 11(1) 62-72 January (2003).
23.Singhose, W., Seering, W., and Signer, N. , Residual Vibration Reduction Using Vector Diagrams to Generate Shaped Inputs. ASME Journal Mechanical Design, 116(2) 654-659 June (1994).
24.Hai, T. H., Fast Servo Bang-Bang Seek Control. IEEE Transactions on Magnetics, 33(6) 4522-4527 November (1997).